Joo-ri Kim-Kaneyama, Department of Biochemistry, Showa University School of Medicine, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8555, Japan. Tel.: +81 3 3784 8116; fax: +81 3 3784 2346. E-mail: firstname.lastname@example.org
Summary. Background: Integrin αIIbβ3 plays key roles in platelet aggregation and subsequent thrombus formation. Hydrogen peroxide-inducible clone-5 (Hic-5), a member of the paxillin family, serves as a focal adhesion adaptor protein associated with αIIbβ3 at its cytoplasmic strand.
Objectives: Hic-5 function in αIIbβ3 activation and subsequent platelet aggregation remains unknown. To address this question, platelets from Hic-5−/− mice were analyzed.
Methods and Results: Hic-5−/− mice displayed a significant hemostatic defect and resistance to thromboembolism, which were explained in part by weaker thrombin-induced aggregation in Hic-5−/− platelets. Mechanistically, Hic-5−/− platelets showed limited activation of αIIbβ3 upon thrombin treatment. Morphological alteration in Hic-5−/− platelets after thrombin stimulation on fibrinogen plates was also limited. As a direct consequence, the quantity of actin co-immunoprecipitating with the activated αIIbβ3 was smaller in Hic-5−/− platelets than in wild-type platelets.
Conclusion: We identified Hic-5 as a novel and specific regulatory factor for thrombin-induced αIIbβ3 activation and subsequent platelet aggregation in mice.
Circulating platelets are activated by vascular endothelial dysfunction, which induces morphological changes and discharge of intracellular granules, leading to platelet aggregation. Platelet aggregation, a key step in thrombus formation, is caused by the activation of integrin αIIbβ3, which was clearly demonstrated in integrin β3-deficient mice . Integrin αIIbβ3 displays low affinity for its ligands (e.g. fibrinogen and von Willebrand factor), under unstimulated conditions. However, following three-dimensional conformational change due to inside-out signaling, integrin αIIbβ3 shifts to an activated form characterized by higher affinity for its ligands [2–4]. Integrin itself possesses no kinase domain, no enzymatic activity and no actin-binding activity . Rather, integrin activation is controlled by a series of molecules forming a complex with integrin at its cytoplasmic short strand through inside-out signaling, which results in the formation of a signaling platform . More than 30 types of proteins (e.g. talin, kindlin and RIAM) are known to be integrin-associated proteins; [6–11] moreover, these proteins appear to play key roles in inside-out signaling. However, the precise mechanisms governing αIIbβ3 activation and the signaling pathway involving integrin and platelet membrane receptors remain unknown.
Hic-5 (hydrogen peroxide-inducible clone-5), the molecule in question in the current study, was isolated as a gene induced by TGF-β or hydrogen peroxide . The Hic-5 gene codes for a protein localized in focal adhesion, which serves as a cellular attachment point to extracellular matrix . Talin is also localized in focal adhesion and binds to Hic-5 in platelets . It is notable that Hic-5 is an adaptor molecule sharing high homology with paxillin, a member of the four LIM domain family . The paxillin family includes three members, paxillin, Hic-5 and leupaxin. These three molecules are expressed in murine platelets; however, only Hic-5 is expressed in human platelets .
Recently, we succeeded in terms of the generation of Hic-5−/− mice; subsequently, our data revealed that the recovery of arterial media following vascular injury is delayed significantly in Hic-5−/− mice . Moreover, our preliminary observations regarding vascular lesions in Hic-5−/− mice suggested that morphological changes are suppressed in those platelets attached to a damaged vascular wall immediately after vascular injury. Therefore, we hypothesized that Hic-5 might play an important role in controlling inside-out and/or outside-in signal(s) in platelet aggregation. Consequently, we examined the role of Hic-5 in platelet aggregation as well as its molecular mechanism in thrombosis and hemostasis.
Materials and methods
Determination of bleeding times
Adult C57BL/6N mice were anesthetized via intraperitoneal injection of pentobarbital (50 mg kg−1 body weight) and intramuscular injection of xylazin (3 mg kg−1). The tail was cut 1 mm from the tip and immersed in saline at 37 °C. Bleeding time was defined as the time at which all visible signs of bleeding from the incision had abated. The experiment was terminated 10 min after the tail was cut. All animal studies were conducted in accordance with the protocols approved by the institutional committee for animal research of Showa University.
To induce thromboembolism, a mixture of collagen (0.5 mg kg−1, equine collagen; NYCOMED, Munich, Germany) and epinephrine (60 μg kg−1, Sigma) was injected into tail veins of wild-type (n = 5) and Hic-5−/− mice (n = 5). Blood and lungs were collected 15 min after the injection and platelet counts were determined. Three mice from each group were histologically analyzed.
Platelet preparation and aggregation
Whole blood was collected from the heart (100 μL mL−1 blood) of anesthetized mice in acidic citrate dextrose, after which it was centrifuged at 220 × g for 10 min. To the supernatant transferred to a new tube, 1 μm prostaglandin E1 and 1 U mL−1 apyrase were added to prevent platelet activation. The mixture was centrifuged at 400 × g for 10 min to sediment a platelet pellet. The pellet was then resuspended in an appropriate volume of modified Tyrode-HEPES buffer at pH 7.4 (10 mm HEPES, 12 mm NaHCO3, 138 mm NaCl, 5.5 mm glucose, 2.9 mm KCl and 1 mm MgCl2) and used for the following assay after addition of 1 mm CaCl2. Platelet aggregation was measured with a platelet aggregometer (PA-200, KOWA, Nagoya, Japan) following stimulation with thrombin (Sigma-Aldrich, St.Louis, MO, USA). The degree of platelet aggregation was expressed as %light transmission and the maximal aggregation (Tmax) was expressed as % light transmission determined 10 min after addition of the stimulus.
To examine αIIbβ3 activation, whole blood was incubated for 10 min at room temperature with PE-conjugated JON/A (emfret, Eibelstadt, Germany) in the presence of ADP (0.1 or 1 μm), Convulxin (10 or 100 ng mL−1) or human thrombin (0.01 or 0.5 U mL−1). JON/A binding to platelets was quantified with a flow cytometer (BD Biosciences, San Jose, CA, USA). Platelets were gated by forward and side scatters. P-selectin surface expression was analyzed via detecting FITC-conjugated CD62 (emfret, Eibelstadt, Germany) binding to platelets using the flow cytometer.
Platelet lysates, which were treated with RIPA buffer (20 mm Tris-HCl, pH 8.0, 0.5% Triton-X 100 and 5 mm EDTA) supplemented with proteinase inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA), were separated by electrophoresis on 10% SDS-polyacrylamide gel and transferred to a polyvinylidine difluoride membrane (Millipore, Bedford, MA, USA). Immunodetection was performed utilizing the primary antibodies against Hic-5 (BD Biosciences, San Jose, CA, USA), paxillin, vinculin, GAPDH, β-actin (these antibodies were purchased from Sigma-Aldrich), integrin β3 (Santa Cruz Biotechnology), kindlin-3 (abcam, Cambridge, UK), ILK (BD Biosciences), followed by the appropriate secondary antibodies conjugated with horseradish peroxidase.
Electron microscopic observation
Immunoelectron microscopy was performed as previously described . Briefly, the sections were stained first with the primary antibodies (anti-Hic-5, BD Biosciences; anti-αIIbβ3 (Leo), emfret; anti-β3, Santa Cruz, Santa Cruz, CA, USA), followed by incubation with their appropriate secondary antibodies (BBI International, Cardiff, UK) conjugated with smaller sized (5 nm) and larger sized (10 or 15 nm) colloidal golds, respectively; subsequently, sections were evaluated with a JEM-1200 EXII electron microscope (JEOL, Tokyo, Japan) at an accelerating voltage of 80 kV. For scanning electron microscopy analysis, cover slips were coated overnight with 1 mg mL−1 human fibrinogen and then blocked for 1 h with 1% BSA in PBS. Platelets were activated with 0.01 U mL−1 thrombin shortly before plating on fibrinogen-, fibronectin- or collagen-coated cover slips. Platelets were allowed to spread for 30 min, after which they were fixed in 2.5% glutaraldehyde in Tyrode’s-HEPES buffer and processed for scanning electron microscopy. Numbers of platelets displaying filopodia longer than 3.5 μm as well as total platelets were counted in 10 separate fields and totaled for each sample (Fig. 1B). The ratios of platelets exhibiting longer filipodia to total platelets were calculated and compared between wild-type and Hic-5−/− platelets. In another set of experiments, platelets were seeded on fibrinogen-coated cover slips in the presence of MnCl2 (3 mm of Mn2+) for analysis of spreading filopodia as above.
Platelets were treated with thrombin for 10 min at room temperature; platelets were then lysed upon the addition of 2 × IP buffer (1% Triton X-100, 150 mm NaCl, 20 mm Hepes, pH 7.4, 1 mm Na3V04, 1 mm NAF) containing a mixture of proteinase inhibitors (Sigma-Aldrich). Dynabeads conjugated with an antibody against αIIbβ3 (Leo) or control IgG were added to lysates (5 μg per sample) and rotated at room temperature for 60 min. Samples were washed with 1 × IP buffer three times and soaked in elution buffer for 10 min. Immunoprecipitated proteins were subjected to 10% SDS-PAGE for immunoblotting or MALDI-TOF. The MALDI-TOF analysis was performed as previously described . Briefly, the excised polyacrylamide gel pieces were incubated with wash solution (50% acetonitrile plus 50 mm ammonium bicarbonate) for 10 min. Disulfide bond cleavage was effected with 10 mm dithiothreitol and 100 mm carbamidomethylation in the presence of 55 mm iodine acetamide. The dried gel pieces were soaked in 5 μL of trypsin solution (12.5 ng μL−1 trypsin and 10 mm ammonium bicarbonate) and incubated overnight. On the following day, sonicated samples were spotted on a stainless steel MALDI carrier tray with matrix. MALDI-time-of-flight (TOF)-MS was conducted on an AXIMA Performance (Shimadzu, Kyoto, Japan). The tryptically-digested proteins were identified using the MASCOT database search engine with National Center for Biotechnology Information and the Swissprot database with a mass tolerance below 0.1 Da of the monoisotopic peaks.
All data are expressed as means ± SD. The two-unpaired t-test (Figs 2B, 3B-F and 1B), the Mann–Whitney U-test (Fig. 2A) and the chi-square test (Fig. 4A) were used to detect differences. A value of P < 0.05 was considered to be significant.
To address the function of Hic-5 in platelets, the hemostatic abilities of wild-type and Hic-5−/− mice were tested via tail bleeding in 8–10-week-old offspring of Hic-5+/− and Hic-5+/− mating under conditions with genotype blind until study completion. Hic-5−/− mice exhibited a pronounced hemostatic defect in comparison to the wild-type background (KO, 7.60 ± 3.40 min vs. wt, 2.72 ± 0.75 min) (Fig. 1A). To determine whether the hemostatic disorder was due to defective thrombopoiesis, platelet counts were compared between wild-type and Hic-5−/− mice. However, Hic-5−/− mice demonstrated platelet counts similar to those of wild-type mice (Fig. 2B). Additionally, expression levels of the proteins, namely, integrin β3, kindlin-3, vinculin, paxillin, integrin-linked kinase and β-actin, were unaltered (Fig. 2C).
To further evaluate the in vivo consequence of Hic-5 deficiency, we examined a model of thromboembolism challenged by collagen and epinephrine. The numbers of platelets in challenged Hic-5−/− mice were significantly larger than those in Hic-5+/+ mice (Fig. 4A). Furthermore, histological analysis of three Hic-5+/+ mice revealed multiple pulmonary thromboembolism in every mouse (Fig. 4B, upper panel). In a sharp contrast, three Hic-5−/− mice did not show any pulmonary thromboembolism (Fig. 4B, lower panel). These data indicate that Hic-5−/− mice are more resistant to thromboembolism than Hic-5+/+ mice.
To determine whether the hemostatic defect and resistance to thromboembolism in Hic-5−/− mice are attributable to the impairment of platelet function, platelet aggregation induced by thrombin was assessed. Representative aggregometric tracing of platelets demonstrated weaker aggregation (lower %light transmission) of Hic-5−/− platelets after thrombin stimulation compared with Hic-5+/+ platelets (Fig. 3A). As a result, Tmax following the addition of thrombin was limited to 27% in Hic-5−/− platelets relative to wild-type platelets (36%) (Fig. 3B). This result was suggestive of a possible dysfunction of platelets in Hic-5−/− mice. Interestingly, Tmax values following the addition of ADP or collagen were not statistically different between Hic-5+/+ and Hic-5−/− platelets (data not shown).
Platelet aggregation occurs consequent to the activation of αIIbβ3. To test whether activation of αIIbβ3 is actually limited in Hic-5−/− platelets, we measured agonist-induced binding of JON/A-PE antibody, which selectively binds to activated αIIbβ3 of mouse platelets . Integrin αIIbβ3 in wild-type platelets was activated in response to thrombin (Fig. 3C,D). In addition, the amount of activated αIIbβ3 in Hic-5−/− platelets was much smaller than that in wild-type platelets (Fig. 3C,D). Thus, Hic-5 deficiency led to limited thrombin activation of αIIbβ3. Hic-5−/− platelets also showed limited activation of integrin αIIbβ3 after stimulation with a PAR4 agonist (Fig. S1D). Notably, other agonists such as ADP, the GPVI receptor agonist Convulxin, the TxA2 analog U46619 and MnCl2 induced αIIbβ3 activation in Hic-5+/+ and Hic-5−/− platelets to the same extents (Fig. 2A). Furthermore, there was no significant difference in the surface expression of integrin αIIbβ3 between Hic-5+/+ and Hic-5−/− platelets (Fig. 2C). P-selectin surface translocation by thrombin or Convulxin in Hic-5−/− platelets was similar to that in Hic-5+/+ platelets (Fig. 3E,F and Fig. S1B). This suggests a selective defect in αIIbβ3-dependent aggregation in response to thrombin rather than an impairment of general signaling pathways in Hic-5−/− platelets.
Hic-5 is a focal adhesion scaffolding protein that binds to integrin α4β1 . Activation of αIIbβ3 was limited in Hic-5−/− platelets (Fig. 3); consequently, we hypothesized that Hic-5 might bind to αIIbβ3 other than integrin α4β1. In order to test this possibility, the association between Hic-5 and αIIbβ3 was evaluated via immunoprecipitation. As shown in Fig. 5(A), Hic-5 co-precipitated with αIIbβ3 in the presence or absence of thrombin involving anti-αIIbβ3 antibody. Moreover, the co-localization of αIIbβ3 and Hic-5 was examined utilizing an immunoelectron microscopic method in platelets from mice (Fig. 5B) and humans (Fig. 5C). Hic-5 and αIIbβ3 were detected by their primary antibodies followed by the secondary antibodies conjugated with smaller sized (5 nm) and larger sized (10 and 15 nm) gold colloids, respectively. The co-localization of αIIbβ3 and Hic-5 was demonstrated by the proximity of these different sized colloids. Association of αIIbβ3 and Hic-5 in both mouse and human platelets suggested the possibility that Hic-5 may modulate the function of αIIbβ3 in platelets. It is also noteworthy that there was no difference in the numbers of gold colloids in resting and activated Hic-5−/− platelet surfaces (data not shown).
An αIIbβ3-associated protein in platelets, Hic-5 probably acts as an adaptor molecule and a scaffold for various functionally interacting molecules. In order to analyze the changes in the components of αIIbβ3 complexes due to Hic-5 deficiency, we compared the αIIbβ3 interacting molecules between wild-type and Hic-5−/− platelets. Platelets were immunoprecipitated with beads conjugated to the anti-αIIbβ3 antibody; subsequently, molecules exhibiting different levels of co-precipitation between wild-type and Hic-5−/− platelets were analyzed by MALDI-TOF-MS. As a result, we found that the amount of actin co-precipitating with activated αIIbβ3 was smaller in Hic-5−/− platelets than that in wild-type platelets (Fig. 6, arrow).
Platelets demonstrate dynamic morphological changes such as the extension of lamellipodia and filopodia due to agonist stimulation. These morphological changes are caused by remodeling of actin cytoskeleton. Thus, actin plays key roles in cytoskeletal construction of platelets. Hic-5 depletion from the αIIbβ3 complex led to limited integrin anchorage to the actin cytoskeletal matrix (Fig. 6). Therefore, we hypothesized that platelet morphological changes might be affected in Hic-5−/− platelets. Wild-type and Hic-5−/− platelets were stimulated with thrombin and placed on a fibrinogen-coated glass slide for analysis with a scanning electron microscope. The findings revealed that the numbers of platelets displaying filopodia longer than 3.5 μm were smaller in Hic-5−/− platelets than in wild-type platelets (Fig. 1A,B), which indicated that Hic-5 deficiency affects platelet morphological changes consequent to actin cytoskeleton remodeling. Moreover, this change was canceled when platelets were placed on a fibrinogen-coated glass slide in the presence of Mn2+ (Fig. 1C). Thus, Hic-5 is not essential for αIIbβ3-dependent outside-in signaling. Finally, the extension of filopodia in Hic-5−/− platelets was not affected by extracellular matrices themselves such as collagen and fibronectin (Fig. S2).
Hic-5−/− mice grew normally, as did wild-type mice with no apparent abnormality . These results suggested that Hic-5 is functionally replaceable with paxillin, another member of the paxillin family, which includes Hic-5, under physiological conditions in mice. Although no difference in paxillin protein expression level was evident between Hic-5−/− and Hic-5+/+ platelets (Fig. 2C), this level of paxillin may be sufficient to compensate for Hic-5 deficiency. Moreover, no difference was observed in the number of platelets between Hic-5−/− and Hic-5+/+ mice; thus, it is unlikely that Hic-5 deficiency in megakaryocytes might affect the platelet production process (Fig. 2B).
During electron microscopic evaluation of arterial restenosis after wire injury in mice,14 we noted as a preliminary result that Hic-5−/− platelets, unlike Hic-5+/+ platelets, formed unstable platelet morphology in thrombosis. This finding led us to examine hemostatic function in Hic-5−/− mice. Indeed, we observed elongation of bleeding time (Fig. 2A) and limited platelet aggregation in Hic-5−/− mice (Fig. 4). Platelet aggregation is a key step in thrombus formation. Moreover, Hic-5 is the sole member of the paxillin family expressed in human platelets, which is indicative of the indispensable roles of Hic-5 in human platelets. Therefore, identification of Hic-5 function in platelet aggregation is essential in order to advance the understanding of platelet dysfunction and subsequent hemorrhagic diseases in human.
Platelet aggregation is caused by the eventual activation of αIIbβ3, the focal adhesion receptor crucial for aggregation . The activation of αIIbβ3 is controlled by the two types of intracellular molecular mechanisms, namely, inside-out and outside-in signaling. Dysfunctions of these mechanisms play a role in Glanzmann’s thrombasthenia and other platelet disorders consequent to quantitative or qualitative abnormality of αIIbβ3. Identification of a novel molecule involved in the regulation of αIIbβ3 activation may lead to a novel therapeutic target.
In this study, we proposed that Hic-5 may be a novel αIIbβ3 regulatory factor based on the findings that Hic-5 demonstrated interaction and co-localization with αIIbβ3 (Fig. 5) and that thrombin-induced αIIbβ3 activation was limited in Hic-5−/− platelets (Fig. 3). Hic-5 is a well-known adaptor protein, which functions as a scaffold for various interacting molecules and participates in integrin signal transduction [13,16,23,24]. In order to identify the changes attributable to Hic-5 deficiency in the αIIbβ3-interacting proteins in platelets, we conducted immunoprecipitation followed by MALDI-TOF-MS (Fig. 6). As a result, we determined that the amount of actin co-precipitating with activated αIIbβ3 decreased in Hic-5−/− platelets. Thus, Hic-5 is likely to act as a scaffold that stabilizes the association between αIIbβ3 and actin. When the scaffold is destabilized consequent to Hic-5 deficiency, the binding between αIIbβ3 and actin may weaken. Integrin αIIbβ3 is known to interact indirectly with actin in the presence of vinculin and talin mediation [25–28]. Vinculin and talin binding to Hic-5 was examined via GST pull-down assay. These data indicated that Hic-5 behaves as an intermediator in conjunction with vinculin and talin in the association between αIIbβ3 and actin .
Complex formation of Hic-5 and CrKL was also noted in studies regarding actin cytoskeleton regulation . CrKL, Wiskott-Aldrich syndrome protein (WASP) and syk play key roles in adhesion and migration of leukocytes. During platelet aggregation, the majority of Hic-5, CrKL, syk and WASP localized in cytoplasm migrates to actin cytoskeleton; moreover, Hic-5 is thought to serve as an essential scaffold for CrKL during the migration process . Moreover, when platelets were stimulated and stretched on fibrinogen, Hic-5 was phosphorylated by protein-rich tyrosine kinase 2 (Pyk2); eventually, Hic-5 localization shifted to the terminal region of actin filaments . These findings were suggestive of a cooperative relationship between Hic-5 and actin in terms of morphological change in platelets due to reassembly of actin cytoskeleton.
In addition, the number of Hic-5−/− platelets characterized by longer filopodia decreased in comparison with wild-type platelets (Fig. 1B). Platelet adhesion and aggregation occur as a result of the tangling of stretched filopodia; this process leads to the formation of clumps of deformed platelets in which each platelet is indistinguishable from adjacent cells. Thus, Hic-5 is believed to play an important role in platelet morphological change and aggregation, which are mediated by actin cytoskeleton reassembly. The aforementioned results suggested that Hic-5 regulates αIIbβ3 activation by thrombin and participates in the process of thrombus formation, including platelet aggregation.
In summary, the current study demonstrated essential roles of Hic-5 in αIIbβ3 activation and platelet aggregation in mice. Paxillin, a molecule belonging to the family that includes Hic-5, is not expressed in human platelets; consequently, Hic-5 deficiency is expected to exert more serious effects on human platelet function relative to mouse platelet function. Therefore, it is possible that the Hic-5 gene could be responsible for platelet disorders in humans with unknown molecular basis. Furthermore, because platelets are involved in both thrombosis and hemostasis, elucidation of pathophysiological implications of Hic-5 in human platelets may provide novel mechanistic insights into more common and fatal human diseases such as stroke and acute coronary syndrome, the leading causes of death in western countries.
Contribution: A. Miyauchi, Jr Kim-Kaneyama, N. Takeda, K. Kou, S. Arita, X.-F. Lei, K. Kou, T. Mino and T. Miyazaki performed experiments; Jr Kim-Kaneyama and A. Miyauchi analyzed results and prepared the figures; Jr Kim-Kaneyama, K. Eto, T. Yoshida, S. Shioda and A. Miyazaki designed the research and wrote the paper; all authors approved the final manuscript.
We gratefully acknowledge A. Sasai for her contribution to basic data collection as the theme for her master’s degree. This work was supported by Grants-in-Aid for Young Scientists (B) (20790076 and 23791090 to K-K. J.), Scientific Research (C) (23591341 to A. M.) from the Japan Society for the Promotion of Science and MEXT-Supported Program for the Strategic Research Foundation at Private Universities, 2012-2016.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interests.