Lysophospholipid receptor nomenclature review: IUPHAR Review 8

Introduction

The biological and pathophysiological functions of the small signalling lipids known as lysophospholipids continues to expand, with roles that involve virtually every vertebrate organ system (Fukushima et al., 2001; Ishii et al., 2004; Choi et al., 2010; Mutoh et al., 2012; Choi and Chun, 2013). The overwhelming majority of effects, and all activities that have led to actual medicines or to compounds that have entered late-stage clinical trials, rely mechanistically on lysophospholipid receptors. All bona fide receptors are of the 7-transmembrane, GPCR class (Table ​(Table11 and Figure ​Figure11).

Figure 1

Lysophospholipid receptors and their intracellular signalling pathways. Lysophospholipid ligands (LPA, S1P, LPI and LysoPS) bind to their specific GPCRs, which activate heterotrimeric G-proteins (defined here by their α subunits) to initiate downstream signalling cascades. R in the chemical structures is a variable acyl side chain.

Various orphan receptor names have been used over the years; however, receptor identities have led to two nomenclatures: the first used in pharmacological fields and supported by the International Union of Basic and Clinical Pharmacologists (IUPHAR), and the second used in genetic or genomic fields, as represented by the Human Genome Organization (HUGO), Gene Nomenclature Committee (HGNC), and the Mouse Genome Informatix (MGI) Guidelines for Nomenclature of Genes, based upon the 2011 International Committee on Standardized Genetic Nomenclature for Mice. We briefly review these lysophospholipid receptors and their names, and suggest use of a hybrid nomenclature wherein protein names are referred to by their original IUPHAR names (Chun et al., 2002; 2010; Davenport et al., 2013), while HGNC nomenclatures are used to identify the human genes, and MGI nomenclatures for mice are extended to cover non-human genes (Table ​(Table2).2). In each subheading of this review, the protein name is followed by the human and non-human gene names. Recent additions to the lysophospholipid receptor family include glycerophospholipid species lysophosphatidyl inositol (LPI) and lysophosphatidyl serine (LysoPS); names for these newer receptors and genes have been proposed, which generally follow the receptor protein and gene for other lysophospholipid receptors and have been incorporated in this review. The names of established receptors and their human and non-human gene names start each subsection, while new receptors are treated under a separate heading.

Lysophosphatidic acid (LPA) receptors

The many effects of LPA are mediated through the six currently recognized LPA receptors, LPA_1–6. These 7-transmembrane GPCRs couple to one or more of the four classes of heterotrimeric G-proteins, commonly defined by their G_α proteins (G_α12/13, G_αq/11, G_αi/o, and G_αs). Less explored is possible signalling through these receptors that do not require heterotrimeric G-proteins (Rajagopal et al., 2005). Activation of these receptors and G-proteins can initiate myriad downstream pathways that in turn, produce a similarly diverse range of biological and pathological effects (Gilman, 1987). The agonists and antagonists for these receptors and their efficacy are summarized in Table ​Table33.

LPA₄/LPAR4/Lpar4

LPA₄ is notable because it shares less than 20% amino acid sequence identity with LPA_1–3 and S1P_1–5, and is phylogenetically far from them and located near the P2Y receptor family (Figure ​(Figure2).2). Identification of LPA₄ was made by screening orphan receptors, including purine receptor families, using calcium mobilization as a readout for ligand-induced signals (Noguchi et al., 2003). P2Y₉ has ∼20% sequence identity to LPA_1–3 (Noguchi et al., 2003), yet it responds to LPA and not to assayed nucleosides or nucleotides (Noguchi et al., 2003). LPAR4 is located on chromosome Xq21.1 and encodes a 370 amino acid protein of ∼42 kDa, with mouse Lpar4 being present on the D-region of chromosome X. Lpar4 gene expression is observed in the brain, heart, lung, skin, thymus, and uterus (Ishii et al., 2009b). It is also developmentally expressed within the embryonic brain branchial arches, limb buds, liver, maxillary processes, and somites (Ohuchi et al., 2008).

Figure 2

Phylogenetic tree of related GPCRs and amino acid sequence identities. (A) A molecular phylogenetic tree of human GPCRs. The selected GPCR protein sequences were analysed for the phylogenetic reconstruction by the ‘All against All’ sequence programme at the Computational Biochemistry Research Group server of the ETH Zürich. (B) Pair-wise matrices comparing amino acid sequences of lysophospholipid receptors. The upper and lower matrices specify identities among lysophospholipid receptors in human and mouse respectively. The amino acid sequence identities are shown in a gray-to-white gradient. The numbers in the boxes were calculated by Clustal Omega (Sievers et al., 2011).

LPA₄ couples to G_α-proteins G_αs, G_αi, G_αq, and G_α12/13 (Lee et al., 2007), the latter of which activates Rho/ROCK to induce neurite retraction and stress fibre formation seen with activation of other LPA receptors (Lee et al., 2007; Yanagida et al., 2007). It can also induce cell aggregation and adhesion through N-cadherin (Yanagida et al., 2007) and was the first LPA receptor activating G_αs activity (Lee et al., 2007) to promote intracellular cAMP accumulation. LPA₄ can transform cells when co-expressed with oncogenic-promoting genes like c-Myc or Tbx2 (Taghavi et al., 2008). It has also been reported to affect immortalized hippocampal progenitor cells (Rhee et al., 2006).

Null mutant mice for Lpar4 do not show overt abnormalities (Lee et al., 2008) aside from some prenatal loss, probably produced by blood vessel defects that result in abnormal haemorrhage (Sumida et al., 2010). Lymphatic vessels and lymph sacs are also affected during development of the circulatory system (Sumida et al., 2010). Osteoblast differentiation is also inhibited based on cell culture analyses in experiments that knocked down LPAR4 (Liu et al., 2010). Cells from Lpar4^−/− mice show reduced cell motility (Lee et al., 2008).

LPA₁/LPAR1/Lpar1

The first receptor identified for any lysophospholipid came from studies on the brain, which identified LPA₁ (Hecht et al., 1996), a receptor that mediates the effects of LPA. LPAR1 encodes a receptor of 364 amino acids, with a molecular mass of ∼41 kDa. The human gene is located on chromosome 9 (9q31.3), and consists of at least five exons. A gene variant of Lpar1 (Lpar1-mrec1.3) lacks a predicted 18 amino acids from the amino terminus (Contos and Chun, 1998); however, its function and significance remain unclear. This receptor couples to three G_α proteins – G_αi/o, G_αq/11, and G_α12/13, which can result in the activation of a range of well-known, downstream pathways that include Akt, Rho, MAPK, and PLC. These pathways in turn can account for many of the cellular responses initiated by LPA₁ such as changes in cell shape through alterations in the actin cytoskeleton, cell migration, adhesion and cell–cell contact, and Ca ^{mobilization (reviewed in Contos} et al., 2000b; Fukushima et al., 2001; Ishii et al., 2004; Choi et al., 2010; Mutoh et al., 2012; Choi and Chun, 2013).

Expression of Lpar1/LPAR1 is widespread, and can be found in most tissues at various stages of development albeit with non-uniform expression (An et al., 1998; Contos et al., 2000b; Ohuchi et al., 2008; Ye, 2008), particularly within the developing nervous system (reviewed in Contos et al., 2000b; Ishii et al., 2004) where it is found in the neuroproliferative ventricular zone (VZ) as well as superficial marginal zone and meninges (Hecht et al., 1996). By birth, the VZ dissipates as does the expression of Lpar1 in this region; however, it reappears in oligodendrocytes that are involved in myelination.

Knockout mice have provided important insights for most of the lysophospholipid receptors, beginning with Lpar1 ^{mice that exhibit ∼50% perinatal lethality (Contos} et al., 2000a) attributable to olfactory deficits that affect suckling as well as possible central mechanisms that show background strain dependence (Weiner et al., 2001; Estivill-Torrús et al., 2008; Matas-Rico et al., 2008). The developing cerebral cortex in particular is affected by LPA signalling including overall organization (Kingsbury et al., 2003), cell survival, migration, proliferation and process outgrowth (Contos et al., 2000b; Fukushima et al., 2000; 2002; Campbell and Holt, 2001; Kingsbury et al., 2003; Yuan et al., 2003).

Effects on the normal development and organization of the brain have pointed towards LPA influences on central nervous system (CNS) disorders. In particular, neuropsychiatric disorders that could arise prenatally and that could involve bleeding, hypoxia and immunological challenge, as proposed for autism and schizophrenia (Hultman et al., 1999; Cannon et al., 2002; Brimacombe et al., 2007; Byrne et al., 2007), could involve LPA signalling. Proof-of-concept for this idea comes from studies of congenital or fetal hydrocephalus (Yung et al., 2011), one of the most common neurological disorders of newborns and young children, wherein models of FH can be rescued by removal of LPA signalling. Schizophrenia-relevant signals include Lpar1 ^{mutant mice that show deficits in pre-pulse inhibition 5-HT levels and glutamatergic synapses (Harrison} et al., 2003; Santin et al., 2009; Musazzi et al., 2010; Roberts et al., 2005), while a variant mutant, maLPA1^{, display a range of other defects (Harrison} et al., 2003; Estivill-Torrús et al., 2008; Santin et al., 2009; Castilla-Ortega et al., 2010).

Glia are also influenced by LPA₁ signalling. Astrocytes express most LPA receptors (LPA_1–5; Shano et al., 2008), and upon treatment with LPA, initiate a wide range of effects in vitro including morphological changes and stabilization of stress fibres (Manning and Sontheimer, 1997; Suidan et al., 1997; de Sampaio et al., 2008) that may contribute to astrogliosis (Sorensen et al., 2003, reviewed in Noguchi et al., 2009). Neuronal differentiation can also be influenced by LPA₁ and LPA₂ (Spohr et al., 2008). Myelinating cells, oligodendrocytes (Allard et al., 1998; Weiner et al., 1998; Yu et al., 2004) and Schwann cells (SCs) all, express LPA₁ and LPA₂ (Weiner et al., 2001; Kobashi et al., 2006) and Lpar1(−/−) mutants show increased survival via the G_αi-PI3K-Akt pathway (Weiner and Chun, 1999) and higher levels of Schwann cell apoptosis within the sciatic nerves (Inoue et al., 2004). Myelinating cells, oligodendrocytes, and Schwann cells all express LPA1 and LPA2, and Lpar1(−/−) mutants show increased survival via the Gai-P13K-Akt pathway and higher levels of Schwann cell apoptosis within the sciatic nerves.

LPA receptors have also been linked to neuropathic pain (Inoue et al., 2004) using an animal model of partial sciatic nerve ligation (PSNL) in Lpar1 ^{mutants, which may involve demyelination (Inoue} et al., 2004; Fujita et al., 2007). Other LPA receptors also appear to participate, including LPA₅ (Lin et al., 2012). Moreover, autotaxin (gene name ENPP2/Enpp2) that converts lysophosphatidylcholine (LPC) into LPA (Inoue et al., 2008a,b) also affects neuropathic pain animal models, such that Enpp2 ^{mice show protection in a PSNL model (Inoue} et al., 2008a). These observations support roles for LPA signalling in neuropathic pain.

LPA signalling is also found to play a role in obesity and fibrosis. LPA signalling can affect both proliferation and differentiation of pre-adipocytes (Valet et al., 1998; Ferry et al., 2003; Simon et al., 2005; Nobusue et al., 2010), and LPA's effects have been observed in adipocytes, including those from db/db mice (type II diabetes obese-diabetic mice; Ferry et al., 2003; Boucher et al., 2005). Fibrosis links to LPA include in the lung, kidney, and liver (Ikeda et al., 1998; Wu and Zern, 2000; Pradere et al., 2007; 2008; Watanabe et al., 2007; Tager et al., 2008). LPA₁ is expressed on both cancer cell lines and in tumours, where it can have a variety of effects, both cancer promoting and inhibiting (Yamada et al., 2004; Yu et al., 2008; Li et al., 2009; Shin et al., 2009). LPA₁ mutations have been reported in an osteosarcoma cell line (Okabe et al., 2010) and in rat lung and liver tumours (Obo et al., 2009).

LPA₂/LPAR2/Lpar2

LPA₂ is encoded by LPAR2 on chromosome 19 (19p12) and encodes 348 amino acids for a calculated molecular mass of ∼39 kDa (Contos and Chun, 2000). It is ∼50% identical at the amino acid level to LPA₁. Lpar2/LPAR2 is expressed at relatively high levels in leukocytes, kidney, testis, and uterus (An et al., 1998; Contos and Chun, 2000). Relatively low levels are present in most other organs, including the brain (Ohuchi et al., 2008). LPA₂ couples to the same heterotrimeric G-proteins as LPA₁: G_αi/o, G_αq/11, and G_α12/13 (Contos et al., 2000b), and like LPA₁, can promote cell migration and survival (Goetzl et al., 1999; Zheng et al., 2000; 2001; Deng et al., 2002; Panchatcharam et al., 2008). LPA₂ may also produce effects via TRIP6, a focal adhesion molecule (Lai et al., 2005; 2007), and both zinc finger or PDZ-domain protein interactions have been reported (Lin and Lai, 2008), along with MAGI3 and Na^{/H exchanger regulatory factor 2 (NHERF) interactions (Lee} et al., 2011). LPA₂ signalling may inhibit EGF-induced migration of pancreatic cancer cells through G_α12/13/Rho (Komachi et al., 2009). SCs up-regulate myelin markers like P0 protein via LPA₂, including after insult by injury, nerve transection, and in PSNL models of neuropathic pain (Weiner et al., 2001; Inoue et al., 2004). It has also been reported to modulate hippocampal excitatory synaptic transmission (Trimbuch et al., 2009). LPA₂, in conjunction with LPA₁, can also alter cerebral cortical architecture in ex vivo cultures after exposure to exogenous LPA (Kingsbury et al., 2003), effects of which are lost in Lpar1^{/Lpar2 mutant mouse cultures.}

Links to cancer have been reported for LPA₂ in promoting neoplasms based upon designed or observed overexpression (Kitayama et al., 2004; Lee and Yun, 2010). LPA₂ signalling has also been associated with cancer metastasis and colon endometrial, mesothelia, and ovarian cancer cells (Shida et al., 2003; Jeong et al., 2008; Hope et al., 2009). Instances of cancer inhibition in pancreatic cells have also been reported (Komachi et al., 2009). This influence may involve regulation of a range of factors including Akt/Erk1/2, COX-2, epithelial growth factor receptor, metalloproteinases, VEGF, and urokinase-type plasminogen activator (Huang et al., 2004; Yun et al., 2005; Estrella et al., 2007; Jeong et al., 2008; Shida et al., 2008). Loss-of-function for LPA₂ generally appears to be protective against tumourgenesis (Masiello et al., 2006; Estrella et al., 2007; Yu et al., 2008; Zhao et al., 2013).

In the immune system, Lpar2 (similar to Lpar1) is expressed in a variety of immunological organs like the spleen and thymus (Ishii et al., 2004; Kotarsky et al., 2006; Oh et al., 2008), and in lymphocytes (Komachi et al., 2009). LPA₂ is expressed in unstimulated T-cells, as compared with LPA₁ that is predominantly within stimulated T-cells that can influence cell survival (Goetzl et al., 1999). In unstimulated T-cells, LPA₂ is upregulated while LPA₁ is downregulated, leading to LPA-induced chemotaxis and inhibition of (Goetzl et al., 2000; Zheng et al., 2000; 2001). In contrast, activated T-cells upregulate LPA₁ and downregulate LPA₂, leading to inhibited chemotaxis, increased proliferation, and increased IL-2 and IL-13 production upon LPA stimulation (Zheng et al., 2000; Rubenfeld et al., 2006). LPA₂ is also expressed on dendritic cells (Panther et al., 2002; Chen et al., 2006).

LPA₃/LPAR3/Lpar3

LPAR3/Lpar3 was identified based upon homology to defined LPA receptor genes and cloned using a degenerate, PCR-based cloning strategy (Bandoh et al., 1999; Im et al., 2000b). LPAR3 (human chromosomal locus 1p22.3-p31.1) encodes a 353 amino acid, ∼40 kDa GPCR, which in mice is ∼50% identical in amino acid sequence to LPA₁ and LPA₂. LPA₃ couples to the heterotrimeric Gα proteins G_αi/o and G_αq/11 to mediate downstream signalling pathways including adenyl cyclase activation, PLC activation and Ca ^{mobilization, and MAPK activation (Ishii} et al., 2000). LPA₃ appears to prefer 2-acyl-LPA containing unsaturated fatty acids (Bandoh et al., 1999; Sonoda et al., 2002).

LPAR3 is expressed in multiple human organs including the brain, heart, lung, ovary, pancreas, prostate, and testis (Bandoh et al., 1999; Im et al., 2000b), as well as mouse lung, testes, kidney, small intestine, spleen, stomach, and heart (Contos et al., 2000b), and during development (Ohuchi et al., 2008). While Lpar3 null mice are viable, they have defects in the immune system reflecting in part LPA₃-specific dependent activation of chemotaxis of immature, but not mature, dendritic cells (Chan et al., 2007). They also have effects on zebra fish body asymmetry (Lai et al., 2012) and probably are involved in effects of the nervous system including those involving pain (Ma et al., 2009) and possibly other modalities. However, the most dramatic effect is on embryo implantation and fertility.

Lpar3 ^null female mutants have a prominent reproductive system phenotype whereby normal embryo implantation is disrupted (Ye et al., 2005). Within the uterus, Lpar3 is specifically expressed in luminal endometrial epithelial cells where it is markedly up-regulated during the brief window of embryo implantation, following which its expression is rapidly down-regulated (Ye et al., 2005). The hormones oestrogen and progesterone influence this expression pattern (Hama et al., 2006), and may play a role in allowing embryos to implant within the uterus. Lpar3 null mutant mice were found to have abnormal, delayed implantation of embryos that included crowding along the uterine horn and subsequent reductions in live births that could be attributed to maternal effects of LPA₃ loss (Ye et al., 2005). Mechanistic studies demonstrated that LPA₃ promotes COX-2 expression; COX-2 is a rate-limiting enzyme for the production of PGs that are known to be important for fertility, although there is evidence that COX–2-independent functions are involved as well (Hama et al., 2007). This may be relevant for the embryo spacing phenotype in Lpar3 ^{mice that could interface with cytosolic PLA}_2α (cPLA_2α) or Wnt/β-catenin signalling, in view of the reminiscent phenotypes in null-mutants for these genes (Song et al., 2002; Mohamed et al., 2005). In addition to this maternal phenotype, combined loss of LPA_1–3 that are expressed in the testis (Ishii et al., 2004; Ye, 2008) results in loss of germ cells and progeric azoospermia (Ye, 2008), adding to the reproductive spectrum of effects produced by LPA receptor loss from reproductive tissues (reviewed in Ye, 2008).

LPA₄/LPAR4/Lpar4

LPA₄ is notable because it shares less than 20% amino acid sequence identity with LPA_1–3 and S1P_1–5, and is phylogenetically far from them and located near the P2Y receptor family (Figure ​(Figure2).2). Identification of LPA₄ was made by screening orphan receptors, including purine receptor families, using calcium mobilization as a readout for ligand-induced signals (Noguchi et al., 2003). P2Y₉ has ∼20% sequence identity to LPA_1–3 (Noguchi et al., 2003), yet it responds to LPA and not to assayed nucleosides or nucleotides (Noguchi et al., 2003). LPAR4 is located on chromosome Xq21.1 and encodes a 370 amino acid protein of ∼42 kDa, with mouse Lpar4 being present on the D-region of chromosome X. Lpar4 gene expression is observed in the brain, heart, lung, skin, thymus, and uterus (Ishii et al., 2009b). It is also developmentally expressed within the embryonic brain branchial arches, limb buds, liver, maxillary processes, and somites (Ohuchi et al., 2008).

Figure 2

Phylogenetic tree of related GPCRs and amino acid sequence identities. (A) A molecular phylogenetic tree of human GPCRs. The selected GPCR protein sequences were analysed for the phylogenetic reconstruction by the ‘All against All’ sequence programme at the Computational Biochemistry Research Group server of the ETH Zürich. (B) Pair-wise matrices comparing amino acid sequences of lysophospholipid receptors. The upper and lower matrices specify identities among lysophospholipid receptors in human and mouse respectively. The amino acid sequence identities are shown in a gray-to-white gradient. The numbers in the boxes were calculated by Clustal Omega (Sievers et al., 2011).

LPA₄ couples to G_α-proteins G_αs, G_αi, G_αq, and G_α12/13 (Lee et al., 2007), the latter of which activates Rho/ROCK to induce neurite retraction and stress fibre formation seen with activation of other LPA receptors (Lee et al., 2007; Yanagida et al., 2007). It can also induce cell aggregation and adhesion through N-cadherin (Yanagida et al., 2007) and was the first LPA receptor activating G_αs activity (Lee et al., 2007) to promote intracellular cAMP accumulation. LPA₄ can transform cells when co-expressed with oncogenic-promoting genes like c-Myc or Tbx2 (Taghavi et al., 2008). It has also been reported to affect immortalized hippocampal progenitor cells (Rhee et al., 2006).

Null mutant mice for Lpar4 do not show overt abnormalities (Lee et al., 2008) aside from some prenatal loss, probably produced by blood vessel defects that result in abnormal haemorrhage (Sumida et al., 2010). Lymphatic vessels and lymph sacs are also affected during development of the circulatory system (Sumida et al., 2010). Osteoblast differentiation is also inhibited based on cell culture analyses in experiments that knocked down LPAR4 (Liu et al., 2010). Cells from Lpar4^−/− mice show reduced cell motility (Lee et al., 2008).

LPA₅/LPAR5/Lpar5

LPA₅ was the fifth LPA receptor to be reported (Kotarsky et al., 2006; Lee et al., 2006), sharing ∼35% homology with LPAR4, while being more dissimilar to LPAR1–3 (Lee et al., 2006). LPAR5 has a chromosomal location of 12p13.31 and encodes a 372-amino acid protein with a molecular mass of ∼41 kDa, and Lpar5 is located on chromosome 6. LPA₅ couples to G_α12/13 and G_αq (Lee et al., 2006) and is expressed broadly, with high expression in dorsal root ganglia (DRG), gastrointestinal lymphocytes, heart, platelets, and spleen (Kotarsky et al., 2006; Lee et al., 2006; Amisten et al., 2008). It is also expressed developmentally in the embryonic mouse brain (Ohuchi et al., 2008).

LPA₅-expressing cell lines can induce both neurite retraction and stress fibre formation in response to LPA via the G_α12/13 pathway, including clear receptor internalization (Lee et al., 2006). It also activates G_αq, Gai, leading to intracellular calcium levels (Lee et al., 2006), while also increasing cAMP accumulation via a non-G_αs mechanism, based upon minigene experiments, which implicates other G-protein involvement (Kotarsky et al., 2006; Lee et al., 2006). LPA₅ signalling also appears to affect intestinal water absorption (Lin et al., 2010) through effects on intestinal epithelial cells, whereby LPA induces Na^{-dependent water absorption via Na/H exchanger 3 (NHE3; see Alexander} et al., 2013c) and the NHERF2 that recruits NHE3 to intestinal microvilli (Lin et al., 2010). This receptor has also been implicated in neuropathic pain models through mechanisms that appear to be distinct from effects mediated by LPA₁ (Lin et al., 2012).

LPA₆/LPAR6/Lpar6

The latest member of the LPA receptor family is LPA₆. LPA₆ is encoded by LPAR6 on chromosome 13 (13q14) and encodes 344 amino acids for a calculated molecular mass of ∼39 kDa. It is a member of the P2Y receptor family like LPA₄, and was known originally by its orphan name P2Y5, which was identified as a human mutation affecting hair growth (Pasternack et al., 2008). Use of a chimeric G_α13 protein enabled detection of LPA₆-mediated cAMP accumulation and Rho-dependent morphological alterations, as well as [^{H]-LPA binding and LPA-induced [S]-guanosine 5′-3-}O-(thio)triphosphate binding (Yanagida et al., 2009). LPA₆ has some preference for 2-acyl-LPA rather than 1-acyl-LPA. The receptor is distinct from the other five in being somewhat refractory to many cell-based tests, as evidenced by the much higher concentrations of LPA required to get a signal (Yanagida et al., 2009). When co-expressed with a promiscuous G_α protein, which activates G_αi, LPA₆ stimulated with LPA increased intracellular Ca^{, reduced forskolin-stimulated cAMP and ERK1/2 activation (Lee} et al., 2009).

LPA₆ was initially identified as being an autosomal dominant genetic factor for hypotrichosis simplex, a complex of rare diseases characterized by familial hair loss in humans. Independent studies identified LPA₆ mutations in hypotrichosis patients (Pasternack et al., 2008; Shimomura et al., 2009; Nahum et al., 2011). Conceptually linked reports have implicated lipase member H, associated with decreased LPA production in culture studies that then fail to activate LPA₆ (Pasternack et al., 2009; Shinkuma et al., 2010). More recent analyses of this receptor by use of a TGFα shedding assay (Inoue et al., 2012) validate it as an atypical, but legitimate, LPA receptor.

Sphingosine 1-phosphate (S1P) receptors

S1P is a pleiotropic bioactive lipid that is an important regulator of many physiological processes including proliferation, migration, survival, and differentiation and plays important roles in disorders of the immune system and CNS (Maceyka et al., 2012). Most of the actions of S1P are mediated by five specific cognate GPCRs, designated S1P₁-S1P₅ (Chun et al., 2010; Blaho and Hla, 2011). These receptors bind S1P and dihydro-S1P with high affinity and there is very little evidence for additional endogenous ligands. We have summarized the experimental pharmacological tools for S1P receptors in Table ​Table44.

S1P₁/S1PR1/S1pr1

S1P₁ was one of the first S1P receptors to be functionally identified (Lee et al., 1998b) and it is the most well studied. Early studies suggested that it might mediate actions of LPA based on its sequence and function (Lee et al., 1998a); however, it is now known to be a selective S1P receptor. S1PR1 is located on chromosome 1 (1p21) and encodes a 382-amino acid of ∼43 kDa that is highly conserved and has 94% sequence identity with the murine receptor. S1PR1 is the only S1PR that couples exclusively to G_αi/o. Although S1PR1 is ubiquitously expressed (Zhang et al., 1999; McGiffert et al., 2002), its most important functions are in the regulation of trafficking of lymphocytes and other haematopoietic cells and vascular development and integrity. Genetic and pharmacological approaches, together with sophisticated intravital staining, have established that S1P₁ controls the trafficking and migration of numerous types of haematopoietic cells, including T and B lymphocytes, NK T-cells, dendritic cells, macrophages, neutrophils, haematopoietic progenitors, mast cells, and osteoclasts (Matloubian et al., 2004; Spiegel and Milstien, 2011; Cyster and Schwab, 2012), in both homeostatic and disease settings. Blood and lymph contain high nM levels of S1P, which form a gradient between the much lower levels in tissues (Pappu et al., 2007; Pham et al., 2010). When S1P₁ on lymphocytes recognizes high levels of S1P in the blood and lymph, egress of the cells from lymphoid organs into the blood is promoted through activation of the G_αi-phosphatidylinositol-3-kinase pathway and the small GTPase Rac (Spiegel and Milstien, 2011; Cyster and Schwab, 2012). Down-regulation or desensitization of S1P₁ enables lymphocytes to subsequently migrate from the blood into tissues (Schwab and Cyster, 2007).

The immunomodulatory drug FTY720/fingolimod, which has been approved by the Food and Drug Administration for the treatment of relapsing forms of multiple sclerosis (MS) (Chun and Hartung, 2010; Chun and Brinkmann, 2011; Cohen and Chun, 2011), is phosphorylated in vivo to FTY720-P, producing the active form of the drug (Brinkmann et al., 2010). FTY720-P is a structural analogue of S1P and an agonist of S1P₁, S1P₃, S1P₄, and S1P₅. However, persistent activation of S1P₁ by FTY720-P causes its internalization and degradation and thus it acts as a functional antagonist (Graeler and Goetzl, 2002; Matloubian et al., 2004; Oo et al., 2007; Brinkmann et al., 2010; Gonzalez-Cabrera et al., 2012). Down-regulating surface expression of S1P₁ on lymphocytes prevents their egress from lymphoid organs and reduces peripheral blood lymphocyte levels (Brinkmann et al., 2010; Gonzalez-Cabrera et al., 2012). Concomitantly, direct CNS actions may be relevant to MS through S1P₁ expressed on astrocytes, since conditional removal of this receptor reduces MS-like disease in animals and attenuates FTY720 activity (Choi et al., 2011). Expression of this and other S1P receptors in the CNS supports other activities relevant to MS, and perhaps other CNS disorders (Gardell et al., 2006; Herr and Chun, 2007; Noguchi and Chun, 2011; Soliven et al., 2011; Mutoh et al., 2012; Choi and Chun, 2013; Groves et al., 2013).

S1P₁ maintains the integrity of the vascular system (Liu et al., 2000; Camerer et al., 2009; Wang and Dudek, 2009; Abbasi and Garcia, 2013), which is critical for homeostasis and to prevent extravasation of plasma during infections, sepsis and anaphylactic shock, which can be life threatening. Blood S1P enhances vascular barrier function by ligation of S1P₁ with subsequent downstream activation of the Rho family of small GTPases, cytoskeletal reorganization, adherens junction and tight junction assembly, and focal adhesion formation (Wang and Dudek, 2009; Abbasi and Garcia, 2013). Depletion of blood S1P in mice induces basal vascular leak and increases lethal responses in anaphylaxis induced by administration of platelet-activating factor or histamine (Camerer et al., 2009). It has been suggested that either S1P continuously activates luminal endothelial S1P₁ to maintain tight cell–cell junctions or alternatively, following entry of S1P into the sub-endothelial space via ‘leaky’ endothelium, dynamic S1P₁ signalling activates abluminal surface S1P₁ to close intercellular gaps. Furthermore, the S1P/S1P₁ axis also attenuates LPS-induced acute lung injury in murine and canine models (Wang and Dudek, 2009; Abbasi and Garcia, 2013). Deciphering the mechanisms by which the S1P₁ signalling pathway regulates endothelial barrier integrity will help our understanding and treatment of acute inflammatory diseases.

The vital role of S1P₁ in vascular maturation and development was demonstrated by knockout of the S1pr1 gene in mice that die in utero because of a defect in the association of mural cells with nascent vessels and incomplete coverage (Liu et al., 2000; Allende et al., 2003). More recently, the role of S1P₁ in angiogenesis, the development of new blood vessels, has been slightly revised. It was shown that S1P₁ in fact acts independently of mural cells in an endothelial cell-autonomous manner to inhibit sprouting angiogenesis (Shoham et al., 2012). Endothelial S1P₁ stabilizes the primary vascular network during development and homeostasis (Gaengel et al., 2012; Jung et al., 2012).

Recently, the crystal structure of S1P₁ fused to T4-lysozyme in complex with an antagonist was solved to 2.8 Å resolution (Hanson et al., 2012). Intriguingly, this receptor has a novel N-terminal fold that blocks access of S1P to the binding pocket from the extracellular environment. Therefore, S1P must gain access by entering laterally between helices I and VII within the transmembrane region of S1P₁. This work provides the first view of the molecular recognition of S1P (Hanson et al., 2012; Rosen et al., 2013) and may aid in the development of S1P₁-specific drugs as well as providing a basis for determining the structure of the other S1P receptors.

S1P₂/S1PR2/S1pr2

Now denoted as S1P₂, this receptor was previously known as Edg-5, H218, AGR16, and lp_B2 and was one of the first to be identified as an S1P receptor (An et al., 1997). The human gene, S1PR2, is located on chromosomal locus 19p13.2 and its sequence is highly conserved across species, with the human receptor containing 353 amino acids and a receptor of ∼39 kDa compared with the murine transcript with 352 (also ∼39 kDa). The S1P₂ gene is expressed in a variety of tissues (Zhang et al., 1999; McGiffert et al., 2002) and can couple to multiple G-proteins, although it most efficiently utilizes G_α12/13 to activate the small GTPase Rho. Thus, S1P₂ typically inhibits motility through inhibition of Rac. S1P₂ has been shown to be involved in S1P-induced cell proliferation, motility and transcriptional activation, usually acting in opposition to S1P₁ (Skoura and Hla, 2009; Chun et al., 2010).

S1P₂ was initially shown to be required for heart development in zebrafish (Kupperman et al., 2000). It was subsequently reported that S1P₂ signals through the G_α13/RhoGEF pathway to promote the migration of myocardial precursor cells (Ye and Lin, 2013), although S1pr2 knockout mice are viable (Ishii et al., 2002), demonstrating species differences. However, these null mutants have multiple severe inner ear defects, leading to deafness and balance problems (Herr et al., 2007; Kono et al., 2007). Using an S1P₂ antagonist, JTE013, it was shown that S1P₂ promotes vasoconstriction of the spiral modiolar artery, which protects the stria vascularis capillary bed of the inner ear from high perfusion pressure. Several other studies have linked S1P₂ to vascular development and remodelling. S1P₂ is induced in endothelial cells undergoing hypoxic stress and mice lacking both S1pr1 and S1pr2 exhibit substantially more vascular defects than S1pr1 knockout alone, suggesting that the two receptors may act coordinately during vascular development (Kono et al., 2004). Experiments in developing zebrafish, which have S1PR homologues and S1P levels in the blood that are higher than the K_D of the receptors, showed similar results. S1pr1 knockdown interfered with the development of the intersegmental vessels, and this phenotype was enhanced when S1pr2 was suppressed (Mendelson et al., 2013).

S1P₂ has also been suggested to play a role in endothelial barrier integrity. In an LPS-induced model of acute lung injury, S1P₂ deletion reduced oedema while activation of S1P₁ with a specific agonist also reduced oedema (Sammani et al., 2010), suggesting that S1P₂ reduces endothelial barrier function in contrast to S1P₁, which enhances it. In mice, S1P₂ can also promote the recovery from anaphylactic shock, at least in part through counteracting the histamine-induced vasodilatation responsible for hypotension (Olivera et al., 2010; 2013). Accordingly, histamine initiates a negative feedback loop, stimulating production of S1P that acts through S1P₂ to increase clearance of histamine by the kidney through excretion.

S1P₂ also plays a role in bone maintenance. Bone is remodelled throughout life, with osteoblasts forming bone and osteoclasts resorbing it. Osteoclast precursor cells migrate dynamically between bone and blood, which is controlled by the balance between S1P signalling through S1P₁ versus S1P₂. While S1P₁ promotes osteoclast migration from bone towards high blood levels of S1P (Ishii et al., 2009a), migration away from bone is negatively controlled by S1P₂ (Ishii et al., 2010). Insight into how the balance of S1P receptor expression controls bone remodelling was provided by the demonstration that calcetriol, the active form of vitamin D that promotes bone growth, reduces S1P₂ expression on osteoclasts (Kikuta et al., 2013). This balance between S1P₁ and S1P₂ that controls traffic of cells into and out of tissues is becoming paradigmatic. Cyster and colleagues showed that S1P₂ promotes the retention of B cells in the germinal centres of lymphoid follicles at the low end of an S1P gradient (Green et al., 2011). Moreover, S1P₂ also plays a role in controlling growth and apoptosis of germinal centre B cells through inhibition of Akt (Green et al., 2011).

S1P also has an important role in muscle regeneration through activation of muscle stem cells called satellite cells (Rapizzi et al., 2008). Saba and colleagues demonstrated that S1P biosynthesis is up-regulated following muscle injury (Loh et al., 2012) and activation of S1P₂, but not S1P₁, promoted muscle regeneration by activating STAT3, which in turn down-regulates the cell cycle inhibitors p21 and p27 allowing for satellite cell growth (Loh et al., 2012). Moreover, Mdx mice, a model for muscular dystrophy, have higher levels of S1P-metabolizing enzymes and lower circulating levels of S1P. However, using a different model of muscle injury induced by bupivacaine, S1P₂ was not found to be involved in muscle regeneration (Danieli-Betto et al., 2010). It was suggested that S1P₃ promoted, while S1P₁ inhibited, muscle regeneration. The conflicting data concerning the specific S1P receptors involved may be due to the different models used or the timing of S1P receptor activation.

S1P₂ has also recently been implicated in promoting metastasis. Using genetic and pharmacological approaches, it was shown that bladder cancer xenografts increased systemic S1P levels. This S1P in turn activated S1P₂, leading to the down-regulation of Brms1, a known suppressor of metastasis (Ponnusamy et al., 2012). Thus, inhibition of systemic sphingosine kinase 1 and production of S1P and/or S1P₂ signalling increased Brms1 expression suppressing lung metastasis (Ponnusamy et al., 2012).

S1P₃/S1PR3/S1pr3

S1P₃, previously known as Edg-3 and lp_B3, was also an early identified S1P receptor (An et al., 1997), with human S1PR3 located at chromosomal locus 9q22.1-q22.2, encoding a 378-amino acid protein of ∼42 kDa, with seven predicted transmembrane domains. It shares 87% identity with the murine S1P₃ receptor.

Like S1P₂, S1P₃ can couple to multiple G-proteins, including G_αi/o, G_αq, and G_α12/13 (Chun et al., 2010), although in cells it most commonly couples to G_αq, leading to the generation of inositol trisphosphate and diacylglycerol with subsequent calcium mobilization and activation of PKC respectively.

Despite fairly broad gene expression (Zhang et al., 1999; McGiffert et al., 2002), global deletion of S1pr3 in mice did not reveal an obvious phenotype or developmental defects (Ishii et al., 2001), although the S1pr2/3 double knockouts have reduced fertility (Ishii et al., 2002). Initially, S1P₃ was reported to be highly expressed in breast cancer models where it plays a positive role in cell migration (Chun et al., 2010). Moreover, increased expression of S1P₃ in oestrogen receptor (ER)-positive tumour samples correlated with decreased disease-free survival times (Watson et al., 2010). One possible explanation for this is the intriguing finding that in breast cancer cells, oestrogen stimulates S1P release and activation of S1P₃ (Sukocheva et al., 2006). This then increases the activity of MMP9, resulting in the release of EGF to signal in an autocrine manner. Additionally, in this system, S1P₃ also activates Cdc42 and decreases degradation of, and increases signalling from, the EGF receptor (Sukocheva et al., 2013). Interestingly, an S1P₃-blocking monoclonal antibody, 7H9, has been developed that blocks the growth of breast cancer tumours in a xenograft model (Harris et al., 2012).

S1P₃ has also been implicated in sepsis. Signalling of the protease-activated receptor 1 on dendritic cells promotes the inflammatory response in sepsis syndrome. Treatment with S1P₃-specific antagonists, as well as S1P₃ deletion, protects from LPS-induced lethal sepsis (Niessen et al., 2008; Sammani et al., 2010). Although activation of S1P₁ increases endothelial barrier enhancement, S1P₃ disrupts it (Sammani et al., 2010). Indeed, recent studies associate increased S1P₃ expression with sepsis and mortality of intensive care patients (Sun et al., 2012). Finally, several studies indicate that S1P₃ is involved in liver fibrosis. S1P, acting through both S1P₁ and S1P₃, promotes the motility of hepatic stellate cells and their differentiation into hepatic myofibroblasts (Liu et al., 2011), and enhances liver angiogenesis associated with fibrosis (Yang et al., 2013).

S1P₄/S1PR4/S1pr4

S1PR4 is located at chromosomal locus 19p13.3, previously known as Edg-6 and lp_C1 (Contos et al., 2002) and encodes a 384-amino acid protein of ∼42 kDa in humans that is highly homologous across mammalian species (Van Brocklyn et al., 2000).

S1P₄ couples to G_αi and G_α12/13 and promotes cell migration (Graler et al., 2003; Kohno et al., 2003). S1P₄ has a restricted tissue distribution and is expressed mainly in haematopoietic tissue, though it was recently reported to be in other tissues, such as the muscle satellite cells, where together with S1P₁, it promotes migration in response to S1P (Calise et al., 2012). Expression of S1pr4 has also been reported in rat lungs, but not in renal or mesenteric arteries, and the S1P₄ agonist {"type":"entrez-protein","attrs":{"text":"VPC23153","term_id":"1643590030","term_text":"VPC23153"}}VPC23153 promoted vasoconstriction of both normotensive and hypertensive pulmonary arteries (Ota et al., 2011). Moreover, expression of S1P₄ in ER-negative breast cancer cells correlated with poorer prognosis (Ohotski et al., 2012).

S1P₄ is also important in megakaryocytes where it is highly induced upon differentiation. Although S1P₄ knockout mice have normal platelet levels, their ability to generate platelets after experimentally-induced thrombocytopenia is delayed, suggesting a role for S1P₄, in thrombopoiesis (Golfier et al., 2010). Also in these mice, T-cell proliferation and cytokine secretion are not significantly altered (Schulze et al., 2011). Interestingly, S1pr4 knockout mice also have differential responses in various models of inflammation with exacerbated Th2-mediated responses, but reduced Th1-mediated responses. These changes were linked to altered dendritic cell functions, including decreased IL-6 production and IL-17 secretion. S1pr4 deletion also decreased neutrophilia, suggesting a potential role for this receptor in neutrophil migration (Allende et al., 2011).

S1P₅/S1PR5/S1pr5

Previously known as Edg-8, lp_B4, and Nrg-1, S1PR5 is located at chromosomal locus 19p13.2 and encodes a highly conserved 398-amino acid protein with a calculated molecular mass of ∼39 kD with tissue expression primarily restricted to brain and spleen (Im et al., 2000a; Malek et al., 2001). Like other S1P receptors, it couples to multiple G-proteins, although in its common role of inhibiting migration and promoting cell retraction, it couples to G_α12/13. S1P₅ knockout mice are viable and fertile. Intriguingly, they show greatly decreased numbers of circulating NK cells (Walzer et al., 2007). Similar to the role S1P₁ plays in T and B cell trafficking, S1P₅ promotes the egress of NK cells from bone marrow and lymph nodes into blood and other tissues. Moreover, S1P₅ is required for NK recruitment to sites of inflammation (Walzer et al., 2007; Jenne et al., 2009). Furthermore, during NK cell differentiation, S1P₅ is expressed, allowing exit from the bone marrow (Mayol et al., 2011). S1P₅ knockout mice also lack circulating Ly6C-negative peripheral monocytes, but have normal levels in the bone marrow (Debien et al., 2013). Interestingly, although S1P₅ is required for egress of these cells, S1P is not a chemoattractant, suggesting that S1P₅ may act during their differentiation.

New lysophospholipid receptors

Efforts to de-orphanize GPCRs led to the identification of putative new members of the lysophospholipid receptor family. These receptors interact with two distinct glycerophospholipids: LPI and LysoPS. Newer technologies to identify receptors, such as the TGFα shedding assay, are being developed and used successfully for both de-orphanization and correction or augmentation of lysophospholipid identities.

LPI receptor: LPI₁/LPIR1/Lpir1 (orphan GPR55)

Orphan receptor GPR55 had originally been reported to be a novel cannabinoid receptor (Lauckner et al., 2008); however, it appears that this receptor may in fact act as a LPI receptor based upon recent evaluations (Kotsikorou et al., 2011; Inoue et al., 2012; Aoki, Inoue and colleagues, unpublished). In view of these data, we consider GPR55 as a provisional LPI receptor with receptor name LPI₁ and gene names LPIR1/Lpir1 for human and non-human genes respectively. LPIR1 is located on human chromosome 2 (2q37) and encodes a 319-amino acid protein (∼37 kDa). It is currently unclear whether this receptor genuinely acts as a cannabinoid receptor, and efforts are underway to better determine the ligand specificity of this GPCR.

Proposed LysoPS receptors

The following receptors have shown activity using a TGFα shedding assay (Inoue et al., 2012), which strongly support their identity as LysoPS receptors; however, this identity should be considered provisional. In addition, the name of the receptors may require future modification: LysoPS_x is utilized here to avoid confusion with lipopolysaccharide that is commonly referred to as LPS. The lysophospholipid LysoPS, has been known as an immune cell stimulus, leading to identification of the first LysoPS receptor from mast cells via de-orphanization of the P2Y family of GPCRs known as GPR34 (Sugo et al., 2006). LyPSR1 is located at chromosomal locus Xp11.4 and encodes a 381-amino acid protein for a calculated molecular mass of ∼44 kD. Receptor identity was confirmed using the TGFα shedding assay (Inoue et al., 2012; Kitamura et al., 2012; Makide and Aoki, 2013), although there is some disagreement in the literature on the veracity of this identity (Ritscher et al., 2012). Genetic deletion of GPR34 does result in immunological dysfunction (Liebscher et al., 2011), consistent with the immunological effects of LysoPS, and combined with positivity in the TGFα assay, its designation as LysoPS₁ appears to be warranted. LysoPS₁ has been implicated in other cell types such as microglia in the brain (Bedard et al., 2007), and has been linked to diseases or disorders, including a form of night blindness (Jacobi et al., 2000) and cancers of both immune (Ansell et al., 2012) and non-immune origin (Yu et al., 2013). Through the use of the TGFα shedding assay as a screening tool, three other receptors were identified, the first of which was another P2Y orphan receptor, P2Y₁₀. LyPSR2 is located on human chromosome X (Xq21.1) and encodes a 339 amino acid protein (∼39 kDa). Consistent with the biological effects of LysoPS on the immune system and data from analyses of LysoPS₁, LysoPS₂ also influences the immune system, and appears to show restricted expression in dendritic cells derived from monocytes (Berchtold et al., 1999) and lymphoid lineages (Rao et al., 1999). LysoPS₃/LyPSR3/Lypsr3, another orphan receptor (formerly GPR174), was identified as a third LysoPS receptor by TGFα assay (Inoue et al., 2012) and independently supported by classical assays (Sugita et al., 2013). LyPSR3 is located near the LPAR4 and LyPSR2 genes (Xq21.1) and encodes a 333 amino acid protein of ∼39 kDa, which shares about 45% identity with LysoPS₂. LyPSR3 has recently been reported as a genetic risk locus for Graves disease (Zhao et al., 2013). During TGFα screening analyses of orphan GPCRs, a mouse cDNA not present in humans, A630033H20, was identified as a LysoPS receptor with predicted homology to LysoPS₂ (Inoue et al., 2012). This gene is located between Lypsr2/p2ry10 and Lypsr3/GPR174 on mouse chromosomal locus Xq21.1, which corresponds to the human P2RY10P2 pseudogene. Therefore, nomenclature for a mouse-specific receptor and consequent gene names is neither proposed nor discouraged. A number of lysophospholipid receptor mutants or variants have been reported, such as the mRec1.3 mutant of LPA₁ (Contos et al., 2000b; Fukushima et al., 2001) or the original sequence for S1pr3 (Edg-3) that was a variant form present in a cancer cell line (An et al., 1997), and there is currently no uniform recommendation for naming these receptor variants, which could be a topic for future nomenclature efforts.

Concluding remarks

This nomenclature review for lysophospholipid receptors incorporates the recommended, as well as the most common uses of protein and gene names. For receptor proteins, the simple use of the cognate ligand immediately followed by a subscript to designate a receptor subtype is easily extended to receptors for other lysophospholipid ligands, as illustrated by the additions of LPI₁ and LysoPS_1–3, as was first used for this family based upon IUPHAR recommendations. To easily differentiate proteins from genes and provide an accurate interface with sequence databases such as ENCODE (Maher, 2012; Skipper et al., 2012), the italicized use of the HGNC and MGI nomenclatures are recommended for human and non-human genes respectively. This nomenclature will accommodate the likely addition of new members to the lysophospholipid receptor family via both de-orphanization and revised receptor identities.