Thematic Review Series: The Immune System and Atherogenesis. Cytokines affecting endothelial and smooth muscle cells in vascular disease

In experimental models of atherosclerosis, the initiation of a high-cholesterol diet rapidly induces expression of specific adhesion molecules at sites of predilection for lesion formation (). Cytokines, such as TNF-α, are potent stimulants of adhesion molecule expression (, , ), and the absence of its signaling receptor can inhibit their expression and leukocyte infiltration of the vessel wall (). Similarly, targeted deletion of a natural inhibitor of interleukin (IL)-1, the IL-1 receptor antagonist (IL-1Ra), effectively increases local concentrations of IL-1 in apolipoprotein E (apoE)^−/− mice, increases mRNA expression of the adhesion molecules vascular cell adhesion molecule (VCAM)-1 and intercellular adhesion molecule (ICAM)-1 in the aorta, enhances mRNA levels of the leukocyte chemokine monocyte chemoattractant protein (MCP)-1, and promotes a 1.9-fold increase in monocyte accumulation (). Mice lacking the IL-1 receptor I have also been shown to be unable to induce E-selectin expression when injected with IL-1 (). Thus, multiple cytokines are sufficient to induce specific adhesion molecule expression.

Transplant atherosclerosis remains the leading cause of graft failure. It is characterized by involvement of the entire wall of the artery, and results in a concentric lesion that often involves long segments of affected arteries (). Although the exact immunologic mechanisms responsible for chronic vascular rejection are not known, evidence is consistent with involvement of alloreactive T lymphocytes and antibodies (). Cytokine regulation of adhesion molecule expression on graft endothelial cells can contribute to lymphocyte recruitment to the graft. Analysis of rejecting murine heterotopic cardiac allografts has shown that endothelial expression of VCAM-1 can be abrogated and ICAM-1 expression reduced by treatment with either a soluble TNF-α receptor or IL-4 antagonists (monoclonal antibody and soluble IL-4 receptor) (). Although either treatment reduced but did not eliminate leukocyte infiltration, the allografts were rejected at the same rate, potentially due to continued expression of multiple cytokines within the graft. So, although TNF-α and IL-4 have clear roles in regulating adhesion molecule expression in allografts, elimination of VCAM-1 expression is not sufficient, nor is VCAM-1 essential, for graft rejection (). The absence of the anti-inflammatory cytokine IL-10 has also been evaluated in mice receiving cardiac allografts. Although targeted deletion of IL-10 led to enhanced leukocyte recruitment and graft rejection, molecular changes in endothelial phenotype were not evaluated ().

For leukocytes to deposit within the intima, they must undergo a sequence of interactions with the endothelium. Initially, this includes tethering and rolling along the endothelial surface until chemokine stimuli cause the rolling leukocytes to arrest and adhere to the endothelium, followed by migration to endothelial junctions and transendothelial cell migration. Cytokines and chemokines can also significantly enhance the function of endothelial cell adhesion molecules at all stages of transendothelial cell migration. Analysis of mice with targeted deletion of macrophage migration inhibitory factor (MIF), using intravital microscopy to examine transendothelial migration in the inflamed cremaster muscle, revealed a significant reduction in P-selectin-dependent leukocyte rolling and adhesion, and reduced entry of leukocytes into the site of inflammation (). The effect of administration of chemokines has also been examined in lesion-prone apoE^−/− mice. Ex vivo perfusion of apoE^−/− carotid arteries has shown a keratinocyte chemokine (KC)/growth-related oncogene (GRO)-α-dependent monocyte arrest, but no effect of inhibition of MCP-1 or its receptor, CC chemokine receptor (CCR)2 (). The KC/GRO-α-dependent monocyte arrest could be inhibited by blockade of either integrin α4β1 or VCAM-1, and the authors have proposed that the arrest is due to chemokine regulation of integrin avidity and adhesiveness, because it is dependent upon CXCR2 signaling () and based upon in vitro chemokine modulation of integrin avidity (). Perfusion of the carotid artery with KC was also able to further enhance monocyte arrest, demonstrating the ability of locally released KC to increase the extent of monocyte arrest (), an effector mechanism also employed by TNF-α (). IL-15 appears to act at a later step by inducing endothelial hyaluronan expression that promotes a CD44-mediated pathway, which enhances transendothelial cell migration (). Thus, cytokines and chemokines can modulate adhesion molecule properties to further enhance leukocyte recruitment.

Endothelial cell apoptosis has the potential to expose the underlying basement membrane, which can lead to thrombosis and further promotion of the inflammatory response. Enhancement of endothelial cell survival and function would therefore be hypothesized to be protective, especially in transplants in which the time between donor organ removal and transfer to the recipient can lead to significant loss of endothelial cells. This possibility is supported by intradermal injection of the anti-inflammatory cytokine IL-11, which protects human microvascular endothelium in severe combined immunodeficient (SCID) mice bearing human skin grafts (). Although IL-11 had no effect on T cell infiltration, T cell activation markers and effector molecules, or endothelial ICAM-1 expression, it was able to significantly delay the time course of graft microvessel loss because of its ability to upregulate survivin, a member of the inhibitors of apoptosis family (). Although IL-11 did not completely prevent allograft rejection, this was the first demonstration in vivo of cytokine regulation of survivin and protection from T cell-mediated endothelial cell injury.

Although the endothelial cell monolayer primarily serves a protective function in normal vessels, allografts place the immunologically competent endothelial cell in contact with circulating immune cells. Endothelial cells express lymphocyte costimulatory molecules, and when induced by interferon (IFN)-γ to express major histocompatibility complex (MHC) class II, they can induce proliferation of allogeneic T cells (). Infusion of proinflammatory IL-18, a member of the IL-1 cytokine family, into apoE^−/− mice induced a 4-fold increase in cells expressing MHC class II, including endothelial cells, and an associated increase in aortic T cells (). MHC class II expression is enhanced by IFN-γ, and the IL-18-mediated increase in MHC class II was not seen in male IFN-γ-null/apoE^−/− mice infused with IL-18, implying a male-specific requirement for IFN-γ. However, a T cell-independent role for IFN-γ was further supported by infusion of IL-18 into SCID/apoE^−/− mice that lacked T cells (). MHC class II expression was increased 3-fold following IL-18 infusion that accompanied a 2-fold induction of IFN-γ produced by macrophages, NK, and vascular cells (). The chemokine MCP-1 can promote another proinflammatory signaling pathway, the CD40/CD40 ligand (CD40L)-coupled signaling that is required for T cell priming and other immune regulation (). Administration of an MCP-1 antagonist to apoE^−/− mice with established lesions decreased CD40 and CD40L expression, including endothelial expression, and reduced T cell infiltration (). However, it is unclear whether this is a direct or indirect effect of MCP-1 blockade.

Although several cytokine and chemokine receptors are expressed on endothelial cells (see Table 1), and some of these have been shown to promote proliferation or migration of endothelial cells in vitro, only TNF-α and KC/GRO-α have been shown to alter large endothelial cell repair in vivo. Administration of a soluble TNF-α receptor following balloon injury of the rat carotid artery decreased intimal lesion formation and accelerated endothelial cell regrowth by 125–140% 1 and 2 weeks after injury by Evan's blue dye labeling of vessel not covered by endothelium (). In contrast, blockade of KC for 3 weeks after wire injury of apoE^−/− carotid arteries with a monoclonal antibody to KC increased neointimal plaque area and decreased endothelial cell regrowth 3-fold, as evaluated by Evan's blue dye labeling and by CD31 and VCAM-1 endothelial cell staining (). These data are compatible with the idea that KC normally has a protective role in accelerating endothelial recovery, whereas TNF-α can inhibit regrowth, although it is unclear whether the effects are direct or indirect.

The roles of particular angiogenic cytokines have not been tested in models of atherosclerosis; however, inhibition of plaque neovascularization has been demonstrated to be sufficient to decrease macrophage accumulation and plaque progression in advanced lesions of atherosclerosis (). IL-2, IL-8, IL-15, and IL-18 have all been shown to induce angiogenesis in vivo (, , , ). Therefore, it will be important to determine the extent to which these cytokines promote angiogenesis within the context of atherosclerotic lesions.

A product of endothelial cells that is a potent anti-inflammatory agent is nitric oxide, and therefore induction or suppression of nitric oxide by cytokines has the potential to enhance or inhibit the inflammatory response. The receptor for IL-10, an anti-inflammatory cytokine, has been shown to be upregulated under proinflammatory conditions in vivo, and subsequent infusion of IL-10 induces nitric-oxide synthase-3, which attenuates expression of proinflammatory IL-12 (). Further, the absence of IL-10 is sufficient to impair endothelial cell-dependent vasorelaxation and is associated with increased superoxide formation, and endothelial impairment is reversed by treatment with superoxide dismutase (). In contrast, blockade of the proinflammatory cytokine IFN-γ in human allografts is sufficient to prevent endothelial cell dysfunction and loss of endothelial nitric oxide expression (). Thus, cytokine stimulation of endothelial cells can both positively and negatively modulate expression of endothelial gene products that control vascular tone and the ability of the vessel to respond to vasodilatory signals.

The accumulation of SMCs within intimal lesions is the combined result of their migration from the media into the intima and their proliferation (). Although proliferation in vivo can be evaluated with antibodies to markers such as proliferating cell nuclear antigen, it is much more difficult to evaluate a specific contribution to migration in slowly progressing diseases such as atherosclerosis. The acute injury model, in which balloon or wire injury is used to denude a normal vessel, has been useful for studying migration, because the intimal lesions that form following injury consist primarily of SMCs, and kinetics of proliferation and migration have been characterized (). Using this approach, the proinflammatory cytokines MCP-1, stromal cell-derived factor (SDF)1α, and CCL11 (eotaxin) have been implicated in the promotion of SMC migration and proliferation.

Femoral artery injury in mice lacking the MCP-1 receptor CCR2 results in smaller intimal SMC-rich lesions and less SMC proliferation (). The same injury in MCP-1-null mice also resulted in a reduction in intimal lesion size, as compared with MCP-1^+/+ littermates, but there was no decrease in the SMC proliferative index (). These data have been interpreted to suggest that MCP-1 may have a more important role in mediating SMC migration, whereas CCR2 may regulate cell proliferation (). Differences have also been noted between vascular repair in normolipidemic versus hyperlipidemic models, raising the possibility of functional alterations in the MCP-1/CCR2 axis with differing levels of hypercholesterolemia (). These effects may also be explained by the recent description of a second MCP-1 receptor (). Blockade of SDF-1α in apoE^−/− mice with a blocking antibody also strongly inhibits the accumulation of SMC in the neointima after vascular injury without any significant change in neointimal macrophage content, an effect mediated to a large degree by recruitment of hematopoietic SMC precursors (). The role of CCL11 (eotaxin) on SMC accumulation following injury has not been directly addressed in vivo. However, CCL11 and its receptor, CCR3, are not expressed in normal artery but are abundant in medial and neointimal SMCs after injury (). In vitro, CCL11 promotes SMC migration (), so it will be interesting to determine whether CCR3 or CCL11 antagonists can inhibit SMC accumulation in injury models. Thus, at least the proinflammatory cytokines MCP-1 and SDF-1α promote SMC intimal accumulation following arterial injury.

Two other models of vascular injury that have been used to examine cytokine effects on lesion development involve either the placement of a silastic cuff around the femoral artery or ligation of the carotid artery. These models are less well characterized and both have a more significant involvement of inflammatory cells in the injury response. Analysis of mice lacking the natural IL-1 inhibitor IL-1Ra that had significantly increased IL-1 concentrations in the cuff model, showed a 2.5-fold increase in intimal thickness comprised primarily of SMCs and a 110% increase in intimal proliferation (). Carotid artery ligation was used to investigate the role of TNF-α, and TNF-α^−/− mice showed reduced SMC accumulation (). Therefore, the proinflammatory cytokines IL-1 and TNF-α both appear to promote SMC accumulation after injury, although the relative contribution of migration versus proliferation is not clear.

Analyses of models of murine atherosclerosis have provided further support for the involvement of the proinflammatory cytokines IL-1β, TNF-α, IFN-γ, and MIF in SMC accumulation. However, the relative effects are not always the same as those seen following injury in the models discussed above. Advanced lesions in apoE-deficient mice (32 weeks) with reduced expression of IL-1Ra (IL-1Ra^+/− mice) showed a small but significant reduction (15%) in α-SMC actin-positive area (), in contrast to enhanced SMC proliferation and accumulation following wire injury in mice on a C57BL/6 background (). The reduction in SMC in lesions of IL-1Ra^+/−/APO^−/− mice was likely due to the different cellular composition in these advanced lesions in which there was an 86% increase in lesion macrophages and expression of other inflammatory cytokines (). Expression of a noncleavable mutant of TNF-α that effectively lowers TNF-α levels because it remains cell-associated seems to alter SMC phenotype in the medial SMCs adjacent to the lesion by reducing α-actin expression, with no marked effect on SMC proliferation (). Thus, the accumulation of SMCs in lesions of atherosclerosis does not appear to be strongly dependent upon the action of IL-1β and TNF-α.

In IFN-γ receptor-null mice, atherosclerotic lesions appear strikingly less cellular, with increased accumulation of extracellular collagen, suggesting that signaling from IFN-γ positively contributes to SMC proliferation (). This possibility is further supported by studies of LDLR-null mice crossed with IFN-γ-deficient mice (). Finally, in a model of transplant atherosclerosis that utilized SCID/beige mice as hosts, deficient in T and B lymphocytes and natural killer cell function, IFN-γ has been shown to induce SMC proliferation, but synergistically with the action of platelet-derived growth factor (). The proinflammatory cytokine MIF appears to be one of the more potent cytokines in promoting SMC proliferation. MIF deletion leads to an ∼80% reduction of SMC proliferation in atherosclerotic lesions of LDLR-null mice (). CX3C chemokine ligand 1 (fractalkine) may also play a role in SMC accumulation, because it promotes SMC migration in vitro and is found in intimal SMCs in human atherosclerotic lesions (). A positive correlation was observed between SMCs expressing the fractalkine receptor CX3CR1 and fractalkine-expressing cells in human lesions (), and targeted deletion of fractalkine on either the LDLR^−/− or apoE^−/− background resulted in decreased SMC accumulation (). However, because both lesion size and macrophage accumulation were also reduced, it is unclear whether the effect on SMCs is direct or indirect.

Negative regulators of SMC accumulation have also been identified in vivo. Among cytokines with anti-inflammatory properties, as defined by their actions on lymphocytes and monocytes, IL-10 is the best characterized. IL-10 directly inhibits mitogen-induced SMC proliferation in vitro (), and its role in vivo has recently been evaluated following injury. In a rabbit model of balloon injury, IL-10 infusion reduced SMC proliferation by 81% (), and a similar reduction was seen with IL-10 administration following balloon injury of the rat carotid artery (). IL-10 also inhibited intimal and medial SMC accumulation in a murine heart transplant model (). Therefore, IL-10 shows inhibitory activity for SMCs in vivo under conditions in which SMCs are the primary intimal cell, and its action may counteract the proatherogenic activity of other cytokines that accumulate in lesions. Lesions in mice overexpressing IL-10 also appear less advanced than those in mice transplanted with wild-type bone marrow, with larger necrotic core area and reduced accumulation of SMCs and ECM, suggesting an antiproliferative SMC activity of IL-10 in murine atherosclerosis as well (). Surprisingly, the proinflammatory cytokine, IL-18, also inhibits SMC accumulation in apoE-deficient mice, and its absence leads to a 2- to 3-fold increase in the proportion of α-SM-actin-positive cells (, ). Thus, although the majority of cytokines and chemokines promote SMC accumulation either directly or indirectly, two cytokines, IL-10 and IL-18, appear to negatively regulate SMC accumulation in lesions.

The pathogenesis of atherosclerosis and restenosis following angioplasty or stent placement includes the abnormal production of ECM proteins by “synthetic” SMCs as well as remodeling of existing ECM components. Disruption and/or modification of SMC interactions with matrix components can significantly influence their responses to locally expressed cytokines and growth factors (), and will further alter the structural properties of the vessel. Several in vivo studies have demonstrated direct or indirect roles for cytokines, including transforming growth factor (TGF)-β, MCP-1, MIF, IL-18, and IFN-γ, on ECM remodeling.

TGF-β has a well-established profibrotic activity that has been confirmed in vivo using different approaches. Injection of neutralizing anti-TGF-β₁ antibody, or a soluble TGF-β receptor that acts as an antagonist, into apoE^−/− mice has demonstrated a significant reduction in lesion collagen content (∼50%), with no apparent effect on SMC accumulation (, ). However, transgenic expression of a dominant negative TGF-β receptor II in T cells led to the development of thicker and more advanced lesions compared with apoE^−/− with intact TGF-β signaling, although disrupted TGF-β signaling was associated with reduced collagen staining (). Reduced accumulation of collagen and other ECM components has also been observed in lesions of MIF-deficient mice, suggesting that MIF, similar to TGF-β, positively contributes to matrix deposition (). However, reduced collagen deposition in MIF-null mice has been partially attributed to their increased expression of matrix proteolytic enzymes, such as cathepsin S and l (). Other in vivo evidence supports a profibrotic effect of both IL-6 and IL-10 (, ).

In contrast, blockade of MCP-1 signaling in apoE^−/− mice, through the expression of an N-terminal-deleted mutant of MCP-1, leads to the development of more stable atherosclerotic plaques with increased SMC content and collagen deposition (). Thus, MCP-1 may be a central mediator in the progression and destabilization of established atherosclerotic plaques, but a direct versus indirect effect has not been determined (). Destruction of elastin lamina within the media has also been observed in IL-1Ra-deficient mice, suggesting that IL-1β signaling may promote the progression of unstable atherosclerotic plaques (). Moreover, at least two in vivo studies indicate that a loss of IFN-γ signaling leads to substantial changes in lesion composition, supporting the notion that IFN-γ antagonists may serve to stabilize atherosclerotic plaques (, ). Similarly, overexpression of IL-18 decreases intimal collagen content in apoE-deficient mice (), whereas overexpression of its endogenous binding protein increases collagen content ().

Thus, TGF-β and MIF promote collagen deposition, whereas MCP-1, IL-1, IL-18, and IFN-γ enhance ECM remodeling. Collagen and elastin degradation, mediated by specific proteolytic enzymes, may facilitate the response of SMCs to the proliferative signaling of different cytokines, and consequently the enlargement of atherosclerotic lesions. Uncontrolled accumulation of SMCs, expressing proinflammatory cytokines, may also perpetuate the local inflammatory response in the arterial wall, leading to the progression and destabilization of advanced atherosclerotic plaques.

Two cytokines, IL-10 and TGF-β, are notable in their ability to significantly inhibit the nuclear factor κB (NF- κB) proinflammatory signaling pathway. NF-κB is a pleiotropic transcription factor that has been linked to atherosclerosis () and has the ability to modulate a wide array of SMC functions (). Activation of NF-κB in neointimal SMCs lining the vessel wall is observed after balloon injury of rat carotid arteries, and this response is significantly inhibited in mice and rats treated with IL-10 (, ). In vivo administration of blocking anti-TGF-β antibody for 9 weeks in atherosclerotic mice is sufficient to induce expression of activated NF-κB in the myocardium (). Thus, both IL-10 and TGF-β are potent inhibitors of the pleotropic NF-κB signaling pathway.

Although cell adhesion molecules expressed on endothelial cells directly mediate leukocyte emigration into the vessel wall, increased expression of ICAM-1, VCAM-1, and P-selectin has also been observed in SMCs after vascular injury (, , ). More significantly, targeted deletion of the TNF-α receptor 1 decreased VCAM-1 and ICAM-1 expression by 50–60% in murine vein graft SMCs and decreased graft neointimal formation (). Cytokines can also decrease adhesion molecule expression in vivo, as shown by the ability of adenoviral expression of IL-10 to inhibit ICAM-1 induction following rat venous injury (). In contrast, in vivo administration of anti-TGF-β antibody for 9 weeks or disruption of TGF-β signaling in T cells did not alter VCAM-1 expression in atherosclerotic lesions at the aortic sinus or in SMCs (, ). Thus, TGF-β does not appear to alter adhesion molecule expression in SMCs during atherogenesis. Because blockade of cytokines such as TNF-α inhibits multiple effects, including the levels of NF-κB and other cytokines, it is not possible to evaluate their specific role in decreasing adhesion molecule expression. However, it is tempting to speculate that SMC adhesion molecules may contribute to retention of inflammatory cells within the vessel wall, and consequently further promote the inflammatory response within lesions.

The importance of the adaptive immune response in atherosclerosis remains controversial, with several studies demonstrating that immunization with specific antigens can protect against disease and others showing that disease-related antigens may be responsible for increased atherosclerosis (). However, atherosclerosis is dramatically enhanced in apoE^−/− mice with loss of the potent immune inhibitor TGF-β resulting from transgenic expression of a dominant negative TGF-β receptor II in T cells (). Inhibition of TGF-β led to an increase in the number of cells, including SMCs, expressing I-A^b region of the MHC in lesions of atherosclerosis (). In contrast, in post-transplant graft atherosclerosis, administration of immune-promoting cytokine IFN-γ restored the weak basal expression of MHC class I antigen by graft SMCs (). T cells recognize SMC MHC antigens, and their dependence on IFN-γ for basal expression indicates that IFN-γ has a physiological role in noninflammatory states. Another proinflammatory cytokine, IL-18, has also been shown to increase the mean number of SMCs expressing MHC II in atherosclerotic lesions of apoE-deficient mice, and appears to act upstream of IFN-γ, inasmuch as administration of IL-18 in IFN-γ-null mice did not alter MHC class II expression (). MCP-1 also promotes antigen presentation and the immune response, as has been demonstrated in apoE-null mice treated with an inactive MCP-1 mutant that showed decreased expression of two crucial regulators of antigen presentation, CD40 and CD40L ().