The plaque lipid lysophosphatidic acid stimulates platelet activation and platelet-monocyte aggregate formation in whole blood: involvement of P2Y₁ and P2Y₁₂ receptors

Oxidized low-density lipoprotein (LDL) and platelets play a central role in the pathogenesis of atherosclerotic cardiovascular diseases. We have previously found that lysophosphatidic acid (LPA) accumulates in mildly oxidized LDL and in human atherosclerotic lesions and is responsible for platelet shape change induced by mildly oxidized LDL and extracts of lipid-rich atherosclerotic plaques.¹  On plaque rupture, LPA becomes available to circulating platelets triggering their activation, which, in turn, might contribute to the formation of intravascular thrombi responsible for acute coronary syndrome, myocardial infarction, and stroke. Besides its accumulation in the vascular intima, oxidatively modified LDL is also present in blood circulation and its plasma levels are elevated in patients with coronary artery disease.^2,3  LPA present in circulating oxidized LDL could contribute to the platelet hyperreactivity observed in these patients.^4-6 

Low nanomolar concentrations of LPA induce shape change of washed platelets through LPA receptor-linked signal transduction pathways that involve the small guanosine triphosphatase (GTPase) Rho and the Rho-kinase–mediated stimulation of myosin light-chain phosphorylation.^7-9  In contrast, high micromolar concentrations of LPA are required to induce aggregation of platelet-rich plasma.^10-12  In the present study we investigated whether and through which mechanism LPA could induce platelet shape change and aggregation in whole blood. We also studied the formation of platelet-monocyte aggregates, an early marker of acute myocardial infarction.¹³  Here we report that LPA is capable of eliciting platelet shape change, aggregation, and platelet-monocyte aggregate formation in whole blood at concentrations that are only slightly above its plasma concentration. Platelet aggregation induced by LPA was inhibited by antagonists of the platelet adenosine diphosphate (ADP) receptors P2Y₁ and P2Y₁₂, but it was insensitive to inhibition by aspirin.

1-Palmitoyl-sn-glycero-3-phosphate (1-acyl-LPA 16:0) was obtained from Alexis (San Diego, CA). 1-Hexadecyl-LPA (1-alkyl-LPA 16:0) was synthesized as described previously.¹⁴  N-palmitoyl-serine-phosphoric acid (NPSerPA) was purchased from Biomol Research Laboratories (Plymouth Meeting, PA). Collagen (Kollagenreagens “Horm”) was from Nycomed Pharma (Unterschleißheim, Germany). Adenosine 3′-phosphate 5′-phosphate (A3P5P), apyrase (catalogue no. A-6535; adenosine-5′-triphosphate diphosphohydrolase, EC 3.6.1.5), 1-α-oleoyl-lysophosphatidic acid (1-acyl-LPA 18:1), acetylsalicylic acid, Arg-Gly-Asp-Ser (RGDS), and bovine serum albumin (A-4503) were obtained from Sigma (Taufkirchen, Germany). AR-C69931MX was a gift from AstraZeneca R & D Charnwood (Loughborough, United Kingdom). MRS2179 was kindly provided by J. Bourguignon and P. Raboisson, (UMR CNRS Strasbourg, France).¹⁵  Refludan was obtained from Pharmion (c/o LOGOSYS Logistic, Darmstadt, Germany). The enzyme immunoassay kit for cyclic adenosine monophosphate (cAMP) measurement was from Assay Designs, distributed by Biotrend (Cologne, Germany). The following monoclonal antibodies directly conjugated to fluorochromes were all purchased from BD Biosciences (Heidelberg, Germany): phycoerythrin (PE)–conjugated anti-CD14 (MϕP9, IgG_2b), fluorescein isothiocyanate (FITC)–conjugated anti-CD41a (HIP8, IgG₁), and FITC-conjugated IgG₁, κ (MOPC-31C). PE-conjugated anti-CD62P (CLB-thromb/6, IgG21) was from Immunotech Coulter (Marseille, France). A blocking monoclonal anti–P-selectin antibody (G-1, IgG₁) was from BenderMedSystems (Vienna, Austria). Red blood cell lysing buffer (Erythrolyse solution) was from Serotec (Oxford, United Kingdom).

Informed consent from the blood donors was obtained in accordance with the protocol of Helsinki. Blood from healthy volunteers not taking any medication was drawn by cubital venipuncture into syringes containing either trisodium citrate dihydrate (3.8% wt/vol) or recombinant hirudin (Refludan; 13 μg/mL; 200 U/mL), and immediately transferred into polypropylene tubes and kept at 20°C. Experiments were completed within 2 hours after blood was drawn. Whole blood platelet aggregation was measured using the single-platelet counting technique described by Heptinstall et al¹⁶  with the following modifications. Aliquots (400 μL) of anticoagulated blood samples were placed into aggregometer cuvettes. The samples were stirred in a LABOR aggregometer (Fa Fresenius, Bad Homburg, Germany) at 400 or 1000 rpm at 37°C for 30 seconds. The inhibitor/antagonist was then added for 30 seconds before stimulation with the different LPA species (0.2-20 μM, from a 1-mM stock solution; LPA was dissolved in albumin buffer, 0.25 mM albumin, 20 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], 150 mM NaCl, pH 7.4). Aliquots (15 μL) before and after agonist addition were removed and placed in 30 μL fixing solution containing 0.16% (wt/vol) formaldehyde as described.¹⁶  After dilution, the samples were counted using a platelet counter (Sysmex Platelet Counter PL-100, TOA Medical Electronics, Kobe, Japan) and percentage aggregation calculated as percentage loss of single platelets compared to baseline count. All platelet counts were done in duplicate.

Aspirin was dissolved in 0.9% NaCl solution at pH 7 (maximum solubility 3.3 g/L), which was drawn together with the anticoagulant into a separate syringe and mixed with whole blood to obtain a final concentration of 1 mM aspirin.

For the measurement of platelet shape change, blood was incubated with or without stirring for 5 minutes at 37°C with buffer (control), 1-acyl-LPA (16:0) or ADP. The reaction was stopped by placing 100 μL blood sample in Erythrolyse solution that contains formaldehyde as fixative, and the samples were processed as described. Platelets labeled with anti-CD41a–FITC antibody were gated. Platelet shape change was determined by flow cytometry collecting 10 000 platelet events and measuring the ratio forward scatter/side scatter.¹⁷  Values in control samples (0.18-0.19) were set to 100%.

Platelets were pretreated with aspirin, washed, and resuspended in buffer (1 × 10⁶ cells/μL) containing apyrase as described previously.¹  Platelet suspensions were incubated at 37°C while stirring (1000g) with iloprost (50 nM) for 5 minutes before exposure to LPA (0.1-80 μM) or epinephrine (10 μM) for 2 minutes. Levels of cAMP were determined with an enzyme immunoassay kit.

Aliquots (400 μL) of anticoagulated blood samples were transferred into aggregometer cuvettes and incubated while stirring at 1000 rpm for 5 minutes, 37°C in the absence or presence of platelet agonists (LPA; ADP; collagen). The reaction was stopped by placing 100 μL of the blood sample in 1 mL Erythrolyse solution and left for 10 minutes at room temperature. The lysed samples were centrifuged and washed with 1 mL phosphate-buffered saline (PBS) at 3000 rpm for 5 minutes. The pellet was incubated for 15 minutes in the dark at room temperature with the following antibodies: anti-CD41a–FITC antibody (5 μL) binding to the glycoprotein IIb to label platelets, isotype-matched IgG₁-FITC (6 μL diluted 1:100) to determine unspecific binding, and anti-CD14–PE antibody (2 μL) to label monocytes. Subsequently, 600 μL PBS was added, and platelet-monocyte conjugate formation was measured by FACScan (Becton Dickinson, Heidelberg, Germany). The measurement was performed by collecting 5000 monocyte events using a combination of forward scatter and side scatter indicating cell size and granularity, respectively. Monocytes were distinguished from other cells by their size and granularity as well as their anti-CD14–PE fluorescence. The percentage of platelet-monocyte conjugates was measured in single parameter histograms of anti-CD41a–FITC fluorescence displaying events from the monocyte gate. The positive analysis region was determined using the IgG₁-FITC isotypic control.

Anticoagulated blood was centrifuged for 20 minutes at 180g and plateletrich plasma (PRP) was collected. Platelet aggregation was measured by recording the light transmission of stirred PRP (1100 rpm) at 37°C by using a LABOR aggregometer (Fa Fresenius).¹⁸  For shape change experiments blood was drawn into acidic-citrate-dextrose (ACD). Shape change was measured by the decrease in light transmission of the stirred (1100 rpm) PRP.

Washed platelets were prepared as described previously,¹⁹  and resuspended in Tyrode buffer containing 2 mM CaCl₂, and 0.35% fatty acid-free human albumin at a density of 3 × 10⁵ platelets/μL, in the presence of 0.02 U/mL of the ADP scavenger apyrase, a concentration sufficient to prevent the desensitization of platelet ADP receptors during storage. Platelets were kept at 37°C throughout all experiments. Aggregation and [³H]-serotonin release were measured as described.^19,20  For P-selectin expression, platelets (1.5 × 10⁸) were incubated for 15 minutes with 0.5 mg anti-CD62–PE antibody, and 0.5 mL Tyrode buffer containing 2 mM CaCl₂ and 0.35% fatty acid-free albumin were added. Mean fluorescence intensity was measured by fluorescence-activated cell sorting (FACS) after collecting 10 000 platelet events. The mean fluorescence intensity of isotype-matched IgG₁-PE was subtracted.²¹ 

Data were analyzed by comparing all corresponding single values in a paired Student t test of unequal variance. A P value less than .05 was considered to be significant.

Acyl-LPA (16:0) LPA elicited platelet aggregation in whole anticoagulated blood in a time- and dose-dependent manner (Figure 1). Platelet aggregation started within 30 seconds and was irreversible. When the effect of 1-alkyl-LPA (16:0) was compared with that of 1-acyl-LPA (16:0), it was found that the alkyl-LPA species was almost 20-fold more potent than the acyl-LPA species in inducing platelet aggregation in whole blood (Figure 1B). The effective concentration (EC₅₀) of 1-alkyl-LPA (16:0) was 0.3 μM, and that of 1-acyl-LPA (16:0) was 5 μM, and platelet aggregation was observed after 0.2 μM 1-alkyl-LPA (16:0) and 2.5 μM 1-acyl-LPA (16:0) (P < .01 and .001, respectively, as compared to buffer). This finding is of possible pathophysiologic significance because we recently identified alkyl-LPA species in the lipid-rich core region of human atherosclerotic plaques.²²  1-Acyl-LPA (16:0) and 1-acyl-LPA (18:1) demonstrated the same platelet aggregating effect (results not shown). In the following results LPA refers to 1-acyl-LPA (16:0) unless otherwise mentioned.

We found that the aggregation response to LPA in whole blood showed considerable variations that were mainly donor dependent. The distribution of the LPA aggregation response was biphasic. Of 56 donors tested, 21 donors showed a platelet aggregation response to 20 μM 1-acyl-LPA (16:0) higher than 50% with a mean ± SD of 70% ± 14%. Others (25 donors) showed aggregation lower than 50% with a mean ± SD of 26% ± 14%. Ten donors had no significant aggregation as compared to control. Five donors were repeatedly (5 to 10 times) tested over a period of 2 years and it was found that the aggregation response in whole blood to LPA, but not to other agonists, such as ADP or collagen, was dependent mainly on the blood donor.

The type of anticoagulant used had no influence on LPA-induced aggregation. The time-course, magnitude, and concentration-response curves of LPA-induced aggregation were similar in hirudin- and citrate-anticoagulated blood and PRP, indicating that the extracellular Ca²⁺ concentration was not critical for the aggregation response induced by LPA (data not shown).

In the presence of fibrinogen, LPA also induced aggregation of washed platelets. In contrast to previous studies of LPA-induced aggregation of isolated platelets, platelets in our study were washed in the presence of prostacyclin and resuspended in buffer containing apyrase to avoid platelet preactivation and desensitization of the ADP-receptor P2Y₁.²³  In this preparation, 1-acyl-LPA (16:0) induced shape change and subsequent aggregation with an EC₅₀ of 8 μM. A concentration of 1 μM 1-acyl-LPA (16:0) only induced shape change, 3μM 1-acyl-LPA (16:0) induced shape change, reversible aggregation (10%) with only 1% to 2% serotonin secretion, higher concentrations (10, 30, and 100 μM) 1-acyl-LPA [16:0]) elicited maximal irreversible aggregation with 9%, 19%, and 36% serotonin secretion, respectively (data not shown and Figure 2A). 1-alkyl-LPA (16:0) was again more potent than 1-acyl-LPA (16:0) in inducing shape change and aggregation of washed platelets (data not shown).²² 

Platelet aggregation in whole blood stimulated by LPA was dependent entirely on ADP-induced activation of the platelet P2Y₁ and P2Y₁₂ receptor as determined by the use of the ADP-scavenging enzyme apyrase and specific ADP receptor antagonists (Figure 2B; Table 1).^15,24,25  The ADP-scavenger enzyme apyrase, the antagonists of the P2Y₁ receptor A3P5P and MRS2179, and the antagonist of the P2Y₁₂ receptor AR-C69931MX completely inhibited platelet aggregation induced by LPA. Platelet aggregation induced by ADP was also inhibited by MRS2179 and AR-C69931MX (Table 1).

Similarly, MRS2179 and AR-C69931MX inhibited the LPA-induced irreversible aggregation of washed platelets and converted it into a reversible aggregation (Figure 2A; Table 2). A combination of both ADP receptor antagonists was only slightly more effective in inhibiting LPA-induced aggregation. Similar results were obtained in PRP. Irreversible platelet aggregation in PRP stimulated by 80 μM LPA was inhibited by apyrase or the P2Y₁ and P2Y₁₂ receptor antagonists and converted into a reversible aggregation. LPA-induced primary reversible aggregation was not inhibited (data not shown). The reduction of LPA-induced aggregation of washed platelets was paralleled by the inhibition of serotonin secretion to less than 2% by AR-C69931MX, MRS2179, or a combination of both ADP receptor antagonists, which was independent of the concentration of LPA (10, 30, or 100 μM) applied (Figure 2A and data not shown).

Remarkably, aspirin did not inhibit LPA-induced platelet aggregation in whole blood (Figure 2B; Table 1), PRP, or washed platelets (Table 2 and data not shown).

To gain further insight into how LPA might induce platelet aggregation in whole blood, platelet shape change was measured in PRP, in unstirred and stirred blood. Shape change is the initial platelet response independent of positive feedback mediators such as released ADP or thromboxane A₂.²⁶  We found that LPA induced shape change in blood and PRP (Figure 3). This response was blood donor independent. Stirring of blood, which might promote ADP release from red cells, did not affect the shape change response in blood (data not shown). Furthermore, shape change induced by 1-acyl-LPA (16:0) showed a similar dose-response curve, whether this response was induced in blood or in PRP, showing the independence of red blood cells. 1-alkyl-LPA (16:0) was again almost 20-fold more potent than 1-acyl-LPA (16:0) in inducing shape change of PRP (Figure 3A). These results indicate that LPA directly induces shape change in whole blood independent of a synergistic interaction with ADP. Further supporting this conclusion is the lack of shape change inhibition by the P2Y₁ receptor antagonist, MRS2179, the P2Y₁₂ receptor antagonist, AR-C69931MX, and the combination of both antagonists (Figure 3B) when shape change was induced by LPA. In contrast, MRS2179 but not AR-C69931MX blocked the ADP-induced shape change in whole blood.

The concentration-response curves of 1-acyl-LPA (16:0) in washed platelets resuspended in the absence of albumin showed about 1000-fold lower EC₅₀ values (2-20 nM; n = 22) as compared to the concentration-response curve in PRP (data not shown). We found albumin in plasma to be the main factor responsible for the large difference in the concentration response curves of shape change in washed platelets and PRP (data not shown). Albumin inhibited the shape change induced by 50 nM 1-acyl-LPA (16:0) with an inhibitory concentration of 50% (IC₅₀) of 6 μM.

Based on the results described and the fact that the dose-response curves of 1-acyl-LPA (16:0) for platelet shape change and aggregation in blood were almost identical, we hypothesized that LPA stimulates signal transduction pathways during platelet shape change that synergize with signaling pathways stimulated by other agonists leading to platelet aggregation in whole blood. It has been shown in the past and rediscovered recently that on coactivation of platelet receptors that couple to different G proteins, their specific signal transduction pathways converge in inducing platelet aggregation.^27-32 

LPA is known to activate in various cell types the G_12/13 protein family that couples to the Rho/Rho-kinase pathway. In the present study we investigated whether LPA might also induce the activation of G_i that leads to a decrease of cAMP levels in platelets. Levels of cAMP in unstimulated platelets were raised from 8 ± 1 pmol/10⁸ platelets (n = 5) to 54 ± 24 pmol/10⁸ platelets (n = 7) by the prostacyclin-analog iloprost. In platelets treated with a large concentration range of LPA (0.1-80 μM) there was no significant change in the concentration of cAMP (n = 7), whereas epinephrine (10 μM) showed a decrease in cAMP to 10 ± 5 pmol/10⁸ platelets (n = 7), respectively. Therefore, platelet LPA receptors are unlikely to couple to G_i. We also have previously obtained evidence that LPA receptors activated during shape change do not couple to G_q.⁷  Therefore, the only G protein activated by LPA during shape change is the G_12/13 family.

Next, we investigated whether LPA receptor-coupled G_12/13 activation might synergize with other platelet receptors that couple to G_i or G_q in inducing platelet aggregation. Stimulation of washed platelets by 1-acyl-LPA (16:0) together with epinephrine, known to selectively activate G_i in platelets, or together with ADP, activating G_i through the P2Y₁₂ receptor, synergistically induced platelet aggregation (Figure 4A-B). Moreover, also 1-acyl-LPA (16:0) and serotonin, which selectively stimulates G_q and Ca²⁺ mobilization in platelets, synergized in inducing platelet aggregation (Figure 4C). Interestingly, the synergistic platelet aggregation of LPA plus serotonin, but not of LPA plus epinephrine, was inhibited by the ADP receptor antagonists indicating a different dependency of the aggregation response on released ADP (Figure 4B-C). We also observed that LPA synergized with ADP, epinephrine, and serotonin in inducing platelet aggregation in blood (data not shown).

To explore the signal transduction pathway stimulated by LPA receptor/G_12/13 activation possibly synergizing with the signaling pathway induced by G_i and G_q activation, we studied the involvement of the Rho/Rho kinase pathway, which is known to mediate LPA-induced shape change in washed platelets.^8,33  Although LPA-induced shape change in whole blood was blocked by pretreatment of platelets with the Rho-kinase inhibitor Y-27632 (Figure 3B), aggregation induced by LPA plus low concentrations of ADP (0.5 and 1 μM) in blood was not inhibited by Y-27632, excluding a role of the Rho/Rho-kinase pathway in the synergistic aggregation response (data not shown).

LPA produced by thrombin or collagen-aggregated platelets might play a role as a positive feedback mediator of platelet activation. Washed platelets produce about 5 ng LPA/10⁹ cells after aggregation with thrombin (2 U/mL) for 2 minutes that will lead to a maximal concentration of 10 nM LPA assuming a complete release of LPA from platelets to the extracellular medium.³⁴  With the aim of analyzing the possible role of endogenous LPA as a positive feedback loop for aggregation, washed platelets were pretreated with the LPA receptor antagonist NPSerPA and then stimulated with different concentrations of collagen and thrombin. NPSerPA (10 μM), which completely blocked shape change induced by 0.1 to 0.5 μM 1-acyl-LPA (16:0) (data not shown), had no effect on thrombin- or collagen-induced platelet aggregation (Table 3). These results indicate that LPA formed by activated platelets does not mediate or support stimulus-induced platelet aggregation.

LPA (20 μM) stimulated the formation of platelet-monocyte aggregates in anticoagulated blood from 32% ± 13.5% (control, stirred samples) to 89% ± 19% (mean ± SD, n = 13; Figure 5). A similar increase was seen after addition of 5 μM ADP (72% ± 25%) and 0.1 μg/mL collagen (75% ± 31%; values are mean ± SD, n = 11 and 8).

The stimulation of platelet-monocyte aggregate formation was dose and time dependent (Figure 5A and data not shown). Platelet-monocyte aggregate formation was significantly increased after 2.5 μM LPA and was maximal after 20 μM LPA (87% ± 12%). Preincubation of blood with a blocking P-selectin antibody completely blocked LPA-induced platelet-monocyte aggregate formation. These results indicate that platelet-monocyte aggregate formation in LPA-stimulated platelets is mediated by P-selectin, an adhesion molecule that is expressed on the platelet surface on platelet activation. LPA-induced platelet aggregation was not inhibited (Table 1), indicating that the P-selectin antibody did not unspecifically inhibit LPA-induced platelet activation.

The P2Y₁ antagonist MRS2179 (100 μM), or the P2Y₁₂ antagonist AR-C69931MX (1 μM), did not significantly inhibit the LPA-induced platelet-monocyte aggregate formation in whole blood (Figure 5B). Only both receptor antagonists together significantly reduced the LPA-induced platelet-monocyte aggregate formation (P < .05), indicating that the concomitant stimulation of both ADP receptors plays a role in the LPA-induced plateletmonocyte adhesion in whole blood. Supporting the results in whole blood, the P2Y₁ antagonist MRS2179 and P2Y₁₂ antagonist AR-C69931MX, either alone or in combination, did not inhibit P-selectin exposure in washed platelets stimulated by LPA (3, 10, 30, and 100 μM; data not shown).

Preincubation of whole blood with aspirin (1 mM) did not inhibit the LPA-induced platelet-monocyte aggregate formation. The fibrinogen receptor antagonist RGDS (5 mM) completely inhibited the LPA-induced platelet aggregation (Table 1), but not platelet-monocyte aggregate formation (66% ± 26% versus 59% ± 30%, mean ± SD, n = 9). The basal levels of plateletmonocyte aggregates were even enhanced after RGDS (59% ± 30% versus 32% ± 13%; n = 8, P < .05).

We found that LPA is capable of inducing platelet aggregation and platelet-monocyte aggregate formation in whole blood at concentrations approaching those found in vivo (0.5-1 μM).³⁵  This finding bears pathophysiologic significance for 2 reasons. First, LPA that accumulates in the lipid-rich core of atherosclerotic lesions and is exposed on plaque disruption might trigger platelet aggregation and thrombus formation at the site of plaque rupture in vivo. Based on our findings of the high biologic potency of 1-alkyl-LPA in whole blood we suggest that the alkyl-LPA species that amounts to 20% of total LPA in the lipid-rich core region of human atherosclerotic plaques might be particularly important in stimulating platelets after plaque rupture.²²  Moreover, LPA present in oxidized LDL could contribute to the platelet hyperreactivity observed in patients with high levels of circulating oxidized LDL suffering from coronary heart diseases.^1,2,4-6  Furthermore, we discovered that platelets taken from different donors showed a different aggregation response to LPA, which might be of clinical significance. Donors with high sensitivity to LPA showed a maximum platelet aggregation of 60% to 80% in whole blood and a significant aggregation in response to a concentration as low as 2.5 μM 1-acyl-LPA and 0.2 μM 1-alkyl-LPA.

With the purpose of exploring the mechanism of LPA-induced platelet aggregation in whole blood, we studied LPA-induced shape change and aggregation in blood, PRP, and washed platelets and used different platelet inhibitors. We provide evidence that LPA directly induced platelet shape change not only in washed platelets, but also in PRP and whole blood. The much higher EC₅₀ of LPA for the induction of shape change in PRP as compared to shape change in isolated platelets is due to the inhibitory effect of plasma albumin on LPA/platelet interaction (our study and Haserück et al¹²  and Packham et al³⁶ ). The dose-response curves for platelet shape change in PRP and blood were identical, indicating that platelet stimuli possibly released from blood cells, such as ADP from red cells, did not synergize with LPA in inducing platelet shape change in blood. Moreover, the LPA-induced platelet shape change in blood was not inhibited by ADP receptor antagonists. The fact that dose-response curves for shape change and aggregation in whole blood were similar suggests that analogous signaling pathways are activated during LPA-induced shape change and platelet aggregation. We found indeed that LPA, which activates G_12/13 during shape change, showed strong synergisms with other platelet stimuli in inducing platelet aggregation. These platelet stimuli were epinephrine that activates G_i, ADP that stimulates Gi and G_q, and serotonin that activates G_q. These results support previous and recent studies showing that different signaling pathways converge in inducing platelet aggregation.^27-32 

To consider the role of Rho-kinase, which is downstream of LPA-induced G_12/13 activation, we used the Rho-kinase inhibitor Y-27632. Although LPA-induced shape change in whole blood was blocked by pretreatment of platelets with Y-27632, aggregation induced by LPA or by the combination of LPA plus ADP was not affected by this drug. These results indicate that Rho-kinase is not involved in the signaling pathway of LPA that mediates platelet aggregation.

Although LPA was able to stimulate a small reversible aggregation of washed platelets independent of secreted ADP, LPA-induced platelet aggregation in whole blood was completely dependent on ADP-mediated activation of the P2Y₁ and P2Y₁₂ receptors. ADP is most likely secreted from platelet-dense granules by the action of LPA alone or in combination with another agonist. LPA on its own induced very little dense granule secretion (1%-2%) in isolated platelets, but LPA together with ADP, once secreted, effectively induced dense granule release. Remarkably, blockade of either the P2Y₁ receptor or the P2Y₁₂ receptor completely inhibited LPA-induced platelet aggregation in whole blood and dramatically reduced LPA-induced dense granule secretion of washed platelets (Figure 2A; Table 2). These results indicate that coactivation of the LPA receptors with either one of the 2 purinergic ADP receptors is sufficient to induce the full platelet response. Both ADP receptors were equally important in mediating the LPA-induced aggregation in whole blood and aggregation and dense granule secretion of washed platelets.

LPA-induced aggregation was similar in hirudin- and citrateanticoagulated blood and PRP. Some platelet agonists such as ADP and platelet-activating factor are known to stimulate thromboxane A₂ formation, thromboxaneA₂-dependent dense granule secretion and the second wave of aggregation under conditions of low concentrations of extracellular Ca²⁺.^26,36  LPA-induced platelet aggregation was independent of platelet cyclooxygenase activity, which might explain the lack of effect of the anticoagulant on LPA-induced platelet aggregation.

The reason why different donors show different aggregation responses to LPA in blood is unclear. It is unlikely to be explained by a different capacity of albumin to bind LPA; otherwise concomitant differences of the LPA-induced shape change in the different donor groups should have been observed. We assume that platelets from some donors reach ADP secretion in response to LPA easier, whereas shape change is similar in platelets of all donors. The released ADP then synergizes with LPA in stimulating platelet aggregation. Based on the results with washed platelets only little dense granule secretion seems to be required for LPA-stimulated aggregation. The process of secretion might not depend solely on LPA but on the presence of an additional agonist present in blood. As illustrated in Figure 4C by the synergism of LPA with serotonin in washed platelets, LPA in synergy with another agonist may similarly cause platelet aggregation in whole blood, which can be blocked by ADP receptor antagonists.

It is likely that LPA stimulates human platelets also in vivo. Indeed the study that first identified LPA as the platelet-activating substance in aged serum of cats reported a systemic effect of LPA on platelets after intravenous application of this phospholipid. It induced thromboembolism.¹⁰  Interestingly, only human and cat platelets are sensitive to LPA. Platelets from other species such as pig, dog, rabbit, rat, and, very importantly, the mouse cannot be activated by LPA in vitro (Schumacher et al¹⁰  and Tokumura et al¹¹ ).

LPA is also formed on platelet aggregation, but it does not seem to be a mediator of platelet aggregation induced by physiologic stimuli.^26,37  Also in our study, the LPA receptor antagonist NPS-erPA did not inhibit collagen- or thrombin-induced platelet aggregation. Therefore, LPA is unique in being a pathophysiologically, but not physiologically, important platelet agonist. Hence, we expect drugs that selectively interfere with platelet activation induced by LPA could be beneficial in preventing and treating cardiovascular diseases, without the major side effects of antiplatelet drugs, that is, bleeding.

LPA also stimulated the formation of platelet-monocyte aggregates in whole blood. The dose-response curve was similar to the one for whole blood platelet aggregation. However, within the same experiment the extent of platelet aggregation and plateletmonocyte aggregate formation often did not correlate, suggesting that these 2 LPA-induced responses are regulated differently. LPA-induced platelet-monocyte adhesion was mediated by P-selectin, which is expressed on the platelet surface after platelet activation. The P2Y₁ antagonist and the P2Y₁₂ antagonist separately did not significantly inhibit LPA-induced platelet-monocyte aggregate formation in whole blood and P-selectin exposure of washed platelets, indicating that LPA can stimulate platelet P-selectin exposure and subsequent platelet-monocyte aggregate formation through a mechanism independent of a synergistic interaction with ADP. Our results with RGDS, a fibrinogen receptor antagonist, that blocked aggregation, but not platelet-monocyte aggregate formation, show that these 2 LPA-induced responses are regulated independently of each other. Platelets respond to LPA either with secretion of ADP from dense granules leading to a synergistic aggregation response, or with P-selectin expression and platelet-monocyte aggregate formation.

The finding that low LPA concentrations stimulate the formation of platelet-monocyte aggregates in whole blood is of possible pathophysiologic relevance. Oxidatively modified LDL, which contains LPA and is elevated in plasma of patients with coronary artery disease, might lead to the formation of platelet-monocyte aggregates, which are a more sensitive indicator of in vivo platelet activation than platelet P-selectin and have been recently recognized as an early marker of acute myocardial infarction.^13,38  Moreover, the adherence of platelets to monocytes stimulates the expression of tissue factor on monocytes, promotes fibrin formation, and might accelerate intravascular thrombosis.³⁹  Plateletmonocyte interaction also stimulates the synthesis of cytokines such as interleukin 6 that increases the hepatic production of C-reactive protein, a risk factor of acute myocardial infarction.^40,41 

As effective LPA concentrations in blood approach those reached in vivo, factors that might neutralize this activation are important. Albumin binds LPA with a high affinity and inhibits the interaction of LPA with platelets. However, another plasma protein, gelsolin, which also binds LPA with a high affinity, did not inhibit the interaction of LPA with platelets in our study (data not shown). Furthermore, ectoenzymes inactivating LPA such as the lipid phosphate phosphatase 1 and 2 expressed on blood cells and endothelial cells might keep the local concentration of LPA low and limit the platelet-activating effect of circulating LPA.

Aspirin did not inhibit LPA-induced platelet aggregation in whole blood, PRP, or washed platelets. It also did not inhibit LPA-induced P-selectin exposure of washed platelets and plateletmonocyte aggregate formation in whole blood. Such a complete independence of platelet aggregation and P-selectin exposure of cylooxyxgenase-derived prostaglandin endoperoxides and thromboxane A₂ is not known for other platelet stimuli, except for ADP. We suggest that ADP receptor antagonists that, in contrast to aspirin, effectively inhibited LPA-induced platelet aggregation might reduce LPA-triggered thrombus formation in vivo. Based on our study we would expect that particularly patients with platelets that are highly sensitive to LPA should benefit from this type of antiplatelet drugs.

The technical assistance of Nicole Wilke is greatly appreciated. The results are part of the doctoral thesis of N.H., at the University of Munich.