Discovery of GPCRs for LPA and S1P

Lysophosphatidic acid (LPA) and sphingosine 1-phosphate (S1P) are two representative lysophospholipid mediators acting on G-protein-coupled receptors (GPCRs). LPA was first identified three decades ago as a factor affecting blood pressure, platelet aggregation, and smooth muscle contraction1, 2, 3, 4, 5 and rediscovered as a mitogenic serum lipid inducer of neurite retraction6, 7. The name 'lyso' originates from lysis of blood cells; thus, the non-specific detergent action of LPA has been doubted. In 1996, the high potency (nmol/L range) and GPCR implications of LPA were finally connected to the discovery of the first LPA receptor (LPA1, formerly known as Edg-2)8. Subsequently, LPA2 and LPA3 were identified as members of the endothelial differentiation gene (Edg) subfamily of GPCRs9, 10, 11. These three LPA receptors (Edg family) share a high homology with each other12.

Recently, the non-Edg family of LPA receptors, LPA4 (GPR23, p2y9), LPA5 (GPR92), and LPA6 (p2y5), were reported as members of the purinergic GPCR cluster13, 14, 15, 16, 17, 18, 19. Using the GPCR-Gα16 fusion expression system, Fujita's group reported GPR87 as another LPA receptor and P2Y10 as a dual receptor for both LPA and S1P20, 21. These results have not yet been confirmed by other research groups22. Instead, lysophosphatidylserine has been suggested to act as a ligand for P2Y10 without confirmation of LPA and S1P as ligands in NIH3T3 cells23.

S1P was initially reported 15 years ago to be a second messenger, mediating an increase in calcium levels due to PDGF and IgE signaling24, 25. The molecular target of S1P in the cytosol has not yet been identified. The initial finding of an S1P-induced calcium increase stimulated research in S1P biology and was linked to its recognition as an intercellular first messenger. The involvement of trimeric G proteins in S1P-induced actions as well as pertussis toxin sensitivity strongly suggested the presence of S1P GPCRs in the plasma membrane26, 27, 28, 29, 30. The discovery of S1P1 (formerly known as Edg-1) in 1998, along with four other receptors (S1P2–5) of the Edg subfamily GPCRs, became a milestone in S1P biology12, 31, 32, 33, 34, 35, 36.

Development of agonists and antagonists for LPA receptors

Prior to receptor cloning, various aspects of LPA receptor structure-activity relationships had been studied: 1) fatty acid chain length and the presence of double bonds, 2) acyl and alkyl linkage, 3) stereo-selectivity on the sn-2 position, and 4) modification of phosphate groups37. Oleoyl LPA (18:1) was considered the optimum ratio to increase Ca2+ levels in A431 cells38. Alkyl LPA showed a better effect than acyl LPA in platelet aggregation assays and an equal effect in other responses39, 40. Several groups reported a lack of stereo-selectivity on the sn-2 position of LPA. Early studies showed that modification of the phosphate group resulted in an absence of LPA responses38, 40.

The initial synthetic approach to LPA response was characterized using platelet aggregation assays. Replacement of the glycerol backbone with amino acids resulted in the production of N-palmitoyl serine phosphatidic acid (NPSPA) and N-palmitoyl tyrosine phosphatidic acid (NPTyrPA)40. These molecules were not active on platelet aggregation in humans, but did demonstrate activity as LPA receptor antagonists in Xenopus41. Later, NPSPA and NPTyrPA were reported as partial agonists of mammalian LPA receptors42, 43. Sugiura et al introduced the ethanolamine-based LPA mimetic, N-acyl aminoethanol phosphoric acid (NAEPA), as an equipotent agonist of platelet aggregation40. Later, research groups from the University of Virginia used NASPA and NAEPA as platforms to synthesize a series of VPC compounds37, 44. The phosphonate analogue of NAEPA lost its platelet aggregation properties40, but methylene phosphonate LPA was equipotent to LPA as an inhibitor of forskolin-driven cAMP accumulation in rat C6 glioma cells45. Jalink et al synthesized various phosphonate analogues along with fatty alcohol phosphates and the methyl ester of LPA (lysophosphatidylmethanol, LPM), but could not show a significant impact of these compounds on Ca2+ increase in A431 cells38. Ironically, these chemicals turned out to be selective or non-selective agonists of cloned LPA receptors (see details below). In the early era of LPA biology, suramin and lysophosphatidylglycerol were used to demonstrate GPCR involvement in LPA responses46 and as an antagonist of LPA-induced Ca2+ responses in Jurkat T cells47, respectively.

LPA GPCR agonists

Since the discovery of the three-Edg family of LPA receptors, the development of selective receptor-subtype agonists and antagonists has accelerated. The optimal chain length and the presence of double bonds have been found to vary depending on receptor subtype. For example, LPA3 showed a preference for unsaturated LPA similar to oleoyl LPA48, whereas LPA6 showed a preference for 2-acyl LPA19. Synthesis of LPA derivatives with phosphonate or thiophosphate groups instead of the phosphate group showed receptor-subtype selective activity similar to 1-oleoyl-2-O-methyl-rac-glycerophosphothionate (OMPT), 1-O-acyl-α-fluoromethylenephosphonate, and α-hydroxymethylenephosphonate LPA analogues as LPA3 receptor selective agonists49, 50, 51, 52. The phosphonate derivatives also provided a path toward the development of phosphatase-resistant long-lasting LPA derivatives53. A dialkyl phosphatidic acid (PA) 8:0 analog with a thiophosphate was reported as a potent and selective LPA3 agonist54, but later, agonistic activity on the LPA5 receptor and antagonistic activity on the LPA1,3 receptors was reported55. Based on computational modeling, dodecyl fatty alcohol phosphate was shown to be a specific LPA2 receptor agonist56, and later, oleoyl-thiophosphate was identified as a pan-agonist (LPA1–3)43. A methylene phosphonate LPA analogue was reported as a selective LPA2 agonist, but it recently was shown to exert agonistic and antagonistic activities on LPA5 and LPA3, respectively52, 57. T-15 (LPA1 agonist) and T-13 (LPA3 agonist) were both synthesized using a carbohydrate scaffold58. Darmstoff analogues were introduced as novel scaffolds for subtype-selective LPA receptor ligands59. Finally, octadecenyl phosphate was shown to be a selective agonist for LPA4 and LPA5 receptors55.

Although alkyl LPA has been shown to exert equipotent activity on each Edg-family LPA receptor compared to acyl LPA60, alkyl LPA was found to be more potent than acyl LPA in platelet aggregation responses3. Therefore, the Edg-family LPA receptors were not able to account for the LPA response in platelets61. Recently, non-Edg (purinergic) LPA receptors have been identified. The LPA5/GPR92 receptor has demonstrated alkyl preference and a presence in platelets55. Farnesyl diphosphate, N-arachidonyl glycine, and carba-cyclic phosphatidic acid were shown to be selective agonists for LPA555, 62, 63. LPA6/p2y5 showed a preference for 2-acyl LPA over 1-acyl LPA19. Alkyl-OMPT, an LPA3 agonist, and 2-linoleoyl LPA showed better agonistic effects on the LPA6 receptor than 2-oleoyl LPA showed19, 64. In contrast, the methyl ester of LPA (LPM) was shown to be a pan-agonist for LPA1–5 although it was less potent than LPA65.

LPA GPCR antagonists

The development of LPA derivatives with a bulky group on the sn-2 position has demonstrated the stereo-selectivity of LPA receptors and has led to the further development of the LPA1/LPA3 selective antagonists VPC12249 and VPC3218366, 67. 2-Pyridyl-containing phosphonate was developed as a non-hydrolyzable LPA3 receptor antagonist68. Dioctylglyceropyrophosphate (DGPP) and Ki16425 were determined to be selective LPA1/LPA3 antagonists following screening of available lipid and non-lipid molecules69, 70. The thiophosphate version of PA 8:0 was shown to be the most potent LPA3 antagonist, whereas the dialkyl PA 8:0 analog with a thiophosphate was shown to be a potent and selective LPA3,5 agonist, as mentioned above54, 55. Tetradecyl-phosphonate was identified as a pan-antagonist (LPA1–3)43. Furthermore, farnesyl phosphate and farnesyl diphosphate potently and selectively blocked the LPA3-mediated Ca2+ increase71, although farnesyl diphosphate is a LPA5 selective agonist, as mentioned above55. T-14, which used a carbohydrate scaffold, was shown to be an LPA3 antagonist58.

Recently, virtual screening identified non-lipid LPA3 antagonist NSC161613, LPA2 antagonist H2L5186303, LPA4,5 antagonist 5987411, and LPA1,6 antagonist 576583455, 72. An LPA analogue of α-bromomethylene phosphonate was reported to act as a pan-antagonist to LPA1–4 and an autotoxin inhibitor52, 73.

Pharmacological tools for LPA GPCR signaling

Commercially available chemicals for studying LPA receptor subtypes are currently in development, although the effects of previously developed chemicals on recently identified non-Edg LPA receptors have not been completely verified (Figure 1). For LPA1 or LPA3 receptor signaling, a combined application of LPA1,3 antagonists such as VPC12249, VPC32193, DGPP, and Ki16425 and LPA3 agonists such as OMPT and α-fluoromethylene phosphonate would be more favorable. For LPA2 receptor signaling, dodecyl phosphate is an adequate LPA2 selective agonist. For LPA5 receptor, farnesyl diphosphate could be used as a selective agonist. For LPA6 receptor, alkyl OMPT, an LPA3,6 agonist, could be used in combination with LPA1,3 antagonists (Table 1).

Figure 1
figure 1

Structures of commercially available agonists and antagonists for LPA GPCR signaling. Sources are Avanti polar lipid, Biomol international, Echelon bioscience, Enzo Life sciences, and Sigma-Aldrich.

Table 1 Agonistic and antagonistic characters of each compound on each LPA receptor. Numbers (nmol/L) mean EC50 or KD values for agonists and IC50 or KI values for antagonists.

Development of agonist and antagonist for S1P receptors

In contrast to LPA, the structure-activity relationship of S1P has a very short history. Using cloned S1P receptors, sphinganine-1-phosphate (dihydro-S1P) and sphingosylphosphorylcholine (SPC) were shown to be equipotent and far less potent on each S1P receptor36, 74, 75, 76.

S1P GPCR agonists

Following the cloning of the S1P receptors, the development of S1P agonists and antagonists began. The importance of the D-erythro configuration of S1P was demonstrated using the cloned receptors77. The linkage of the immune modulator FTY720 to S1P receptors, however, boosted this area of research and opened a new direction for S1P biology78, 79, 80. Lymphopenia induction by inhibiting lymphocyte egress from lymphoid organs was shown to be mediated through the S1P1 receptor81. High-throughput screening (HTS) of an available chemical library showed that SEW2871 acted as an in vivo active heterocyclic S1P1 selective agonist81, 82 and compound 26 was synthesized as a potent 3,5-diphenyl-12,4-oxadiazole S1P1 agonist83. Later, using ultra-HTS, 3,5-diaryloxadiaxole (CYM5181) and dicyclohexylamide were found to be selective agonists for S1P1 and S1P3, respectively84. Using computational modeling, CYM-5442 was developed as an S1P1 selective agonist that was more potent than CYM518185. AUY954, an aminocarboxylate analogue of FTY720, was also introduced as an S1P1 selective agonist86. VPC01091, a cyclized analogue of FTY720, was shown to act as an orally active S1P1 agonist and an S1P3 antagonist87. KRP-203 is a pro-drug immune modulator similar to FTY720; the phosphorylated form of KRP-203 was shown to be a selective S1P1 agonist88, 89. Constrained azacyclic analogues of FTY720 showed selective agonist activities on S1P4 and S1P5 receptors90. Finally, phytosphingosine-1-phosphate was shown to act as a potent and selective agonist on the S1P4 receptor76.

S1P GPCR antagonists

Suramin was temporarily used as an S1P3 antagonist75, 91. Human S1P5 was also reported to be sensitive to suramin and its analogue NF02392.

Following screening of an available chemical library, JTE-013, a pyrazopyridine derivative, was identified as an S1P2 antagonist93, 94. Modification of the FTY720-phosphate structure led to the development of VPC23019 and VPC25239 as selective S1P1/S1P3 antagonists95. As mentioned above, VPC01091 is an orally active S1P1 agonist and S1P3 antagonist87. W146, hexyl phenyl amide phosphonate, was found to be a selective S1P1 antagonist96. VPC44116, an octyl analogue of W146 and γ-aminophosphonate analogue of VPC23019, antagonized lymphopenia and lung permeability via the S1P1 receptor97. SB64146 was reported to act as an inverse agonist on the S1P1 receptor98. Ascotricins A and B were isolated from a cultured broth of a fungus identified as Ascotricha chartarum and shown to inhibit the S1P1 receptor and S1P-mediated HUVEC migration99. Sankyo Co synthesized compound lead 2 (CL2), 2-(4-ethoxyphenoxy)-5-(3-octadecyl-5-oxo-4,5-dihydro-1H-pyrazol-1-yl) benzenesulfonate, which antagonized the S1P1>S1P3>S1P2 receptors100. Human S1P1 receptor-selective antagonist and agonist effects of a rat monoclonal antibody (4B5.2) in vivo were also reported101. Using a 3D database search, BML-241, 2-alkylthiazolidine-4-carboxylic acid, was found to act as an S1P3 antagonist, but its selectivity and potency were not recapitulated in CHO-K1 cells expressing the S1P3 receptor102, 103. A pharmacophore-based design of an S1P3 antagonist with a 3,4-dialkyoxybenzophenone scaffold was suggested104.

Pharmacological tools for S1P GPCR signaling

Commercially available tools for studying S1P receptor subtypes are highlighted in Figure 2. For S1P1 receptor signaling, CYM-5442 or SEW2871, both potent selective S1P1 agonists, and W146, a selective S1P1 antagonist, should be sufficient to elucidate S1P1 receptor involvement. S1P2 receptor signaling could be dissected using JTE-013, an S1P2 selective antagonist. For S1P3 GPCR signaling, a combined application of an S1P1,3 antagonist (VPC23019) and S1P1 antagonist (W146) or S1P1 agonist (CYM-5442) could be useful. Phytosphingosine 1-phosphate, an S1P4 selective agonist, could be used to study S1P4-mediated signaling. S1P1,3 antagonist (VPC23019)-insensitive, S1P2 antagonist (JTE-013)-insensitive, S1P4 agonist-non-responsive, and S1P- or FTY720-phosphate-sensitive signaling might be interpreted as S1P5 receptor or unidentified S1P receptor signaling (Table 2).

Figure 2
figure 2

Structures of commercially available agonists and antagonists for S1P GPCR signaling. Sources are Avanti polar lipid, Biomol international, Cayman chemical, Echelon bioscience, Enzo Life sciences, Sigma-Aldrich, and Tocris.

Table 2 Agonistic and antagonistic characters of each compound on each S1P receptor. Numbers (nmol/L) mean EC50 or KD values for agonists and IC50 or KI values for antagonists.

Development of drugs acting on lysophospholipid GPCRs

Prior to the molecular cloning of GPCRs for LPA and S1P, a number of studies were conducted with the primary goal of determining the functions of lipid mediators at both the cellular and the organ level. These functions included platelet aggregation, smooth muscle contraction, and cell proliferation, among others1, 2, 3, 4, 5, 6, 105. The discovery of GPCRs allowed signal transduction studies to proceed in cells over-expressing the receptors and in receptor knock-down transgenic mouse models106, 107, 108, 109. The discovery of the pathophysiological significance of LPA and S1P, particularly on each receptor subtype, would contribute favorably to new drug development. In developing new medications acting on LPA or S1P receptors, subtype selectivity would be a major issue along with potency and efficacy to avoid side effects and ensure drug safety.

FTY720 was initially developed as an immune modulator for organ transplant patients110. At present, this compound is under clinical study for the treatment of multiple sclerosis111. Amira Pharmaceuticals reported LPA1 selective antagonist (4′-{4-[(R)-1-(2-chloro-phenyl)-ethoxycarbonylamino]-3-methyl-isoxazol-5-yl}-biphnyl-4-yl)-acetic acid, AP2966, which showed good therapeutic potential in idiopathic pulmonary fibrosis and good pharmacokinetic profiles, including oral bioavailability112. Pfizer global research and development introduced the S1P1 selective agonists PF-A and PF-B. These agonists resulted in lymphopenia in rats and monkeys similar to FTY720 and reduced collagen-induced arthritis113. Using a different approach, Lpath Inc. recently introduced monoclonal antibodies against S1P and LPA. Humanized anti-S1P monoclonal antibody (mAb) sonepcizumab blocked the tumorigenic effect of S1P produced by cancer cells as well as the angiogenic effect induced during pathological angiogenesis114, 115. The LPA receptor pan-antagonist (LPA1–4) and autotoxin inhibitor, α-bromomethylene phosphonate LPA analogue, was shown to be an excellent anti-cancer agent116, 117.

Closing remarks

Following the initial discoveries of LPA activity 30 years ago and S1P activity 15 years ago, a dark dawn era in lysophospholipid biology occurred due to the lack of identifiable targets. The discovery of the Edg-family GPCRs for LPA and S1P shined light on this field. The first meeting that focused on lysophospholipid biology was a New York Academy of Science meeting conducted in 1999 at Rockefeller University118. Since 2001, FASEB summer research conferences on lysophospholipids have been held biannually. Lysophospholipid receptor nomenclature has been systematically assigned119, and more LPA receptors (purinergic or non-Edg LPA receptors) are being discovered each year. Every two years, we have had exciting findings including the linkage of FTY720 to the S1P receptor, discovery of an autotoxin as a LPA-producing lysoPLD, lysolipid-sensitive proton-sensing GPCRs (OGR1 subfamily), and finally, the development of new chemicals. This review integrates accumulated information regarding pharmacological tools for lysophospholipid GPCR signaling to compare their characteristics and provides valuable information such as available chemical sources. Using a combination of receptor expression patterns in each organ or cell, these pharmacological tools might prove useful in defining the pathophysiological impact and significance of lysophospholipids. As mentioned above, the control of lysophopholipid functions using specific agonists or antagonists will contribute toward novel drug development. At a 2009 FASEB meeting in Carefree (Arizona, USA), we began to see several of the chemicals described here, in addition to FTY720, being used in clinical applications112, 113, 114, 116. We now know more about many things than ever before. It is very likely that in the near future, the agonists/antagonists for LPA or S1P receptors will be on the market commercially and that there will be a section on lysophospholipid GPCRs in every basic pharmacology textbook.