Insulin-like growth factor-1 potentiates platelet activation via the IRS/PI3Kα pathway

Vascular damage leads to the rapid recruitment of platelets to the site of injury. Platelets adhere to the newly exposed von Willebrand factor (VWF) and collagen in the extracellular matrix, leading to subsequent platelet activation and the secretion of dense and α granule content into the extracellular media. Platelet granules contain a range of autocrine agonists that can activate or modulate the function of other platelets to form aggregates, including ADP, serotonin, and possibly Gas6^1,2  and tachykinins.³  Another factor that is found in the α granules in platelets is insulin-like growth factor-1 (IGF-1),^4,5  a growth factor involved in cell proliferation, differentiation, and migration.⁶  The presence of IGF-1 at the site of vascular injury is essential in the wound-healing process,⁷  and local delivery by platelets may therefore support the healing process

IGF-1 is a peptide hormone whose concentration in plasma increases in response to growth hormone, largely due to increased production by the liver.⁸  The effect of IGF-1 is modulated by multiple IGF-binding proteins (IGFBPs), which bind IGF-1 and thereby serve as transporter proteins and storage pools.⁶  The receptor for IGF-1 has a wide tissue expression and is closely related to the insulin receptor. The insulin receptor is involved in plasma glucose homeostasis and therefore signals mainly to metabolic responses, whereas the IGF receptor couples primarily to mitogenic responses. Both receptors are heterodimers composed of two ligand-binding α subunits linked by disulphide bridges to two transmembrane β subunits that possess ligand-stimulated tyrosine kinase activity. Ligand binding and subsequent activation leads to the phosphorylation of the major substrates Shc and the insulin receptor substrate (IRS) proteins IRS-1 and IRS-2. Tyrosine phosphorylation of IRS proteins on specific residues creates binding sites for Src homology (SH2)–containing proteins such as the tyrosine phosphatase SHP-2 and the regulatory p85 subunit of phosphoinositide-3 kinase (PI3K), thereby increasing their enzyme activity and/or localizing them to their substrates. Activation of PI3K leads to a rapid rise in PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃, which results in the recruitment of protein kinase B (PKB) to the plasma membrane and subsequent phosphorylation of PKB on Thr308 by PDK1 and Ser473 by second kinase, which is likely to be the mTOR/Rictor complex.⁹  Subsequent activation of PKB is involved in numerous cellular responses to insulin and IGF-1.^10,11 

Insulin resistance and altered blood IGF-1 levels have been implicated in the development of cardiovascular disease.^12,13  Insulin resistance is described as the reduced ability of insulin to lower glucose concentrations in the blood, and is closely associated with obesity and diabetes.¹³  Increased free IGF-1 levels and decreased levels of IGFBP-1 are found in serum of obese subjects compared with lean subjects, whereas total IGF-1 levels are within the normal range.¹⁴  Raised local levels of IGF-1 are also linked to the accumulation of vascular smooth muscle cells (VSMCs) in restenotic plaques, which frequently occur after angioplasty,¹²  and to microangiopathy (in kidney and eye) in patients with diabetes.¹⁵  Furthermore, patients with acromegaly, a condition characterized by high growth hormone and IGF-1 levels, have a significantly increased risk of cardiovascular disease.¹⁶ 

In contrast, low levels of IGF-1 have also been associated with atherosclerotic disease, such as ischemic heart disease,¹⁷  myocardial infarction,¹⁸  and ischemic stroke.¹⁹  Reduced levels of total IGF-1 are found in patients with type II diabetes, especially patients with poor metabolic control and prolonged duration of the disease.^20,21  The association of low IGF-1 concentrations with atherosclerotic disease may partially be explained by the ability of IGF-1 to prevent apoptosis, promote cell survival, and stimulate regeneration of cardiac myocytes.²²  Low IGF-1 concentrations may therefore promote apoptosis, resulting in cardiomyocyte dysfunction and plaque instability.

Platelets play an essential role in hemostasis and contribute significantly to the pathogenesis of cardiovascular disease.²³  They express high levels of the IGF-1 receptor on their plasma membrane,²⁴  and local increases in free IGF-1 concentrations may therefore play an important role in the autocrine and paracrine regulation of platelet function. To obtain an insight into the role of IGF-1 in hemostasis and pathologic conditions such as cardiovascular disease, it is essential to delineate the physiologic role of IGF-1 and the intracellular signaling pathways that are involved in its modulation of platelet function.

Approval was obtained from the United Bristol Healthcare Trust institutional review board for these studies. Informed consent was obtained in accordance with the Declaration of Helsinki.

The anti–IRS-1 (C-20) antibody, the anti-IGF receptor antibody (C-20), and the PKBα (B1) antibody were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The rabbit polyclonal antibody against the p85 regulatory subunit of PI3K, the anti–IRS-2 antibody, and the anti-phosphotyrosine antibody 4G10 were obtained from Upstate Biotechnology (Lake Placid, NY). The phosphospecific antibody against Ser473 was from Cell Signaling Technology (New England Biolabs, Hitchin, United Kingdom). The PAR-1 agonist SFLLRN was from Bachem Bioscience (Weil am Rhein, Germany). Wortmannin, LY294002, UO126, and GF109203X (BIM1) were all from Tocris (Avonmouth, United Kingdom). The P2Y12 antagonist AR-C69331-MX and the IGF receptor inhibitor NVP-AEW541 were generous gifts from AstraZeneca (Alderley Park, United Kingdom) and Novartis Pharma AG (Basel, Switzerland), respectively. Indomethacin, apyrase, human IGF-1, adenosine 3′5′ diphosphate (A3P5P), and rapamycin were from Sigma (Poole, United Kingdom). The PI3K inhibitor ZSTK474 and the mouse monoclonal neutralization antibody against the IGF receptor αIR3 were from Calbiochem (La Jolla, CA). The PI3K isoform-specific inhibitors GTX-155, GTX-221, IC87114, PI-103, and PIK-75 were kindly provided by Prof Bart van Haesebroeck (Ludwig Institute for Cancer Research, London, United Kingdom), Dr Peter Shepherd (University of Auckland, New Zealand), and Prof Kevin Shokat (University of California, San Francisco, CA). Luciferase was from Chrono-log (Labmedics, Manchester, United Kingdom). The IV.3 antibody against the immunoglobulin (IgG) receptor FcγRIIa was kindly provided by Prof Jonathan Gibbins (University of Reading, United Kingdom). F(ab′) fragments of the IV.3 antibody were generated using the Immunopure kit (Pierce, Tattenhall, United Kingdom). All other reagents were as previously described.²⁵ 

Freshly drawn venous blood from healthy volunteers who claimed not to have taken any medication in the previous 14 days was anticoagulated with 0.04-vol 4% trisodium citrate. Blood was drawn under local ethics committee agreement. The blood was acidified with 0.14 vol acidic citrate dextrose (120 mM sodium citrate, 110 mM glucose, and 80 mM citric acid), and platelet-rich plasma (PRP) was obtained by centrifugation at 180g for 17 minutes. Platelets were subsequently pelleted by centrifugation at 650g for 10 minutes in the presence of 10 μM indomethacin and resuspended in HEPES-Tyrode buffer, pH 7.2 (145 mM NaCl, 5 mM KCl, 0.5 mM Na₂HPO₄, 1 mM MgSO₄, and 10 mM HEPES), supplemented with 0.1% glucose [wt/vol]) and 10 μM indomethacin.

Platelets (1.3 × 10⁸ platelets/mL) were incubated for 10 minutes with a range of inhibitors, or for 3 minutes with A3P5P, AR-C69331, and apyrase. IGF-1 was added 5 minutes before stimulation with the PAR-1 agonist SFLLRN, ADP, collagen-related peptide (CRP), or the thromboxane analog U464119, and aggregation was recorded at 37°C under continuous stirring at 1200 rev/min in a Chrono-log 490-4D Aggregometer (Labmedics). For IGF receptor neutralization studies, platelets were incubated for 3 minutes with saturating concentration of the F(ab′) fragment IV.3 to block the IgG receptor FcγRIIa, followed by incubation for 3 minutes with the IGF receptor neutralization antibody IR3 and subsequent stimulation with SFLLRN. The optimal saturating concentration of F(ab′) fragment was determined by its ability to block platelet aggregation stimulated by cross-linking of the IV.3 antibody on platelets.

Platelets (4 × 10⁸ platelets/mL) were incubated for 10 minutes with a range of inhibitors, or for 3 minutes with A3P5P, AR-C69331, and apyrase, and subsequently stimulated under nonstirring conditions with the indicated concentration of IGF-1, thrombopoietin (TPO; 50 ng/mL), leptin (100 ng/mL), or SFLLRN (5 μM). For IGF receptor neutralization studies, platelets were incubated for 3 minutes with saturating concentration of the F(ab′) fragment IV.3, followed by incubation for 3 minutes with various concentrations of the IGF receptor neutralization antibody αIR3 and subsequent stimulation with 100 nM IGF-1 for 5 minutes. At different time intervals after stimulation 100-μL aliquots were taken and added to 33 μL 4 × Laemmli sample buffer. Alternatively, the sample (250 μL) was extracted 1:1 (platelet volume–volume) in ice-cold Nonidet P-40 (NP40) extraction buffer (50 mM Tris [pH 7.5] containing 1% NP40, 1 mM EDTA, 120 mM NaCl, 50 mM NaF, 40 mM β-glycerophosphate, 1 mM benzamidine, 1 μM microcystin, 10 mM sodium orthovanadate, and 1 μg/mL each of pepstatin, leupeptin, and antipain). Cell extracts were centrifuged at 10 000g for 10 minutes at 4°C, and the infranatant was taken for subsequent analysis.

The IGF receptor, IRS-1, IRS-2, or p85 was immunoprecipitated by rotating 250 μL total cell extract with 1 μg antibody and 20 μL protein A–Sepharose (50% wt/vol) at 4°C. The protein A–Sepharose beads were isolated by centrifugation and washed 3 times in NP40 extraction buffer. Subsequently, Laemmli sample buffer was added, and proteins were separated by SDS–polyacrylamide gel electrophoresis (PAGE) for immunoblotting.

Proteins were separated by SDS-PAGE using 6% or 7.5% gels and transferred to polyvinylidene difluoride membranes. The membranes were blocked in 10% (wt/vol) bovine serum albumin dissolved in Tris-buffered saline with 0.1% Tween (TBS-T; 20 mM Tris, 137 mM NaCl, and 0.1% Tween) and subsequently incubated with primary and secondary antibodies, which were diluted in TBS-T containing 5% (wt/vol) bovine serum albumin. Blots were washed at least 5 times for 5 minutes in TBS-T after each antibody incubation. Membranes were developed using an enhanced chemiluminescence (ECL) detection system (GE Healthcare, Amersham, United Kingdom). Primary antibodies were used at a concentration of 1 μg/mL. Horseradish peroxidase–conjugated secondary antibody (Amersham, Little Chalfont, United Kingdom) was diluted 1:10 000 for all antibodies.

The release of ATP from 250 μL platelets (1.3 × 10⁸ platelets/mL) was measured simultaneously with aggregation using a PICA (platelet-ionized calcium aggregometer; Chrono-log) at 37°C, with constant stirring (1200 rev/min) and a gain setting of 0.005. Platelets were incubated with firefly luciferase/luciferin reagent (5 μL; Chrono-log) for 2 minutes before stimulation with 0.5 μM of the PAR-1 agonist SFLLRN. The release of ATP was recorded as an increase in light output and calibrated by the subsequent addition of 1 nmol ATP standard.

Even though platelets have been demonstrated to express the IGF receptor on their plasma membrane as early as 1989,²⁴  their functional role and the signaling mechanisms that are stimulated by IGF-1 are still largely unknown. One of the first events after IGF stimulation of cells is autophosphorylation of the β subunits on tyrosine residues, leading to activation of its kinase activity and phosphorylation of downstream substrates. To investigate whether the IGF receptor is phosphorylated upon IGF-1 stimulation of platelets, the IGF receptor was immunoprecipitated and subsequently immunoblotted with a generic antiphosphotyrosine antibody. Figure 1 A demonstrates that the IGF receptor is dose-dependently phosphorylated in the presence of increasing concentrations of IGF-1. Phosphorylation was detectable at IGF concentrations as low as 1 to 5 nM and reached a plateau at 50 to 200 nM. Furthermore, phosphorylation of the IGF receptor was very rapid, reaching a maximum within 1 minute after stimulation (Figure 1B). These results demonstrate that IGF-1 stimulates strong and rapid phosphorylation of the IGF receptor on human platelets.

The major substrates of the insulin receptor and IGF receptor in other cell types include IRS-1 and IRS-2. Tyrosine phosphorylation of YXXM motifs in these substrates create binding sites for SH2 domain–containing proteins such as the regulatory subunit (p85) of PI3K, leading to its activation. To investigate whether IGF-1 stimulates similar pathways in platelets, IRS-1 and IRS-2 were immunoprecipitated and subsequently immunoblotted for phosphotyrosine and the presence of coimmunoprecipitated p85 PI3K. Figure 1C-D show that IGF-1 stimulation of platelets results in rapid tyrosine phosphorylation of IRS-1 and IRS-2, respectively, and their association with p85. In contrast to IGF receptor phosphorylation (Figure 1B), tyrosine phosphorylation of IRS-1 was transient and declined after 30 to 60 minutes of stimulation (Figure 1C). The association between p85 and IRS-1 was also rapid and relatively sustained, although binding was significantly reduced 60 minutes after stimulation (Figure 1C). Similarly, IGF-1–stimulated IRS-2 tyrosine phosphorylation and the association between p85 and IRS-2 was transient with reduced levels of phosphorylation and p85 binding present at 15 to 60 minutes (Figure 1D).

These findings were further supported by the reverse experiment where p85 was immunoprecipitated followed by immunoblotting for phosphotyrosine (Figure 1E). IGF-1 stimulated the rapid interaction between p85 and two tyrosine-phosphorylated proteins of around 160 and 170 kDa, corresponding to the molecular weight of IRS-1 and IRS-2, respectively, with the latter showing the highest intensity. The major 170-kDa tyrosine-phosphorylated protein that associates with p85 is indeed IRS-2, as the band runs at the same height as IRS-2 (Figure 1F, compare lane 2 and lane 6) and overlaps with the IRS-2 reprobe (Figure 1F bottom panel). Together these results demonstrate that IGF-1 stimulates tyrosine phosphorylation of both IRS-1 and IRS-2 and their subsequent binding to the p85 subunit of PI3K in human platelets.

One of the kinases that is activated downstream of PI3K is PKB, a member of the AGC kinase family that plays an important role in many of the metabolic and mitogenic effects of insulin and IGF-1, respectively. IGF-1 stimulation of platelets resulted in dose-dependent phosphorylation of PKB on Ser473 (Figure 2A). Phosphorylation was detected at an IGF-1 concentration as low as 5 to 10 nM (Figure 2A), although the lowest dose at which PKB phosphorylation was observed varied slightly between donors. IGF-1 stimulated rapid but transient PKB phosphorylation that was maximal within 2 minutes after stimulation (Figure 2B). Interestingly, phosphorylation of PKB closely followed tyrosine phosphorylation of IRS-2 and its association with p85 (compare Figure 2B with Figure 1D).

To compare the level of PKB phosphorylation in the presence of IGF-1 to other platelet modulators, platelets were stimulated with maximal concentrations of IGF-1, TPO, leptin, and the PAR-1 agonist SFLLRN, respectively. Figure 2C shows that IGF-1 stimulates PKB phosphorylation to an extent comparable with the PAR-1 agonist SFLLRN, a strong activator of PKB, whereas PKB phosphorylation was weak and absent after TPO and leptin stimulation, respectively.

The main route to PKB activation in platelets by various platelet modulators involves PI3K, although a minor role has been reported for the PLC/PKC pathway in thrombin-induced Ser473 phosphorylation.²⁶  To investigate whether the PI3K pathway is involved in IGF-1–stimulated PKB phosphorylation, platelets were incubated with the PI3K inhibitors wortmannin and LY294002 and subsequently stimulated with IGF-1. Figure 2D shows that IGF-1–stimulated PKB phosphorylation is completely blocked in the presence of these PI3K inhibitors (Figure 2D lanes 3 and 4), demonstrating an essential role for PI3K. In contrast to thrombin-mediated Ser473 phosphorylation,^26,27  inhibition of the PKC pathway with BIM1 had no effect on IGF-1–stimulated PKB phosphorylation (Figure 2D; lane 6). Furthermore, the MEK pathway inhibitor UO126 and the mTOR inhibitor rapamycin had no effect on PKB phosphorylation (Figure 2D lanes 5 and 7, respectively). As ADP plays an important contributory role to PKB phosphorylation by thrombin,²⁷  platelets were also incubated with the ADP receptor antagonists AR-C69331-MX/A3P5P and apyrase prior to IGF-1 stimulation. None of these compounds had an effect on IGF-1–stimulated PKB phosphorylation (Figure 2D lanes 8 and 9). These results together demonstrate an essential role for PI3K in IGF-1–stimulated PKB phosphorylation.

As IGF-1 is a strong activator of the IRS/PI3K pathway in platelets, it was important to investigate the effect of IGF-1 on platelet function. Incubating platelets with IGF-1 (1-400 nM) by itself had no effect on platelet aggregation (results not shown). Figure 3 shows the aggregation responses of platelets to a variety of platelet agonists. Preincubation with IGF-1 dose-dependently increased platelet aggregation in response to the PAR-1 agonist SFLLRN (Figure 3A). In addition, IGF-1 increased platelet aggregation in response to other platelet agonists such as CRP (Figure 3B), ADP (Figure 3C), and the thromboxane analog U46419 (Figure 3D), suggesting that IGF-1 has a general effect on platelet function. Interestingly, the potentiatory effect of IGF-1 was very rapid, as maximal potentiation was already achieved after 5 seconds of preincubation with IGF-1 (Figure 3E, compare 5-second and 5-minute preincubation). Furthermore, addition of IGF-1 5 seconds after stimulation with SFLLRN was still able to significantly potentiate platelet aggregation (Figure 3E).

One of the mechanisms by which IGF-1 may increase platelet responses is by positive effect on granule secretion, thereby further amplifying the platelet response by the release of ADP. IGF-1 indeed significantly increased dense granule secretion in response to SFLLRN, as measured by the release of ATP (Figure 4A). The potentiatory effect of IGF-1 on SFLLRN-mediated platelet aggregation (Figure 4B left panel), however, was retained in the presence of a mixture of ADP receptor antagonists (Figure 4B middle and right panels), demonstrating that the effect of IGF-1 is largely independent of secreted ADP.

To evaluate the contribution of the PI3K pathway to the potentiatory effect of IGF-1 on platelet function, platelets were pretreated with either vehicle (DMSO; Figure 5A) or a range of PI3K inhibitors (Figure 5B-I). IGF-1 significantly potentiated platelet aggregation in the presence of DMSO at both 0.5 μM SFLLRN (Figure 5Ai) and 0.7 μM SFLLRN (Figure 5Aii). The potentiatory effect of IGF-1 on SFLLRN-induced aggregation was completely absent in the presence of the generic PI3K inhibitor wortmannin (Figure 5B) and the novel PI3K inhibitor ZSTK474²⁸  (Figure 5D), whereas it was significantly reduced in the presence of LY294002 (Figure 5C). As expected, wortmannin, LY294002, and ZSTK474 inhibited IGF-1–stimulated PKB phosphorylation in a dose-dependent manner (Figure 5B-Diii). To investigate which PI3K isoforms are involved in IGF-1–mediated potentiation of platelet aggregation and PKB phosphorylation, platelets were subsequently incubated with the p110α-selective inhibitor PI-103,²⁹  the p110α-specific inhibitor PIK-75,^29,30  the p110β inhibitors GTX-155 and GTX-221,³⁰  and the p110δ inhibitor IC87114.³⁰  Interestingly, IGF-1–mediated potentiation was blocked in the presence of optimal concentrations of the p110α inhibitors PI-103 (Figure 5E) and PIK-75 (Figure 5F), whereas it was still largely intact in the presence of the p110β inhibitors GTX-155 (Figure 5G) and GTX-221 (Figure 5H) and the p110δ inhibitor IC87114 (Figure 5I). Both PI-103 and PIK-75 dose-dependently inhibited IGF-1–stimulated PKB phosphorylation, with complete inhibition reached at a concentration of 0.5 μM and 1 μM, respectively (Figure 5Eiii-Fiii). In contrast, the p110β inhibitor GTX-155 and GTX-221 only partially inhibited PKB phosphorylation at a concentration of 0.5 μM and 0.2 μM, respectively, with no further inhibition detected at higher concentrations (Figure 5Giii,Hiii). Similarly, the p110 δ inhibitor had a weak effect on IGF-1–mediated PKB phosphorylation (Figure 5Iiii). These results together indicate an essential role for the p110α isoform of PI3K (PI3Kα) in IGF-1–mediated potentiation of platelet aggregation and PKB phosphorylation.

As IGF-1 is present in the α granules of human platelets and is released upon platelet activation, it may play a supporting role in the autocrine regulation of platelet function. In order to investigate whether this is the case, platelets were incubated with various concentrations of the IGF receptor inhibitor NVP-AEW541, a recently developed small-molecule inhibitor from Novartis.²⁸  This inhibitor potently reduced IGF-1–stimulated IGF receptor phosphorylation and PKB phosphorylation, with complete inhibition achieved at a concentration of 0.5 μM (Figure 6Ai). NVP-AEW541 had no effect on SFLLRN-stimulated total tyrosine phosphorylation (results not shown), indicating that tyrosine kinases such as the Src family of kinases are not affected. Interestingly, preincubating platelets with NVP-AEW541 significantly reduced platelet aggregation in response to the PAR-1 agonist SFLLRN (Figure 6Aii), suggesting a role for IGF-1 in autocrine regulation of platelet function. Furthermore, the αIR3 antibody dose-dependently inhibited IGF-1–stimulated IGF receptor phosphorylation and downstream signaling to PKB, with complete inhibition at 5 μg/mL (Figure 6A). Incubation of platelets with αIR3 significantly reduced PAR-1–mediated platelet aggregation (Figure 6Bii), an effect very similar to that of the IGF receptor inhibitor NVP-AEW541 (Figure 6Aii). Together, these results demonstrate that blocking IGF-1 signaling in platelets significantly inhibits platelet aggregation in response to the PAR-1 agonist SFLLRN.

Altered levels of IGF-1 have been implicated in the increased risk of cardiovascular disease as is seen in conditions of insulin resistance. As platelets play an important role in the pathogenesis of cardiovascular disease,²³  an effect of IGF-1 on platelet function may be involved in this process. Platelets indeed express high levels of the IGF receptor on their plasma membrane.²⁴  Changes in the levels of IGF-1–binding proteins and increased local concentrations of IGF-1, further contributed to by secretion of IGF-1 from the α granules,^4,5  may therefore positively regulate platelet function and contribute to cardiovascular disease.

Although a single report had previously shown IGF-1 to increase platelet aggregation in response to collagen, thrombin, and ADP,³¹  the underlying functional and signaling mechanisms that are involved in this process were not studied. In the present study, I therefore investigated the effect of IGF-1 on platelet function and the signaling pathways that are involved in this process. I demonstrate that IGF-1 stimulates (1) rapid and dose-dependent tyrosine phosphorylation of the β subunit of the IGF receptor, (2) transient tyrosine phosphorylation of IRS-1 and IRS-2, (3) association of IRS-1 and IRS-2 with the SH2-containing p85 subunit of PI3K, and (4) transient phosphorylation of PKB on Ser473. Furthermore, I show that this signaling pathway, and more specifically the p110α isoform of PI3K (PI3Kα), is essential for the potentiatory effect of IGF-1 on platelet aggregation, as generic PI3K inhibitors and the PI3Kα inhibitors PI-103 and PIK-75 prevent IGF-1–stimulated PKB phosphorylation and potentiation of SFLLRN-mediated aggregation. More importantly, blocking antibodies and the IGF receptor inhibitor NVP-AEW541 significantly inhibit platelet aggregation in response to SFLLRN, implicating an important role for IGF-1 in autocrine regulation of platelet function.

The IGF receptor contains three tyrosine residues in the activation loop—Tyr1131, Tyr1135, and Tyr1136—which become phosphorylated upon IGF-1 stimulation. IGF-1 induced a dose-dependent phosphorylation of the IGF receptor that was rapid (< 1 minute) and sustained (1-60 minutes) (Figure 1A,B). In contrast, tyrosine phosphorylation of both IRS-1 and IRS-2 was transient with decreased phosphorylation at 15 to 60 minutes after stimulation (Figure 1C,D). Interestingly, a recent paper also reported transient tyrosine phosphorylation of IRS-1 in human platelets in response to insulin, whereas insulin receptor phosphorylation was sustained.³²  As insulin is known to negatively regulate IRS-1 tyrosine phosphorylation by serine phosphorylation on a variety of residues by kinases such as PKC, ERK, JNK, mTOR, and S6K, a similar negative feedback mechanism may be in place in IGF-1 stimulation.^33,34  Phosphorylation of IRS-1 and IRS-2 on YXXM motifs creates binding sites for the regulatory p85 subunit of PI3K. IGF-1 stimulation of platelets indeed results in rapid but transient association of both IRS-1 and IRS-2 with the p85 subunit of PI3K. The association of p85 with IRS-1, however was more prolonged than the interaction between p85 and IRS-2. Interestingly, the major tyrosine-phosphorylated protein present in p85 immunoprecipitates is IRS-2, suggesting that IGF-1 signals to p85 mainly via phosphorylation of IRS-2. This is in agreement with a recent study in myocytes that elegantly showed that the IGF receptor mainly signals via IRS-2, whereas the insulin receptor signals via IRS-1.³⁵  However, we can not at present rule out the possibility that IRS-2 becomes more heavily tyrosine phosphorylated by IGF-1 than IRS-1. Binding of the p85 subunit of PI3K to IRS-1 and IRS-2 leads to the activation of the catalytic PI3K subunit p110, resulting in downstream activation of PKB. IGF-1 stimulation of platelets indeed resulted in a rapid (<2 minutes) dose-dependent increase in PKB phosphorylation on Ser473. Stimulation of platelets with a maximal concentration of IGF-1 resulted in PKB phosphorylation comparable with that found in the presence of 5 μM PAR-1 agonist SFLLRN. In contrast to PKB phosphorylation by SFLLRN or thrombin,^26,27  however, IGF-1–stimulated PKB phosphorylation was independent of PKC and ADP secretion, demonstrating that IGF-1 directly and potently signals to PKB in human platelets. Interestingly, PKB phosphorylation by IGF-1 in human platelets was of a transient nature and closely followed p85 binding to IRS-2 (Figure 1D,E). Together, these results demonstrate that IGF-1 stimulates the association of p85 with IRS-2, and to a lesser extent with IRS-1, leading to downstream activation of PI3K and phosphorylation of PKB.

PI3K plays an important role in platelet activation.³⁶  Here, I show that PI3K is also involved in the potentiatory effect of IGF-1 on platelet aggregation. The generic PI3K inhibitors wortmannin, LY294002, and the recently identified inhibitor ZSTK474²⁸  all blocked PKB phosphorylation and strongly inhibited potentiation of aggregation by IGF-1. The catalytic isoforms that interact with the p85α regulatory subunit of PI3K are p110α, p110β, and p110δ. All these isoforms are expressed in platelets, although the expression level of p110δ is extremely low.³⁷  Using isoform-specific inhibitors, I demonstrate that p110α is the isoform involved in the effect of IGF-1 on aggregation and PKB phosphorylation. PI-103, a p110α-selective inhibitor,²⁹  significantly reduced the effect of IGF-1 on aggregation and blocked IGF-1–stimulated PKB phosphorylation. However, this finding does not rule out a role for p110β, as the specificity of PI-103 for p110α over p110β is mimimal (5-fold in vitro).³⁰  Platelets were therefore also treated with the more specific p110α inhibitor PIK-75^29,30  (44-fold selectivity for p110α over p110β),³⁰  which completely blocked IGF-1–mediated potentiation of platelet aggregation and PKB phosphorylation, thereby confirming an essential role for p110α in the effect of IGF-1. In contrast, the effect of IGF-1 on platelet aggregation was still largely intact in the presence of the p110β inhibitors GTX-155³⁸  and GTX-221,³⁰  and the p110δ inhibitor IC87114,³⁰  although a small reduction in PKB phosphorylation was apparent. Together, these results demonstrate an essential role for p110α in IGF-1–mediated potentiation of aggregation and PKB phosphorylation.

IGF-1 did not stimulate platelet aggregation or secretion by itself but potentiated platelet responses to a variety of agonists. These results show that activation of the PI3K pathway by itself is not sufficient to stimulate platelet activation, but instead promotes activation by other stimuli. A similar mechanism has previously been described for TPO, a cytokine that signals via JAK2, and increases platelet aggregation via a PI3K-dependent pathway.³⁹  Furthermore, recent studies suggest that platelet modulators such as Gas6 and Ephrin B also signal via PI3K.³⁶  These results indicate that activation of PI3K is involved in increasing the platelet response to a variety of agonists.

Although PKB is strongly activated by IGF-1 in human platelets, it is presently unclear whether it is involved in the potentiatory effect of IGF-1 on platelet aggregation. The effect of IGF-1 on platelet function is very rapid (within seconds), whereas IGF-1–stimulated phosphorylation of PKB only becomes detectable after 30 to 60 seconds of stimulation. Although we cannot rule out that very low levels of PKB activation below the detection limit are sufficient for the maximal potentiatory effect of IGF-1, these results strongly indicate that PKB is not involved in IGF-1–mediated potentiation of aggregation.

As IGF-1 clearly can modulate platelet function, one of the important questions is whether this modulation plays a physiologic important role, contributing to normal hemostasis and pathologic conditions such as cardiovascular disease. Although the IGF-1 concentration in circulation is relatively high, the majority of IGF-1 is bound to IGF-binding proteins that protect them from degradation and are involved in transport and modulation of IGF-1 function.⁴⁰  To make matters more complicated, in addition to inhibiting the effect of IGF-1 on cells, IGF-binding proteins can also potentiate the effect of IGF-1 depending on cell type and experimental conditions.⁴⁰  Furthermore, the physiologic effect of IGF-1 on cells can be regulated by increased local activity of proteases that break down IGF-binding proteins, leading to a local increase in IGF-1 concentration. The concentration of free IGF-1 that platelets are exposed to during thrombus formation is presently unknown. Platelets secrete IGF-1 from the α granules, thereby increasing the local concentration of IGF-1. The free IGF-1 concentration available to modulate platelet function, however, may be further increased by thrombin formation, which cleaves IGFBP-3 and IGFBP-5.^41,42  Furthermore, the secretion and/or surface expression of metalloproteases (MMPs) and members of the metalloprotease-disintegrin family (ADAM) on platelets may contribute to the free IGF-1 concentration, as they both have been reported to degrade IGF-binding proteins.⁴³  To demonstrate a role for IGF-1 in platelet function, I used two different approaches to inhibit IGF-1 signaling in platelets: inhibition of the kinase activity with the small-molecule inhibitor NVP-AEW541 from Novartis,⁴⁴  and blocking the IGF receptor with the neutralization antibody αIR3.⁴⁵  The IGF receptor inhibitor NVP-AEW541 is a selective inhibitor capable to distinguish between the IGF receptor and the closely related insulin receptor (IC₅₀ of 0.086 μM and 2.3 μM, respectively).⁴⁴  Platelet aggregation in response to the PAR-1 agonist SFLLRN was significantly reduced in the presence of NVP-AEW541, especially at lower concentrations of SFLLRN. Similar results were obtained with the αIR3 antibody, a neutralizing IGF receptor antibody which has been extensively been used in the last 20 years to block IGF receptor signaling.⁴⁵  These results clearly show that IGF-1 is involved in autocrine regulation of platelet function, especially at lower doses of platelet agonist.

In conclusion, these results demonstrate that IGF-1 potentiates platelet aggregation via the IGF receptor/IRS/PI3Kα pathway. This study further emphasizes the important role of IGF-1 in autocrine platelet regulation, and implicates that changes in the insulin/IGF axis resulting in increased local concentrations of IGF-1 may contribute to cardiovascular disease.

The author would like to thank Prof Jeremy Tavaré, Prof Jeff Holly, and Roger Hunter, University of Bristol, for useful discussions; Prof Bart van Haesebroeck, Dr Peter Shepherd, and Prof Kevin Shokat for the generous gifts of PI3K isoform-specific inhibitors; and Prof Jonathan Gibbins for the generous gift of the IV.3 antibody.

This work was supported by the British Heart Foundation (Research Fellowship no. FS/04/027 awarded to I.H.).

Contribution: I.H. designed and performed research, collected and analyzed data, and wrote the paper.

Conflict-of-interest disclosure: The author declares no competing financial interests.

Correspondence: Ingeborg Hers, Department of Biochemistry, School of Medical Sciences, University of Bristol, University Walk, Bristol BS8 1TD, United Kingdom; e-mail:i.hers@bris.ac.uk.