Diminished Thrombogenic Responses by Deletion of the Podocalyxin Gene in Mouse Megakaryocytes

Podocalyxin (Podxl) is a type I membrane sialoprotein of the CD34 family, originally described in the epithelial glomerular cells of the kidney (podocytes) in which it plays an important function. Podxl can also be found in megakaryocytes and platelets among other extrarenal places. The surface exposure of Podxl upon platelet activation suggested it could play some physiological role. To elucidate the function of Podxl in platelets, we generated mice with restricted ablation of the podxl gene in megakaryocytes using the Cre-LoxP gene targeting methodology. Mice with Podxl-null megakaryocytes did not show any apparent phenotypical change and their rates of growth, life span and fertility did not differ from the floxed controls. However, Podxl-null mice showed prolonged bleeding time and decreased platelet aggregation in response to physiological agonists. The number, size-distribution and polyploidy of Podxl-null megakaryocytes were similar to the floxed controls. Podxl-null platelets showed normal content of surface receptors and normal activation by agonists. However, the mice bearing Podxl-null platelets showed a significant retardation in the ferric chloride-induced occlusion of the carotid artery. Moreover, acute thrombosis induced by the i.v. injection of sublethal doses of collagen and phenylephrine produced a smaller fall in the number of circulating platelets in Podxl-null mice than in control mice. In addition, perfusion of uncoagulated blood from Podxl-null mice in parallel flow chamber showed reduced adhesion of platelets and formation of aggregates under high shear stress. It is concluded that platelet Podxl is involved in the control of hemostasis acting as a platelet co-stimulator, likely due to its pro-adhesive properties.


Introduction
Podocalyxin (Podxl), also known as Myb-Ets transformed progenitor (MEP), is a ,160 KDa, highly sialylated and sulfated, membrane type I mucoprotein of the CD34 family [1]. This protein was first reported in the epithelial cells, podocytes, lining the luminal face of kidney glomeruli [2]. The podocytes are injured in many glomerular diseases, including minimal change disease (MCD), focal segmental glomerulosclerosis (FSGS), collapsing glomerulopathy, diabetic nephropathy, membranous glomerulopathy, crescentic glomerulonephritis, and lupus nephritis [3][4][5]. Those inflammatory processes are accompanied by loss of the normal architecture, ''effacement'', of the luminal surface of podocytes and loss of the urinary filtration slits [3,6]. Moreover, total podocalyxin knock out mice die of anuria within the first few hours after birth [7]. Podxl has also been found in vascular endothelium, lung, brain, multipotent hematopoietic precursors and platelets [8][9][10], and in several types of tumors in which the expression levels appear to correlate with the metastatic capacity [11][12][13]. Despite the consistent data indicating the important role played by Podxl in maintaining a normal renal function, almost nothing is known about its physiological role in extrarenal tissues. Podxl is exposed on the plasma membrane of platelets upon stimulation by agonists [14] suggesting that it could cooperate in maintaining a normal hemostatic responsiveness. Moreover, the loss-of-function mutation (plt1 mice) in the glycoprotein-Nacetylgalactosamine-3-galactosyltransferase (C1GalT1), essential for the synthesis of extended mucin-type O-glycans, produced underglycosylation of podocalyxin and glycoprotein Ib (GPIb) and caused thrombocytopenia and kidney disease, suggesting a contribution of the abnormally glycosylated Podxl to this pathology [15]. In agreement with these observations, we have recently reported a diminished adhesion of sialic acid deficient Oglycomutant CHO-Podxl cells compared with cells expressing normal sialylated Podxl [16]. These antecedents suggest that Podxl, like other mucin proteins, could play a role in cell adhesion/migration. In fact, a particular glycosylation pattern of Podxl could be responsible for the ability of high endothelial venules to bind L-selectin from other cells [17][18][19]. Moreover, CHO cells expressing recombinant human Podxl showed, Pselectin and integrin-dependent, enhanced adhesion and motility over immobilized ligands and increased cellular interactive capacity [16].
The present work aimed at investigating the functional role of Podxl in platelets. For this purpose, we used the Cre-LoxP gene targeting methodology to generate mice with specific ablation of the podxl gene in megakaryocytes. Mice with null-Podxl megakaryocytes showed prolonged bleeding time, decreased platelet aggregation in response to agonists, retarded ferric chlorideinduced closure of the carotid artery, diminished in vivo thrombogenesis and decreased adherence of platelets under flow. These observations suggest that removal of the podxl gene from megakaryocytes perturbs the control of hemostasis.

Animals
Mice of the C57BL/6 strain, maintained under controlled conditions of light and temperature, were used in all the experiments. Podxl lox/lox :: Control or Podxl lox/lox :: Pf4-Cre littermates used in this study were generated on a mixed C57BL/6-129sv genetic background and were backcrossed for 10 generations with C57BL/6 mice. All animal experiments were done in such a way as to minimize the animal suffering, according to relevant national and international guidelines (''Guide for Care and Use of Laboratory at the Institute of Laboratory Animal Resources'').
The ethics committee of the Center for Biological Research (CSIC) and the grant review board (ANEP) of the Spanish Ministry of Science and Technology specifically approved this study in accordance with the guidelines of the European Community Council Directive 86/609 EEC.

Generation of a conditional (floxed) podocalyxin allele
The DNA sequence of the podocalyxin gene was obtained by digestion of the murine genomic BACS 449o15 and 465p05 with BamH I, EcoR I and Xba I. DNA fragments containing the podxl gene were identified by southern blot and cloned. Digestion of clone 465p05 with Sac I released a 4,381 bp fragment comprising exons 3 to 8 of podxl that was cloned into the Sac I site of pCR 2.1-TOPO.
A conditional podxl transgene was built in three steps into the pBS II vector using the pLox plasmids [20], as shown in Fig. 1. The pLoxL contained the left homology arm (exons 3 and 4 of podxl); the pLoxC encompassed exons 5 to 7 of podxl flanked by LoxP sites and neo (PGK-NEO) and tk (HSV-TK) selection cassette; the pLoxR contained the right homology arm (exon 8 and part of the 39-UTR of podxl) and a counterselection cassette for the diphteria toxin A chain (PGK-DT).
The left homology arm was generated by Sac I digestion of the genomic DNA fragment containing the podxl gene and amplification with oligonucleotide primers LoxL-S (59-CCGAATTCCGCCTGCCTCAGCC-39) to introduce an EcoR I site, and LoxL-AS (59-GGATGCATTATCATGCCT-CAGCTGG-39) to replace the site AlwN I by a Nsi I site. The PCR product was cloned into AccepTor-Vector (AccepTor Vector Kits, Novagen). The pLoxL vector was modified by means of adapters to clone the PCR product into the EcoR I and Nsi I sites.
The genomic DNA fragment containing exon 8 and 39-UTR of the podxl gene was amplified using oligonucleotides LoxR-S (59-TCATGCATGCTTGTCTCTAACGCTG-39) to replace a Nsi I site for Nhe I, and LoxR-AS (59-GCTCTAGA GTGCAGTTGGGGTG-39) that replaces a Sac I site for a Xba I. The PCR product was cloned in the AccepTor-Vector. We introduced an adapter into the pLoxR vector to generate Nsi I site for subcloning of the podxl right homology arm.
To generate the ''floxed'' fragment (exons 5 to 7 of podxl) we amplified a DNA fragment comprising exons 5 to 7 from the Sac I genomic DNA fragment using primers LoxC-S (59-TACTG-CAGCTGGGTCCAGAGCCTTA-39) designed to replace the restriction site AlwN I by a Pst I site, and LoxC-AS (59-AGCTGCAGTGACAGCGGGGACAAGA-39) to replace an Nhe I site by Pst I site. The PCR product was cloned into AccepTor-Vector, released by digestion with Pst I, and cloned into the Nsi I site of the vector pLoxL. The pLoxC vector was digested with Not I and Sse I to clone the left homology arm and ''floxed'' fragments upstream of the neo and tk selection cassette. Finally, the Not I-Asc I fragment was subcloned upstream of the right homology arm in the pLoxR vector.
The final construction was sequence-verified using the following primers: Generation of mice with a floxed podxl gene R1 E14 (129/Sv) embryonic stem (ES) cells, kindly donated by Dr. Nagy, were cultured according to their indications [21]. Twenty-five mg of the podxl construct were electroporated into ES cells. Positive selection was achieved with G418 (Sigma) at a concentration of 250 mg/mL and the resistant ES cell clones were screened for homologous recombination by PCR. The modified ES cells were expanded and transfected with the p-MC-Cre plasmid in order to remove the selection cassette (neo and tk). Finally, ES cells carrying the ''floxed'' podxl allele were used for morula aggregation to generate chimeric mice according to previous protocols [22]. Genotyping was carried out by PCR. Two founders were subsequently used for further analysis.
To generate mice with a restricted deletion of Podxl in platelets, the Podxl floxed mice were crossed to Pf4-Cre mice [23] (kindly donated by Dr. Skoda, Basel).

Hemograms and tail bleeding time
Hemograms were determined in blood-EDTA samples withdrawn from the submaxillary vessels, using an automatic hemocytometer model Abacus Junior (Diatron, Diatron Messtechnik GmbH, Austria).
To determine the tail bleeding time the animals were anesthetized with isofluorane. The tail was transected at 5 mm from the tip and it was immediately immersed in 36uC phosphate buffered saline (PBS), and the time until cessation of bleeding recorded.

Ferric chloride-induced closure of the carotid artery
The carotid artery blood flow was determine in mice anesthetized with isofluorane, while maintained the body temperature at 37uC on a heated blanket. The temperature was continuously recorded using a rectal miniprobe. The carotid artery was exposed and the blood flow measured with a 0.5 mm micro-flow probe (0.5 PBS) and the TS420 flow module (Transonic Systems, Inc., Ithaca, NY). The data was continuously recorded with a data acquisition system PowerLab 8/30, model ML870 (AD Instruments Ltd., Chalgrove, Oxfordshire OX44 7RW, UK).
To determine the carotid artery closure time, a 1 mm 2 paper fragment soaked on 2% ferric chloride was placed for 2 minutes on the exposed carotid artery while the blood flow was continuously monitored. We considered the closure time to be that at which the blood-flow was interrupted for 3 minutes.

Preparation and activation of platelets
Citrated blood samples were obtained by cardiac puncture under isofluorane anesthesia. The blood was centrifuged at 100 g for 5 minutes at room temperature. Platelet-rich plasma (PRP) was removed and centrifuged at 1300 g in the presence of 0.1 mg/mL prostacyclin and 0.02 U/mL apyrase (Sigma) for 5 minutes at room temperature, washed 3 times with modified Tyrode's buffer containing 0.1 mg/mL prostacyclin and 0.02 U/mL apyrase. The sedimented platelets were finally resuspended in Tyrode's buffer containing 0.02 U/mL apyrase and rested for 30 minutes at 37uC prior to stimulation.
To determine the extent of platelets activation, 25 mL of washed platelets (10 6 approximately) suspended in Tyrode's-buffer containing 1 mM CaCl 2 were stimulated with ADP, thrombin or collagen. Activation was assessed by determining P-selectin exposure and aIIbb3 activation by flow cytometry (cytometer model FC500) with a twocolor analysis protocol using 5 mL of Dylight 649 anti-GPIba moAb (Emfret Analytics) and 5 mL of FITC-anti-P-selectin (Beckton and Dickinson) or PE-anti-activated aIIbb3 (JON/A, Emfret Analytics) moAbs. The extent of platelet activation was further assessed by measuring the surface exposure of phosphatidylserine, as indicated by the binding of FITC-Annexin V (Immunostep, Salamanca, Spain).

Determination of prothrombin time (PT) and activated partial thromboplastin time (aPTT)
The concentration of PT and aPTT were determined in plasma samples in a coagulometer (Start4, Diagnostica Stago) according to the protocols of the manufacturer. Briefly, PT was determine by adding 100 mL of Neoplastin (Roche) to 50 mL of plasma; to determine aPTT, 50 mL of plasma were mixed with 50 mL of aPTT reagent (Roche) in the presence of CaCl 2 .

Platelet aggregation assays
The aggregation of platelets in whole blood collected in the presence of PPACK (40 mg/mL) was performed measuring the changes in impedance (Ohms) produced by physiological agonists in a Chrono-Log instrument. Samples were incubated at 37uC under constant stirring and, after 5 minutes, 10 mM ADP or 5 mg/ mL collagen (Chrono-Log) were added. The analysis of data was performed 8 minutes after addition of agonists. The results are means 6SEM of seven different observations in each group.
Collagen-induced aggregation of isolated platelets was measured in a final volume of 800 mL of PRP diluted to 1:2 with PBS buffer.

Platelet adhesion and spreading assays
To determine the adherence of platelets to immobilized ligands 100 mL of 5610 8 platelets/mL were added to 96-well microtiter flatbottomed plates coated with fibrinogen (5 mg/mL) and incubated at 37uC for 30 minutes. Non-adherent platelets were removed by aspiration and wells were washed 3 times with 200 mL of PBS buffer.
To study the platelet spreading 100 mL of a platelet suspension containing 5610 7 /mL platelets was added to microtiter plates coated with fibrinogen (5 mg/mL) and allowed to adhere for 1 hour at 37uC. Non adherent cells were removed by washes with PBS buffer. The adhered platelets were fixed with 3.7% PFA, permeabilized with 0.2% Triton X-100 in PBS buffer, washed twice with PBS and stained for F-actin with FITC-conjugated phalloidin for 20 minutes and washed at least 3 times with PBS. Fluorescence was visualized using an Olympus IX-50 inverted microscope with a monochromatic light source and images taken with a DP70 digital camera (Olympus). The images were analyzed using the ImageJ software [24].

In vitro perfusion experiment
To analyze the formation of platelet thrombi we perfused whole blood diluted to 1:2 with Tyrode's buffer containing 1% of serum albumin, through parallel plastic chambers (Ibidi, Germany) coated with fibrinogen. Blood was perfused by aspiration with a syringe pump (Harvard Apparatus) at an estimated shear stress of 5 dyne/cm 2 (mslide III 0.1 Luer) or 20 dyne/cm 2 (m-slide I 0.1 Luer). Images were taken with an Olympus digital camera. The quantification of platelets adhered to immobilized substrates was carried out with the ImageJ software [24]. We first generated a list of the total number of particles without setting any limits to size or circularity. The particles were sorted by size. We had previously estimated the size of a platelet to be ,900 pixels. We, then, sorted the aggregates in the groups according to their sizes: small, between 1000-5000 pixel; intermediates, 5000-10000 pixel; large, 10000-30000, and, finally, very large aggregates, greater than 30000 pixel.
To assess adhesion of platelets onto immobilized collagen under flow conditions, ,25610 6 washed platelets suspended in Tyrode's containing 1 mM CaCl 2 were labeled with calcein-AM (10 mg/ mL) (Molecular Probes, Invitrogen). The labeled platelets were perfused through a parallel flow chamber (m-slide VI 0.1 Luer) precoated with 100 mg/mL of collagen (Chrono-log) at an estimated shear stress of 20 dyne/cm 2 . Non-adhered platelets were removed by washing the chamber with PBS for 5 minutes. To quantify the number of platelets adhered onto collagen the perfusion channel was washed with 50 mL of DMSO and its fluorescence determined. As a blank we used DMSO passed through a collagen coated cell-free perfusion chamber.

Studies of thrombogenesis in vivo
Mice were anesthetized with isofluorane and the yugular veins exposed. A small sample of blood (#100 mL), extracted from one of the yugular veins, was dropped into a tube containing EDTA and the number of platelets at zero determined in a hemocytometer analyzer Abacus Junior (Diatron Messtechnik GmbH, Austria). Through the same vein we injected a mixture of collagen/epinephrine (150 ng collagen type I (Chronolog) and 15 ng epinephrine (SIGMA) per g of mice body weight). In preliminary experiments we could observe that this proportionate amount of reagents produced a significant drop in the number of circulating platelets without causing the death of the animal. At the indicated time after the injection, blood withdrawn from the yugular vein not previously injected was processed as indicated above and the number of platelets determined.

Detection of podxl gene deletion and Podxl protein
The correct recombination of the podxl gene was verified by PCR analysis of genomic DNA extracted from isolated megakaryocytes. The presence of Podxl protein in megakaryocytes was analyzed by western blot using specific monoclonal antibodies [25].

Megakaryocyte quantification
Mice were perfused through the heart with PBS followed by 4% PFA in PBS and the femurs were immersed in 4% PFA-PBS overnight. The fixed femurs were decalcified and pictures of whole longitudinal sections were taken under light microscope at 206 using a DP70 digital camera mounted on as inverted microscope (IX50, Olympus). The area of megakaryocytes was calculated using the ImageJ software and the cells classified in 5 different groups according to their size. A total of two hundred megakaryocytes from either control or Podxl-null mice were analyzed.

In vitro differentiation of megakaryocytes
Bone marrow cells were flushed from femurs and tibias with Tyrode's buffer, dissociated and cultured in IMDM with 5% fetal bovine serum, 50 ng/mL TPO, 2 mM L-glutamine, 50 U/mL penicillin, and 50 mg/mL streptomycin. Megakaryocytes were  Bleeding time was measured in animals anesthetized by isofluorane inhalation. A small fragment of the tail (approx. 5 mm) was excised and the tail immersed in isotonic buffer at 37uC, recording the time until bleeding stopped. The carotid closure time was determined by placing a small piece of paper (,1 mm 2 ) soaked with 2% FeCl 3 onto the artery for two minutes and the blood flow was continuously monitored. A closure time was defined as the time needed for the blood flow be stopped at least for three minutes. recovered by passing the suspension through a discontinuous density gradient of albumin.

Megakaryocytes ploidy
TPO-differentiated megakaryocytes were labeled with a FITCconjugated anti-CD41 antibody. The cells were fixed in 70% ice-cold ethanol and stained with propidium iodide solution (50 mg/mL) in the presence of RNAse A (100 mg/mL). The ploidy distribution was determined by 2-color flow cytometry (FACS EPICS XL).
The percentage of megakaryocytes forming proplatelets was determined by differential interference contrast (DIC) microscopy using a Leica inverted microscope (objective 406/0.55; Leica Microsystems). Several photographs of the cell were taken and analyzed. Megakaryocytes were examined from 3 independent cultures.

Results
The ablation of the podxl gene in megakaryocytes resulted in no apparent phenotype and the rates of growth and fertility were normal. To assess the presence of the podxl gene, we amplified genomic DNA from megakaryocytes of either control (podxl lox/lox :: Control) or Podxl-null mice (podxl lox/lox :: Pf4-Cre). Floxed allele yielded a fragment of 1,700 bp whereas the deleted allele generated a fragment of 450 bp ( Fig. 2A). By western blot we detected a main band of ,165 KDa in lysates from control megakaryocytes, presumably native Podxl, and other immunoreactive bands of smaller size. In contrast, no immunoreactive signals were detected in lysates from Podxl-null mice (Fig. 2B, left panel). The absence of immunoreactive bands suggests that the smaller bands detected in the control could be partially glycosylated Podxl. Podxl-null megakaryocytes showed no reactivity against anti-Podxl, whereas the control cells showed a strong staining all over the plasma membrane (Fig. 2B, right panel). The pattern of phalloidin staining indicates a similar distribution of the cytoskeleton in both conditions. Hemograms showed no statistical differences between the controls and Podxl-null mice (results not shown).

Number and function of megakaryocytes
To determine the number and shape of megakaryocytes, longitudinal sections of decalcified femurs were stained with Giemsa and the number and areas of megakaryocytes from several sections were analyzed using the ImageJ software. In vitro differentiated megakaryocytes were stained with Giemsa and analyzed as described in Methods. The number of megakaryocytes per bone marrow surface area in the Podxl-null mice did not differ from the control. To assess the maturation of megakaryocytes we distributed the megakaryocytes in groups according to sizes ranging from 10 to 60 mm. The number of megakaryocytes in control and podxl-null mice showed no significant differences, neither in decalcified bone marrow sections nor in megakaryocytes differentiated in vitro (Figs. 3A,B). Differences in the size-distribution of megakaryocytes in cultured cells or in bone marrow most likely are the result of different states of maturation in each case (Figs. 3C,D).
The adherence of cultured megakaryocytes onto immobilized fibrinogen was similar in controls and Podxl-null megakaryocytes.  Transformation of megakaryocytes into proplatelets was observed in both cases (Figs. 3E,F). In agreement with these data, the polyploidy of Podxl-null cultured megakaryocytes did not differ from the control (Figs. 3G,H).

Bleeding and experimental thrombosis
In view that Podxl seems to enhance the cellular motility and cellular interactions [16,26] we studied whether the loss of Podxl would alter the platelet function. We found that bleeding time in Podxl-null mice increased by 87% over the control values, and that the carotid closure time induced by ferric chloride was almost double in Podxl-null than in control mice (Table 1). Moreover, the acute systemic thrombosis induced by (i.v.) injection of a mixture of collagen and phenylephrine produced a significantly smaller reduction in the number of circulating platelets in Podxl-null mice than in the controls ( Table 2).
The platelet content of aIIbb3, GPIb-V-IX, GPVI and integrin alpha2 chain (CD49b) receptors in Podxl-null mice was similar to that of controls (Table 3) but the aggregation of Podxl-null platelets, analyzed in PPACK-uncoagulated whole blood, showed diminished (30-40%) responses to either ADP or collagen (Fig. 4A). However, no significant differences were detected between control and Podxl-null mice in the collagen-induced aggregation of isolated platelets (Fig. 4B).
The agonist-induced activation of washed platelets, assessed by measuring the binding of the activation-sensitive antibody JON/A or the surface exposure of P-selectin, was similar in Podxl-null and control platelets (Fig. 5). Moreover, the surface exposure of phosphatidylserine, as indicated by the binding of annexin V-FITC after collagen (5 mg/mL) or thrombin (1 U/mL) activation, was also similar in Podxl-null and control platelets (results not shown).

Platelet adhesion under static conditions and under flow
Figs. 6A,B show that static adhesion stimulated by 10 mM ADP was similar in control and Podxl-null washed platelets, and that in both cases platelets showed morphological features of activation.
For adhesion experiments under flow, citrate-uncoagulated blood diluted with Tyrode's buffer was perfused in plastic parallel flow chambers coated with fibrinogen at an estimated shear stress of 5 dyne/cm 2 (Figs. 6C,F,G) or 20 dyne/cm 2 (Figs. 6D,E). The plates perfused with Podxl-null mice blood showed a statistical significant decrease in the number of aggregates.
Similar results were obtained when calcein-labeled platelets were perfused under high shear stress ($20 dyne/cm 2 ) in a parallel flow chamber coated with collagen. In five different experiments, the adhesion of Podxl-null platelets to collagen was approximately 40% of the control values (Fig. 7).

Discussion
Podocalyxin (Podxl) is a heavily sialylated and sulfated plasma membrane sialoprotein. To investigate the functional role of Podxl in platelets we have generated mice with a megakaryocyte-restricted deletion of the podxl gene. The ablation of the podxl gene was verified by PCR amplification of genomic DNA from megakaryocytes differentiated in vitro. The efficiency of Pf4-Cre in the recombination of floxed genes was further verified by measuring Podxl protein in megakaryocyte lysates. The platelet Podxl-null mice showed normal rates of growth and fertility with no apparent phenotypical traits. The blood cell counts, number of bone marrow megakaryocytes, polyploidy and proplatelet formation did not show statistical differences between the Podxl-null mice and the controls. Despite this apparent normality, the Podxl-null mice showed prolonged bleeding time, reduced platelet aggregation in response to ADP or collagen, retardation of the ferric chloride-induced closure of the carotid artery and a significant decreased in the thrombogenic response in vivo induced by (i.v.) injection of collagen and phenylephrine. Moreover, adherence and formation of aggregates onto fibrinogen or collagen-coated plates in a whole blood perfusion system was reduced in Podxl-null platelets compared to the controls.  According to these observations, the removal of Podxl from megakaryocytes altered the control of hemostasis. To note, overexpression of Podxl in megakaryocytes and platelets [27] has been recently reported to produce just the opposite effects, this is, decreased bleeding time and increased agonist-induced platelet aggregation. Thus, these observations reinforce the idea that Podxl may contribute to the control of hemostasia.
Whatever the mechanism of Podxl action, one should expect a decreased aggregation of platelets to be caused either by defective or deficient platelet receptors. However, the platelet content of fibrinogen receptor (aIIbb3) and von Willebrand factor receptor (GPIb-V-IX) in Podxl-null platelets was similar to the controls. Moreover, the agonistinduced activation of washed platelets, as indicated by surface exposure of P-selectin, binding of the activation-dependent moAb JON/A to aIIbb3, or membrane exposure of phophatidylserine, did not show differences between control and Podxl-null platelets.
Since the number of platelet receptors and their ability to respond to agonists in vitro were normal in the Podxl-null mice, then, what is causing a decreased efficiency of the hemostatic control in the absence of Podxl?. At least two simultaneous processes are involved in controlling the hemostasia: firstly, the formation of platelet clots; secondly, the formation of fibrin clots by the plasma coagulation sequence. A role for Podxl in fibrin clot formation appears to be excluded in view that Podxl-null mice showed similar plasma values of PT and aPTT than the control group.
The adhesion and aggregation of platelets and the thrombogenic occlusion of vessels can take place in mice lacking vWF and fibrinogen [28]. Thus, factors other than interaction of the main ligands, vWF and fibrinogen, with their respective platelet receptors must be involved in the formation of a platelet clot. Moreover, GPIba contributes to arterial thrombosis by mechanisms independent of its main ligand, vWF [29]. The extracellular protein fibronectin was postulated to be one of these factors, since reduced plasma fibronectin produced defective thrombus formation [30]. On the other hand, both plasma and platelet proteins seem to contribute to vWF/fibrinogen-independent platelet aggregation [31]. More than 20 cytoplasmic proteins have been reported to interact with the carboxyterminal (CT) domains of either aIIb or b3 glycoproteins [32]. Plasma membrane or transmembrane proteins like CD40L, Gas6, semaphorin 4D (CD100) or Platelet Endothelial Aggregation Receptor 1 (PEAR1), have been reported to be involved in the control of thrombus formation and/or stabilization. These antecedents highlight the redundancy of the system controlling the hemostasia and give support to the possibility that Podxl was one more of the factors controlling the hemostasia acting either directly or through interactions with other membrane proteins. It should also be considered that Podxl could influence the coagulation through interactions with glycoproteins of the coagulation cascade.
Podxl is heavily sialylated and changes in sialic acid content of platelets have been reported to alter platelet function [33,34]. Even if the loss of platelet Podxl would not change significantly the total content of sialic acid, the possibility should be considered that the absence of specific protein interactions could alter the intracellular platelet signaling. Moreover, the actions of Podxl in enhancing adhesion and motility were abolished in mutated cells in which the O-sialylation was impeded [16].
Podxl could also act in controlling aIIbb3 signaling pathway through cytoplasmic interactions. The CT domain of Podxl is linked to the cytoskeleton through the scaffold protein EBP50/NHE-RF (ezrin-binding protein/Na + H + exchange regulatory factor) [18,35] and colocalizes with vinculin [16]. On the other hand, talin, a cytosolic protein with a FERM domain is specifically involved in the activation and linking of integrin aIIbb3 to the actin cytoskeleton [35][36][37]. Talin binds vinculin with high affinity [37][38][39][40]. Vinculin and talin are two pivotal components of the focal adhesions that may indirectly activate integrins. Both, Podxl and aIIbb3, colocalize with vinculin at the leading edges of the cell and at regions of intercellular contacts [16]. Thus, it is tempting to speculate that a decrease functional efficiency of aIIbb3 in Podxl-null mice could arise from a functional imbalance of FERM and/or scaffold proteins.
Stable transfection of human Podxl in CHO cells enhances cell adhesion, motility and cellular interactions in a selectin and integrindependent manner, suggesting that Podxl could act as a co-activator of platelet aIIbb3 [16]. On high endothelial venules, Podxl is involved in lymphocyte homing through L-selectin interactions [18]. Since the Podxl effects on cell adhesion are selectin-dependent, the possibility should also be considered that Podxl-selectin interactions would be a significant factor in the control of hemostasis.
To conclude, mice with null-Podxl megakaryocytes generated by the Cre/LoxP gene targeting methodology showed perturbed hemostatic reponses: increased bleeding time, diminished aggregation of platelets in response to agonists, retarded ferric chlorideinduced closure of the carotid artery, decreased systemic thrombosis and decreased adhesion and formation of platelet aggregates under parallel flow conditions. Since Podxl-null platelets showed normal receptors content and normal activation by agonists in vitro, the perturbation of hemostasis in the absence of Podxl could be the result of a decreased efficiency of platelet adhesiveness produced by the loss of sialic acid and/or selectindependent interactions.