Bleeding diathesis in platelet talin–deficient mice
Despite normal platelet counts in adult Tln1^fl/fl Cre⁺ mice (Table S1), these mice showed dramatically impaired hemostasis. In a tail bleeding assay, Tln1^fl/fl Cre⁺ mice bled continuously for the 10-min duration of the assay, whereas Tln1^fl/fl Cre^– mice stopped bleeding an average of 4.6 min after tail resection (Fig. 2 A). Platelet talin deficiency was also associated with spontaneous bleeding. Lethal spontaneous internal hemorrhage, most often localized to the abdominal cavity, was observed in 8% of 1–2-d-old Tln1^fl/fl Cre⁺ mice (Fig. 2 B). By 3 wk of age, 45% fewer Tln1^fl/fl Cre⁺ than Tln1^fl/fl Cre^– mice were alive. In addition, 12 out of 76 Tln1^fl/fl Cre⁺ mice were found dead between 3 and 9 wk of age compared with 3 out of 139 Tln1^fl/fl Cre^– mice (Fig. 2 B). Tln1^fl/fl Cre⁺ mice that survived to 8 wk of age had a 95% incidence of gastrointestinal bleeding compared with 7% of Tln1^fl/fl Cre^– littermates as judged by an assay for fecal blood (Fig. 2 C). Gastrointestinal bleeding in Tln1^fl/fl Cre⁺ mice was associated with profound anemia as manifested by significantly reduced red blood cell counts and hemoglobin concentration (Fig. 2 D and Table S1).

View larger version (19K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Reduced survival and perinatal hemorrhage in Tln1fl/fl Cre+ mice. (A) Tln1fl/fl Cre+ mice have prolonged bleeding times in a tail bleeding assay. Time to cessation of bleeding after tail resection was recorded for up to 10 min, at which time bleeding was stopped by cauterization. (B) Example of 1-d-old Tln1fl/fl Cre+ pup found dead with visible internal hemorrhage (arrows). Table showing reduced survival of Tln1fl/fl Cre+ mice. The number of animals obtained from Tln1fl/fl Cre2 x Tln1fl/fl Cre+ breeding at 3 wk of age is shown. (C) Incidence of fecal blood in Tln1fl/fl Cre+ and Tln1fl/fl Cre2 mice at 8–10 wk of age was determined by a guaiac-based hemoccult assay. (D) Peripheral red blood cell counts from 10-wk-old Tln1fl/fl Cre+ and Tln1fl/fl Cre2 littermates.

Platelet talin is required for thrombus formation
Thrombus formation in mice with talin-deficient platelets was examined in vivo by ferric chloride–induced injury of the common carotid artery. Complete occlusion of the carotid arteries of Tln1^fl/fl Cre^– mice occurred 7.0 ± 0.9 min after injury, whereas none of the Tln1^fl/fl Cre⁺ mice tested showed reduced flow during the 20-min assay (Fig. 3). Histological examination of the carotid arteries of these animals after the thrombosis experiment indicated a similar extent of ferric chloride–induced vessel injury in Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– mice (Fig. S2, available at http://www.jem.org/cgi/content/full/jem.20071800/DC1). Collectively, our results show that deletion of talin1 in platelets leads to markedly impaired hemostasis and thrombus formation in vivo.

View larger version (101K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Impaired thrombus formation in Tln1fl/fl Cre+ mice in vivo and ex vivo. (A) Time to occlusion of the carotid artery was determined with a Doppler flow probe after a 3-min application of 10% ferric chloride. The experiment was stopped 20 min after injury in all animals. (B) Adhesion of Tln1fl/fl Cre+, Tln1fl/fl Cre2, and b3(L476A) platelets to collagen and subsequent thrombus formation in flowing blood were analyzed by epifluorescence and confocal microscopy at a shear rate of 1,500 s21. Heparinized whole blood containing 10 mM mepacrine to render platelets fluorescent was perfused over glass coated with fibrillar type I collagen for 2 min. Tln1fl/fl Cre2 platelets adhere to the collagen-coated surface and form thrombi (larger aggregates of bright fluorescence). In contrast, Tln1fl/fl Cre+ platelets form only transient contacts resulting in sparse coverage of platelets on the surface. b3(L746A) platelets form a monolayer that cover much of the surface but do not form thrombi, as seen by the lack of highly fluorescent aggregates that form with Tln1fl/fl Cre2 platelets. Images shown are single frames from a real-time recording (Video S1). Bar, 20 mm. (C) Percent of the collagen-coated surface covered with platelets was calculated as the number of fluorescent pixels (due to adhesion of fluorescently labeled platelets) divided by the total number of pixels (representing the total surface). *, P < 0.0005; NS, not significant. (D) Quantification of the volume of the thrombi formed on the collagen-coated surface after perfusion for 2 min with blood from Tln1fl/fl Cre+, Tln1fl/fl Cre2, and b3(L746A) mice. Confocal serial Z-section reconstructions of the platelet thrombi were used to calculate the thrombi volume as described previously (reference 32). *, P < 0.0005.

Platelet talin is required for integrin-mediated platelet adhesion to collagen and platelet aggregation
We examined platelet adhesion and aggregation in vitro to directly assess the effects of talin deficiency on these integrin-dependent processes. In static adhesion assays, the talin-deficient platelets showed a marked defect in adhesion to soluble type I collagen (Fig. S1 A, available at http://www.jem.org/cgi/content/full/jem.20071800/DC1). Furthermore, like β3(L746A) platelets (18), talin-deficient platelets manifested profoundly impaired aggregation in response to stimulation with ADP or PAR4 peptide (Fig. S1 B). Thus, the talin-deficient platelets manifest virtual absence of platelet functions mediated by both 2β1 and IIbβ3 integrins, thus accounting for their profound hemostatic defect.

Talin is required for activation of platelet 2β1 and IIbβ3 integrins
Agonist-induced increase in integrin IIbβ3 affinity (activation) is required for platelet aggregation (19). Indeed, β3(L746A) mice, in which IIbβ3–talin interactions are disrupted, have impaired IIbβ3 integrin activation and platelet aggregation (18). To test the requirement of talin for the activation of IIbβ3, we measured binding of FITC-labeled fibrinogen to the surface of washed Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelets by flow cytometry. Stimulation of Tln1^fl/fl Cre^– platelets with a combination of ADP/epinephrine (100 µM each) or PAR4 peptide (1 mM) led to an increase in the amount of bound fibrinogen. In contrast, the amount of agonist-induced fibrinogen binding was greatly reduced in Tln1^fl/fl Cre⁺ platelets (Fig. 4 A and Fig. S3, which is available at http://www.jem.org/cgi/content/full/jem.20071800/DC1). In the presence of 0.5 mM MnCl2, however, Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelets bound similar amounts of fibrinogen, indicating that the IIbβ3 present on Tln1^fl/fl Cre⁺ platelets is capable of binding fibrinogen if activated exogenously (Fig. 4 B and Fig. S3). Thus, with regards to IIbβ3 activation, platelet talin deficiency phenocopies the β3(L746A) mutation in which the β3 integrin–talin interaction is disrupted. Collectively, these data firmly establish talin as a critical regulator of IIbβ3 integrin activation in vivo.

View larger version (21K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Impaired agonist-induced activation of b3 and b1 integrins in Tln1fl/fl Cre+ platelets. (A) The amount of FITC-labeled fibrinogen bound to platelets from Tln1fl/fl Cre+ or Tln1fl/fl Cre2 mice was measured by flow cytometry and expressed as the amount of fibrinogen bound to platelets in each group relative to the amount of fibrinogen bound to platelets in the presence of 0.5 mM MnCl2. *, P < 0.001. (B) Fibrinogen binding to Tln1fl/fl Cre+ and Tln1fl/fl Cre2 platelets was similar in the presence of 0.5 mM MnCl2. (C) Activation of b1 integrin in Tln1fl/fl Cre+, Tln1fl/fl Cre2, and b3(L746A) platelets after stimulation with 1 mM PAR4, 100 mm ADP, and 100 mM epinephrine was measured by binding of the conformation-sensitive antibody 9EG7 and expressed relative to total b1 integrin surface expression measured by the conformation-insensitive antibody HMb1-1. *, P < 0.05; **, P < 0.0005.

Thus, the principle that talin is required for activation of β1 and β3 integrins, which was suggested by in vitro studies (20), applies in vivo. Furthermore, the central role of talin in integrin function in invertebrates (11, 12, 21) extends to vertebrates. A β3(L746A) mutation that selectively disrupts the ability of β3 integrin to bind talin leads to impaired agonist-induced activation of platelet IIbβ3 (18). Together with the present finding of impaired agonist-induced activation of IIbβ3 in talin-deficient platelets, these data show that talin binding to integrin β cytoplasmic domains is a final common step in integrin activation in vivo, and that disruption of this interaction has a profound impact on integrin-dependent adhesive functions in mammals.

Talin-deficient platelets are nearly completely deficient in hemostatic function. It is noteworthy that the pathological bleeding observed in Tln1^fl/fl Cre⁺ mice is absent in β3(L746A) mice despite having comparable impairments in IIbβ3 activation and platelet aggregation (18). One obvious explanation for this more severe phenotype in the platelet talin-deficient animals is the impairment in the activation and function of platelet β1 integrins. Furthermore, in Drosophila, talin deletion results in marked weakening of the connections of integrins with the actin cytoskeleton; hence, a defect in the connection of β1 and β3 integrins to the actin cytoskeleton may also contribute to the severe phenotype observed. Deletion of platelet β1 integrins or point mutations that would disrupt β1–talin interactions can result in defects in platelet function and in hemostasis (4, 22). Given the hemostatic defects that result from lack of platelet β3 or β1 integrins, and our finding that platelet integrin function is virtually completely dependent on talin, it is likely that the hemostatic defect in Tln1^fl/fl Cre⁺ mice is ascribable largely to the lack of integrin function.

Morphology and surface receptor expression is normal in talin-deficient platelets
To examine the effect of deleting talin on platelet structure, we examined Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelet morphology by electron microscopy. Platelet shape, granules, mitochondria, open canalicular system, and microtubule coils appeared similar in Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelets (Fig. 5 A). Tln1^fl/fl Cre⁺ mice had slightly larger platelets than Tln1^fl/fl Cre^– mice as judged by flow cytometry (forward scatter, 13.9 ± 0.2 vs. 15.3 ± 0.4 arbitrary units, Tln1^fl/fl Cre^– vs. Tln1^fl/fl Cre⁺, P < 0.005) and by measurement of the area of at least 140 randomly selected electron microscopic platelet profiles (0.67 ± 0.03 µm² vs. 0.97 ± 0.4 µm², Tln1^fl/fl Cre^– vs. Tln1^fl/fl Cre⁺ , P < 0.005). As noted above, the Tln1^fl/fl Cre⁺ mice have active gastrointestinal bleeding, suggesting that the slightly increased platelet size could be due to an increased proportion of circulating young platelets. In contrast to β3 integrin null platelets (17), talin-deficient platelets had normal fibrinogen content, suggesting that talin-dependent activation of IIbβ3 in mature megakaryocytes is not required for fibrinogen uptake (Fig. 5 B). In addition, the surface expression of several adhesion receptors, as measured by flow cytometry, was not significantly different in Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelets (Fig. 5 C). Of particular importance, the quantity of surface P-selectin was similar on both resting and stimulated Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– platelets, confirming the presence of platelet granules in the absence of talin. Furthermore, both platelet genotypes exhibited a similar fourfold increase in P-selectin surface expression in response to ADP/epinephrine/PAR4 peptide stimulation, confirming that the mutant platelets were capable of responding to platelet agonists. Collectively, these data demonstrate that talin is dispensable for the formation of platelets that can respond to platelet agonists and manifest a normal complement of granule constituents and adhesion receptors. Thus, the observed adhesion defects in Tln1^fl/fl Cre⁺ mice are ascribable to loss of talin-mediated integrin functions and not a general disruption of platelet structure and function.

View larger version (86K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Platelets from Tln1fl/fl Cre+ mice are structurally normal and express normal levels of surface receptors. (A) Electron micrographs showing normal structural features of Tln1fl/fl Cre+ platelets. Platelets from Tln1fl/fl Cre+ (top left) and Tln1fl/fl Cre2 (bottom left) mice both display a discoid shape characteristic of resting platelets and similar granular contents. Insets show microtubule coils in platelets from both Tln1fl/fl Cre+ and Tln1fl/fl Cre2 mice. Equatorial section of a Tln1fl/fl Cre+ platelet (top right) shows circumferential microtubule coil (arrow heads). Bars, 1 mm. (B) Platelet fibrinogen content was determined by Coomassie blue staining of platelet lysates separated by SDS-PAGE. 5 mg of purified human fibrinogen served as a marker for the prominent protein band corresponding to fibrinogen in the platelet lysate samples. Consistent with previous reports (reference 17), fibrinogen content is reduced in platelets from b3 integrin null mice. (C) Surface expression of integrin aIIb, a2, b1, b3, and P-selectin was measured by flow cytometry. For P-selectin expression, platelets were incubated with or without PAR4/ADP/epinephrine (1 mm/100 mM/100 mM) for 5 min before the addition of an FITC-conjugated P-selectin antibody.

Here, we have shown that platelet talin is essential for platelet-dependent hemostasis because it is required for the function and activation of β1 and IIbβ3 integrins. Despite Tln1^fl/fl Cre⁺ mice having normal platelet counts, these animals exhibited both lethal spontaneous bleeding and resistance to induced thrombosis. These in vivo findings were ascribable to profound defects in the function of multiple platelet integrins, as platelets from Tln1^fl/fl Cre⁺ mice failed to adhere to collagen or to form platelet-rich thrombi ex vivo. In vitro studies documented the profound impairment of integrin-mediated adhesion and platelet aggregation in talin1-deficient platelets and showed that these platelets were deficient in the agonist-induced activation of both 2β1 and IIbβ3 integrins, despite maintaining the capacity to respond to the agonists as indicated by surface display of P-selectin. Thus, talin is necessary for the activation of 2β1 and IIbβ3 integrins in vivo, and platelet talin is absolutely required for hemostasis because it is necessary for the adhesive functions of these integrins.

	MATERIALS AND METHODS

Top ABSTRACT RESULTS AND DISCUSSION MATERIALS AND METHODS REFERENCES

 
Generation of mice.
Conditional talin1 knockout mice were generated by introducing loxP sites flanking coding exons 1–4 of the Tln1 gene by gene targeting. Targeting of the Tln1 locus was confirmed by Southern blot of EcoR1-digested genomic DNA hybridized with a 5' cDNA probe. Mice were genotyped by PCR using the following primers indicated in Fig. 1 A: primer a: 5'-aagcaggaacaaaagtaggtctcc-3' and primer b: 5'-gcatcgtcttcaccacattcc-3'. Mice homozygous for the Tln1 floxed allele (Tln1^fl/fl) on a mixed C57BL/6-Sv129 genetic background were crossed with PF4-Cre (Cre⁺) mice on a C57BL/6 background (15). To obtain mice with talin1-deficient platelets, Tln1^fl/fl Cre⁺ males were bred with Tln1^fl/fl Cre^– females. In all experiments, Tln1^fl/fl Cre⁺ mice were compared with Tln1^fl/fl Cre^– sex-matched littermates. The generation of β3(L746A) mice has been recently described (18). Mice were housed in the University of California, San Diego, animal facility, and experiments were approved by the university's Institutional Animal Care and Use Committee.

SDS-PAGE.
For examination of platelet talin protein content, washed platelets were lysed by adding 1 vol of 2X modified RIPA buffer (300 mM NaCl, 100 mM Tris, pH 7.4, 0.2% SDS, 2% Triton X-100, 2% sodium deoxycholate, 2 mM PMSF, 2 mM NaVO4, 2 mM NaF, 2 mM EDTA, and complete protease inhibitor; Roche), and samples were clarified by centrifugation at 13,000 g for 10 min at 4°C. Laemmli buffer containing 10 mM DTT was added to 10 µg of protein lysates, and samples were boiled for 5 min before being separated on 6% Tris-glycine gels (Invitrogen) and stained with Coomassie blue. For analysis of platelet fibrinogen content, platelets were lysed in Laemmli buffer in the absence of any reducing agent and separated on a 6% Tris-glycine gel. Fibrinogen was identified as a Coomassie blue–stained band migrating with an apparent molecular mass of 340,000.

Immunofluorescence.
After lysing red blood cells with RBC lysis buffer (155 mM NH4Cl, 10 mM KHCO₃, and 0.1 mM EDTA) bone marrow cells from the femurs of 6–9-wk-old mice were fixed in 3.7% formaldehyde/PBS and applied to fibrinogen-coated (100 µg/ml) glass slides by Cytospin preparation (Thermo Fisher Scientific). Cells were permeabilized with 0.1% Triton X-100/PBS containing 5% BSA for 1 h at room temperature and incubated with talin antibody 8d4 (1:50 dilution; Sigma-Aldrich) overnight at 4°C. After washing with PBS, cells were incubated with 2.5 µg/ml FITC-conjugated anti-CD41 (BD Biosciences) and 5 µg/ml Alexa-568 goat anti–mouse IgG for 2 h at room temperature. After washing, slides were mounted with coverslips using Vectashield anti-fade media (Vector Laboratories) and observed on a Leica DM LS fluorescence microscope. Images were captured with a spot color digital camera (National Instruments) using manual exposure settings that were identical for Tln1^fl/fl Cre⁺ and Tln1^fl/fl Cre^– samples.

Hemostasis assays.
The presence of fecal blood was detected with a guaiac-based hemoccult detection assay (Helena Laboratories) on freshly obtained stool samples.

Tail bleeding assays were performed by resecting 1 mm of the tail, followed by immersion in 37°C isotonic saline as described previously (17). All experiments were terminated at 10 min by cauterizing the tail.

Platelet isolation and functional assays.
Washed platelets were obtained as described previously (30). Soluble fibrinogen binding was measured by incubating platelets for 20 min with 150 µg/ml FITC-labeled fibrinogen, followed by fixation with 1% formaldehyde for 10 min at room temperature. Specific fibrinogen binding was determined by subtracting the amount of fibrinogen bound in the presence of 5 mM EDTA. Bound fibrinogen was detected with a FACScan flow cytometer (Becton Dickinson). β1 integrin activation was measured by the binding of an FITC-labeled confirmation-sensitive β1 integrin antibody 9EG7 (BD Biosciences), or the confirmation-insensitive PE-conjugated β1 antibody HMβ1-1 (BD Biosciences). Washed platelets were incubated with or without agonists for 10 min at room temperature, followed by the addition of 3 µg/ml of either 9EG7 or HMβ1-1 for 20 min and detected by flow cytometry. Similarly, surface expression of P-selectin was measured by the binding of FITC–anti–P-selectin (BD Biosciences) following the same protocol as described above for measuring β1 integrin activation. Surface expression of β3, IIb, and 2 integrins were measured by flow cytometry with the following antibodies: FITC–anti-CD61, FITC–anti-CD41 (BD Biosciences), and PE–anti-2 integrin (eBioscience).

For analysis of static adhesion of platelets to collagen, 96-well plates (Immulon HB2; Dynex Technologies) were coated with 2 µg of acid-soluble type I collagen from rat tail (Sigma-Aldrich) in 100 µl PBS overnight at 4°C. After two washes with PBS and blocking with 5% BSA/PBS for 2 h at room temperature, 5 x 10⁶ washed platelets suspended in platelet incubation buffer were added to each well and allowed to adhere for 1 h at room temperature. Wells were then washed three times with platelet incubation buffer, and adherent platelets were quantified by acid-phosphatase assay (18). Percent platelet adhesion was calculated as the number of adherent platelets relative to the number of platelets in wells that were not washed (total platelets per well). Platelet aggregation was performed as described previously (18) using platelet-rich plasma (PRP) diluted to a platelet concentration of 3 x 10⁸ platelets/ml with platelet-poor plasma.

Ex vivo adhesion to collagen.
Adhesion and thrombus formation in flowing blood was performed and analyzed as described previously (31, 32).

Transmission electron microscopy.
Blood was drawn by cardiac puncture into 0.1 vol of 0.13 M sodium citrate. After adding 1 vol modified Tyrode's buffer (140 mM NaCl, 2.7 mM KCl, 0.4 mM NaH₂PO₄, 10 mM NaHCO₃, 5 mM dextrose, and 10 mM Hepes) samples were centrifuged for 5 min at 200 g at room temperature to obtain PRP. The PRP was incubated for 30 min at 37°C before fixing by the addition of 1 vol of 2X fixative (3% gluteraldehyde and 6% paraformaldehyde in 0.2 M cacodylate buffer plus 10% sucrose, pH 7.4) and incubated for 15 min at room temperature. Platelets were centrifuged at 700 g for 5 min and resuspended and stored overnight in 1X fixative. Samples were processed as described previously (33), and images were obtained with a JEOL 1200 EX II electron microscope.

Ferric chloride–induced thrombosis.
Ferric chloride–induced thrombosis was performed as described previously (18) by applying a 1.2 X 1.2–mm piece of filter paper soaked in 10% ferric chloride to each side of the common carotid artery of a mouse under isoflurane anesthesia.

Statistics.
Statistical analyses of mouse survival and spontaneous death were performed with ² and Fisher's exact tests, respectively. The statistical significance of all other data were determined using Student's t test. A p-value of <0.05 was considered statistically significant. All error bars represent standard error of the mean.

Online supplemental material.
Video S1 shows adhesion of fluorescently labeled platelets to fibrillar type I collagen in flowing blood. Time shown is from the beginning of flow over the collagen-coated surface. Video S1 is available at http://www.jem.org/cgi/content/full/jem.20071800/DC1.