PACSIN2 regulates platelet integrin β1 hemostatic function

Abstract

Background: Upon vessel injury, platelets adhere to exposed matrix constituents via specific membrane receptors, including the von Willebrand factor receptor glycoprotein (GP)Ib-IX-V complex and integrins β1 and β3. In platelets, the Fes/CIP4-homology Bin-Amphiphysin-Rvs protein PACSIN2 associates with the cytoskeletal and scaffolding protein filamin A (FlnA), linking GPIbα and integrins to the cytoskeleton.

Objectives: Here we investigated the role of PACSIN2 in platelet function.

Methods: Platelet parameters were evaluated in mice lacking PACSIN2 and platelet integrin β1.

Results: Pacsin2^-/- mice displayed mild thrombocytopenia, prolonged bleeding time, and delayed thrombus formation in a ferric chloride-mediated carotid artery injury model, which was normalized by injection of control platelets. Pacsin2^-/- platelets formed unstable thrombi that embolized abruptly in a laser-induced cremaster muscle injury model. Pacsin2^-/- platelets had hyperactive integrin β1, as evidenced by increased spreading onto surfaces coated with the collagen receptor α2β1-specific peptide GFOGER and increased binding of the antibody 9EG7 directed against active integrin β1. By contrast, Pacsin2^-/- platelets had normal integrin αIIbβ3 function and expressed P-selectin normally following stimulation through the collagen receptor GPVI or with thrombin. Deletion of platelet integrin β1 in Pacsin2^-/- mice normalized platelet count, hemostasis, and thrombus formation. A PACSIN2 peptide mimicking the FlnA-binding site mediated the pull-down of a FlnA rod 2 construct by integrin β7, a model for integrin β-subunits.

Conclusions: Pacsin2^-/- mice displayed severe thrombus formation defects due to hyperactive platelet integrin β1. The data suggest that PACSIN2 binding to FlnA negatively regulates platelet integrin β1 hemostatic function.

Keywords: PACSIN2; bleeding; filamin A; integrin β1; platelets.

Figure 1.. Characterization of Pacsin2^−/− platelets and megakaryocytes.

(A) Immunofluorescence analysis of fibrinogen (green) and β-tubulin (red) in Pacsin2^+/+ and Pacsin2^−/− platelets on BSA. Scale bars, 2 μm. (B) Surface area of resting platelets. Results represent mean ± SD of 239 Pacsin2^+/+ and 122 Pacsin2^−/− and were compared by using the unpaired Student t test (P = .0223). (C) Biotin NHS was injected into Pacsin2^+/+ and Pacsin2^−/− mice and blood samples were collected at the indicated time points. (D) CMFDA-labeled Pacsin2^+/+ and Pacsin2^−/− platelets were injected into Pacsin2^+/+ mice and blood samples were collected at the indicated time points. The percentage of biotin- or CMFDA-positive platelets in circulation was determined by flow cytometry, compared to <2 min post injection (100%). Results are mean ± SD of 3-5 independent experiments. (E) Pacsin2^+/+ and Pacsin2^−/− platelet lysates corresponding to 2 mg of protein were subjected to SDS-PAGE and probed for the indicated proteins. Results are representative of 3 independent experiments. (F) Top. Seven-mm-thick femur bone marrow sections from Pacsin2^+/+ and Pacsin2^−/− mice were immunostained for resident MKs and vasculature using anti-GPIbα (green) and anti-laminin (red) antibodies, respectively. Scale bars, 30 μm. Sections shown are representative of 7 mice per genotype. Bottom. Transmission electron microscopy analysis of freshly isolated bone marrow Pacsin2^+/+ and Pacsin2^−/− MKs. Scale bars, 2 μm. (G) Bone marrow MK counts (n = 7 Pacsin2^+/+ and 8 Pacsin2^−/−; P = .3287).

Figure 2.. Increased bleeding time and thrombus formation defects in Pacsin2^−/− mice.

(A) Tail-bleeding time of Pacsin2^+/+ and Pacsin2^−/− mice (n = 12 in each group). Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (log-rank P = .0040). (B) Time to occlusion after FeCl₃-induced injury. Pacsin2^+/+ (n = 15) and Pacsin2^−/− (n = 16) carotid arteries were exposed to 10% FeCl₃ for 3 min, and arterial flow rates were measured. Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (log-rank P = .0071). (C) Injection of 80 x 10⁶ control Pacsin2^+/+ platelets into Pacsin2^−/− mice normalized their median time to occlusive thrombosis (n = 6 in each group). Pacsin2^+/+ and Pacsin2^−/− mice were injected with anti-GPIbβ (red) and anti-fibrin (green) antibodies, and the cremaster arteries were interrogated with a 3i Ablate! laser during fluorescence, real-time, and intravital video microscopy. (D) Representative still images. Arrows represent blood flow direction. Time in s. Scale bars, 20 μm. Platelet (E) and fibrin (F) accumulation at the site of injury was measured by fluorescence intensity. PPACK-anticoagulated whole blood from Pacsin2^+/+ and Pacsin2^−/− mice was labeled and perfused on a type I collagen-immobilized surface at an arterial shear rate of 1500 s⁻¹. (G) Representative still images at 3 min. Bars represent 100 mm. (H) Fluorescence intensity at 3 minutes. Results represent mean ± SD and were compared by using the unpaired Student t test (n = 9 Pacsin2^+/+ and 11 Pacsin2^−/−; P = .0127). (I) Dwell time of individual Pacsin2^+/+ and Pacsin2^−/− platelets. Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (n = 60 in each group; log-rank P = .0463).

Figure 3.. Hemostatic functions of Pacsin2^−/− platelets.

Pacsin2^+/+ and Pacsin2^−/− platelets were activated for 2 min at 37°C with CRP (A,B) or thrombin (C,D), incubated with FITC-labeled anti-mouse CD62P antibody (A,C) or Oregon Green 488-labeled fibrinogen (B,D), and analyzed by flow cytometry. Results are expressed as a percentage of positive platelets and represent mean ± SD of 3-6 independent experiments. Aggregation of Pacsin2^+/+ and Pacsin2^−/− platelets was determined by light transmission under stirring conditions at 37°C in response to 5 μg/mL (E) or 10 μg/mL (F) of collagen. Graphs are representative of 4 independent experiments. Pacsin2^+/+ and Pacsin2^−/− blood samples were analyzed using the ROTEM assay. Clotting time (G) and clot formation time (H) of Pacsin2^+/+ and Pacsin2^−/− blood samples. Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (n = 6 in each group; log-rank P = .0046 and .5265, respectively). Maximum clot firmness (MCF) (I), α-angle (J), and A10 (K). Results represent mean ± SD of 6 independent experiments. (L) Washed Pacsin2^+/+ and Pacsin2^−/− platelets were added to human platelet-poor plasma in ACD containing 5 mM CaCl₂ and 0.2 U/mL thrombin in siliconized cuvettes (n = 9 in each group).

Figure 4.. Hyperactive integrin β1 in Pacsin2^−/− platelets.

(A) Pacsin2^+/+ and Pacsin2^−/− platelets were activated with 1 μg/mL CRP and left to adhere to 100 μg/mL fibrinogen or 20 μg/mL GFOGER for 30 min at 37°C. Fixed platelets were stained for phosphotyrosine (4G10, green) and F-actin (phalloidin, red) and analyzed by immunofluorescence. Immunofluorescence micrographs representative of four independent experiments. Scale bars, 10 μm. (B) Surface area of 388 Pacsin2^+/+ and 635 Pacsin2^−/− platelets on fibrinogen (P = .8586). (C) Surface area of 56 Pacsin2^+/+ and 40 Pacsin2^−/− platelets on GFOGER (P < .0001). Results represent the mean ± SD and were compared by using the unpaired Student t-test. Pacsin2^+/+ and Pacsin2^−/− platelets were incubated with rat anti-mouse active integrin β1 antibody 9EG7 (n = 5 in each group; P = .0057) (D) or hamster anti-mouse total integrin β1 antibody (n = 7 Pacsin2^+/+ and 8 Pacsin2^−/−; P = .7609) (E) and analyzed by flow cytometry. (F) An integrin β7 peptide containing the FlnA repeat 21 binding site (residues 752-770) was conjugated to SulfoLink agarose resin and incubated with His-tagged FlnA repeats 16-23 (left panel) or 21 (right panel) in the presence (+) or absence (−) of a PACSIN2 peptide (P2) mimicking the FlnA repeat 20 binding site (residues 171-189) as indicated. Complexes were subjected to SDS-PAGE and probed for His (FlnA). (G) Top. Hypothetic model on the role of PACSIN2 in β-integrin regulation by FlnA. In full-length FlnA, a β-strand in incompletely folded repeat 20 masks the β-integrin binding site on repeat 21. PACSIN2 binding to repeat 20 initiates major conformational changes in FlnA, resulting in the unmasking of repeat 21 and binding of β-integrin. Bottom. Alignment of GPIbα and β-integrin FlnA-binding motifs (blue highlight), flanked by endocytic NPxY/F motifs in integrins (yellow highlight). Integrin β1 has an optimal FlnA-binding motif, while integrin β3 contains an acidic glutamate at position 749, instead of a neutral serine, and lacks a C-terminal flanking proline at position 757 (orange).

Figure 5.. Platelet-specific integrin β1 deletion normalizes the thrombus formation defects of Pacsin2^−/− mice.

(A) Tail-bleeding time of Itgb1^Plt−/− Pacsin2^+/+ and Itgb1^Plt−/− Pacsin2^−/− mice (n = 12 in each group). Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (log-rank P = .6749). (B) Time to occlusion after FeCl₃-induced injury. Itgb1^Plt−/− Pacsin2^+/+ and Itgb1^Plt−/− Pacsin2^−/− carotid arteries were exposed to 10% FeCl₃ for 3 min, and arterial flow rates were measured (n = 12 in each group). Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (n = 16 in each group; log-rank P = .9368). Itgb1^Plt−/− Pacsin2^+/+ and Itgb1^Plt−/− Pacsin2^−/− mice were injected with anti-GPIbβ (red) and anti-fibrin (green) antibodies, and the cremaster arteries were interrogated with a 3i Ablate! laser during fluorescence, real-time, and intravital video microscopy. Platelet (C) and fibrin (D) accumulation at the site of injury was measured by fluorescence intensity. Clotting time (E) and clot formation time (F) of Itgb1^Plt−/− Pacsin2^+/+ and Itgb1^Plt−/− Pacsin2^−/− blood samples using the native ROTEM assay. Results were estimated by the Kaplan-Meier method and were compared by using the log-rank test (n = 3 in each group; log-rank P = .6939 and .6068, respectively). Maximum clot firmness (MCF) (G), α-angle (H), and A10 (I). Results represent mean ± SD of 3 independent experiments. (J) Washed Itgb1^Plt−/− Pacsin2^+/+ and Itgb1^Plt−/− Pacsin2^−/− platelets were added to human platelet-poor plasma in ACD containing 5 mM CaCl₂ and 0.2 U/mL thrombin in siliconized cuvettes (n = 5 in each group).