Human Disabled-2 regulates thromboxane A2 signaling for efficient hemostasis in thrombocytopenia

Abstract

Understanding platelet protein functions facilitates better assessment of platelet disorders. Megakaryocyte lineage-restricted human Disabled-2 knock-in (hDAB2-KI) mice are generated to delineate the functions of hDab2, a regulator of platelet function, in the control of bleeding associated with thrombocytopenia. Here we show that hDab2-KI mice with thrombocytopenia display decreased bleeding time when compared to the control mice. hDab2 augments thromboxane A₂ (TxA₂) mimetic U46619- but not other agonists-stimulated granule secretion, integrin activation, and aggregation at a lower platelet concentration in vitro. Binding of hDab2 to phosphatidic acid (PA) facilitates formation of the PA-hDab2-AKT complex leading to an increase in U46619-stimulated AKT-Ser473 phosphorylation and the first wave of ADP/ATP release. Consistent with these findings, hDab2 expression in platelets from patients with immune thrombocytopenic purpura is positively correlated with U46619-stimulated ATP release, which in turn inversely correlated with their bleeding tendency. hDab2 appears crucial in regulating bleeding severity associated with thrombocytopenia by a functional interplay with ADP/ATP release underlying TxA₂ signaling.

Fig. 1. Generation of megakaryocyte lineage-restricted hDAB2-KI mice.

a Targeting the conditional hDAB2 gene expression cassette into the Rosa26 locus of ESCs. The pRosa26PA-c-hDAB2-KI plasmid containing the conditional hDAB2 gene expression cassette flanked with Rosa26 genomic sequences (green and red lines) was transfected into ESCs. Insertion of the expression cassette into the wild-type Rosa26 locus was obtained by homologous recombination after positive and negative selection with neomycin and diphtheria toxin, respectively. Stop cassette flanked by two LoxP sites containing the sequences of PGK promoter, neomycin resistant gene (Neo), and 4X polyadenylation signal (4XpA) prevented constitutive HA-hDAB2 gene expression. DTA, diphtheria toxin fragment A. b Representative data for the wild-type (+, 11.5 kb) and fl (6 kb) allele are shown. Genomic DNA from 96 ESCs were analyzed by Southern blot. There were 84 ESCs with wild-type (+/+) and 12 ESCs with heterozygous (fl/+) Rosa26 locus containing one wild-type allele and one allele with the conditional hDAB2 gene expression cassette. EcoRV-digested genomic DNA was hybridized with a S1 probe as shown in panel a for genotyping. c Illustration of the genomic positions related to the generation of Rosa26^fl/fl, PF4-Cre, and hDAB2-KI mice. hDAB2-KI mice were generated when the stop cassette was removed by in vivo Cre-LoxP recombination through mating Rosa26^fl/fl mice with PF4-Cre mice. d Tail genomic DNA from the representative mice were analyzed by PCR for genotyping. Genotyping was performed in over ten independent experiments. The positions of the six PCR primers (P1, P2, P3, P4, P5 and P6) used in the study are indicated in panels a and c. The PCR products of 447, 320, and 450 bps correspond to the wild-type Rosa26 allele (+), the locus-specific hDAB2-KI allele (fl), and the PF4-Cre gene (Cre), respectively. NTC, no template control. e The lysates of Rosa26^fl/fl and hDAB2-KI platelets were analyzed by Western blot using antibodies against Dab2, HA-tag, and β-actin, respectively. Representative data from 3 biological replicates are shown. Source data are provided as a Source Data file.

Fig. 2. hDAB2-KI mice display normal appearance, platelet biogenesis, hemostasis and thrombus formation.

a The appearance for Rosa26^fl/fl and hDAB2-KI mice. b, c Representative TEM images (32,000 X) of platelets from Rosa26^fl/fl and hDAB2-KI mice (n = 3 mice per genotype), with quantification of the number of α- and δ-granules per platelet shown in panel c. A total of 144 platelets from Rosa26^fl/fl mice and 142 platelets from hDAB2-KI mice was present in the 10 random fields selected for quantification (n = 10). Each data point represents the average number of α- and δ-granules per platelet obtained from one random field. The experiments were repeated two times independently with similar results. α, α-granule; δ, δ-granule. Scale bar = 1 μm. d The content of ADP and ATP of washed platelets was analyzed and quantified using the GloMax 20/20 luminometer. Data represent the mean ± SEM of 3 biological replicates each consisting of 3 technical replicates. e Bleeding time for the Rosa26^fl/fl (n = 19) and hDAB2-KI (n = 17) mice was plotted and the horizontal lines represent the median of tail bleeding time. Unpaired two-sided Student’s t-test was used for statistical analysis of the data in panels (c–e). p < 0.05 was considered statistically significant. Exact p values for each group and Source data are provided as a Source Data file. f The percentages of mice as described in e that were rebled within 2 min after blood flow stop are shown. g, h Representative images for Rosa26^fl/fl (n = 12) and hDAB2-KI (n = 16) mice at 0 and 900 sec following FeCl₃ treatment are shown. a, arterioles; v, venules. Scale bar = 200 μm. The occlusion time of the Rosa26^fl/fl (n = 12 for arterioles and n = 11 for venules) and hDAB2-KI mice (n = 11 for arterioles and n = 14 for venules) was recorded and analyzed by the two-sided Mann-Whitney U test. p < 0.05 was considered statistically significant. Data represent the mean ± SEM. Source data are provided in Source Data file.

Fig. 3. hDAB2-KI mice with thrombocytopenia display a reduced bleeding time.

a, b Rosa26^fl/fl (n = 18) and hDAB2-KI (n = 20) mice were intraperitoneally injected with anti-CD41 antibody (anti-CD41) daily for 2 days to induce ITP. Platelet number and bleeding time analyses were performed 24 h after final injection of anti-CD41. For analysis at the recovery phase (RP), Rosa26^fl/fl (n = 17) and hDAB2-KI (n = 13) mice were injected with anti-CD41 daily for 2 days and experimentation was performed 72 h after final injection of anti-CD41. Platelet number was plotted and data represent the mean ± SEM (left of panel a). Unpaired two-sided Student’s t-test was used for statistical analysis. A bleeding time assay was performed and the time when bleeding stopped was recorded. A bleeding time longer than 600 sec was set as 600 sec. Bleeding time was plotted and the horizontal lines represent the median of tail bleeding time (right of panel a). The percentage of mice with bleeding exceeding 600 sec in each group was indicated. The effect of hDAB2-KI on the bleeding time under thrombocytopenia and at RP was analyzed by the log-rank test (panel b). p < 0.05 was considered statistically significant. Exact p values for each group and Source data are provided as a Source Data file.

Fig. 4. hDab2 selectively augments U46619-stimulated platelet activation at a lower platelet concentration of 1.8 × 10⁸/ml in vitro.

a Spreading of Rosa26^fl/fl and hDAB2-KI platelets on collagen was performed. The platelets were then labeled with FITC-conjugated phalloidin and recorded by fluorescence microscopy under a high power field (HPF) of 1000 X magnification. The images of spreading platelets are shown. The number and area of platelet spreading/HPF were determined. Data represent the mean ± SEM of 3 biological replicates each consisting of 3 random HPFs. scale bar = 10 μm. b Washed platelets were stimulated by the indicated soluble agonists and platelet aggregation was recorded by an AggRAMTM system (Helena Laboratories). Representative aggregation traces in response to collagen (Col), thrombin (Th), PAR4 peptide (PAR4), ADP, and U46619 (U46) at the indicated concentration of platelets and agonists are shown. Arrows indicate the point of adding the agonist into the reaction. The percentage of light transmission at the end of the aggregation assays was recorded. The number of biological replicates for stimulation of 2.4 x 10⁸/ml and 1.8 × 10⁸/ml platelets by high and low concentrations of agonists was 4, 6, 4, and 4, respectively for Col; was 6, 8, 4, and 6, respectively for Th; was 3, 5, 3, and 5, respectively for PAR4, was 3, 5, 3, and 6, respectively for ADP; was 6, 4, 9, and 9 for U46. Data represent the mean ± SEM. c Resting or U46619-stimulated (0.25 μM) platelets were incubated with Alexa Fluor 488-conjugated fibrinogen followed by flow cytometry analysis. Data represent the mean ± SEM of 3 biological replicates. Asterisks above the bars indicate statistically significant between groups based on unpaired two-sided Student’s t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Exact p values and Source Data are provided in Source Data file.

Fig. 5. hDab2 regulates TxA₂ signaling leading to reduced bleeding time of hDAB2-KI mice with thrombocytopenia.

a Rosa26^fl/fl mice (n = 3 for each group) were treated daily for 7 days with solvent vehicle DMSO (2.5%) or low-dose aspirin (25 mg/kg) by oral gavage. Blood samples were collected for analysis of serum TxB₂ concentration. Data represent the mean ± SEM of 3 biological replicates each consisting of 3 technical replicates. b Rosa26^fl/fl and hDAB2-KI mice were treated with vehicle control (n = 5), aspirin (25 mg/kg, n = 5), anti-CD41 antibody (n = 13) and anti-CD41 antibody with aspirin (n = 8), respectively. Bleeding time was recorded and the percentage of mice with bleeding time exceeding 600 sec of each group was indicated. A bleeding time longer than 600 sec was set as 600 sec. c The effect of low-dose aspirin on the bleeding time of hDAB2-KI mice under thrombocytopenia was analyzed by the log-rank test. p < 0.05 was considered statistically significant. d Blood samples of Rosa26^fl/fl (n = 4 for vehicle control and n = 3 for anti-CD41 antibody) and hDAB2-KI (n = 5 for vehicle control and n = 3 for anti-CD41 antibody) mice were collected for analysis of the serum TxB₂ concentration. e 2,3-dinor-TxB₂ and 2,3-dinor-6-keto-PGF_1α present in the urine samples of Rosa26^fl/fl (n = 6 for vehicle control and n = 5 for anti-CD41 antibody) and hDAB2-KI (n = 7 for vehicle control and anti-CD41 antibody) mice were determined by LC-MS/MRM and were normalized by the urinary level of creatinine. The two-sided Mann-Whitney U test was used for statistical analysis. ns, no significance. f Washed platelets were incubated with the FITC-conjugated anti-TP receptor antibody and analyzed by flow cytometry. Data represent the mean ± SEM of 4 biological replicates. g Supernatants from PAR4 peptide (PAR4, n = 3 biological replicates)- or collagen (Col, n = 4 biological replicates)-stimulated Rosa26^fl/fl and hDAB2-KI platelets were collected for analysis of TxB₂ concentration. Asterisks above the bars in panel a, d, f and g indicate statistically significant between groups based on unpaired two-sided Student’s t-test (*p  <  0.05, **p  <  0.01). Data represent the mean ± SEM. Exact p values and Source Data are provided in Source Data file.

Fig. 6. Augmentation of TxA₂-stimulated ADP/ATP release by hDab2 enhances platelet aggregation and reduces bleeding in mice with thrombocytopenia.

a Resting (n = 6 biological replicates) and U46619-stimulated (0.25 μM, n = 3 biological replicates) platelets (1.8×10⁸/ml) were incubated with PE-conjugated anti-CD62P antibody followed by flow cytometry analysis (left panel). Supernatants from U46619-stimulated (0.25 μM) platelets (1.8×10⁸/ml) were collected. ADP and ATP release was analyzed and quantified using the GloMax 20/20 luminometer (right panel). Data represent the mean ± SEM of 5 biological replicates. b, c hDAB2-KI platelets (1.8×10⁸/ml) were pre-incubated with the indicated concentrations of apyrase (n = 4, n = 3, n = 3 and n = 3 biological replicates for low to high concentration of apyrase) for 1 min and then stimulated with U46619 (0.25 μM). The Rosa26^fl/fl (n = 3 biological replicates for each ADP concentration) and hDAB2-KI (n = 17, n = 9, n = 11 and n = 17 biological replicates for low to high concentration of ADP) platelets (1.8 × 10⁸/ml) were stimulated with U46619 (0.25 μM) and the indicated concentrations of ADP simultaneously. Platelet aggregation was recorded by using the AggRAMTM system (Helena Laboratories). Representative traces for platelet aggregation are shown. Arrows indicate the point of adding agonists into the reaction. Data represent the mean ± SEM. Asterisks above the bars in panel a and c indicate statistically significant between groups based on unpaired two-sided Student’s t-test (*p < 0.05, **p < 0.01, ***p < .001). d, e hDAB2-KI mice were treated with vehicle control (n = 5), clopidogrel (3 mg/kg, n = 5), anti-CD41 antibody (n = 14) and anti-CD41 antibody with clopidogrel (n = 12), respectively. A bleeding time assay was performed and the time when bleeding stopped was recorded. A bleeding time longer than 600 sec was set as 600 sec. Bleeding time was plotted and the horizontal lines represent the median of tail bleeding time (panel d). The percentage of mice with bleeding exceeding 600 sec in each group was indicated. The effect of clopidogrel on the bleeding time of hDAB2-KI mice under thrombocytopenia was analyzed by the log-rank test (panel e). p < 0.05 was considered statistically significant. Exact p values and Source Data are provided in Source Data file.

Fig. 7. Correlation study of hDab2 expression in platelets, U46619-stimulated ATP release, and the bleeding score of ITP patients.

a–c The peripheral blood from ITP patients (n = 11) was collected for analysis of platelet hDab2 expression and U46619-stimulated ATP release. The bleeding score was obtained at the initial stage of diagnosis. One-sided Pearson’s correlation test was performed to evaluate the strength of correlation for the indicated variables. The strength of correlation was determined by the value of the correlation coefficient (r). The data dots that were close to each other were marked in red to indicate the presence of two data points in the plot. Source Data are provided in Source Data file.

Fig. 8. hDab2 regulates TxA₂-induced AKT-Ser473 phosphorylation and enhances TxA₂-induced first wave of ADP/ATP release.

a, b ATP release of agonist-stimulated Rosa26^fl/fl and hDAB2-KI platelets was detected by a Model 700 lumi-aggregometer (Chrono-Log). Representative traces for ATP release are shown. The amount of U46619-induced first wave (1^st) and second wave (2^nd) ATP release, and the maximal amount of ATP release (total) by U46619 (n = 6 biological replicates), PAR4 peptide (PAR4, n = 3 biological replicates) and collagen (Col, n = 3 biological replicates) were plotted. Data represent the mean ± SEM. c hDAB2-KI platelets were pre-incubated with 0.5 μM U73122, 500 nM wortmannin (Wort), 0.1 μM Cangrelor (Can) or 1.25 μM VU0155069 (VU1) and then stimulated with U46619 (0.25 μM), PAR4 peptide (0.15 mM) and collagen (2.5 μg/ml), respectively. ATP release was detected by a Model 700 lumi-aggregometer (Chrono-Log). Representative traces of ATP release are shown. Data represent the mean ± SEM. The number of biological replicates for control group and pretreatment of U73122, Wort, Can, and VU1 followed by U46619 stimulation was 6, 3, 4, 4, and 3, respectively; followed by PAR4 stimulation was 7, 3, 6, 4, and 4, respectively; and followed by collagen stimulation was 7, 3, 3, 3, and 3, respectively. d, e Washed platelets were stimulated by U46619 for 15 sec. Platelet lysates were analyzed by Western blot using the indicated antibodies. The relative level of phosphorylation was quantified by ImageJ software. Data represent the mean ± SEM of 4 biological replicates. Asterisks above the bars in panel b, c and e indicate statistically significant between groups based on unpaired two-sided Student’s t-test (*p  <  0.05, **p  <  0.01, ***p  <  0.001). ns, no significance. Exact p values and Source Data are provided in Source Data file. The representative Western blot images were derived from different experiments. f Lysates from resting or U46619-stimulated hDAB2-KI platelets were enriched by immunoprecipitation using the anti-HA antibody. The immunoprecipitated proteins were analyzed by Western blot using the anti-Dab2 and anti-AKT antibodies, respectively. Representative data from 4 biological replicates are shown.

Fig. 9. hDab2 is a PA-binding protein.

a A membrane strip spotted with 15 types of lipids (100 pmol) was incubated with GST (8 nM, left panel) or GST-Dab2N (8 nM, right panel). The binding of recombinant proteins to lipids was immunodetected by using anti-GST antibody. GST or GST-Dab2N proteins were spotted directly on the nitrocellulose membrane as the positive control (PTC) for immunodetection. Representative data for at least 2 independent experiments are shown. b The indicated compositions of liposomes were incubated with GST-Dab2N, GST, GST-SidC-3C (a PI4P-binding protein) or GST-PLC-δ1-PH (a PIP2-binding protein), respectively. Proteins that were present in the supernatant (S) and the liposome pellet (P) fractions were analyzed by Western blotting using the anti-GST antibody. Representative data from at least 2 biological replicates are shown. c Interactions of PA with immobilized GST-Dab2N were analyzed by SPR. Response units were used to quantify the bound protein fraction at increasing PA concentrations. Data represent the mean ± SEM of 4 biological replicates for 0, 0.17, 0.37, 0.74, 0.93, and 1.56 nM PA and 3 biological replicates for 1.25 nM PA. The equilibrium dissociation constant (K_D) was calculated by using the built-in non-linear regression curve fit and the one-site binding model. Source Data are provided in Source Data file.

Fig. 10. AKT is recruited to PA through an hDab2-AKT interaction.

a Illustration of the PA consensus binding motifs in GST-Dab2N. b The nitrocellulose membranes spotted with 1500 pmol PA were incubated with HA-Dab2-20-45 (20-45) and HA-Dab2-167-186 (167-186) peptides, respectively. The signal of PA binding in the 20-45 peptide group was arbitrarily set as 1. Data represent the mean ± SEM of 3 biological replicates. c The mutational sites of GST-Dab2-4M and GST-Dab2-5M. d, e The GST (n = 4 biological replicates), GST-Dab2N (n = 8 biological replicates), GST-Dab2-4M (n = 4 biological replicates) and GST-Dab2-5M (n = 3 biological replicates) proteins were incubated with 2 mM reconstituted liposomes in the absence or presence of PA. Proteins that were present in the supernatant (S) and the liposome pellet (P) fractions were analyzed by Western blotting using anti-GST antibody. Data represent the mean ± SEM. f The indicated proteins bound to the PA were immunodetected by using anti-GST antibody. The indicated recombinant proteins were spotted directly on nitrocellulose membrane as a positive control (PTC) for equal protein loading. The signal of binding to PA was quantified by ImageJ software. The signal of PA binding in the GST-Dab2N group was arbitrarily set as 1. Data represent the mean ± SEM of 5 biological replicates. g The binding of the indicated proteins to the PA was immunodetected by using anti-GST antibody. Representative data for 3 biological replicates are shown. h The nitrocellulose membrane was directly incubated with His-AKT-FL (lane 1), spotted with PA and incubated with AKT (lane 2), spotted with PA and incubated with AKT and GST-Dab2-FL (lane 3), or spotted with PA and incubated with AKT and GST-Dab2N (lane 4), respectively. The binding of AKT to the lipids was immunodetected by using anti-AKT antibody. Representative image for 3 biological replicates are shown. The mean signal in lane 2 was arbitrarily set as 1. Data represent the mean ± SEM of 3 biological replicates. Asterisks above the bars in panel b, e, f and h indicate statistically significant between groups based on unpaired two-sided Student’s t-test (**p  <  0.01, ***p  <  0.001). Exact p values and Source Data are provided in Source Data file.