OUP user menu

Platelet-released phospholipids link haemostasis and angiogenesis

Denis English, Joe G.N. Garcia, D.N. Brindley
DOI: http://dx.doi.org/10.1016/S0008-6363(00)00230-3 588-599 First published online: 16 February 2001


Considerable attention has focused on identifying mediators of neovascularization at sites of growth and abnormal tissue development. By contrast, mediators of angiogenesis at sites of injury and wound repair are not well defined but factors generated during blood coagulation (haemostasis) are attractive candidates. In addition to proteins generated, activated and released during the activation of clotting cascades, platelet-derived lipid mediators are now known to play a key role in many aspects of the angiogenic response. The first indication of lipid mediator involvement in angiogenesis was the discovery that lysophosphatidate (LPA), phosphatidic acid (PA) and sphingosine 1-phosphate (SPP) are high affinity agonists for G-protein coupled EDG (endothelial differentiation gene) receptors. The prototype for this family, EDG-1, was cloned from genes expressed when endothelial cells were activated to assume an angiogenic phenotype in vitro. The subsequent finding that SPP is a high affinity ligand for EDG-1 led Spiegel, Hla and associates (Lee et al., Science 1998;279:1552–1555) to hypothesize that platelet-released phospholipids play an important role in angiogenesis. These investigators and others demonstrated that SPP, LPA and phosphatidate (PA) induce many important endothelial cell responses associated with angiogenesis, including liberation of endothelial cells from established monolayers, chemotactic migration, proliferation, adherens junction assembly and morphogenesis into capillary-like structures. Although these studies indicated the potential involvement of platelet-derived phospholipids in angiogenesis, their physiological importance was not established. However, recent work demonstrates that <80% of the potent endothelial cell chemoattractive activity generated in human serum during clotting — an activity necessary for optimal angiogenesis — results from platelet-derived SPP. Other factors released from platelets during clotting, including LPA and PA, exert profound effects on endothelial cells that contribute unique aspects to the angiogenic response. These combined studies establish that SPP and other platelet-derived lipid mediators provide a novel link between haemostasis and angiogenesis.

  • Angiogenesis
  • Hemostasis
  • Platelets
  • Endothelial receptors

Time for primary review 25 days.

1 Introduction

Biologically active lipids that are released from platelets exert a variety of profound effects on cells and have been useful for defining biochemical events which couple cellular stimulation to functional activation, or signal transduction [1–4]. These lipids, including lysophosphatidic acid (LPA), sphingosine 1-phosphate (SPP) and phosphatidic acid (PA), exert their effects by ligation of cell surface receptors. While several families of receptors that respond to these lipids have been identified, much attention has recently been directed toward defining the role of the EDG (endothelial differentiation gene) family of receptors in cellular responses to platelet-released phospholipids. The prototype of this family, EDG-1, was cloned by Hla and Maciag [5] from genes expressed in endothelial cells stimulated to assume an angiogenic phenotype in vitro. Its structure was consistent with that of a G-protein-coupled receptor. Its ligand, while unknown, presumably played a key role in the angiogenic response.

While several years would lapse before the natural ligand for EDG-1 was identified, ligands for other members of the EDG family of receptors were identified earlier. The seminal report in this regard was the cloning by Hecht et al. [6] of a uniquely expressed gene, ventricular zone gene-1 (vzg-1), in neocortical neuroblast cell lines. cDNA for vzg-1 hybridized to an mRNA transcript that encoded a protein with a predicted molecular mass of 41–42 kDa. Overexpressed in the cell from which it was cloned, vzg-1 induced serum-dependent cell morphogenic differentiation, an effect attributed to LPA present in serum. Indeed, vzg-1 overexpression increased specific LPA binding to these neuroblasts, suggesting a role for LPA signalling in cortical neurogenesis.

The reports described above opened the door to the identification of a family of receptors for bioactive phospholipids; receptors that, once occupied, evoked numerous cellular functions, including morphogenic differentiation, proliferation, adherence and chemotaxis. An and Goetzl [7] quickly identified vzg-1 as a homologue of EDG-1, and cloned human EDG-2 as a functional LPA receptor. EDG-2 is widely distributed in mammalian tissue, with high levels of expression in cardiovascular tissue, cells of the central nervous system, gonadal tissue and gastrointestinal tract (see Table 1). Several months later, these investigators cloned human EDG-3 and its homologue, H218, and identified them as high affinity receptors for SPP [8]. EDG-3 is widely expressed, with highest levels observed in endothelial cells, leukocytes and cardiovascular tissue. H218 was subsequently identified as EDG-5 and displays a more restricted pattern of expression (Table 1). An and Goetzl [9] then cloned human EDG-4, which responded to both LPA and PA, but showed no activity when challenged with SPP. EDG 4 was found to be 46% identical and 72% similar in amino acid sequence to human EDG-2, and it is primarily expressed in leukocytes and testes. Recent reports indicate the existence of three additional receptors with EDG homology sequences, EDG-6, 7 and 8 [10–12]. EDG-6 and -8 bind SPP while EDG-7 responds exclusively to LPA. Other lysophospholipid receptors that appear not to be related to the EDG receptor family include PSP24, cloned by Tigyi and associates [13] from Xenopus oocyte cDNA. The influence and mechanism of action of SPP and related lipids on cell function and their coupling to G-proteins, as well as a discussion of the relationship between various members of the EDG receptor family, has recently been summarized in an excellent review by Pyne et al. [14]. The results of various studies carried out to date suggest that members of the EDG family of receptors play important roles in conveying responses transmitted by extracellular phospholipids and lysophospholipids, especially the biologically active lipid phosphates released from platelets.

View this table:
Table 1

Ligands, tissue distribution and G-protein coupling of receptors for platelet-derived phosphorylated lipid growth factorsa

ReceptorHigh affinityLow affinityTissue distributionG-protein
EDG-1SPPLPA ???UbiquitousGi
EDG-2LPAPACV, GI, CNS, gonadal tissueGi, G12/13
EDG-3SPPLeukocytes, EC, other tissueGi, Gq, G12/13
EDG-4LPA<PALeukocytes, testicular tissueGi, Gq, G12/13
EDG-5SPPCV, CNS, gonadal, PI, AECGi, Gq, G12/13
EDG-6SPPLeukocytes, BM, spleen, Thy, LymphGi
EDG-7LPACV, GI, testicular, prostate???
EDG-8SPPSPCSpleen, brainGi
  • a Abbreviations used: CV, cerebrovascular tissue; GI, gastrointestinal tissue; AEC, aortic endothelial cells; CNS, central nervous system tissue; EC, endothelial cells; Lymph, lymphocytes; SPP, sphingosine 1-phosphate; LPA, lysophosphatidic acid; PA, phosphatidate.

2 Postulated role of EDG receptors in angiogenesis

Thus, by early 1998, multiple, EDG-related receptors that respond to platelet-derived biologically active lipid phosphates, including LPA, SPP and PA had been identified [15]. Paradoxically, the natural ligand for the prototype of this receptor family, EDG-1, remained undefined, as did the involvement of EDG receptors and their agonists in endothelial cell activation and angiogenesis. This gap closed rapidly later that year, when three groups identified SPP as a high affinity agonist for EDG-1 [16–18]. Prominent among these is the study by Spiegel, Hla and associates [16] demonstrating that overexpression of EDG-1 conferred upon HEK cells a unique ability to differentiate when exposed to purified SPP. Cells transfected with EDG-1 also responded with morphogenic differentiation upon exposure to foetal bovine serum, but serum treated with activated charcoal or extracted with butanol to remove biologically active lipids was inactive. Of several lipids present in serum, including LPA, PA, arachidonic acid and its derivatives, and SPP, only SPP was effective in restoring reactivity to charcoal-stripped serum. Expression of EDG-1 also conferred upon cells the ability to bind SPP specifically with a disassociation constant of approximately 10 nM. Neither LPA, nor PA, was effective in competing for these specific binding sites. The authors therefore postulated a key role for SPP and related lipids in angiogenesis. The hypothesis was based on the high affinity and striking specificity of the EDG-1 receptor for its ligand, the origin of the genetic material used to clone the cDNA 8 years earlier, and the striking morphogenic changes induced upon receptor ligation. However, at the time, the effects of SPP on endothelial cell angiogenic responses had not been assessed.

In conclusion, the above results set the stage for the definition of the role of EDG receptors and their response to platelet-released phospholipids in the induction of angiogenesis. Thus, while early studies by Hla and associates demonstrated that these receptors were expressed in endothelial cells undergoing angiogenic differentiation, later results would show that SPP and other platelet-derived lipids were high affinity ligands for members of this unique family of cellular receptors. A thorough analysis of the effects of platelet-derived phospholipids on endothelial cell structure and function would confirm the involvement of EDG receptors and their ligands in angiogenesis.

3 Effects of platelet-derived phospholipids and on endothelial cells

The first study to examine the influence of platelet-derived phospholipids on endothelial cell activation was reported by Natarajan et al. in 1994 [19]. The authors demonstrated that sphingosine (Fig. 1), in a dose and time-dependent manner stimulated the hydrolysis of phosphatidylcholine in bovine pulmonary artery endothelial cells (BPAEC) by activation of phospholipase D, an effect that was not abated by inhibition of protein kinase C. Among various sphingolipids, only SPP mimicked the effect of sphingosine on endothelial cell PLD, resulting in accumulation of PA. While the effects of sphingosine on activation of endothelial cell PLD were not inhibited by pretreatment of cells with pertussis toxin, the effect of pertussis toxin on responses to SPP was not assessed. This distinction becomes important in light of subsequent studies showing that many of the effects of SPP on endothelial cells are, indeed, pertussis toxin sensitive, but can be mimicked by sphingosine acting by inducing divergent signalling pathways.

Fig. 1

Metabolism and inter-relation of various sphingolipids. The figure emphasizes the intracellular metabolic relationship of sphingomyelin, ceramides, sphingosine and sphingosine 1-phosphate (SPP). Also shown are the some of the metabolic steps that are involved in the synthesis and degradation of the bioactive sphingolipids.

A subsequent study also focused on the influence of sphingosine on vascular function. Noting that little was known about the effects of sphingomyelin metabolites (see Fig. 1) on the function of the intact endothelium, Murohara et al. [20] examined the vascular effects of sphingomyelin, sphingosine and sphingomyelinase in vitro. Strikingly, sphingosine-mediated contraction of coronary rings was abolished in rings denuded of endothelium. In contrast, sphingomyelin did not alter endothelium-dependent relaxation. Thus, the effects of sphingosine on vasoconstriction were mediated by vasoconstrictor substances released from the endothelium, possibly including PA. The authors concluded that the sphingomyelin pathway could be an important mediator of vascular function.

In 1996, Meyer zu Heringdrof et al. [21] examined the effect of G-protein coupled sphingolipid receptors on endothelial cell Ca2+ mobilization. Although the receptor for SPP was not defined at the time, these insightful investigators demonstrated that SPP effected a rapid and transient increase in intracellular Ca2+ concentrations as a result of mobilization of Ca2+ from intracellular stores. Since this response was markedly abated by preincubation of cells with pertussis toxin, the investigators concluded that the response was driven by a receptor associated with the Gi/o family of G-proteins. That a novel signalling pathway was responsible for this effect was evident from the finding that SPP failed to activate phospholipase C in bovine endothelial cells. In addition, phospholipase D was not activated in endothelial cells exposed to SPP in this system. Whereas these observations contrast with those discussed above by Natarajan et al. [19], the results indicate strongly that extracellular SPP is able to activate endothelial cells through a pertussis toxin-sensitive receptor. Of various related lipids, only sphingosylphosphorylcholine (SPPC) induced a similar increase in Ca2+ mobilization in BAEC and this effect, like that induced by SPP, was pertussis toxin sensitive. However, while SPP induced half maximal effects on Ca2+ mobilization at 0.8 nM, the concentration of SPPC required to attain half maximal Ca2+ mobilization was 260 nM. The authors concluded that both SPP and SPPC are novel and potent endothelial cell agonists that exert their effects by receptor-dependent mechanisms. Since SPP is released from activated platelets [22,23], Meyer zu Heringdrof et al. [21] prophetically surmised that SPP represents a novel mediator of platelet interactions with the endothelium, a conclusion that would be confirmed by many subsequent studies. The mechanism of SPP release from activated platelets is not entirely clear. Platelets avidly accumulate sphingosine from the extracellular medium and convert it to SPP via the action of sphingosine kinase [22] (see Fig. 1). A large portion of platelet SPP is released upon cellular activation, potentially as a result of activation of protein kinase C [23]. However, further study will be required to fully discern the events involved in platelet release of SPP.

In 1997, Panetti and Mosher [24] reported that LPA was a potent endothelial cell mitogen. Results obtained with LPA were contrasted with the known endothelial cell mitogen, FGF-b. The authors found that LPA and FGF-b stimulated endothelial cell proliferation by ligating distinct receptors that elicit different signalling pathways. Maximal stimulation was achieved with doses of LPA ranging from 1 to 30 μM; higher doses were required for optimal stimulation of human fibroblasts. It is now known that the requirement for μM LPA concentrations to activate cells compared to low nM LPA concentrations required to saturate its EDG receptors result from the effects of mM extracellular Ca2+ present in biological fluids and incubation media [25]. Ca2+ chelation of LPA thereby decreases its effective concentration and ability to stimulate cell signalling. The work of Panetti and Mosher [24] therefore shows that endothelial cells appeared to be responsive to LPA, an effect that may have physiological relevance. Indeed, the effects of LPA were inhibited by preincubation of the cells with both thrombospondin-1 and thrombospondin-2, known regulators of angiogenesis. Since thrombospondin-2 does not activate latent transforming growth factor beta-1 (TGF-β1), the authors concluded that inhibition of angiogenesis by thrombospondins was not a result of regulation of TGF-β1 activation. The intriguing possibility that thrombospondins exert their anti-angiogenic effects in vivo by attenuating endothelial cell responses to LPA was left open.

Fuentes et al. [26] found, in 1997, that serum albumin induced a rather large yet transient release of Ca2+ from internal stores of HUVECs. This albumin effect was abolished by methanol extraction and thereby attributed the action to a polar lipid bound to the protein. Indeed, purified LPA induced similar effects. Moreover, albumin obtained from plasma was much less effective than albumin obtained from serum in inducing the response. Thus, the authors concluded that blood coagulation could play a key role in regulating vascular tone and capillary permeability by effecting LPA release from platelets. Experiments from our group, discussed below, would later expand and validate this important conclusion.

In 1998, Chua et al. [27] demonstrated that LPA markedly elevated endothelin-1 expression in rat aortic endothelial cells. mRNA levels for endothelin-1 (ET-1) increased within an hour after addition of LPA and subsequently declined. LPA stimulated AP-1 transcription factor binding as a result of activation of ET-1 gene activity. Thus, the authors concluded that increasing ET-1 by LPA could augment and prolong the vasoconstrictive actions of this lipid mediator.

In conclusion, the above studies demonstrated that sphingosine, a precursor of SPP which is, in fact, rapidly converted into SPP by the action of specific kinases in many cells (see Fig. 1) potently activates phospholipase D with resultant generation of PA in vascular endothelial cells. PA or related lipids released as a result of this reaction may have important vasoactive effects, mediating sphingosine-induced vasoconstriction. At extremely low concentrations, SPP effected mobilization of Ca2+ from intracellular stores by ligating a pertussis toxin-sensitive receptor; a similar effect was induced by SPPC at higher concentrations. LPA induced endothelial cell mitogenesis through a thrombospondin-sensitive pathway, a pathway that differed from that evoked by FGF-b in induction of mitogenesis. LPA induced endothelial cell gene expression and when bound to serum albumin, was effective in promoting Ca2+ mobilization form intracellular stores in HUVECs. The fact that albumin obtained from serum but not plasma was effective in induction of Ca2+ mobilization was the first evidence of a role for clotting-released lipids in endothelial cell activation.

4 Adhesion molecule expression; a prelude to angiogenesis

Xia et al. [28] examined the role of the sphingosine kinase pathway (see Fig. 1) in adhesion molecule expression induced by tumour necrosis factor-α (TNFα) in human umbilical vein endothelial cells. The investigators found that TNFα exerted its effects on both E-selectin and vascular adhesion molecule-1 expression through the sphingosine kinase signalling pathway. Indeed, treatment of endothelial cells with purified TNFα activated sphingosine kinase with consequent generation of SPP within the cell. Exogenous SPP, but not ceramide or sphingosine, potently induced adhesion molecule expression in human endothelial cells. In fact, exogenous SPP was able to mimic the effects of TNFα on endothelial cells, leading to extracellular signal regulated kinase (ERK) and NF-κB activation. Thus, the sphingosine kinase pathway is centrally involved in mediating the effects of TNFα on endothelial cells as a result of generation of endogenous SPP. These results would strengthen later studies examining the role of SPP in induction of endothelial cell angiogenic responses. They also open the possibility that paracrine signalling mechanisms, wherein activation of sphingosine kinase in endothelial cells results in potentiation of the cells' reactivity as a result of SPP generation, play an important role in angiogenesis. These interactions could have wide ramifications in the development of new blood vessels at sites of tumor growth and metastasis. Studies to date have not clarified how SPP differentially exerts effects through external EDG receptors versus its action within the cell. The mechanism by which the polar lipid, SPP, might be released from endothelial cells to stimulate the same or other cells through EDG receptor activation is also not clearly established.

Recently, Rizza et al. [29] examined the influence of LPA and other platelet-derived lipids on leukocyte/endothelial interactions, and focused on the functional consequences of increased expression of endothelial cell adhesion molecules. Treatment of human aortic endothelial cells for 4 h with 10 μM LPA, a concentration similar to that found in human serum [15], increased their ability to bind monocytes and neutrophils as a result of increased expression of E-selectin and vascular cell adhesion molecule-1 (VCAM-1) at the cell surface. To identify the receptor that mediated this effect of LPA, the investigators assessed EDG receptor expression in human aortic endothelial cells. These cells expressed EDG-1, -3, -4, and -5 as well as PSP24. While the PSP24 antagonist N-palmitoyl-serine and N-palmitoyl-tyrosine completely inhibited LPA-induced VCAM expression, E-selectin expression was barely affected, suggesting the involvement of an additional LPA receptor. The authors proposed that the second receptor may be EDG-1, since when expressed in HEK293 cells, EDG-1 caused ERK activation in response to LPA. In addition, exposure of endothelial cells to SPP, a high affinity EDG-1 receptor agonist [16–18], resulted in increased expression of E-selectin. The ability of LPA to induce VCAM-1 and of SPP to increase E-selectin in these cells was abolished by pretreatment with pertussis toxin, suggesting that both PSP24 and EDG-1 transmit signals leading to endothelial cell adhesion molecule expression through activation of a Gi-dependent pathway.

In conclusion, platelet-derived phospholipids have dramatic effects on endothelial cells. TNF-α increases adhesion molecule expression as a result of increasing intracellular SPP via its effects on sphingosine kinase. These effects were mimicked by exogenous SPP, which, in addition to increasing adhesion molecule expression, activated ERK and NF-κB. LPA and related lipids increased the adherence of endothelial cells to monocytes and neutrophils by increasing adherence molecule expression. Both PSP24 and EDG-1 were implicated as receptors involved in induction of these effects, which were abated by preincubation of cells with pertussis toxin. Thus, pertussis toxin sensitive receptors likely play a key role in the upregulation of endothelial cell adhesion molecules effected by LPA and related lipids, an early and important event in the angiogenic response.

5 Apoptosis

Endothelial cell apoptosis is thought to exert key regulatory influences on angiogenesis, and may be responsible for the ability of certain promising therapeutic compounds, including endostatin and angiostatin, to inhibit the response. Two recent studies showed that SPP, either generated exogenously or produced by activation of endothelial cell sphingosine kinase (Fig. 1), has a profound protective effect on endothelial cell apoptosis. Hisano et al. [30] found that sphingosine as well as ceramide effectively induced apoptosis in HUVECs. However, SPP was found to be an effective endothelial cell survival factor that prevented apoptosis induced by growth factor withdrawal and stimulated HUVEC DNA synthesis. Under the conditions used, sphingosine was rapidly incorporated into HUVECs and sequentially converted into ceramide and sphingomyelin (Fig. 1) in a time-dependent manner. However, SPP generation from sphingosine within HUVECs was weak and transient, suggesting that HUVECs are not able to supply the survival factor, SPP, but receive it from activated platelets. A markedly different situation may exist when HUVECs or other endothelial cells are stimulated with agents that activate sphingosine kinase, such as TNFα. Thus, Xia et al. [31] demonstrated that endogenous SPP generated in TNFα-stimulated HUVECs protected the cells from TNFα-induced apoptosis. Conversely, N,N-dimethylsphingosine, a competitive sphingosine kinase inhibitor, potently sensitized HUVECs to TNFα-induced apoptosis. In conclusion, activation of sphingosine kinase by TNFα and perhaps other endothelial agonists has important consequences, resulting in both enhanced expression of adhesion molecules as discussed above and protection of the cells from apoptosis.

6 Angiogenic responses to SPP and other platelet-released phospholipids

In late 1999, several groups directed their attention to defining the role of SPP in endothelial cell angiogenic responses as a result of the identification of SPP as a high affinity agonist for EDG-1, a gene induced during angiogenesis [5]. Hla and colleagues [32] studied the effects of SPP on endothelial cell adherens junction assembly and morphogenic differentiation into capillary-like structures, two in vitro correlates of angiogenesis that are well recognized. Northern blot analysis demonstrated that HUVECs expressed both EDG-1 and EDG-3, whereas expression of EDG-5 was undetectable. EDG-1 expression was estimated to be about 16-fold more abundant than that of EDG-3, the predominant SPP receptor expressed in human embryonic kidney cells. SPP (1 nM) induced robust Ca2+ mobilization in HUVECs, a response inhibited by more than 90% by pretreatment with pertussis toxin. This indicates that the response was driven by ligation of EDG-1, a Gi-protein-coupled receptor, rather than EDG-3, which couples to the pertussis toxin resistant G-proteins, Gq and G12/13 (Table 2). SPP also effectively activated ERK in a dose-dependent manner (10–500 nM); this response was inhibited by pretreatment of cells with either pertussis toxin or the MAP kinase inhibitor, PD98059. Later studies would show that the latter response is probably not required for SPP-induced migration of endothelial cells, but may play a key role in the induction of other angiogenic responses by SPP (see below).

View this table:
Table 2

Effect of platelet-derived phospholipids on endothelial cells

Principal investigator(s)Year [Ref.]LipidCellular effect(s)Cell typeCharacteristics
Natarajan and Garcia1994 [19]SPPActivation of PLDBPAEC
Meyer zu Heringdrof1996 [21]SPPCa2+ MobilizationBAECPertussis toxin sensitive
Panetti and Mosher1997 [24]LPAMitogenesisBAECInhibited by thrombospondins
Fuentes1997 [26]LPACa2+ MobilizationHUVECLPA bound to albumin
Chua1998 [27]LPAET-1 upregulationRAECBlocked by suramin and
PKC inhibitors
Rizza, Tigyi, Berliner1999 [29]LPAIncreased adherenceHAECEffect attributed to PSP24
and EDG-1
Hla1999 [32]SPPHUVECEffects attributed to EDG-1 and
Morphogenesis/vascularEDG-3. Stress fiber assembly and
junction assembly/in vivocortical actin structures induced by
angiogenesisSPP treatment. In vivo, SPP acts in
conjunction with protein angiogenic
factors to induce optimal response
Spiegel1999 [34]SPPDirected migration,HUVECPertussis toxin sensitive effects
chemokinesis, tube formationattributed to EDG-1 binding
in vitro, proliferation.No effect with LPA
English, Brindley, Garcia1999 [40]PAPermeability disruptionBPAECPertussis toxin sensitive
English, Brindley, Garcia1999 [35]SPP/LPAChemotaxisBPAECMigration to SPP inhibited by
antisense to EDG-1. LPA/PA have
little effect on migration. LPA effective
in stabilizing monolayer permeability.
Kwon1999 [36]SPPChemotaxis/angiogenesisHUVECMigration inhibited by pertussis toxin.
Activation associated with activation
of ERK/p38 MAPK
Kwon2000 [49]SPPChemotaxisHUVECInhibited by inhibitors of PLC but not
PKC. Migration associated with
p125FAK tyrosine phosphorylation
English, Brindley, Garcia2000 [46]SPPChemotaxis,BPAEC,Effects mimicked by serum; serum
capillary formation,HUVEC,activity traced exclusively to SPP
in vivo angiogenesisHMVECreleased from platelets during clotting
Goetzl and An2000 [51]SPP/LPAWound healingBAEC/HUVECPertussis toxin sensitive; distinct
receptors involved for LPA, SPP
Panetti and Mosher2000 [52]SPP/LPAChemokinesisFBHEResponse inhibited by pertussis
toxin, inhibitors of Rho and inhibitors
of PI 3′-kinase
Okajima2000 [54]SPPProliferationHAECResponses inhibited by pertussis
migrationtoxin and suramin. Migration inhibited
by p38 MAP kinase inhibitors

Treatment of endothelial cells with exogenous SPP dramatically increases actin stress fiber formation and the development of cortical actin structures. While these responses were receptor dependent, as evidenced by the failure of microinjected SPP to induce them, they were not suppressed by pretreatment of cells with pertussis toxin. Inhibition of the intracellular low molecular weight G-protein, Rho, by intracellular injection of C3 exoenzyme abolished SPP-induced stress fiber formation, but had little effect on cortical actin. Microinjection of dominant/negative Rac abolished both responses. Thus, similar to responses observed in fibroblasts [33], Rac appears to act upstream of Rho in induction of endothelial cell cytoskeletal changes. SPP increased the localization of VE-cadherin as well as α-, β- and γ-catenin at cell–cell junctions, suggesting that SPP effectively induced the formation of endothelial cell adherens junctions. Treatment of cells with related lipids (Fig. 1), including sphingosine, sphingomyelin, ceramide and ceramide-1-phosphate, or instillation of SPP into the cytoplasm by microinjection had no effect.

These results demonstrate that the angiogenic responses to SPP resulted from ligation of specific cell surface receptors. Co-immunoprecipitation confirmed that SPP increased the levels of α-catenin and VE-cadherin polypeptides in both γ and β-catenin immunoprecipitates. Antisense oligonucleotides, designed to inhibit expression of both EDG-1 and EDG-3, inhibited SPP-induced VE cadherin localization at cell–cell junctions. In addition, microinjection of EDG-1 antisense oligonucleotides attenuated the formation of cortical actin structures in SPP treated HUVECs, a consequence of activation of the Rac pathway. By contrast, Rho-dependent stress fiber formation was specifically attenuated by antisense to EDG-3. SPP was strikingly effective in promoting HUVEC morphogenesis into capillary-like structures, as assessed using a Matrigel assay system, and this response was inhibited by pretreatment of cells with either pertussis toxin, C3 exoenzyme, VE-cadherin blocking antibodies or antisense oligonucleotides to both EDG-1 and EDG-3. Finally, SPP enhanced the ability of vascular endothelial growth factor (VEGF) and FGF-b to induce angiogenic responses in subcutaneously implanted Matrigel in athymic mice. Like its in vitro counterparts, potentiation of angiogenesis in vivo by SPP was attenuated by EDG-1 and EDG-3 antisense oligonucleotides, indicating an involvement of both of these SPP receptors in the response.

Microscopic examination revealed that vessels formed in Matrigel plugs containing SPP and protein growth factors were intact and well developed, in contrast to the rather leaky vessels generated in plugs seeded with protein growth factors alone. Consistent with this observation, Garcia and colleagues (unpublished observations) found that SPP is strikingly effective in increasing the integrity of endothelial cell monolayers in vitro.

In conclusion, SPP may act in conjunction with protein growth factors to effect optimal angiogenic responses. Endogenous production of SPP by thrombotic platelets could thus be a key aspect of regulation of the angiogenic process. Activation of the EDG-1 and EDG-3 pathways by SPP or related agonists may have important therapeutic implications, facilitating neovascularization during wound healing, or after ischemic injury. Conversely, EDG-1 and EDG-3 antagonists may be useful to limit angiogenesis in order to prevent the growth of tumours, or to decrease diabetic retinopathy.

7 Chemotaxis

Directed migration of endothelial cells is an important aspect of the angiogenic response, allowing repositioning of cells after their liberation from existing vessels. In December, 1999, three groups reported that SPP is a potent and specific chemoattractant for endothelial cells, which exerted its effects at low yet physiologically relevant concentrations [34–36]. Previous studies had investigated the effects of SPP on cell migration, but only inhibitory effects had been observed [37–39]. Spiegel and co-workers [34] demonstrated that SPP induced migration of both human and bovine endothelial cells in a pertussis toxin-sensitive manner, implicating the involvement of a Gi-coupled cell surface receptor. These effects were induced by concentrations of SPP as low as 10 nM and were mimicked by sphinganine 1-phosphate (dihydro-SPP) whereas LPA was without effect. Checkerboard analysis revealed that SPP induced migration by enhancing both random movement, or chemokinesis, as well as by inducing a directional response. SPP was effective in inducing the proliferation of endothelial cells in a pertussis toxin-sensitive manner, and stimulated capillary-like tube formation of endothelial cells grown on collagen gels, an in vitro correlate of angiogenesis. The authors concluded that SPP plays an important role in angiogenesis by binding specific Gi-coupled receptors.

Studies from our laboratories indicated that the receptor involved in SPP-induced chemotaxis is, in fact, EDG-1 [35]. Thus, purified SPP was shown to induce a potent chemotactic response of BPAECs and this response was specifically inhibited by pretreatment of cells with phosphorothioate antisense oligonucleotides designed to limit the expression of EDG-1. Under optimal conditions, physiologically relevant concentrations of SPP induced chemotaxis at rates 3- to 10-fold greater than responses induced by either VEGF or FGF-b. PA, a biologically active lipid that we previously demonstrated to disrupt the integrity of endothelial cell monolayers [40], was also an effective endothelial cell chemoattractant, inducing responses similar to those induced by the protein growth factors. By contrast, LPA was without effect. Chemotaxis induced by SPP was markedly inhibited by the tyrosine kinase inhibitors genistein (50 μM), herbimycin A (10 μM) and PP2 (10 μM), but inhibitors of phosphatidylinositol 3′-kinase were without effect. Our study further demonstrated that LPA in concentrations ranging from 1 to 10 μM effectively stabilized the integrity of endothelial monolayers, a result consistent with those of earlier studies by Alexander et al. [41]. In conclusion, the results provide a firm foundation for our hypothesis that the phospholipid growth factors, PA, SPP, and LPA operate in tandem to first liberate endothelial cells from established vessels, induce their directional migration and finally stabilize newly formed vessels.

Kwon and associates [36] also reported the ability of SPP to induce endothelial cell migration and stimulate DNA synthesis in a dose-dependent manner. Consistent with the earlier results of Hla et al. [32], SPP was effective in promoting tube formation of HUVECs on Matrigel and in enhancing angiogenic responses in vivo. Exposure of HUVECs to SPP resulted in activation of ERK as well as p38 MAP kinase in a pertussis toxin-sensitive manner. The authors reported that the MEK inhibitor, U0126, markedly attenuated SPP-induced tube formation, but migration was not blocked by inhibition of either ERK or p38 MAPK. However, we note that results from our laboratories demonstrate profound inhibition of SPP-induced endothelial cell migration by inhibitors of p38 MAP kinase (unpublished observations). Thus, the involvement of ERK and other members of the MAP kinase family in SPP induced endothelial cell migration awaits clarification. While Kwon and associates [36] concluded that SPP may act as an important modulator of platelet-induced angiogenesis, they were careful to note that future studies would be necessary to determine the role of SPP in platelet-induced angiogenesis under physiological conditions.

Thus, as discussed above, while the release of SPP from activated platelets has been demonstrated, SPP is also rapidly degraded by phosphatases and a lyase (Fig. 1) (see Ref. [42]). In the case of extracellular SPP, work from one of our laboratories (Waggoner and Brindley, unpublished results) has demonstrated that extracellular SPP is dephosphorylated relatively specifically by lipid phosphate phosphatases (phosphatidate phosphohydrolase-Type 2). At present, there are three major isoforms of lipid phosphate phosphatases, each of which has three conserved phosphatase domains [43,44]. The structure of LPPs is consistent with that of an exoenzyme which has its active site on the exterior surface of the plasma membrane [45]. In contrast, the lyase is more likely to act on intracellular SPP.

In conclusion, by late 1999, purified SPP was found to exert a number of effects on endothelial cells, effects that were consistent with the proposed role of the lipid in angiogenesis. Thus, SPP had been shown to be a potent endothelial cell chemoattractant, which possessed the ability to induce morphogenic differentiation of endothelial cells into capillary-like structures in vitro. In addition, SPP was demonstrated to be an effective ‘survival’ factor, which protected endothelial cells from stimulus-dependent apoptosis. Perhaps most importantly, SPP enhanced the ability of protein growth factors to induce angiogenic responses in vivo, an effect due to ligation of members of the EDG family of receptors. Thus, SPP appeared to possess all the properties of a physiologically relevant angiogenic factor. However, due to the polar nature of SPP, extracellular export of the compound is extremely difficult to detect [41], and neutralizing antibodies to the lipid are not presently available. Thus, no reports had directly traced biological activities of serum or other tissue extract to the presence of SPP. Therefore, while the reports described above implicate platelet-released SPP as a potentially important angiogenic agent, it physiological significance remained unclear.

8 Physiological relevance of SPP-induced angiogenic responses

As a result of these considerations, we undertook an investigation to define agonists generated in serum during clotting that induce angiogenic responses of endothelial cells, as assessed by permeability changes, morphogenic differentiation and chemotaxis. We initially directed our attention toward defining the factor or factors in serum that are responsible for its potent chemoattractive activity for endothelial cells [46]. These studies demonstrated that the chemotactic activity of serum was markedly greater than that which could be attributed to its content of the known protein endothelial cell chemoattractants, VEGF and FGF-b. Furthermore, the endothelial cell chemotactic activity of serum was remarkably heat-stable, extractable by activated charcoal and recovered quantitatively in butanol extracts, properties which are consistent with a lipid origin of the activity. Finally, this potent endothelial cell chemotactic activity was present in markedly diminished quantities in plasma obtained form anti-coagulated blood, but reappeared in plasma clotted in the presence, but not in the absence, of platelets. Taken together, these observations led us to hypothesize that platelet-derived phospholipids contributed markedly to the angiogenic potential of serum.

Several approaches were used to define the factor responsible for this activity. Using several carefully calibrated thin layer chromatography systems, we fractionated butanol extracts of human serum and tested fractions for their ability to restore chemotactic activity to charcoal-stripped serum. Chemotactic activity consistently co-eluted with SPP from thin layer plates after chromatography in four different solvent systems, directly implicating platelet-derived SPP in this angiogenic response. Serum chemotactic activity was quantitatively recovered in these eluates, leaving little doubt that SPP accounted for nearly all of the activity generated during blood clotting. By contrast, several other bioactive lipids expected to be present in serum, including LPA, LPE, 12-hydroxyeicosatetranoic acid (HETE), 15-HETE, sphingosine, sphingomyelin, ceramide, ceramide-1-phosphate, PA, arachidonic acid and platelet activating factor, were not able to restore chemotactic activity to charcoal-treated serum. These experiments thus directly linked the angiogenic activity of serum to its content of SPP and provided conclusive evidence that platelet-released SPP was responsible for this activity.

Our study demonstrated that SPP exerted its effect optimally in conjunction with a serum co-factor that stabilized endothelial cell interactions with the substratum [46]. This factor, which appears to be fibronectin, acts in conjunction with low levels of SPP to induce potent angiogenic responses. After SPP-induced migration, endothelial cells proliferated avidly and formed multicellular structures suggestive of early blood vessel formation. SPP was effective in restoring the ability of charcoal-stripped serum to promote differentiation into capillary-like structures in the in vitro Matrigel assay system. Finally, SPP was strikingly effective in promoting angiogenesis in the developing chick embryo chorioallantoic membrane (CAM, see Fig. 2), and markedly enhanced the ability of FGF-b to induce blood vessel formation in the avascular mouse cornea (Fig. 3). In conclusion, the results of this study directly demonstrate that blood coagulation initiates angiogenesis through the release of SPP, a potent endothelial cell agonist that exerts its effects by activating a receptor-dependent process.

Fig. 2

Angiogenic response to SPP embedded in gelatin filters placed upon the developing chicken chorioallantoic membrane. Filters with (1) or without (2) 1 μmol SPP absorbed to them were positioned upon the developing CAM at day 8; incubation was continued for an additional 4 days, at which time the membranes were removed, fixed with formaldehyde and examined microscopically (20×) and photographed. Blood vessel formation is evident radiating toward the center of the filter containing SPP, which is darker than the filter at the center of the filter without the angiogenic lipid due to vascularization throughout the filter.

Fig. 3

Potentiation of angiogenesis in the avascular mouse cornea by SPP. Sulfacrate supplanted Hydron pellets containing approximately 4 μg bovine serum albumin (BSA) in the presence of suboptimal levels of FGF-b were implanted into corneal micropockets in the absence or presence of 10 nmol SPP. Hydron pellets, evident as the white spots near the center of the photographs, containing BSA alone were used as negative controls. Blood vessel formation was assessed by slit lamp photography 5 days later. In the two experiments shown, FGF-b alone induced a noticeable angiogenic response, as evident by blood vessel formation from the limbus to the pellet. However, almost no response was elicited by SPP in the absence of other factors. However, the combination of the two factors induced a vigorous response, resulting in vascularization of the cornea throughout the area below the white pellet and extending to the limbus (reproduced, with permission, from Figure 6 of Ref. [46]).

9 Signalling pathways involved in SPP-induction of endothelial cell migration

In the experiments described by English et al. [46], pertussis toxin was consistently effective in inhibiting migratory responses induced by SPP, implicating the involvement of a Gi-linked cell surface receptor, presumably EDG-1. In a subsequent investigation, we examined the possibility that these receptor-dependent effects were mediated by βγ-subunits liberated from the receptor coupled G-protein, since earlier studies had implicated βγ-subunits in the directed migration of leukocytes upon ligation of G-protein-coupled receptors [47]. To examine this possibility, we intended to examine SPP-induced chemotactic migration of endothelial cells transfected with high levels of the C-terminal domain of the β-adrenergic receptor kinase-1 (βARK), in order to limit the effects of liberated βγ-subunits. However, early results were limited by low transfection efficiencies; in many cases, less than 30% of the exposed cells expressed marker genes in the transient transfection system used. Low transfection efficiencies have limited genetic manipulation of endothelial cells for functional analysis by other groups as well. Therefore, we developed a method to increase transfection efficiencies of endothelial cells for functional analyses based on co-transfection of cells with a plasmid containing green fluorescent protein [48]. By using fluorescent activated cell sorting, we were able to isolate a population of which essentially every cell was transfected with the plasmid of interest. These studies demonstrated that the SPP-induced chemotaxis of endothelial cells was mediated by βγ-subunits liberated after receptor ligation. It is presently not clear how the βγ-subunits exert their effect on chemotactic migration, and experiments are in progress to assess several possibilities. Moreover, the method developed by Kovala et al. [48] should find wide use in defining signalling pathways in endothelial cells and in other cells in which low transfection efficiencies limit precise definition of the effects of genetic manipulation.

In our system, inhibitors of PI 3-kinase had little effect on SPP-induced endothelial cell migration [35] and therefore the nature of the signal transduction process is not fully understood. In a recent study, Lee et al. [49] implicated the Gi-linked phospholipase C pathway in SPP-induced endothelial cell migration and focal adhesion kinase tyrosine phosphorylation. As discussed above, we have observed potent inhibition of SPP-induced endothelial cell migration with inhibitors of tyrosine kinases, including specific inhibitors of members of the src family. We, and others, are presently investigating the sequence of biochemical events involved in induction of migration, morphogenesis and other angiogenic responses by SPP compared to those evoked by other lipids, including sphingosine and sphingosylphosphorylcholine [50] and by protein growth factors such as VEGF and FGF-b. These studies promise to yield important information regarding the mechanisms by which these physiologically relevant agonists induce angiogenic responses in vivo. The results should also facilitate the development of therapeutic interventions to modulate these responses.

10 Wound healing, migration and LPA

In a recent report, An and co-workers [51] evaluated the ability of LPA and SPP to stimulate endothelial wound healing in a cell culture system. The model consists of stimulated repair of disrupted endothelial monolayers. Using HUVECs and bovine aortic endothelial cells (BAECs) which express the LPA receptor, EDG-2, and the SPP receptors, EDG-1 and EDG-3, the authors observed that both LPA and SPP were effective in stimulating closure of wounded monolayers. In addition, the two major components of wound healing, migration and proliferation, were stimulated individually by the two lipids. These effects were partially blocked by pertussis toxin. LPA and SPP also stimulated intracellular Ca2+ mobilization and ERK phosphorylation. Since LPA and SPP did not cross desensitize each other in induction of these responses, the authors concluded that distinct receptors were involved. Thus, the authors concluded, LPA and SPP induce important endothelial cell functions through the activation of signalling pathways through distinct G-protein coupled receptors, functions that may play a key role in wound healing and angiogenesis.

While the ability of LPA to promote endothelial cell migration in the study described above [51] is at variance with results from other groups [35,37,46], a recent report also demonstrates the ability of LPA to induce chemotaxis of endothelial cells. Thus, Panetti and Mosher [52] found that both SPP and LPA were effective chemoattractants for foetal heart endothelial cells. These agonists stimulated phosphorylation of ERK and enhanced paxillin localization to focal contacts. Inhibitors of Rho, Gi and PI 3-kinase each attenuated migration in this system, but the ERK kinase inhibitor, PD98059, caused only a minimal decrease in migration. Thus, SPP and LPA stimulate migration in this system by a mechanism that requires a balance between Gi, Rho and cytoskeletal remodeling.

In the study described above [52], the reason that the results do not follow the pattern discerned by other investigators may result from the source of the endothelial cells used. For example, foetal endothelial cells may react differently than adult cells, thereby accounting for the ability of PI 3-kinase inhibitors to blunt migration and the ability of LPA to stimulate it. Moreover, as mentioned above, the involvement of ERK activities in SPP- (and LPA-) induced endothelial cell migration awaits clarification. Previous studies with the known endothelial cell protein chemoattractant, VEGF, demonstrated that the p38 MAP kinase inhibitor, SB203580, but not the inhibitor of ERK, PD98059, blocked the response [53]. However, although Kwon et al. [36] found a dissimilar pattern of inhibition of SPP-induced migration with these agents, our unpublished results are consistent with those obtained with VEGF, indicating that similar downstream signalling pathways are involved. A recent report lends support to our findings. Thus, Kimura et al. [54] studied the role of MAP kinases in SPP-induced migration of endothelial cells. As in other reports, SPP-induced migration was pertussis toxin-sensitive, mimicked by dihydro-SPP and associated with the activation of both ERK and p38 MAP kinase. While SPP-induced DNA synthesis and ERK activation were inhibited by PD98059 (an ERK kinase inhibitor), the p38 MAP kinase inhibitor SB203580 had little effect. In contrast, cell migration was markedly attenuated by SB203580, but not by PD98059. The authors concluded that SPP is a novel angiogenic factor that exerts its effects through ligation of the cell surface receptors, EDG-1 and EDG-3. As stressed above, the role of MAP kinases in angiogenic responses induced by SPP remains to be clarified.

11 Conclusions

Although platelet-released phospholipids have been shown to exert a variety of effects on endothelial cells, the finding that SPP accounts for nearly all of the endothelial cell chemotactic activity of blood serum, while unexpected, confirms the physiological relevance of many of these effects. SPP exerts its angiogenic effects optimally in conjunction with factors that promote the interaction of endothelial cells with the substratum, such as fibronectin, and in the presence of protein growth factors, including VEGF and FGF-b. This recent work emphasizes that bioactive lipids play a major physiological role in controlling angiogenesis. This knowledge needs to be integrated with what is known concerning the role of the angiogenic protein growth factors in order to understand how new blood vessel formation is achieved. Studies are now being performed to discern the biochemical basis of these interactions, since a thorough understanding of the events involved will facilitate therapeutic manipulation of angiogenesis to promote wound healing, limit tumour growth or restore vascularization to ischemic tissue.


This work was supported by NIH grant # RO1 HL 61751 awarded to DE and DNB, PO1 HL 58064 awarded to JGNG and DE, a grant from the Methodist Hospital Cancer Center, the Methodist Heart Institute and the Methodist Schowalter Foundation and a generous donation from the Phi Beta Psi Sorority awarded to DE.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
View Abstract