Inflammation is a response that protects tissues affected by foreign pathogens or physical trauma. Inflammation is initiated through the carefully orchestrated movement of inflammatory cells (neutrophils, monocytes/macrophages, and lymphocytes) following the initial cellular insult. The trafficking of these cells to the inflammatory nidus is regulated by numerous cell surface receptors and ligands, along with the release of chemokines, cytokines, vasoactive amines, and bioactive lipid mediators. Lipid mediators, primarily comprised of sphingolipids and eicosanoid derivatives, were once thought to be just inert members of the bilipid cell membrane, but are now known to play a decisive role not only in proinflammatory cell movement but also paradoxically in its resolution as well.
Eicosanoid lipids are derived from the activity of phospholipase A2 on the 20-carbon membrane phospholipid arachidonic acid (AA).2 Multiple divergent metabolic pathways use AA as their substrate. One pathway metabolizes AA via the lipoxygenase pathway to leukotrienes (LTs) and lipoxins. A second major pathway forms the PGs and thromboxanes from AA via the cyclooxygenase pathway. The prostanoids and lipoxins will not be discussed in this review, because they have been the subjects of numerous recent reviews (1, 2, 3, 4). The sphingolipids, derived from sphingomyelin, include ceramide and its derivatives, sphingosine, and sphingosine 1-phosphate (S1P), and the novel immunomodulatory sphingosine analog 2-amino-2[2-(4-octylphenyl)ethyl]-1–3-propanediol hydrochloride (FTY720), which has recently been reported to serve as an agonist for leukocyte migration with potential clinical applications (5). This review focuses on the biosynthetic pathways, cellular distribution, receptors, and mechanisms of action of, as well as the interactions among, the LT and sphingolipid bioactive mediators, particularly in reference to their role in leukocyte trafficking.
LTs are a family of eicosanoid lipid mediators derived from the metabolism of AA. First described in 1937 as the slow-reacting substances of anaphylaxis, these compounds are now known as the cysteinyl LTs (cysLTs) LTC4, LTD4, and LTE4 (6). Synthesis of LTs can be divided into two pathways: one to create cysLTs and another to create LTB4 (Fig. 1). Both pathways share the common intermediate, the short-lived epoxide LTA4. AA is oxygenated by 5-lipoxygenase (5-LO) to form hydroperoxide 5(S)-hydroperoxy-6-trans-8,11,14-cis-eicosatetranoic acid, which is subsequently dehydrated, yielding LTA4 (7).
5-LO is a 72- to 80-kDa monomeric soluble protein containing one nonheme iron believed necessary for catalysis (8, 9). Active mainly in myeloid cells, including monocytes, B lymphocytes, granulocytes, and mast cells, 5-LO requires Ca2+ and is stimulated by ATP, phosphatidylcholine, and lipid hydroperoxides (reviewed in Refs.10 and 11). The Ca2+ stimulation is believed to translocate the soluble cytosolic 5-LO to the membrane by association with 5-LO activation protein (FLAP) (12, 13). FLAP, an 18-kDa membrane-associated protein with a leukocyte expression similar to that of 5-LO, is expressed on T lymphocytes but not on erythrocytes or endothelial cells. FLAP has been postulated to present or transfer AA to 5-LO for enzymatic degradation (14, 15). Although not fully understood, the inhibition of FLAP by the compound MK-866 disrupts LTA4 production, despite the presence of 5-LO (16). The molecular details behind the 5-LO/FLAP interaction, including their sites of action and their role in binding AA, are unknown.
LTA4 undergoes further transformation by one of two pathways depending on cellular makeup and enzyme presence: 1) hydrolysis to LTB4 by the zinc metalloenzyme LTA4 hydrolase or 2) glutathione conjugation to LTC4 by LTC4 synthase. LTA4 hydrolase is a 69-kDa widely distributed protein found in almost all mammalian tissue (17). Its product, LTB4, is a powerful chemotactic agent that, in contrast to its cysLT counterparts, has no direct role in bronchoconstriction or pulmonary vasoconstriction (18). LTB4 has two known extracellular receptors, BLT1 and BLT2. BLT1R has been characterized as a 43-kDa G-protein-coupled receptor (GPCR) expressed only in inflammatory cells (neutrophils predominantly) and with a high affinity for only LTB4. (19, 20). The second LTB4R, BLT2, has only recently been described and, although a GPCR similar to its BLT1 homolog, the BLT2R is ubiquitously expressed in all mammalian tissues (20, 21, 22). Both receptors are found on the same chromosome, but intriguingly, the open reading frame of the BLT2R is within the promoter region of the BLT1R (19). Activation of both BLTRs serves as a powerful stimulus for in vitro leukocyte chemotaxis, especially of neutrophils (23, 24). In vivo experiments have shown LTB4 to increase neutrophil rolling and adhesion and egress into the extravascular space through increased expression of adhesion proteins (integrins and selectins) (25). Additional actions of LTB4 include stimulation of IL-5 in T lymphocytes, chemotactic effects on IL-5-activated eosinophils, neutrophil secretion of superoxide anion radicals, and antiapoptotic effects on neutrophils (26, 27, 28, 29). LTB4, along with LTC4 and LTD4, has been shown to promote eosinophil survival by inhibiting apoptosis (30).
LTC4 is formed by conjugation of LTA4 with the tripeptide glutathione through the catalysis of LTC4 synthase. Stimulated by divalent cations and phosphatidylcholine, LTC4 synthase is an 18-kDa protein with a wide tissue distribution (31, 32). After LTC4 synthesis, the multidrug transporter ATP-binding cassette (Abc)c1 (formerly known as MRP-1) actively transports LTC4 out of the cell, where LTD4 and LTE4 are formed through the elimination of glutamine and glycine, respectively, by γ-glutamyl transpeptidase and dipeptidase (33, 34, 35). Failure to express Abcc1 has been shown to increase intracellular accumulation of LTC4 (36). LTC4 has been shown in animal models to be relatively short-lived with rapid conversion to LTD4 and LTE4 (6, 37). The transport of LTC4 by Abcc1 has also been shown to regulate dendritic and T cell migration to peripheral lymph nodes (5, 38). Antagonism of Abcc1 or 5-LO activity inhibits dendritic and T cell migration, which is restored after the addition of exogenous cysLTs (5, 38).
The cysLTs have two described GPCR-type cell surface receptors, cysLT1 and cysLT2. The potency of each ligand was determined through intracellular calcium mobilization with a rank order of LTD4 > LTC4 > LTE4 for cysLT1 and LTC4 = LTD4 > LTE4 for cysLT2 (39, 40). The cysLT1R is highly expressed in lung, spleen, and peripheral blood leukocytes (especially eosinophils) (41, 42). The cysLT2R is likewise expressed on these cells and is also highly expressed in the heart, adrenal glands, and brain (42). The significance of these differences is unclear. Activation of cysLT1Rs has been shown to elicit bronchospasm from the contraction of the bronchial smooth muscle cells that are prevalent in the asthmatic population (14). The cysLTs have also been shown to act through the receptors of both cysLTs to induce pulmonary vasoconstriction (43). This action is mediated through activation of cysLT2Rs on the vascular smooth muscle and of cysLT1Rs on the vascular endothelial cells. Activation of the cysLT1Rs has also been shown to increase the microvascular permeability in the airways through either a mechanism of endothelial cell contraction or an increase in the vascular endothelial hydrostatic pressure (44, 45). CysLT1R antagonism has been shown to block this response in animal models, but not in humans (46). In addition, cysLT1R activation also serves as a chemotactic stimulus for eosinophils and to increase mucus production (47, 48). Clinically, modification of the LT pathway through the use of 5-LO inhibitors (zileuton) or cysLT1R antagonists (montelukast or zafirlukast) (currently, no specific cysLT2R antagonist exists, although Bay u933 blocks both cysLTRs) have played an important role in the management of asthmatic patients (49, 50, 51). Although much is know about the LT pathway, there remain many areas requiring further definition. Future investigations into the compartmentalization of the LT intermediates, the binding mechanism of AA with the FLAP/5-LO interaction, and whether other eicosanoids play a role as lipid mediators may lead to novel discoveries with significant clinical implications.
Sphingolipids, such as ceramide, sphingosine, and S1P, were originally thought to have only the rudimentary task of maintaining the integrity of the bilipid cell membrane (Fig. 2). However, these products of sphingolipid metabolism have been shown to have critical roles in cell migration, proliferation, and survival. Sphingosine is formed by the metabolism of the sphingomyelin derivative ceramide through the action of the enzyme ceramidase (52). Although both sphingosine and ceramide have similar cellular actions, including the induction of apoptosis, the arrest of cell growth and proliferation, neither plays a direct role in cell migration (53, 54).
S1P is the bioactive derivative of the phosphorylation of sphingosine by sphingosine kinase (SpK). Although seven cloned isoenzymes of SpK have been described, only two forms, SpK1 and SpK2, each with different tissue distribution and temporal actions, are predominantly found in humans and mice (55). SpK1 is a 42.4-kDa protein found in the cytosol of lung, spleen, and liver cells (55). SpK2, a proline-rich 65.6-kDa protein, is expressed in liver and heart cells (56). Both kinases are activated by the external stimuli PDGF, nerve growth factor, TNF-α, and IL-1β, driving the conversion of sphingosine to S1P (57, 58, 59, 60). Recent evidence shows that SpK is secreted by endothelial cells (61). This suggests that S1P can be made both intracellularly and extracellulary. Once formed, S1P is metabolized by S1P lyase to ethanolamine phosphate and hexadecanal, or dephosphorylated back to sphingosine by a lipid phosphohydrolase (62). The lipid phosphohydrolase family is composed of S1P phosphohydrolase (SPP) and the type 2 lipid phosphohydrolase. SPP includes three known mammalian (one murine/two human) homologs and two yeast counterparts and is specific to S1P, dihydrosphingosine 1-phosphate, and phytosphingosine 1-phosphate (63, 64, 65, 66). The first described human SPP isoenzyme is ubiquitously expressed unlike the second, which is limited to the endoplasmic reticulum of brain, heart, colon, kidney, small intestine, and lung cells (67). The type 2 lipid phosphohydrolase has a wide expression and substrate specificity, including S1P, lysophosphatide, and ceramide 1-phosphate (68, 69, 70, 71, 72).
The balance between the production and metabolism of S1P is the driving force behind the so-called sphingolipid rheostat of the cellular environment. This rheostat determines the cumulative effect on cell migration, proliferation, and apoptosis profiles of the concentration and receptor gradient between sphingosine and ceramide with S1P. High levels of sphingosine and/or ceramide compared with S1P will inhibit cell proliferation and migration and promote apoptosis (72). Reversing this balance will allow S1P to promote migration and proliferation. An example of this balance is seen in LT production within mast cells. If the balance is shifted to sphingosine, inhibition of LT production occurs; however, if S1P is predominant, increased LT synthesis results (73).
Similar to other lipid mediators, S1P has a dual role as both an extracellular ligand, via the S1PRs, and as an intracellular ligand, via an undefined second messenger system. Formerly known as the endothelial differentiation gene receptors (Edg), the S1PRs exist as five subtypes: S1P1 (Edg-1), S1P2 (Edg-5), S1P3 (Edg-3), S1P4 (Edg-6), and S1P5 (Edg-8). S1P1–3 have a wide tissue distribution, S1P4 is mainly found in lymphoid tissues and platelets, and S1P5 is confined to the nervous system. Although all five subtypes have different tissue distribution and mechanisms of action, they all have an effect on cell migration (whether positive or negative) (53). Activation of S1P1 by S1P causes cell migration through the activation of the guanine triphosphatase Rac, modifying the actin network to form lamellipodia and the filopodia on the leading edge of the migrating cell (74, 75, 76, 77). Homozygous deletion of the murine S1P1 gene leads to intrauterine vascular hemorrhage and death by embryonic day 13.5 due to failure of pericyte and smooth muscle cell migration, despite an intact vasculature (75). Activation of S1P2 inhibits induction of Rac and instead activates the guanine triphosphatase Rho and its subsequent stress fiber assembly, thereby inhibiting cell migration (78). Homozygous deletions of the S1P2 analog in the miles apart gene of the zebra fish causes abnormal heart development due to faulty myocardial cell migration. Stimulation of the S1P3R by S1P has a known antiapoptotic effect and is thought to aid in the promotion of cell migration (78). The roles of S1P4 and S1P5 are poorly defined, although S1P5 is found predominantly in oligodendrocytes and astrocytes and is presumed to play a role in the development of the nervous system (79). The seemingly contrary role of S1P as an agonist and an inhibitor of cell migration can be best explained as dependent on cellular concentrations of S1P and its respective receptors. Evidence shows that higher concentrations of S1P inhibit smooth muscle migration, whereas at lower concentrations, migration is induced, suggesting that different receptor subtypes are concentration sensitive (80).
Evidence also shows that the activation of S1PRs plays an important role sequestering T and B lymphocytes in secondary lymphoid organs by promoting egress from peripheral blood. For example, S1P enhances in vitro migration to the CCR7 ligands CC chemokine ligand (CCL)19 and CCL21 by T lymphocytes pretreated with the ligand (5). The role of S1P as an intracellular second messenger is not as well defined as its extracellular role. Although intracellular S1P is a known inhibitor of apoptosis and an inosine triphosphate-independent regulator of calcium homeostasis, the exact mechanisms and targets of these actions are not known (81). The mechanisms regulating sphingosine or S1P transport from the cytoplasm to the extracellular space are not known. Sphingosine, and perhaps S1P, can be transported out of the cytosol, similar to platelet activating factor, via the Abc family of proteins, in particular Abcb1 (formerly multidrug resistance) (5).
FTY720 is a synthetic derivative from ISP-1 (myriocin), a metabolite from the traditional Chinese herb Iscaria sinclarii (Fig. 2). A novel immunomodulator, FTY720 has been shown in numerous animal models to prolong allograft transplant survival and to reduce inflammation seen in a murine model for multiple sclerosis (82). Recent evidence shows that FTY720 is also effective in human renal transplantation, prolonging graft survival, although further clinical trials are currently underway (83). Originally postulated to act via apoptosis, FTY720 has been recently shown to prolong allograft survival by sequestering lymphocytes in secondary lymphoid organs, thereby preventing them from migrating to the inflammatory nidus of the transplanted organ (84, 85). Recent studies show that a single oral dose of FTY720 in mice depletes peripheral blood and splenic T lymphocytes, which become sequestered in peripheral lymph nodes and Peyer’s patches (5, 86, 87). Despite causing lymphocyte sequestration, FTY720 does not increase susceptibility to systemic viral infection and does not affect T lymphocyte priming or activation (85).
The mechanism of action behind FTY720 has only recently been elucidated. Because FTY720 is a structural homolog of sphingosine, it was proposed that it might share some of its properties and also function as a substrate for the sphingolipid enzyme cascade. In vivo studies have confirmed this hypothesis by showing FTY720 to be phosphorylated to a phosphate ester (P-FTY720), likely by SpK (87, 88). P-FTY720, analogous to S1P, also appears to be degraded by the sphingosine lyases and phosphohydrolases, although the exact isoenzyme and mechanism remain unclear. (87) That FTY720 was shown to elicit apoptosis at higher doses, but cell migration at lower doses, is analogous to the balance between the apoptotic influence of sphingosine and the promigratory one of S1P.
Because P-FTY720 is a homolog of S1P in both structure and in function, it was hypothesized that they may also share receptors. FTY720-driven lymphocyte trafficking is pertussis toxin sensitive, suggesting that P-FTY720 acts at a GPCR similar to S1PR (87). In vitro experiments have confirmed this by demonstrating that P-FTY720 binds with higher affinity than S1P itself to four of the five S1PRs (87, 88). The only receptor that displayed no response to P-FTY720 was S1P2, which in other studies antagonizes cell migration.
Interactions among lipid mediators
These investigations demonstrate that there is degeneracy in the bioactive lipid families, with multiple ligands for each receptor or multiple receptors for each ligand. In addition, there are nonhomogeneous and overlapping cellular and tissue distributions for these receptors. A number of interactions may take place among different members of the bioactive lipid mediator families, including autocrine, paracrine, and endocrine effects, and precise physiologic outcomes depend on the cellular constituents of an inflammatory infiltrate.
There are also interactions between the members of different lipid families that have recently been elucidated by investigations on leukocyte migration. FTY720 enhances T lymphocyte migration to the CCR7 chemokine ligands CCL19 and CCL21 and to the CCR2 ligand monocyte chemoattractant protein-1 (5, 89). These observations, coupled with experiments that show that dendritic cell migration from the periphery was dependent on both of these chemokines and on the activity of the lipid transporters Abcb1 and Abcc1, led to additional experiments aimed at clarifying the molecular mechanism behind the action of FTY720 (5, 38, 90). These experiments suggest a model in which FTY720 is taken into cells, phosphorylated by SpK intracellularly, and effluxed through the Abcb1R, and then activates S1PRs in an autocrine or paracrine function. S1PR activation then enhances 5-LO activity, production of cysLTs, efflux of cysLTs by Abcc1, and autocrine or paracrine activation of cysLTRs. CysLTR activation sensitizes CCR7 and promotes chemotaxis to CCL19 and CCL21 (5). Thus, FTY720 promotes T cell peripheral lymph node migration by activating the Abcb1 sphingolipid transporter and the Abcc1 LTC4 transporter and enhancing 5-LO activity and cysLT production (5). This model of a multidrug transporter and chemokine-dependent pathway for lymph node homing depends on both sphingolipids and LTs. Ongoing work in our laboratories suggests that additional chemokine ligands and receptors are likely involved. Further, the PG lipid PGE2 may also influence CCR7-directed dendritic cell migration by increasing receptor expression (91).
Since the discovery of bioactive lipid mediators over 50 years ago, much critical progress has been made in defining their mechanisms of actions and how they can be harnessed for clinical applications. However, much more information is needed to define the intracellular receptors for sphingolipids and their derivatives; the cellular loci and mechanisms of action of the lipid receptors; the interactions between 5-LO and FLAP with AA in the production of LTs; the existence of other sphingosine analogs, similar to FTY720, that promote cell migration; and other novel, bioactive lipid mediators. Other important issues that remain to be resolved include the roles of bioactive lipid mediators in adaptive vs innate immunity and whether they play a role as autocrine, paracrine, or endocrine ligands.
Abbreviations used in this paper: AA, arachidonic acid; LT, leukotriene; cysLT, cysteinyl LT; S1P, sphingosine 1-phosphate; 5-LO, 5-lipoxygenase; FLAP, 5-LO activation protein; GPCR, G-protein-coupled receptor; SpK, sphingosine kinase; SPP, S1P phosphohydrolase; Edg, endothelial differentiation gene receptor; CCL, CC chemokine ligand; Abc, ATP-binding cassette; FTY720, 2-amino-2[2-(4-octylphenyl)ethyl]-1–3-propanediol hydrochloride; P-FTY720, phosphorylated FTY720.