It is increasingly recognized that atherosclerosis is a complex chronic inflammatory disease rather than a mere phenomenon of lipid deposition on the vascular wall [1,2]. In endothelial cells (ECs), proinflammatory stimuli, including hypercholesterolemia, hyperglycemia and smoking, trigger the expression of adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1) and selectins, which mediate the attachment of circulating monocytes and lymphocytes to ECs. Chemokines and other cytokines, which are produced by vascular wall cells, elicit the infiltration of adherent leukocytes into the intima. Within the intima, monocytes differentiate into macrophages and engulf modified low density-lipoprotein (LDL) through scavenger receptor-mediated endocytosis, leading to conversion of macrophages into lipid-laden macrophages, foam cells. Macrophages amplify the inflammatory responses through the release of numerous cytokines and growth factors. T cells also enter lesions and amplify the local inflammatory responses by producing proinflammatory cytokines. The cytokines and growth factors secreted by these leukocytes as well as ECs direct migration of vascular smooth muscle cells (SMCs) into lesions. In the intimal lesions, SMCs proliferate under the influence of various growth factors and release collagens and other extracellular matrices, expanding the lesions. In advanced atherosclerotic lesions, increased inflammatory activities diminish the collagen content and increase procoagulant activity, leading to plaque rupture and acute coronary thrombosis.

Sphingosine-1-phosphate (S1P) is a blood borne, lysophospholipid mediator that exerts pleiotropic activities including cell proliferation, survival, migration, cell shape and cell-cell adhesion in a variety of cell types [3-6,7]. S1P was originally shown to be released from activated platelets [8] and to be present in the plasma at around 10⁻⁷∼10⁻⁶ mol/L, largely in a form bound to plasma proteins, albumin and high density-lipoprotein (HDL) [8-10]. In agreement with this, plasma S1P levels were highly correlated with HDL concentrations [10]. A number of investigations provided evidence for the notion that HDL possesses an atheroprotective activity [11-13]. It is established that HDL plays an essential role in cholesterol efflux from cholesterol-laden macrophages as a cholesterol acceptor and cholesterol transport to the liver [13]. Besides this role of HDL, it is suggested to exert atheroprotective effects through the HDL-bound S1P [14-16]. Previous studies showed that HDL-bound S1P and other sphingolipids mediated the atheroprotective actions of HDL, including EC survival and proliferation, stimulation of endothelial nitric oxide (NO) synthase (eNOS), inhibition of endothelial expression of VCAM-1 and intracellular adhesion molecule (ICAM)-1, and inhibition of monocyte chemoattractant peptide (MCP)-1 production in SMCs [14-16]. In contrast, other studies showed that S1P exhibits proatherogenic activities. For example, the proinflammatory cytokine tumor necrosis factor-α (TNF-α) activated EC, as evaluated by the expression of adhesion molecules such as E-selectin and VCAM-1, through a sphingosine kinase (SphK), which is an S1P-synthezing enzyme (see below for details) [17]. HDL inhibited the expression of these adhesion molecules by suppressing SphK [18].

Although many reports have suggested that S1P may be involved in atherosclerosis, it remained undefined whether S1P is proatherogenic or antiatherogenic and by what mechanisms S1P modifies atherosclerosis [19]. Recent studies including ours have addressed the in vivo roles of S1P receptor subtypes in atherosclerosis using mouse models of atherosclerosis [20-23]. In this review, we will focus on the receptor subtype-specific, stimulatory and inhibitory effects of S1P on atherosclerosis and discuss the possibility of using S1P receptors as a novel therapeutic target for atherosclerosis.

FTY720-P acts on S1P₁ and induces immunosuppression by sequestering lymphocytes, particularly T lymphocytes, in secondary lymphoid organs and decreasing circulating lymphocytes. T lymphocytes are involved in the initiation and progression of atherosclerosis [1,2]. EC-derived NO is an atheroprotective mediator [60] and upregulated by endothelial S1P₃ and probably S1P₁ [61,62]. Therefore, it is rational to hypothesize that FTY720 may have an impact on atherosclerosis through its immunosuppressing and eNOS-stimulating effects.

Two groups tested the pharmacological actions of FTY720 on the initiation and progression of atherosclerosis in atherosclerotic models, apoE^−/− mice and LDLR^−/− mice [20,21]. Keul et al. demonstrated that oral FTY720 administration (1.25 mg/kg body weight/day) for 20 weeks resulted in more than a 50% reduction of atherosclerotic lesion volumes in apoE^−/− mice fed a high fat (Western) diet without any influence on plasma lipid concentrations [21]. The reduction of atherosclerotic lesions was accompanied by decreases in macrophage density and collagen deposition in the lesions but not changes in the density of CD3⁺ T lymphocytes or SMCs. As expected, FTY720 administration induced lymphopenia, indicating that FTY720 induced immunosuppression by effectively downregulating lymphocyte S1P₁. In contrast, S1P₃-mediated, NO-dependent vasodilator response of aortae isolated from FTY720-administered mice to acute FTY720 challenge remained intact, suggesting that chronic FTY720 administration did not downregulate endothelial S1P₃ or compromise S1P₃-mediated eNOS activation. In addition, they showed that FTY720-P treatment of isolated aortic segments and cultured SMCs potently inhibited thrombin-induced MCP-1 release. Thrombin-induced MCP-1 response was abolished in tissues and cells from S1P₃-null mice. The gene expression of cytokines including IL-12, IL-10, IL-4 or IFN-γ in isolated peritoneal macrophages was not different between mice receiving FTY720 and vehicle. Based on these observations, they suggested that FTY720 inhibited atherosclerosis mainly by suppressing monocyte/macrophage recruitment to atherosclerotic lesions through mechanisms involving S1P₃-mediated, probably NO-dependent inhibition of MCP-1 production.

Nofer et al. has demonstrated in high cholesterol diet (HCD)-fed LDLR^−/− mice that intraperitoneal injection of FTY720 three times a week (0.04 or 0.4 mg/kg body weight/day) for 16 weeks reduced atherosclerotic lesions in a dose-dependent manner [20]. FTY720 substantially decreased CD3⁺ T lymphocytes in lesions but not affect macrophage density, smooth muscle density or collagen content. Moreover, FTY720 inhibited necrotic core formation. FTY720 only at the high dose lowered the peripheral blood lymphocyte with a preferential decrease in T cells, particularly CD4⁺ helper T cell subset. The plasma level of the T cell cytokine IFN-γ was reduced with diminished concanavarin A-induced in vitro mitogenesis of lymphocytes from mice receiving either the low or high dose of FTY720, suggesting that FTY720 attenuated Th1 immune responses. Analogous to lymphocytes, the plasma levels of the macrophage-derived cytokines, TNF-α, IL-6 and IL12, were reduced in mice receiving FTY720 although macrophage accumulation in plaques was not altered in FTY720-treated mice. These observations collectively suggested that chronic administration of FTY720 attenuates development of atherosclerosis through the inhibition of functions of T cells and macrophages.

Recently, it was shown that in apoE^−/− mice on a normal diet, hypercholesterolemia is induced by treatment with a relatively higher dose of FTY720 (3 mg/kg/day) for 12 weeks, which possibly counteracts its anti-atherogenic effect on immune cell distribution [98].

These studies indicate that FTY720 effectively inhibits atherosclerosis, without affecting blood lipid profiles, at the doses which inhibit T cells activity as evaluated with circulating lymphocyte numbers and T cell-specific cytokine production as markers [20,21]. Moreover, the reduced plasma levels of the cytokines, which are abundantly produced by macrophages, suggest that FTY720 directly or indirectly via the regulation of lymphocytes inhibits macrophage activity. In addition, FTY720 may inhibit mobilization of monocytes/macrophages into lesions through mechanisms involving the attenuation of chemokine production. Accumulated evidence indicates that FTY720-induced sequestration of T cells in lymphoid organs and resultant lymphopenia occurs as a result of downregulation of lymphocyte S1P₁, i.e. the functional antagonism of S1P₁ [38,94-96]. It remains unclear whether chronic administration of FTY720 similarly downregulates macrophage S1P₁ in vivo and consequently induces significant functional changes of macrophages. Besides the immune cells as targets of FTY720, this compound may have non-immune cell targets, including ECs and SMCs, in inhibiting atherosclerosis. FTY720-P activates eNOS to stimulate NO production in ECs via S1P₃ and probably S1P₁. In this respect, it is noted that chronic FTY720 administration did not impair FTY720-P-induced vasodilation, which suggest that S1P₃ and/or S1P₁ in ECs were not downregulated [21]. It is likely that FTY720-P activates endothelial S1P₃ and S1P₁ as a functional agonist, leading to increased release of NO. Likewise, S1P₃ and S1P₁ may mediate inhibition of the expression of cytokines including the monocyte chemoattractant MCP-1, leading to inhibition of monocytic infiltration into lesions.

Systemic administration of non-selective immunosuppressive drugs will probably not be useful for the treatment of atherosclerosis because of adverse effects including serious infections [99]. Although FTY720 at a low dose did not reduce circulating lymphocytes [20], it decreased the plasma levels of T cell-specific cytokines, which suggest that FTY720 at the low dose might be accompanied by immune suppression. It is necessary to fully dissect the molecular mechanisms underlying the anti-atherogenic effect of FTY720 at various doses.

Three major S1P receptors, S1P₁, S1P₂ and S1P₃, are expressed in ECs, SMCs and monocytes/macrophages. Among these, S1P₂ but not S1P₁ or S1P₃ in SMCs and ECs mediate inhibition of chemoattractant-directed cell migration [54,58,100]. This unique functional property of S1P₂ can be accounted for by the distinct signaling capacity of S1P₂; differently from S1P₁ and S1P₃, which are G_i-coupled receptors, S1P₂ couples mainly to G_12/13 to result in Rho activation, Rho-dependent Rac inhibition and PTEN stimulation, leading to chemorepulsion. Our recent observations [22,101] showed that ECs express a significant level of S1P₂ in vivo although commonly employed HUVECs do not express an easily detectable level of S1P₂. S1P₁ and S1P₃ stimulate eNOS in ECs and inhibit leukocyte adhesion to ECs. In contrast, the Rho- ROCK pathway, which is activated preferentially by S1P₂, is reported to participate in an inflammatory response [55,102]. In addition, S1P₂ mediates the opposite effect on migration of monocytes/macrophages to that of S1P₁ and S1P₃. These observations raised an intriguing possibility that S1P₂ may have a distinct role in atherosclerosis from S1P₁ and S1P₃.

We have studied the role of S1P₂ in atherosclerosis by using S1P₂-deleted and non-deleted apoE^−/− mice [22]. The en face plaque area in spread aortae was dramatically reduced (approximately70%) in homozygous knockout (S1P₂^−/−) mice compared with S1P₂^+/+ mice after 16 weeks of HCD. The plaque area in heterozygous knockout (S1P₂^+/−) mice was intermediate between S1P₂^−/− and S1P₂^+/+ mice, indicating that S1P₂ has a gene dose-dependent proatherogenic effect. In the plaques of S1P₂^−/− mice, the macrophage density was decreased compared with S1P2^+/+ mice whereas SMC density in the plaques was increased in S1P₂^−/− mice. Consistent with our data, Hla and colleagues very recently showed that S1P₂-deficiency markedly inhibited atherosclerosis in apoE^−/− mice [23]. The mRNA expression levels of the proinflammatory cytokines TNF-α, IL-6, IFN-γ and MCP-1, and the adhesion molecule VCAM-1 were reduced in the aortae of HCD-fed S1P2^−/− mice compared with S1P2^+/+ mice, whereas the phosphorylation of eNOS was increased in the aorta of S1P2^−/− mice [22]. The mRNA expression of S1P₁, S1P₃, and the S1P synthesizing and degradation enzymes including SphK-1, SphK-2, SPL and SPP1 in the aorta was not different between S1P2^+/+ and S1P₂^−/− mice. Thus, the atherosclerotic lesion is reduced in S1P₂^−/− mice with the diminished inflammatory activity and the increased atheroprotective NO activity.

Our previous study using β-galactosidase (LacZ)-knockin mice at the S1P₂ locus, in which LacZ gene expression is driven by endogenous S1P₂ promoter, showed that S1P₂ is expressed in ECs and SMCs of normal blood vessels in a variety of organs and the bone marrow (BM) [100]. In the atherosclerotic lesion in the aortic sinus of LacZ-knockin apoE^−/− mice fed HCD, macrophages, ECs, and intimal and medial SMCs were found to express S1P₂ [22]. The role of S1P₂ in BM-derived cells for atherosclerosis was studied by analyzing BM-chimera mice [22]. The deletion of S1P₂ in BM cells markedly reduced atherosclerotic lesions compared with control. Thus, S1P₂ in BM-derived cells, most likely monocytes and macrophages, play the critical role in atherosclerosis. The study by Skoura et al. [23] supported the importance of S1P₂ in BM-derived cells in atherosclerosis.

The deletion of S1P₂ has a substantial impact on macrophage functions including cholesterol accumulation, cytokine production and migration (Figure 2) [22]; deletion of S1P₂ in macrophages markedly inhibits accumulation of modified LDL through both a substantial decrease in uptake of oxidized LDLs and a modest increase in cholesterol efflux. These effects are accompanied by decreases in scavenger receptor expression (CD36 and scavenger receptor-A) and increases in cholesterol efflux transporter (ABCA1 and ABCG1) expression. These effects of S1P₂-deficiency lead to inhibition of foam cell formation and reductions in atherosclerotic lesions. Second, deletion of S1P₂ in macrophages inhibits the proinflammatory responses by suppressing Rho-ROCK-NF-κB signaling pathway, which is essential for the expression of the proatherogenic gene products including cytokines such as TNF-α and the scavenger receptor CD36. The ABCA1 mRNA expression is negatively regulated by ROCK through mechanisms involving LXR downregulation. Third, S1P₂ possesses a profound influence on transmigration of monocytes/macrophages into atherosclerotic lesions. S1P₂ mediates inhibition of macrophage migration toward a higher concentration of S1P (chemorepulsion) whereas S1P₁ mediates stimulation of migration toward a higher concentration of S1P (chemotaxis). The blood S1P concentration is estimated to be much higher than that in tissues. The chemokines including MCP-1, which are produced in atherosclerotic lesions, attract monocytes to lesions. Therefore, S1P₂ in monocytes very likely promotes macrophage transmigration into lesions. In fact, intravenously infused S1P₂^+/+ macrophages more robustly transmigrated into the vascular wall compared with S1P₂^−/− macrophages. The number of total monocytes and activated monocytes (CD11b⁺Ly6C^hi) in the peripheral blood did not differ between S1P₂^+/+ and S1P₂^−/− mice, suggesting that S1P₂-deficiency did not affect mobilization of monocytes to peripheral blood or their activation. These stimulatory effects of S1P₂ on modified LDL accumulation, cytokine production, and transendothelial migration underlie the proatherogenic action of macrophage S1P₂. Besides monocytes/macrophages, bone marrow-derived mast cells express S1P₂, which stimulates degranulation [42]. Mast cells are implicated in plaque progression and destabilization, and therefore may contribute to the proatherogenic effect of S1P₂ [103]. In our study, activated mast cells were reduced in the aortic wall of S1P₂^−/− mice. The role of mast cell S1P₂ in atherogenesis remains to be clarified.

S1P₂^−/− ECs display altered phenotypes compared with wild-type ECs (Figure 2) [22,100]. eNOS and its product NO have atheroprotective properties. Consistent with the observation that eNOS phosphorylation is increased in the aortae of S1P₂^−/− mice compared with S1P₂^+/+ mice, S1P₂^−/− ECs show stimulation of eNOS phosphorylation in response to S1P stimulation whereas S1P₂^+/+ ECs exhibits a decrease in eNOS phosphorylation in response to S1P. In S1P₂^−/− ECs, S1P₃ and probably S1P₁ mediates eNOS phosphorylation most likely through stimulation of the well known eNOS activating protein kinase Akt. In contrast, S1P induces inhibition of Akt in S1P₂^+/+ ECs through S1P₂-mediated, ROCK-dependent PTEN stimulation [55], which results in inhibition of Akt and consequently eNOS [22]. S1P₂^−/− ECs also showed suppression of the expression of the proinflammatory cytokines including MCP-1 and GM-CSF [22], as in the aortae [104,105]. Because both MCP-1 and GM-CSF, powerful chemoattractants for monocytes, are the NF-κB target genes, the inhibited cytokine response to S1P in S1P₂^−/− ECs is due to diminished ROCK-dependent NF-κB activation. Thus, S1P₂ in ECs could participate in atherosclerosis by regulating adhesion molecule expression, cytokine production and consequently monocyte/macrophage flux, platelet activation and thrombus formation, and intimal cell proliferation through Rho-ROCK-PTEN-mediated Akt-eNOS regulation and Rho-ROCK-NF-κB-mediated regulation of proinflammatory gene expression.

S1P₂ deletion induces alterations of the phenotypes in SMCs (Figure 2) [22]; S1P₂^−/− SMCs show enhanced proliferation in the presence of serum. This could be mediated probably at least partially by Akt stimulation due to loss of ROCK-mediated PTEN stimulation in S1P₂^−/− SMCs. S1P₂^−/− SMCs also exhibit loss of chemorepulsion to S1P. These phenotypes of S1P₂-deficient SMCs might bring about a higher SMC density in atherosclerotic lesions of S1P₂^−/− mice.

The lesions in S1P₂^−/− mice may be stabilized compared with S1P₂^+/+ mice because of a lower macrophage density and higher SMC density in lesions of S1P₂^−/− mice. S1P₂^−/− macrophages display resistance to apoptosis induced by TNF-α and cycloheximide. Stimulated survival of S1P₂^−/− macrophages may also favor plaque stabilization although the relationship between macrophage apoptosis and atherogenesis is also complex [106].

These observations in S1P₂-deleted mice raised the intriguing possibility that pharmacological S1P₂ blockade could afford therapeutic efficacy for atherosclerosis. JTE-K1 is a selective S1P₂ antagonist [107]. We tested the effect of the systemic administration of JTE-K1 into HCD-fed S1P₂^+/+apoE^−/− mice for eight weeks. Oral administration of JTE-K1 (12.5 mg/kg twice daily) by gavage reduced the en face plaque area approximately by 60% with the lower density of macrophages and higher density of SMCs in atherosclerotic lesions compared with the vehicle control, thus recapitulating the phenotypes of S1P₂^−/− mice [22]. The treatment of isolated macrophages with JTE-K1 suppressed uptake of DiI-acLDL and stimulated cholesterol efflux, confirming the effectiveness of S1P₂ blockade at the cellular level. The administration of JTE-K1 did not affect food intake or body weight gain in mice over eight weeks. Mice receiving JTE-K1 did not exhibit ataxia or tilting of the trunk due to vestibular dysfunction that had been reported in S1P₂^−/− mice, or any other discernible abnormality.

Current therapy for human atherosclerotic lesions focuses on reducing the concentration of plasma LDL-associated cholesterol in the blood mainly by administering HMG CoA reductase inhibitors and other drugs [108-110]. Lowering blood cholesterol concentration leads to inhibition of the accumulation of modified LDL in the subendothelial layer, and consequently inhibition of atherosclerotic lesion formation. Although statins also exert plaque stabilizing and anti-inflammatory effects [99,108-110], no therapy to directly target foam cell formation in the face of elevated circulating LDL is currently available.

In genetic mice models for atherosclerosis, pharmacological blockade of chemoattractant receptors that are expressed in monocytes, including leukotriene B4 receptor and RANTES receptors, effectively inhibited plaque formation [111,112]. However, pharmacological blockage of leukocytes recruitment to inflammatory sites may be associated with the side effect of diminished defence mechanisms against infectious pathogens. HDL-associated S1P and the sphingosine mimetic FTY720 seem to exert atheroprotective actions via S1P₁ and S1P₃. In the case of FTY720, it might be necessary to separate the favorable EC-protective effect from the immunosuppressive effect.

In contrast to S1P₁ and S1P₃, S1P₂ is a clearly proatherogenic receptor to act on the three cell types, macrophages, ECs and SMCs [22]. The activity of S1P₂ seems to be diverse compared with chemokine receptors; S1P₂ is involved in modified LDL uptake, cytokine production, migration, eNOS activation, and apoptosis in these cells. Systemic long-term administration of a selective S1P₂-blocker can recapitulate the favorable effects of S1P₂-deficiency without overt toxicity. As human monocytes and macrophages express S1P₂ as in mice [83], S1P₂ likely plays a similar role in human atherosclerotic lesion formation. Therefore, a selective S1P₂-blocker could have clinical benefit as a new therapeutics for atherosclerosis. The combination of an S1P₂-blocker and statins may be of great use. Thus, S1P receptors, particularly S1P₂, are promising therapeutic targets for atherosclerosis.