Clinical Science

Review article

Immunological aspects of atherosclerosis

Kevin J. Woollard


Cardiovascular disease is the leading cause of death in several countries. The underlying process is atherosclerosis, a slowly progressing chronic disorder that can lead to intravascular thrombosis. There is overwhelming evidence for the underlying importance of our immune system in atherosclerosis. Monocytes, which comprise part of the innate immune system, can be recruited to inflamed endothelium and this recruitment has been shown to be proportional to the extent of atherosclerotic disease. Monocytes undergo migration into the vasculature, they differentiate into macrophage phenotypes, which are highly phagocytic and can scavenge modified lipids, leading to foam cell formation and development of the lipid-rich atheroma core. This increased influx leads to a highly inflammatory environment and along with other immune cells can increase the risk in the development of the unstable atherosclerotic plaque phenotype. The present review provides an overview and description of the immunological aspect of innate and adaptive immune cell subsets in atherosclerosis, by defining their interaction with the vascular environment, modified lipids and other cellular exchanges. There is a particular focus on monocytes and macrophages, but shorter descriptions of dendritic cells, lymphocyte populations, neutrophils, mast cells and platelets are also included.

  • atherosclerosis
  • immunology
  • leucocyte biology
  • monocyte
  • macrophage


Cardiovascular disease is the leading cause of death in many countries, resulting in approximately 16–17 million deaths each year [1,2]. The underlying pathological process is atherosclerosis, a slowly progressing chronic disorder of large and medium-sized arteries that can lead to intravascular thrombosis [3]. The pathology is characterized by a chronic inflammatory process at the arterial wall, occurring at predilection sites with disturbed laminar flow [4]. It can be initiated by endothelial dysfunction and structural alterations, including the absence of a confluent luminal elastin layer and the exposure of proteoglycans, which permit subendothelial accumulation of LDLs (low-density lipoproteins) [5,6]. Interactions of modified LDLs within the luminal endothelium, as well as with resident cells located in the subendothelial space, initiate the formation of chemotactic gradients that attract further leucocytes (e.g. monocytes, neutrophils, lymphocytes and possibly circulating stem cells) into the arterial wall from the luminal blood [5,7,8]. A complex cascade of mechanisms, including the recruitment and migration of leucocytes through the activated endothelium in predisposed regions of the intima, as well as receptor-dependent accumulation of lipids within intimal cells, leads to the formation of foam cells (Figure 1). This culminates in the appearance of fatty streaks in the arteries, altered intima and formation of a lipid core [5,6,9]. The atherosclerotic plaque develops, in which the lipid necrotic core becomes separated from the arterial lumen by a fibrous cap. The plaque then progresses into a complicated lesion with a lipid necrotic core at the center and large numbers of leucocytes at the shoulders of the lesion, including neovascularization of capillaries that may originate from the adventitial vasa vasorum [5]. As the plaque matures, there is an increased possibility of breakdown, which is accompanied by the formation of thrombi on the plaque's damaged surface [10]. Consequently as the damage occurs, there is an increased risk of thrombi breaking off and travelling to the lung, heart or brain [10].

Figure 1 Summary of plaque milieu

Monocytes (mainly Gr1high/CD14+) are recruited into the intima and may differentiate into heterogeneous macrophages that accumulate haemolysed debris or cholesterol lipids and form foam cells in the sub-intima. Whether they differentiate into specific ‘M1-like’ or ‘M2-like’ macrophages from specific recruited monocyte subsets and/or proliferate is unknown. Monocytes may also give rise to DCs in the plaque. Use of vasa vasorum and neovascularization within the plaque leads to increased trafficking of leucocytes. Egress (cell efflux) of leucocytes from plaques has been observed. Deposition of matrix components and recruitment of smooth muscle cells give rise to the fibroproliferative progression of the plaques. Apoptosis of macrophages/foam cells creates a necrotic core. Thinning and erosion of the fibrous cap in unstable plaques through matrix degradation by proteases ultimately results in plaque rupture and platelet thrombus. Further recruitment of lymphocyte subsets, platelets, neutrophils and other inflammatory leucocytes may help development of the vulnerable plaque. Circulating lipids (LDL) can become modified and potentially bind autoantibodies to oxLDL, which may also occur in the plaque milieu.

In the early 19th century, the term arteriosclerosis was introduced by Jean Lobstein, and the emergence that inflammation plays a role in atherosclerosis dates to original observations by Virchow and von Rokitansky [11,12]. In the 1970s, Ross demonstrated that leucocyte adhesion to the endothelial surface was an early feature of atherosclerosis [13]. These observations, along with knowledge that modified LDL contributes to foam cell formation and atherosclerotic disease activity, focused early reports on monocytes and macrophages, major players in innate immunity, which express functional scavenger receptors [14,15]. Scavenger receptors recognize modification-specific epitopes of (among others) SRA-1 and SRA-2, SR-B1, LOX-1, MARCO (macrophage receptor with collagenous structure), CD36 and PSOX21 [16]. Although these receptors serve a vital role as mediators of intracellular cholesterol accumulation, their detailed importance in atherosclerosis remains unclear, and gene-knockout studies of hypercholesterolaemic mice have provided contradictory results [15,16]. In 1973, Steinman and Cohn [17] provided a new scientific specialty with their discovery of the DC (dendritic cell), a vital antigen-presenting cell in the immune system. Since then, DCs have proved to be critical sentinels for atherosclerosis [18].

Components of adaptive immunity are also present in human atherosclerotic lesions, and several studies have indicated a central role for antigen-specific adaptive immune responses in atherogenesis [19]. Studies using mouse experimental models of atherosclerosis, such as ApoE (apolipoprotein E)−/− or LDLR (LDL receptor)−/− mice, in combination with mice deficient in both B-cells and T-cells, have demonstrated a key role for adaptive immunity in atherosclerosis [19]. Lymphocyte-deficient animals, for example, lacking recombination-activating gene-1 or -2, or mice with severe combined immunodeficiency under atherosclerotic settings, have been shown to aggregate atherosclerosis in some cases and inhibit it in others, discussed in [20].

Activated endothelium is characterized by adhesion molecule expression and reduced barrier function that can mediate the recruitment of leucocytes into atherosclerotic-prone sites in the arterial wall [21]. These include expression of adhesion molecules at sites of turbulent flow that enable recruitment of inflammatory cells [4]. Interestingly, in vivo reports using intravital microscopy indicate that neutrophils are responsible for the majority of transient interactions between leucocytes and endothelial cells covering atherosclerotic lesions [22,23]. Along with the detection of leucocytes in atherosclerotic lesions, a crucial question is how these cells enter atherosclerotic plaques. Observations from intravital fluorescence microscopy showing adhesion of neutrophils to atherosclerotic lesions and subsequent migration suggest a transluminal route [24]. However, further work indicates that monocytes may infiltrate large arteries via adventitial or intimal microvessels [25]. There are relatively fewer studies investigating molecular mechanisms of in vivo arterial leucocyte infiltration, so that the vast majority of in vivo concepts of recruitment in atherosclerosis are extrapolated from microvascular recruitment models, which points towards the need for development of cellular resolution techniques to investigate in vivo leucocyte recruitment related directly to atherosclerosis.

Altogether there is overwhelming evidence for the underlying importance of immunology in atherosclerosis. A previous review has provided a concise understanding of the immunological roles of immune cell susbets in atherosclerosis, their recruitment and metabolic functions [26]. The present review provides an overview and description of the immune cell response in atherosclerosis, specifically monocytes and macrophages, but also neutrophils, lymphocytes, mast cells and platelets, by defining their response to the vascular environment, modified lipids and other cellular interactions.



During inflammation, blood monocytes can migrate from blood to lymphoid and non-lymphoid tissues in response to tissue-derived signals caused, for example, by infection or tissue damage [27]. They can phagocytose other cells and toxic molecules [such as oxLDL (oxidized LDL)], produce inflammatory cytokines, and can differentiate into inflammatory dentritic cells, macrophages or ultimately foam cells (directly relevant to atherogenesis) [28,29]. During atherogenesis, blood monocytes are recruited into the intima and subintima [30,31]. Upon encountering fatty deposits, via scavenger or other lipid receptors, these cells can take up modified LDLs (and other lipid species), resulting in activation, macrophage polarization and accumulation in the forming lesion [16]. At an early stage of the process, monocytes differentiate into foam cells to form early plaques, termed fatty streaks, in the intima [32] (Figure 1). Interestingly, this process occurs in atherosclerosis-prone areas of the arterial tree and is likely to involve cellular responses to changes in fluid dynamics [4].

Depletion of monocytes in experimental mouse models decreases the development of atherosclerosis, but wholesale ablation of monocytes and macrophages is not a viable therapeutic option, because of their essential role in immunity [33]. Studies have documented heterogeneity among human monocytes [34], but the discovery and characterization of monocyte subsets in the mouse progressed investigation into the relevance of monocyte heterogeneity in atherosclerosis [35,36].

A number of studies have used histology to identify the cellular components of plaques, but adoptive transfer and fate-mapping strategies to study monocyte recruitment (and egress) from plaques, and laser microdissection to directly investigate the functions of plaque macrophages, have been reported [3739]. However, the relationship between blood monocytes and plaque macrophages is an unresolved issue that is difficult to address due to their plasticity and swift response to their tissue microenvironment [40].

Morphologically, monocytes are larger than lymphocytes and contain a horseshoe or kidney-like nucleus. They are phagocytic and can develop into macrophages and DCs in vitro and in vivo. In humans and mice, at least two monocyte subsets have been described. In humans they are usually classified according to CD14 and CD16 expression levels (see extended discussion below). In mice, monocytes can be defined by their expression of Ly6c, which is an epitope of Gr-1. Ly6chigh (Gr-1+) monocytes are CCR (CC chemokine receptor) 2high CX3CR1low, CD62L+, whereas Ly6clow (Gr-1) monocytes are CCR2low, CX3CR1high and CD62L [34,40].

Including recent nomencaulture labelling of M1/M2 macrophages (see the Macrophages section below), originally, Ly6chigh monocytes were called ‘inflammatory’ as they accumulate in the peritoneum in response to inflammatory signals. Ly6clow monocytes are called ‘resident’ because of their accumulation in reported tissues in the steady state, or termed ‘patrolling’ because of their ability to crawl along the endothelial interface in the steady state [41,42]. Others have proposed to call Ly6chigh monocytes, in orthology to human monocytes, ‘classical’ and Ly6clow monocytes ‘non-classical’, a nomenclature that, perhaps convenient for some, offers no functional insight and most likely just adds confusion [43].

In the steady state, the proportion of circulating Ly6chigh and Ly6clow monocytes is similar [41]. In murine models of atherosclerosis, mice deficient in ApoE develop large and complex lesions when fed a high-fat diet. Hypercholesterolaemia can induce monocytosis in the bone marrow, blood and spleen [44,45]. Ly6chigh cells increase in number preferentially (whether there is an intermediate functional subset in mice requires further investigation). Recent work has shown that hypercholesterolaemia induces proliferation of HSPCs [HSC (haemopoietic stem cell) progenitors] involving cholesterol efflux via ApoE and ABC (ATP-binding cassette) (ABCA1 and ABCG1) transporters [46]. HSPCs secrete ApoE that can bind to proteoglycans on the cell surface and mediate cholesterol efflux using ABCA1 and ABCG1 linking directly to HDLs (high-density lipoproteins). ApoE deficiency impairs cholesterol efflux, increases membrane cholesterol content as well as the surface expression of the GM-CSF (granulocyte-macrophage colony-stimulating factor) receptor, leading to accelerated atherosclerosis [46,47]. Ly6chigh monocytes have been reported to accumulate in the growing atheroma preferentially via CCR2 and CX3CR1 [44,48,49]. Ly6clow monocytes have been reported to infiltrate atherosclerotic lesions (less frequently) via CCR5, but subset-specific differences in atherosclerotic recruitment still require further studies [50]. The bone marrow contains specialized haemopoietic niches; however, monocyte subsets can also be found in the spleen as part of a reservoir that can be mobilized in response to systemic inflammatory stimuli such as from MI (myocardial infarction) and recently shown to be crucial in accelerated atherosclerosis post-MI [51].

Human peripheral blood monocytes do show heterogeneity similar to the existence of subsets in mice [34,41,52]. Human monocytes were identified by their expression of CD14. Subsequent identification of differentially expressed antigenic markers further identified CD16 [53]. Similar to mouse peripheral blood, at least two principle subsets can be defined. CD14highCD16low monocytes typically represent ~85–95% of the monocytes in healthy individuals against CD14lowCD16high monocytes (referenced as CD14dim). These subsets, like those in mice, differ in many respects, including their expression of adhesion molecules and chemokine receptors [5456]. CD14highCD16low monocytes express (among others) CCR2, CD62L and CD64 and have consequently been associated with the inflammatory subset in the mouse. CD14dim monocytes lack CCR2 and have higher levels of major histocompatibility complex II and CD32. Both subsets express the fractalkine receptor, CX3CR1, but CD14dim monocytes express much higher levels. It is, however, important to note that there is considerable heterogeneity in the CD16+ population, with a minor population expressing both CD14 and CD16 [56]. Therefore in human blood it would seem there are three key populations of circulating monocytes: CD14high(CD16low), CD14highCD16high and CD14dim(CD16high).

By performing principal component analysis and by comparing the three subsets to each other and to murine subsets, we have concluded that CD14high and CD14highCD16high monocytes cluster with the murine Ly6chigh monocytes, whereas CD14dim monocytes cluster with Ly6clow monocytes [57]. That study contested the importance of using CD16 as a discrimination marker and may suggest using CD14 as an improved marker for describing human monocyte subsets [58]; however, a consensus of monocyte subset referencing is needed.

Regardless of nomenclature, human blood monocyte populations describe subset-specific effector functions. In response to LPS (lipopolysaccharide) CD14highCD16high, monocytes secrete TNF (tumour necrosis factor) α, IL (interleukin)-1β and IL-6, whereas CD14highCD16low cells preferentially produce CCL (CC chemokine ligand) 2, IL-10, IL-8, reactive oxygen species and high levels of IL-6. CD14dim monocytes, which express low levels of CCR2, but higher levels of CX3CR1 exhibit patrolling behaviour in vivo and resemble murine Ly6clow cells. Interestingly, stimulation of CD14dim monocytes with TLR (Toll-like receptor) 7 and TLR8 agonists selectively up-regulated cytokine expression, through specific signalling pathways involving MEK1 (mitogen-activated kinase/extracellular-signal-regulated kinase kinase 1) [57]. Overall these observations argue against describing monocyte subsets by phenotype or nomenclature rather than function, as each monocyte subset responds specifically to its environment in vivo and in vitro. The description of subset-specific functions opens a plethora of questions regarding roles in atherosclerosis; for example, do patrolling CD14dim monocytes mediate distinct functions in atherosclerosis? Recent imaging studies do show Ly6clow monocytes directly interacting with LDL in atherosclerotic plaques [59]. Most importantly; is there a subset-specific fate of blood monocytes in atherosclerosis? These questions will be explored as the area of monocyte heterogeneity expands.

Specific changes in blood monocyte numbers could be a useful clinical biomarker in cardiovascular disease. Monocytosis has been found to predict cardiovascular events in some studies [6063], but not in others [64,65]. Positive correlation of CD14dim cells with plasma cholesterol and triacylglycerol levels was shown in hypercholesterolaemic patients with coronary heart disease [66]. Circulating numbers of CD16high monocytes seem to correlate with body mass index, insulin resistance (diabetes) and intima-media thickness [67]. Weight loss after gastric bypass surgery in morbidly obese patients is associated with a significant reduction of CD16high cells [67]. Interestingly CD14highCD16high monocytes, but not total monocyte counts, predict cardiovascular events in patients with chronic kidney disease and end-stage renal disease on dialysis, a patient population at increased risk of atherosclerotic complications [68,69]. Additionally, in patients with symptomatic coronary artery disease compared with healthy controls, the percentage of CD16high monocytes was increased [70]. Treatment therapies with statins decreased the percentage of CD16high monocytes [71,72]. Finally, CD16high monocytosis after stent placement in patients with MI positively correlated with restenosis [73]. In these examples (not all of them have been described) it would appear CD16high monocytes, in particular CD14highCD16high, correlate with inflammatory disease. Whether this subset is indeed a blood biomarker that can be used for clinical diagnosis, will require careful analysis.

Monocytes as a drug target in atherosclerosis

Interest has focused on targeting Ly6chigh monocytes, which are regarded as the primary inflammatory monocytes; indeed, targeting functionally distinct monocyte subsets may provide attractive atherosclerotic targets without affecting immunological surveillance and innate immunity [41]. Leucocytes, including some monocytes, accumulate to sites of vascular inflammation through a well-described series of distinct adhesion steps, which can be targeted [74]. Initially, chemokines attract monocytes via chemokine receptors. Antagonizing the chemokine receptor CCR2 via nanoparticle-mediated transfer of short interfering RNA decreases Ly6chigh monocytes in peripheral tissues and reduces inflammation associated with atherosclerosis [75]. Selectins and their ligands can then mediate leucocyte capture and rolling, a crucial step to be blocked by antibodies [21]. PSGL-1 (P-selectin glycoprotein ligand-1) on leucocytes interacts with E- and P-selectins on activated endothelium and platelets (P-selectin) [76]. It is worth noting that Ly6chigh monocytes express PSGL-1 at higher levels compared with Ly6clow monocytes, providing a hypothesis for specific accumulation in the growing lesion [34,41]. A soluble version of the PSGL-1 ligand (soluble P-selctin), has also been reported to be raised in plasma from vascular disease patients, which may be involved with leucocyte activation, recruitment and thrombotic events [77,78]. This complicates the idea of blocking either PSGL-1 or the membrane-bound receptor P-selectin [76]. Pharmaceutically inhibiting Mac-1:CD40L, VLA-4:VCAM-1 (vascular cell adhesion molecule 1) or LFA-1:ICAM-1 (intercellular adhesion molecule 1) reduces monocyte adhesion and macrophage lesion number [79,80]. However, this concept of a single sequence of events leading to recruitment, firm adhesion and migration of monocytes is currently restricted to Ly6chigh monocytes (human CD14high), as the mechanisms for patrolling Ly6clow (human CD14dim) endothelial interactions have yet to be fully explored, but do involve LFA-1 integrin [42,57].


Along with monocytes, macrophages are phenotypically and morphologically heterogeneous, a contributing factor to the debate on how to define them [81]. Similar to monocyte heterogeneity, macrophage populations participate in many vital biological processes. From the perspective of monocyte fate mapping, one must ask whether specific subsets are restricted to give rise to specific macrophage populations with defined functions, or whether the subset differences vanish once monocytes accumulate in tissue. Functionally distinct human macrophages can be derived from monocytes in vitro. Briefly, classical macrophage activation, which involves culture of cells with GM-CSF, LPS or IFNγ (interferon γ) yields M1 named macrophages which are TNFα, IL-1β, IL-6, IL-12 and iNOS (inducible nitric oxide synthase) producers [52]. Alternative activation involves culture with IL-4, IL-10 or IL-13 and generates M2 named macrophages which are defined by their high expression of anti-inflammatory arginase 1, IL-10 and CD206 (see further description below) [52]. A major limitation to studying monocyte and macrophage biology is a lack of homology between species; human cells lack F4/80 and Ly6c antigens and, although M1/M2 definitions are convenient for in vitro studies, should be used with caution in the in vivo setting. In the steady state, most organs contain their own particular macrophages, many of which may not derive from blood monocytes [34,8183] and have specific gene signatures which separate them from each other and other myeloid populations [84]. Precisely how monocyte subsets contribute to macrophage and DC populations in the steady state or pathological settings is largely unknown. There are descriptions of Ly6chigh monocytes that extravasate into tissue, differentiate into TNF- and iNOS-producing DCs (Tip-DC), M1-like (classically activated) macrophages, and phagocytose pathogens, produce antibacterial products, and mediate inflammation and proteolysis [52]. Likewise, Ly6clow cells in response to inflammation can extravasate into tissue and have been shown to differentiate into M2-like macrophages, involved in wound repair, tissue remodelling and expression of chemokines [56,85].

In relation to atherosclerosis, whether both Ly6chigh and Ly6clow differentiated monocytes specifically mediate macrophage foam cell formation in response to lipids requires investigation. Simplistically, human macrophages in atherosclerotic lesions can be divided into two types, M1 and M2, that express (for example) MCP-1 (monocyte chemoattractant protein 1) and mannose receptor respectively [86]. As discussed above, M1 macrophages require IFNγ for initial activation and LPS, CpG or TNF and IL-1 for secondary activation [85,87,88]. M1 macrophages can release IL-6 and TNF and express high MHC II and CD80/86 to direct Th (T-helper) 1 activation [89]. In reported opposition to M1 macrophages, M2 macrophages are thought to be anti-inflammatory cells that are critical in the resolution of injury [52,87,90]. Furthermore M2 macrophages have been further subdivided into three groups: M2a, M2b and M2c [87]. M2a are non-cytotoxic cells activated by IL-4 and IL-13 and participate in tissue repair and extracellular matrix deposition [85]. Arginase expression is enhanced in M2a activated mouse (not human) macrophages, and in contrast with M1 activated macrophages, the expression of iNOS and production of nitric oxide is reduced [87]. M2B-cells are activated by LPS and release anti-inflammatory IL-10, as well as inflammatory cytokines [90,91]. IL-10, TGF (transforming growth factor) and glucocorticoids have been shown to activate M2c cells, which produce high levels of IL-10. Therefore M2b and M2c cells have been described as regulatory macrophages due to their high secretion of IL-10 [87,92]. Again it must be repeated that these classifications are based on in vitro activation and thus would most likely not apply to in vivo functional biology. Finally, the haptoglobin–haemoglobin scavenger receptor CD163 has also been associated with an anti-inflammatory macrophage. In separate studies, macrophages within intimal lesions showed strong positivity for CD163, whereas foamy plaque macrophages were CD163low [93]. Distinct populations of CD163+ macrophages have also been identified in haemorrhaged atherosclerotic plaques, which were DRlow and were unlike classical lipid necrotic core macrophages [94]. These types of macrophages are thought to suppress the impact of haemorrhage on atherosclerotic progression [95].

Overall, macrophages are a key feature of all stages of atherogenesis where they have a significant impact on lesion progression (Figure 1). Fundamentally, as part of their innate immune role, macrophages infiltrate developing lesions and respond to specific TLR ligands and can phagocytose modified LDL [9]. This can result in the secretion of cytokines, chemokines and toxic oxygen and nitrogen radicals that not only direct and amplify the local immune response, but may lead to localized tissue injury. Ultimately macrophages may cause plaque destabilization, rupture and thrombosis, highlighting their destructive role [96]. However, as macrophages are heterogeneous, they could develop functions that facilitate atherosclerotic tissue repair, remodelling and restoration of normal tissue homoeostasis [85]. As it has been difficult to distinguish the specific activation phenotypes at the single cell level in vivo, the extent of macrophage heterogeneity and polarization signals in atherosclerotic plaques still require careful investigation.

Macrophages are thought to originate from HSCs, but some macrophages develop in the embryo before the appearance of definitive HSCs, these yolk-sac-derived macrophages populate several tissues and depend on specific transcription factors for their development and survival [81,97,98]. Recently Schulz et al. [82], using fate mapping and conditional knockout approaches, supported the concept of two lineages of macrophages in mice: (i) derived from the bone marrow, phenotyping F4/80lowCD11bhi macrophages (and DCs); and (ii) derived from the yolk sac, phenotyping F4/80hi macrophages. The fate and functions of these yolk-sac tissue-resident macrophages in the adult, such as in atherosclerotic aortic tissue, remain to be investigated.

Dendritic cells

Antigen-presenting DCs are an essential part of the immune system because they provide sentinels of tissues for foreign antigens (and danger signals), including delivering of antigen to localized populations of lymphocytes and co-ordinating the immune response [99101]. Both the adaptive and the innate immune systems require DC participation [101]. DCs act as professional APCs (antigen-presenting cells) that recognize foreign antigens and display them bound to MHC molecules on their surface [100,102,103]. DCs isolated from the mouse aorta are just as effective at presenting foreign substances and stimulating killer cells as those in other immune organs, providing evidence that DCs in the arterial wall possess the same antigen-presenting properties [104,105]. In the walls of healthy arteries, DCs reside in the subendothelial space of the tunica intima and the tunica adventitia where they can mediate early recognition of danger signals [7,105107]. DCs also reside in carotid arteries, coronary arteries, aortic root, arch and descending aorta of experimental animals; these findings facilitated studies of the functional significance of DCs in atherosclerosis [103105,108111] (Figure 1).

DCs belong to the myeloid lineage of blood cells [34]. Depending on specific markers, phenotypic differences between macrophages and DCs are not simple [112]. CD11c has been a reported marker of (mouse) DCs, but it is also up-regulated by monocytes and macrophages [50,113]. A previous study has shown that surface markers for DCs, including co-stimulatory molecules (CD80, CD86), can be expressed by tissue macrophages [114]. However, in mice, DCs are currently typically characterized by CD11c expression, and inducible overexpression of MHC II, CD80, CD86 and CD40 [18,114]. Phenotypically DCs have a morphologically distinct dendritic shape. DCs may represent a unique and central subset of professional APCs capable of activating naive lymphocytes. In secondary lymphoid organs of mice, specific phenotypes of DCs have been defined and are reviewed elsewhere [18,115]

The involvement of DC apoptosis in atherosclerosis has been investigated using mice overexpressing an apoptosis inhibitor (Bcl-2) under the control of the CD11c cell-specific promoter [116]. This model achieved expansion of DCs and enhanced T-cell activation in vivo with a shift toward Th1 cells and increases in anti-oxLDL antibody sera (IgG2c type). Th1 activity is expected to promote atherosclerosis; however, this model described no increased atheroma when transplanted into either LDLR−/− or ApoE−/− mice. This paradoxical result was possibly due to a reduction in cholesterol with DC expansion. Thus DCs may have a role in the clearance of cholesterol.

Local proliferation of DCs has also been demonstrated in the aorta and in secondary lymphoid organs [117,118]. Along with proliferation, arterial DC numbers are also affected by egress. Monocyte-derived DCs can emigrate from the arterial wall and atherosclerotic plaques at the early stages of atherosclerosis; however, their emigration from developed atherosclerotic lesions is significantly impaired in mice with dyslipidaemia [119,120]. Defective egress of DCs from the aorta, altered trafficking toward lymph nodes and/or proliferation in situ could be another mechanism of excessive accumulation of DCs in atherosclerotic lesions.

Overall the functions of DCs in atherosclerosis is still not well characterized. However, data using mice lacking CX3CR1, CCL2 and CCR5 support the importance of DCs in atherosclerosis with correlative reduced atherosclerosis and DC numbers [49,121,122]. CX3CR1 deficiency also decreases the survival of Gr1low monocytes, which could correlate with decreased macrophage and DC aortic content and decreased foam cell formation [123]. Deletion of molecules involved in antigen presentation and DC migration also reduces plaque mass [124,125]. Finally, a recent study by Choi et al. [126] demonstrated a pivotal inhibitory DC population; namely CD103+ classical DCs which were associated with atherosclerosis protection. Along with elegant in vivo DC imaging studies, these studies point to the importance of DC populations in atherosclerosis [127].

In summary monocytes, macrophages and DCs have complex roles in mediating atherosclerotic plaque formation, progression and phenotype (Figure 1). Their emerging heterogeneity and subset-specific interaction with atherogenic lipids, cellular debris and other immune cells make them ideal targets for immunotherapy which may be utilized for novel anti-atherosclerotic drug development. Indeed, data inhibiting subset-specific monocyte recruitment looks promising in experimental atherosclerosis. Finally, modulating macrophage apoptosis or proliferation may also prove to be an effective therapeutic target [128].



Simplistically, the main function of B-cells is to secrete sera antibodies of various isotypes. B-1 cells are derived from committed precursors [129]. They are found predominantly in peritoneal and pleural cavities. B-1 cells participate in systemic immunity by producing IgM antibodies of the serum. B-2 cells are the main population produced in the bone marrow [129]. B-2 cells can produce IgG antibodies after activation by specific antigens and differentiate into plasma cells; other functions include antigen presentation, cytokine production and lymphoid tissue organization [129]. B-cells are found in atherosclerotic plaques of mice and humans, localized in the aortic adventitia [130132]. Adventitial B-cells have been described to form small lymphoid follicles that may contribute to atherosclerosis [133,134]. Both IgG and IgM antibodies against oxLDL have been described (see further description below) [135,136]. A significant decrease in the number of B-cells following splenectomy in mice or interference with B-1 cell IgM production accelerated atherosclerotic lesion development, which suggests a protective role for B-cells in plaque formation [135,137,138]. However, descriptions of pro-atherogenic roles of B-cells have been discussed previously [139,140].

IgM and IgG antibodies are present within atherosclerotic plaques [141]. Specific auto-antibodies against oxLDL have been found in the circulation of mice and humans [142,143], whereas IgM antibodies have been shown to protect against atherosclerosis in mice and rabbits [144]. Injection of IgG1 antibodies against oxLDL epitopes has reduced atherosclerosis in experimental mouse models [145]. Indeed, bone marrow chimaeras of B-cell-deficient bone marrow into LDR−/− mice increased atherosclerosis [137]. In addition, splenectomy aggravates atherosclerosis in ApoE−/− mice and transfer of spleen B-cells into splenectomized mice reduces atherosclerosis [135]. However, splenectomy does not only affect B-cell biology (see the Monocytes section above). The situation is maybe more complex in humans, with studies showing a positive or negative correlation or no correlation between auto anti-LDL titres and atherosclerosis or its clinical correlates [146149]. Regardless, solid phase IgG and IgM seem to activate macrophage populations in vitro [150]. Much more work is needed to elucidate the role and deposition of automimmune antibodies to oxidized and modified LDL in atherosclerosis.

B-cells also exhibit functions that directly modulate immune responses in atherosclerotic plaques. B-cells can phagocytose and internalize their antigen [151,152]. Indeed, compared with matched wild-type mice, expression of CD80 and CD86 on B-cells was higher in atherosclerotic plaques of young ApoE−/− mice [153]. Moreover, a positive correlation between the level of CD80+ B-cells and the severity of carotid atherosclerosis was found in humans [154]. Furthermore, B-cells can modulate immune responses through immunomodulatory cytokine production [155].


Initial studies of the role of T-cells in the development of atherosclerosis have focused on the role of Th1 cells, which produce IFNγ, and that of Th2 cells, which can produce IL-4 [156]. Importantly, IFNγ is present in human lesions which can induce higher expression of MHC II, enhanced protease and chemokine secretion, up-regulation of adhesion molecules, induction of pro-inflammatory cytokines, and enhanced activation of macrophages and endothelial cells [157]. Mice deficient in IFNγ or its receptor have a lower lesion burden, and mice that receive IFNγ had an increased lesional size [158160]. The roles of other T-cell subsets, including those of Treg (T-regulatory) cells have been addressed (see below) [161]. Th2 and Th17 cells have also been discussed to influence atherosclerotic changes [162]. Clonal expansion of memory effector and CD8+ T-cells has been demonstrated in lesions from humans and ApoE−/− mice [163]. Indeed lymphocyte counts seem to be inversely correlated with coronary heart disease and its complications [164].

Treg cells

Several studies have demonstrated a protective effect of various subsets of Treg cells in models of atherosclerosis. Foxp3+ expressed on Treg cells has been found in the plaques of mice and humans [165]. Transfer of Foxp3+ T-cells has also been shown to be protective against atherosclerosis [166]. IL-10 or TGFβ (cytokines partly responsible for the protective functions of Treg cells), have been shown to be anti-atherogenic. Genetic depletion or blockade of IL-10 or TGFβ with neutralizing antibodies accelerated lesion development and aggravated vascular inflammation. Reportedly pathogenic Th1 and Th2 responses are exacerbated to the disadvantage of Treg cell responses [167]. Moreover, T-cell restricted inactivation of the TGFβ signalling pathway specifically led to uncontrolled T-cell proliferation, inflammation, autoimmune disease and increased atherosclerosis [167]. Treg cells are detected in much lower amounts in atherosclerotic plaques than in other chronically inflamed tissues, such as in the skin of patients with psoriasis, where Treg cells can represent up to approximately 25% of T-cells [168]. These findings indicate an impairment of local tolerance against antigens in atherosclerotic plaques. Moreover, ApoE−/− mice exhibit a lower number of Treg cells in the spleen compared with matched control mice [169]. Although Treg cells can inhibit T-cell activation and thereby modulate atherosclerosis, they might also provide substantial protection against atherosclerosis by interacting with APCs. Treg cells inhibited foam-cell formation in vitro and promoted macrophage differentiation towards M2-like cells, dependent on TGFβ and IL-10 [170].

NKT (natural killer T)-cells

Recent research into atherogenesis has further elucidated the role of NKT-cells [171]. NKT cells are an intermediate of the innate and adaptive immune systems who share similarities with both NK (natural killer) cells (see below) and T-cells [171]. NKT cells are modulatory to CD1d, a molecule with antigen-presenting properties [171]. NKT cells express the NK1.1 marker protein in mice and CD161 and CD56 in humans [171]. They also seem to express CD4 [171]. NKT cells have been found in atherosclerotic plaques, albeit in small numbers [171]. In plaques, NKT cells can be activated by signals, such as oxLDL, and can release cytokines of both Th1 and Th2 [171]. This is supported by studies in which NKT−/− mice displayed fewer atherosclerotic lesions [171].

NK cells

NK cells are a subset of cytotoxic lymphocytes found in human atherosclerotic lesions [171]. Interestingly, defective cytolysis by NK cells in perforin−/− mice does not affect atherosclerosis in a double knockout model using LDLR−/− [172]. Moreover, LDLR−/−Lystbeige mutant mice, that have defective protein release from cytoplasmic granules, display reduced atherosclerosis. In contrast, combined LDLR−/−Lystbeige with Rag1−/− mutant mice demonstrate increased atherogenesis and lipid sera levels [172]. NK cells are still present in these models, therefore other NK cell functions may be involved. In another bone marrow chimaera model, LDLR−/− mice using cells from transgenic mice expressing Ly49A under the control of the granzyme-A promoter, the absence of fully functional NK cells resulted in 70% reduction of atherosclerotic lesion formation [173]. However, effects from contaminating cellular populations in these experimental models cannot be excluded.

In summary, lymphocytes make up a heterogeneous population of effector cells that are involved in atherosclerosis (Figure 1). They can be recruited to evolving plaques and have cytotoxic, regulatory and inflammatory functions. Importantly, B-cells can secrete autoantibodies to modified LDLs, which may have roles in promoting or inhibiting atherosclerosis.


Neutrophils have been detected in aortic lesions of primates, humans and mice and have been correlated with cardiovascular incidence and prognosis [14,22,174]. However, it is worth noting that mice have a myeloid-biased immune system compared with humans, denoting another significant difference between mice and human immune physiology, which should always be taken into consideration [175]. Neutrophils represent approximately 2% of plaque leucocytes and accumulate in regions with high monocyte density (in mice) [22]. Neutrophils use the CCL5 receptors (CCR1 and CCR5) to enter atherosclerotic plaques, which correlates significantly with plaque size [24]. Disrupting CXCR4–CXCL12 interaction increases peripheral neutrophil counts and modulates lesion formation, whereas neutrophil depletion predominantly reduces early plaque size [24,174]. A disturbed lipid balance facilitates neutrophil recruitment and infiltration to lesions via TLR signalling and pro-inflammatory gene expression, upon specific activating signals [176]. Hyperlipidaemia can trigger neutrophilia by stimulating granulopoiesis and bone marrow egress, involving CXCL1 plasma overexpression [24]. However, direct roles for neutrophil lipid scavenging and metabolism have been limited; there are reports of HDLs inhibiting neutrophil activation and adhesion [177], although the direct role of neutrophil activation by proatherogenic lipid species requires further investigation.

Interestingly, neutrophil depletion specifically decreases Ly6chigh monocyte recruitment [178]. Neutrophils can rapidly become apoptotic in the intima, attracting macrophages for scavenging [174]. Signals released by damaged neutrophils may help resolve inflammation, where RNA may enhance TLR3 signaling to delay inflammatory lesion development, through increased atherosclerosis in TLR3−/−ApoE−/− mice [179]. Furthermore, neutrophils activated through TLRs, Fc receptors or cytokine receptors can release nuclear content that forms a scaffold containing antimicrobial proteins, named NETs (neutrophil extracellular traps) [180]. NETs have been identified in association with human and murine atherosclerotic lesions [181]. Finally, neutrophils may link coagulation and inflammation; evidence shows that neutrophil serine proteases with externalized nucleosomes promote intravascular thrombus growth in vivo via tissue factor- and factor XII-dependent coagulation, perhaps contributing to plaque thrombosis [182].

In summary, neutrophils respond rapidly to their microenvironment, releasing cytotoxic reagents, including NET formation. Their role in atherosclerosis is becoming apparent, demonstrating neutrophil-specific responses in inflammatory plaque phenotypes (Figure 1). More research is needed to define further neutrophil interactions with other known pro-atherosclerotic agents, including responses to modified lipid species.


Mast cells

Mast cells are derived from bone marrow cells and circulate in the blood as mast cell precursors to be recruited to particular tissues and organs, such as lung and skin, where they mature into tissue mast cells [183]. In the 1950s the first paper to describe a role for mast cells in cardiovascular diseases was published [184]. Studies showing the colocalization of mast cells expressing the pro-angiogenic factor bFGF (basic fibroblast growth factor) with intraplaque neovessels emerged [185]. Human coronary artery specimen analysis reported subendothelial mast cells in proximity to microthrombi [186]. Interestingly, mast cell degranulation has been shown to mediate in vivo uptake of LDL by macrophages in the peritoneal cavity of rats [187]. Indeed, mast cell systemic activation aggravates atherosclerotic lesion formation in the brachiocephalic artery in ApoE−/− mice [188].

Mast cells operate in the first line of host defence against pathogens such as parasites and bacteria: various pathogens have been detected in human atherosclerotic lesions, such as Chlamydia pneumoniae and Aggregatibacter actinomycetemcomitans that can activate mast cells in vitro [189]. Moreover, levels of IgE, a significant mast cell activator, are elevated in patients with unstable angina pectoris or dyslipidaemia and this may have a direct link to atherosclerosis by mediating arterial cell apoptosis, although the pathogenesis is yet to be fully elucidated [190,191].

Interestingly, mast-cell-derived chymase degrades cholesteryl ester transfer protein activity, indicating a novel role for mast cell products in lipid metabolism [192]. Specifically, PLTP (phospholipid transfer protein) transfers phospholipids from liposome donor particles to acceptor HDL particles and facilitates pre-HDL formation. Chymase degrades PLTP fragments, thereby reducing PLTP pre-HDL activity and inhibiting cholesterol efflux from macrophages [193].


Anucleated cells of 1–2 μm have been described since the mid-19th century for their primary physiological role in sensing damaged vessel endothelium and accumulating at sites of injury to initiate blood clotting [194]. The mechanism of thrombus formation can be divided into four steps: platelet tethering, activation and firm adhesion, aggregation and platelet recruitment, and thrombus establishment [195]. The recognition that platelets modulate immune and inflammatory responses is now well established [196]. It is not known whether platelets exert their possible atherosclerotic inflammatory functions by continuously interacting with the endothelium and leucocytes or, over time, cause asymptomatic thrombi promoting leucocyte recruitment. Platelets are well equipped to facilitate leucocyte recruitment to sites of injury (and inflammation), and platelet expression of complement receptors and interactions with bacterial pathogens have been described previously [197,198]. Platelets store and release antibacterial proteins and, upon activation, produce significant amounts of cytokines and chemokines which are released from the α-granules [199]. Proteomic studies indicate that thrombin-stimulated platelets release more than 300 distinct proteins [200]. Several of these secreted proteins have been identified in atherosclerotic lesions [194], albeit not necessarily correlated with platelet expression.

In summary, mast cells operate as the first line of host defense against pathogens. These pathogens are found in atherosclerotic plaques where mast cell functions are poorly described (Figure 1). Mast cell activators including IgE are known to modulate experimental atherosclerosis; however, their role in human atherosclerosis still remains to be elucidated. Platelet roles in atherosclerosis have been described for a number of years; ‘atherothrombosis’ is an established theme in atherosclerotic research (Figure 1). Emerging platelet chemokine, cytokine and microparticle formation will undoubtedly mediate novel interactions in atherosclerosis biology and require further investigation.


Inflammatory cells have largely been considered detrimental in the pathogenesis of atherosclerosis and formation of the vulnerable plaque (summary of plaque milieu shown in Figure 1). The argument for the role of monocyte-derived macrophages is considerable. They express scavenger receptors, modifying and oxidizing lipoproteins and can accumulate cholesteryl esters. Macrophages express high levels of matrix metalloproteinases, pro-inflammatory cytokines and chemoattractants and can produce tissue factors. This data originated from studies of mice deficient in functional macrophages crossed with atherosclerosis-prone mice, resulting in significantly inhibited atherosclerosis. This implies that differentiation of monocytes into macrophages is critical for atherogenesis. However, as the knowledge of immunology has advanced, it has become clear that atherosclerosis involves activation of both cellular innate and humoral immunity, which balances the inflammatory response to lipid deposition in the arterial wall. Moreover, the publication on the heterogeneous phenotypes of monocytes and macrophages with independent effector functions have now complicated the concept that macrophages are not just ‘bad’ in the pathogenesis of atherosclerosis. Immunological research will help to identify these novel subset-specific effector functions in atherosclerosis. This in turn will help develop pharmaceutical targets for inhibiting vulnerable atherosclerosis and its pathological end points. Indeed we now have the tools, at least in experimental models, to alter the course of atherosclerosis through immune modulation. Despite these exciting advances, the best strategies for the treatment of established atherosclerosis remain a challenge and many immunological therapeutic interventions have not been tested in humans; this must be addressed in order to bring immune-related modulation of atherosclerosis into clinical reality.


K.J.W. is currently funded by a British Heart Foundation fellowship.

Abbreviations: ABC, ATP-binding cassette; APC, antigen-presenting cell; ApoE, apolipoprotein E; CCL, CC chemokine ligand; CCR, CC chemokine receptor; DC, dendritic cell; GM-CSF, granulocyte-macrophage colony-stimulating factor; HDL, high-density lipoprotein; HSC, haemopoietic stem cell; HSPC, haemopoietic stem cell progenitor; IFNγ, interferon γ; IL, interleukin; iNOS, inducible nitric oxide synthase; LDL, low-density lipoprotein; LDLR, LDL receptor; LPS, lipopolysaccharide; MI, myocardial infarction; NET, neutrophil extracellular trap; NK, natural killer; NKT, natural killer T; oxLDL, oxidized LDL; PLTP, phospholipid transfer protein; PSGL-1, P-selectin glycoprotein ligand-1; TGF, transforming growth factor; Th, T-helper; TLR, Toll-like receptor; TNF, tumour necrosis factor; Treg, T-regulatory


View Abstract