D. J. Loskutoff, Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, VB-3, La Jolla, CA 92037, USA. Tel.: +1 858 784 7125; fax: +1 858 784 7353; e-mail: email@example.com
Summary. Obesity is associated with elevated levels of leptin in the blood. Elevated leptin is a risk factor for thrombosis in humans, and leptin administration promotes platelet activation and thrombosis in the mouse. The current study examines the effect of leptin on human platelets, and provides initial insights into the nature of the leptin receptor on these platelets. Leptin potentiated the aggregation of human platelets induced by low concentrations of ADP, collagen and epinephrine. However, the response varied significantly between donors, with platelets from some donors (approximately 40%) consistently responding to leptin (responders) and those from other donors (approximately 60%) never responding (non-responders). Western blotting and reverse transcriptase-polymerase chain reaction (RT-PCR) experiments showed that platelets from both groups only express the signaling form of the leptin receptor, and that responder platelets express higher levels of this receptor than non-responders. Ligand-binding assays demonstrate specific, saturable binding of leptin to platelets from both groups with apparent Kd values of 76 ± 20 nm for responders and 158 ± 46 nm for non-responders. Thus, the decreased sensitivity of non-responder platelets to leptin does not result from the absence of the signaling form of this receptor, but may reflect differences in its level of expression and/or affinity for leptin. These preliminary studies demonstrate that platelets are a major source of leptin receptor in the circulation, and suggest that leptin-responsive individuals may have a higher risk for obesity-associated thrombosis than non-responsive individuals.
Obesity is an independent risk factor for atherosclerosis, thrombosis, stroke and myocardial infarction , and is thus a major health problem in westernized societies. In spite of this, the nature of the pathways/factors that promote these disorders in obesity are largely unknown. Recent studies suggest that in addition to metabolic derangements  and alterations in the fibrinolytic and coagulation systems , elevated leptin may be a risk factor for cardiovascular disease in this condition .
Leptin is the 167-amino acid protein product of the ob gene, and it regulates food intake and energy expenditure in mammals . However, leptin receptors are widely expressed [6,7], raising the possibility that leptin may have broad effects. In fact, it is now clear that leptin influences a variety of physiological and pathological processes  including angiogenesis  and vascular disorders . In this regard, leptin-deficient ob/ob mice [11,12], and wild-type (WT) mice treated with a leptin-neutralizing antibody , develop unstable thrombi and an attenuated thrombotic response to arterial injury, and they exhibit a defect in platelet aggregation. Leptin also promotes thrombosis in the mouse, and it potentiates the aggregation of murine platelets in vitro in a leptin receptor-dependent manner. Although some studies have demonstrated a potentiating effect of leptin on human platelet aggregation [14,15], others were unable to confirm this observation .
This report shows that there are two distinct populations of healthy, non-obese donors, those whose platelets consistently respond to leptin (responders), and those that do not (non-responders). Leptin responsiveness neither results from the presence or absence of plasma factors that stimulate or inhibit the response, nor does it correlate with body mass index (BMI). Rather, it is an inherent property of the platelets themselves.
Preparation of human platelets
Whole blood was drawn from healthy volunteers who had normal BMIs (males, 24.4 ± 2.17 kg m−2; females 23.2 ± 2.8 kg m−2) and had abstained from medications known to affect platelet function for at least 2 weeks prior to the blood sampling. Standard hematologic approaches were employed for the collection of blood and the preparation of platelet-rich plasma (PRP), platelet-poor plasma (PPP) and washed platelets . Platelets in PRP were diluted with autologous PPP to a standard concentration of 108 mL−1 and immediately employed in the aggregation experiments. Washed platelets were used at a concentration of 5 × 108 mL−1 for Western blotting, and 109 mL−1 for RNA extraction, for pull-down assays, and for ligand-binding experiments.
Platelet aggregation studies
Platelet aggregation was optically monitored using a platelet aggregometer (Sienco Inc., Wheat Ridge, CO, USA). Briefly, 400 μL of PRP was stirred in an aggregometer cuvette, the appropriate agonist was added, and aggregation was monitored for 5 min at 37 °C as the change in light transmission of the PRP sample, compared with that of PPP. The platelet agonists used were adenosine diphosphate (ADP), type 1 collagen and epinephrine (Sigma, St. Louis, MO, USA). To determine the effect of leptin on platelet aggregation, aliquots of PRP were pretreated with the indicated concentrations of human recombinant leptin (Peprotech Inc., Rocky Hill, NJ, USA) for 10 min at room temperature prior to addition of the aggregating agents.
The effect of leptin on platelet aggregation was also monitored by electron microscopy. In these experiments, platelets (3 × 108 mL−1) were aggregated as above, and after 5 min, 200 μL of the PRP was fixed and processed as described . Platelets were observed using a Philips CM100 (FEI, Hillsborough, OR, USA) electron microscope at 80 kV, and photographed using Kodak SO163 film (Kodak, Rochester, NY, USA).
Washed platelets were lysed by the addition of an equal volume of 2x sodium dodecyl sulfate (SDS) sample buffer  containing 10%β-mercaptoethanol. The lysates were analyzed by Western blotting using a polyclonal antibody that detects both the large and the lower molecular weight forms of the leptin receptor (H-300; Santa Cruz Biotechnology, Santa Cruz, CA, USA), followed by horse radish peroxidase (HRP)-conjugated goat antirabbit IgG. Immunoreactivity was detected by enhanced chemiluminescence (ECL) (Amersham Biosciences, Piscataway, NJ, USA). Quantitation was achieved by densitometric scanning of the film using an Alpha Innotech (San Leandro, CA, USA) imaging system.
RNA isolation and real-time RT-PCR
Total RNA was extracted from washed platelets by employing the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and analyzed by real-time polymerase chain reaction (PCR) as described , employing specific primers (Invitrogen) for the long form of the leptin receptor (i.e. for the intracellular domain): forward 5′-GCCAACAACTGTGGTCTCTC-3′, and reverse 5′-AGAGAAGCACTTGGTGACTG-3′. Quantification of the leptin receptor gene, expressed as relative mRNA level compared with a control (i.e. a sample of human platelet cDNA used as an internal standard in order to normalize the experimental variability between RT-PCR experiments), was calculated after normalization to β-actin .
Binding of leptin to platelets
Human recombinant leptin was radioiodinated with Na-125I (Perkin Elmer, Shelton, CT, USA) to a specific activity of 3900 cpm ng−1 using Iodo beads (Pierce Chemical Co., Rockford, IL, USA). Washed platelets were incubated with the indicated amounts of the 125I-labeled leptin for 2 h at 4 °C in the presence or absence of a 50-fold final molar excess of unlabeled leptin. Except as specifically indicated, leptin–receptor complexes on platelets were stabilized by cross-linking with 3,3′-Dithiobis[sulfosuccinimidyl propionate] (DTSSP) (Pierce) for 30 min at room temperature. The cross-linking reaction was quenched by the addition of Tris buffer (pH 7.5) to a final concentration of 20 mm. For ‘pull-down’ experiments, the treated platelets were washed by centrifugation, lysed in ice-cold 10 mm 3-[(3-chloromidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) detergent (Sigma), and then incubated overnight at 4 °C with the rabbit antileptin receptor polyclonal antibody in the presence of protein G-agarose beads (Amersham). The beads were washed and extracted into reducing sample buffer, and the extracts were fractionated by SDS-polyacrylamide gel electrophoresis (PAGE). The gels were dried and 125I-labeled leptin bands were detected by autoradiography. For ligand-binding experiments, the treated platelets were layered over 300 μL of 20% sucrose in HB buffer (20 mm HEPES, pH 7.4, 150 mm NaCl, 2 mm CaCl2), and centrifuged for 2.5 min in a Beckman microfuge (Fullerton, CA, USA) to separate bound from free 125I-labeled leptin. The tube tips were removed with a razor blade and counted in a gamma counter. Molecules of leptin bound per platelet were calculated from the specific activity of the ligand. Specific binding was defined as the difference in the cpm bound to the platelets in the presence or absence of the 50-fold final molar excess of unlabeled leptin. Dissociation constants (Kd), Scatchard plots and linear correlation coefficients were determined from specific binding isotherms using a nonlinear curve-fitting function (GraphPad Prism Software; GraphPad, San Diego, CA, USA). Ligand-binding experiments were also performed as above using Meg-01 cells (106 mL−1).
Human megakaryoblastic cells (Meg-01) were maintained in RPMI medium (Invitrogen) supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, 1% glutamine and 1% sodium pyruvate.
All calculations were performed using Prism 3.02 software. Statistical significance between two groups was determined by the Student's t-test.
Effect of leptin on platelet aggregation
Experiments were performed to determine the influence of leptin on the aggregation of human platelets. Although leptin alone had no effect (not shown), it consistently potentiated the ADP-mediated aggregation of platelets from some donors (Fig. 1A,B). Unexpectedly, leptin had little effect on the aggregation of platelets from other donors (Fig. 1C,D). Thus, there appear to be two distinct populations of donors, those who respond to leptin (responders) and those that do not (non-responders). Donors were considered to be responders if the weak, reversible aggregation induced by low concentrations of ADP was consistently (i.e. on different days) converted into irreversible aggregation by leptin in a dose-dependent manner (Fig. 1B). Donors were considered to be non-responders if leptin consistently failed to potentiate ADP-mediated aggregation, even when high concentrations (i.e. 2 μg mL−1) of leptin were employed (Fig. 1D). Preliminary analysis of the responsiveness of platelets from 56 different donors indicated that approximately 40% were responders and 60% were non-responders (data not shown). The effects of leptin on ADP-mediated platelet aggregation were confirmed by electron microscopy (Fig. 2). Responsive platelets were only minimally aggregated when they were stimulated with low doses of ADP alone (Fig. 2A), and most still contained numerous electron dense granules (not shown). However, when responsive platelets were pretreated with leptin, the addition of the same low concentrations of ADP resulted in the formation of large, densely packed aggregates (Fig. 2B), and most of the platelets were now completely degranulated (not shown). This potentiating effect of leptin was not observed when platelets from non-responsive donors were employed (Fig. 2C,D). Leptin also potentiated the aggregation of responsive platelets by collagen (Fig. 3A) and epinephrine (Fig. 3B), but it had no effect on the aggregation of non-responsive platelets by these same agonists (Fig. 3C,D).
Effect of plasma on the responsiveness of platelets to leptin
Mixing experiments were performed to determine whether platelet responsiveness was determined by the presence of plasma factors that either enhance or decrease their sensitivity to leptin. In these experiments, PRP from responders was mixed with equal amounts of PPP from non-responders, and vice versa, and then the effect of ADP and leptin on the aggregation of platelets in the various mixtures was determined. Fig. 4 shows that when responsive platelets were mixed with PPP from either responsive or non-responsive donors, they remained responsive (Fig. 4A), and vice versa (Fig. 4B). These results demonstrate that plasma factors do not determine the leptin-responsiveness of platelets from different donors, and suggest that this responsiveness is an inherent property of the platelets themselves.
Presence of leptin receptor on platelets
The biological functions of leptin are mediated through its membrane-associated receptor , and it is now clear that there are multiple splice variants of this receptor, one of which lacks the cytoplasmic domain and cannot signal in response to leptin [7,19,20]. Thus, it is possible that responsive platelets express the long, signaling form of the leptin receptor while non-responsive platelets express the short, non-signaling isoform. This does not seem to be the case, as the lower molecular weight, non-signaling form was not detected in platelet extracts prepared from either responsive or non-responsive donors by Western blotting (Fig. 5A). In fact, a single leptin receptor isoform with the same approximate molecular mass (200 kDa) was present in both extracts. Figure 5 does show that non-responder platelets contain less than half of the receptor antigen (Fig. 5A,B) and mRNA (Fig. 5C) than present in responder platelets. Thus, the decreased sensitivity of non-responder platelets to leptin may result, at least in part, from the presence of lower levels of the signaling isoform of this receptor.
Binding of leptin to platelets
Two approaches were taken to define the affinity constants associated with the binding of leptin to its receptor on platelets. In the first, platelets from responsive donors were incubated with 125I-labeled leptin, lysed with CHAPS, and then incubated with an antileptin receptor antibody in order to co-immunoprecipitate labeled leptin bound to its receptor on the platelet surface. These initial studies clearly demonstrated that leptin could be ‘pulled down’ in complex with its receptor on platelets (Fig. 6A, lane 1). However, because the amount pulled down varied considerably between experiments (not shown), we repeated the experiments in the presence of the cross-linker DTSSP . This treatment significantly increased the amount of leptin recovered in the samples (Fig. 6A, compare lanes 1, 3). Under these conditions, binding was dose-dependent (Fig. 6A, compare lanes 3–5) and specific as it was competed by the addition of a 50-fold excess of unlabeled leptin (Fig. 6A, compare lanes 1 vs. 2 and 3 vs. 6).
Ligand-binding experiments (Fig. 6B) show that leptin binding is dose-dependent and saturable, and that platelets from responders again (see Fig. 5) bind approximately 2.5-fold more leptin than platelets from non-responders. Representative Scatchard plots for these interactions are shown in Fig. 6C. The linear correlation coefficient for responders and non-responders were 0.941 and 0.977, respectively. Nonlinear curve fitting analysis of specific isotherms from several experiments gave an average Kd of 76 ± 20.5 nm for responder platelets, with 2.4 ± 0.3 × 103 leptin molecules maximally bound per platelet, and 158 ± 46 nm for non-responder platelets, with 0.9 ± 0.09 × 103 leptin molecules maximally bound per platelet. The relatively high Kd values determined for platelets in the current study were not the result of technical problems (e.g. partial inactivation of leptin during the iodination procedure), since the same 125I-labeled leptin binds to Meg-01 cells with an apparent Kd of 4.5 ± 1.8 nm (Fig. 6D), well within the normal range [22,23]. Thus, the binding of leptin to its receptor on platelets is of relatively low affinity.
The biological spectrum of the adipocyte hormone leptin is now known to extend far beyond the regulation of food intake and body weight . In particular, leptin appears to represent a novel, direct link between obesity and cardiovascular disease [10,24,25], and studies in the mouse suggest that platelets may be a critical component of this link [11,12]. As platelet activation plays a key role in the formation of arterial thrombi  and in many of the inflammatory processes that accompany the development of atherosclerotic plaques [27,28], we sought to determine the influence of leptin on human platelet function. We thus examined the effect of leptin on the activation of platelets from 56 different donors. Unexpectedly, we found that the effects varied between donors, with leptin consistently promoting the ADP-mediated aggregation of platelets from approximately 40% of the donors, but having no apparent effect on platelets from the other donors (Figs 1 and 2). Repeated independent measurements performed on different days with three different agonists (ADP, type-1 collagen and epinephrine) gave similar results (Fig. 3). Thus, there are two distinct populations of donors in the subjects we examined, responders and non-responders. These differences in leptin responsiveness may help to explain conflicting observations in the literature regarding the effects of leptin on human platelets [11,14–16]. The observation that inhibition of circulating leptin protects against thrombosis  raises the possibility that non-responsive donors may also be protected from such acute events. This possibility remains to be tested.
A variety of experiments were performed to determine why platelets from some donors respond to leptin while those from other donors do not. Leptin responsiveness appears to be blunted in platelets from obese subjects , raising the possibility that the non-responsive platelets may be from obese, leptin-resistant donors. However, no obese individuals were employed in our study group, and plasma leptin and soluble leptin receptor levels were within the normal range for healthy individuals (data not shown). Thus, it is unlikely that the non-responsive platelets originate from leptin-resistant donors. Leptin responsiveness did not correlate with age or gender, and mixing experiments (Fig. 4) showed that plasma does not contain molecules that influence responsiveness. Experiments were therefore performed to investigate the role of platelets, and in particular, of leptin receptors on platelets, in this response. Except for one very preliminary study , the functional and biochemical properties of platelet leptin receptors, and their structural relationship to leptin receptors on other cells, have not been reported.
Western blotting experiments (Fig. 5) show that both responsive and non-responsive platelets only express the long, signaling form of the leptin receptor, and that neither express the short non-signaling form. Thus, the responsiveness of platelets to leptin is not determined by the presence or absence of specific isoforms of this receptor. However, leptin receptor expression by non-responder platelets was less than half of that detected in responders. The lower expression levels may contribute to the decreased sensitivity of these platelets to leptin.
It is also possible that non-responder platelets express a lower-affinity variant of the leptin receptor compared with responders. Again, the kinetic constants for the binding of leptin to its receptor on platelets have not been reported. Figure 6 shows for the first time that leptin actually binds to its receptor on human platelets, and that it does so in a specific and saturable manner (Fig. 6A,B). Responder platelets bound 2.5-fold more radiolabeled leptin than non-responders, consistent with the amount of leptin receptor actually detected in these platelets by Western blotting (Fig. 5). The number of leptin receptors expressed on the platelet membrane (2.4 ±0.3 × 103 molecules per platelet for responders, 0.9 ±0.09 × 103 molecules per platelet for non-responders; Fig. 6C) is similar to the abundancy of other platelet receptors [29,30]. The differences in the apparent Kd values (76 ±20.5 nm for responders; 158 ± 46 nm for non-responders) suggest that platelets from non-responders may also express a lower affinity variant of the leptin receptor. However, additional experiments with more samples are necessary to verify this conclusion.
The affinity of leptin for its receptor on platelets is lower than that reported for the brain [22,23] or observed here for cultured Meg-01 cells (Fig. 6). Although the biochemical basis of this relative low affinity is unknown, the same Kd values were obtained using platelets extracted into CHAPS (data not shown). Thus, the receptor on the platelet surface does not appear to be masked or constrained by membranes. It should be noted that low-affinity binding sites for leptin were also detected on the surface of cultured endothelial cells  and in human brain capillary plasma membranes . Importantly, both endothelial cells  and platelets (Fig. 5) express a relatively large (approximately 200 kDa) form of the receptor, possibly the result of extensive glycosylation. In this regard, there are 20 N-glycosylation sites on the human leptin receptor , and one of these sites occurs in a motif (WSXWS) that is common in cytokine receptors. This site is critical for the correct folding of the binding region in the interleukin-2 receptor , raising the possibility that the decreased affinity may result from changes in N-glycosylation.
In summary, our data show that the potentiating effect of leptin on human platelet aggregation is mediated by the actual binding of leptin to the long, signaling form of its receptor on the platelet surface, and that platelets do not express the short, non-signaling variant of this receptor. These preliminary observations also suggest that differences in the concentration, and possibly the affinity, of the signaling receptor on platelets may determine, at least in part, their responsiveness to leptin. It is likely that differences in additional platelet components (e.g. downstream intracellular signaling molecules) may also contribute to their responsiveness.
This work was supported in part by NIH grants HL47819 and HL 59549 to D.J.L., by NIH grant M01 RR00833 to the General Clinical Research Center for the collection of blood, by a grant from the MURST (Ministero dell'Universita’ e della Ricerca Scientifica e Tecnologica) to G.G. and by a fellowship from the American Heart Association to C.D. The authors thank M. Wood, J. Neels, P. Sartipy and S. Konstantinides for fruitful discussions, and M. McRae for preparing the manuscript. This is manuscript number 15893-CB from The Scripps Research Institute.