Factor XI Messenger RNA in Human Platelets

The human liver FXI cDNA was a gift from Dr Dominic Chung (University of Washington, Seattle, WA). TaKaRa long range DNA polymerase for PCR was from Pan Vera Corp (Madison, WI). A Thermo-sequenase radiolabeled terminator sequencing system for DNA sequencing was from Amersham Life Sciences (Arlington Heights, IL). Guanidinium isothiocyanate extraction buffer was from Promega Corporation (Madison, WI). A Ready-To-Go T-Primed First-strand Kit for RT of RNA was from Pharmacia Biotech (Piscataway, NJ). Omniscript RT for PCR with or without RT (Qiagen, Valencia, CA). Oligonucleotide primers were prepared by the DNA core facility of the Vanderbilt University Cancer Center. SeaKem LE agarose and NuSieve agarose were from FMC Bioproducts (Rockland, ME).

The MEG-01 human megakaryoblastic cell line (ATCC CRL-2021)21 and the human erythroleukemia cells line HEL 92.1.7 (ATCC TIB 175)22 were from the American Tissue Type Collection (Rockville, MD). The CHRF-288-11 human megakaryoblastic cell line23 was a gift from Dr Michael Lieberman (University of Cincinnati, Cincinnati, OH). The 293 human fetal kidney fibroblast cell line transfected with a full-length FXI cDNA derived from liver was described previously.24 25 RPMI 1640 medium was from Mediatech (Herndon, VA), and Dulbecco’s modified Eagle medium (DMEM) and Fischer’s media were from Life Technologies (Grand Island, NY).

The procedure for obtaining human blood specimens was approved by the Institutional Review Board of Vanderbilt University. Platelets were isolated either from peripheral blood or expired platelet-pheresis products from the Vanderbilt Hospital blood bank by differential centrifugation. Briefly, blood (120 mL) was obtained from the antecubital vein of healthy normal individuals into a 1/10th volume of 3.8% sodium citrate anticoagulant. Anticoagulated blood or the platelet-pheresis product was placed in 50 mL polypropylene centrifuge tubes and erythrocytes and leukocytes were pelleted by centrifugation at 300g for 10 minutes in a Sorval RC5 Superspeed centrifuge fitted with an SS-34 rotor (Sorval, Inc, Wilmington, DE). The top half of the platelet rich supernatant was gently aspirated and placed into fresh centrifuge tubes using polypropylene pipettes, and centrifugation was repeated as above. The top half of the platelet-rich plasma was aspirated from the tube and used as the source of platelets for RNA purification. The platelet count was determined using a Technicon H-3 hematology analyzer (Bayer-Miles, Tarrytown, NY). A typical platelet preparation from fresh blood contains 5 to 10 × 10⁹ platelets and from a pheresis product 1 to 5 × 1011 platelets. Samples of purified platelets were stained with Wright’s stain and examined under a microscope. Rare leukocytes or erythrocytes (<1 per 10,000 platelets) were detected in some platelet preparations, whereas others had no detectable cells. Platelets were pelleted at 3,500g for 20 minutes, dissolved in guanidinium thiocyanate extraction buffer (5 to 10 × 10⁹ platelets/mL of guanidinium thiocyanate), and stored at 4°C pending RNA isolation.

Leukocytes were purified as follows: Twenty mL of peripheral blood collected into a 1/10th volume of 3.8% sodium citrate was diluted 1:1 with RPMI 1640 medium. Eight mL of the diluted blood was layered onto 4 mL of Ficoll Hypaque (Lymphoprep Ficoll Hypaque, Nycomed Pharma AS, Oslo) in a 15 mL tube followed by centrifugation at 800g for 30 minutes in a DPR-6000 centrifuge (Damon/International Equipment Co Division, Needham Heights, MA). The pellet (erythrocytes and granulocytes) and the intermediate layer halfway down the Ficoll Hypaque (monocytes and lymphocytes) were carefully collected and subjected to a second purification on Ficoll Hypaque as above. The erythrocyte/granulocyte fraction (1 mL in volume) was brought up to 15 mL with 150 mmol/L NH₄Cl, 10 mmol/L KHCO₃ to lyse red cells, then centrifuged at 1,000g for 5 minutes in a Beckman TJ-6 bench-top centrifuge to pellet the granulocytes. The monocyte/lymphocyte fraction underwent centrifugation in a similar manner. Microscopic evaluation of the granulocyte and mononuclear cell preparations showed contamination with some platelets (<1 per 10 leukocytes) with the contamination being somewhat greater for the mononuclear cell preparations than the granulocytes. The majority of platelets do not sediment through Ficoll Hypaque and remain in the upper layer. Leukocytes were dissolved in guanidinium thiocyanate extraction buffer (1 × 10⁶ cells/0.2 mL extraction buffer). Normal human bone marrow, 5 mL, was obtained from a donor harvest for bone marrow transplantation. The cells were pelleted by centrifugation at 3,000g for 10 minutes and the pellet was dissolved in 2 mL of guanidinium thiocyanate extraction buffer. All cell lysates were stored at 4°C pending RNA purification.

Cell lines were maintained in a 37°C fully humidified incubator with a 5% CO₂ atmosphere. MEG-01 was grown in RPMI 1640 containing 10% heat-inactivated fetal calf serum (FCS), 2 mmol/L L-glutamine, 1 mmol/L sodium pyruvate, 1.5 gm/L sodium bicarbonate, 10 mmol/L HEPES, and penicillin/streptomycin/amphotericin B. HEL 92.1.7 cells were grown in RPMI 1640 containing 10% heat-inactivated FCS, 2 mmol/L L-glutamine, and penicillin/streptomycin. CHRF-288-11 was grown in Fischer’s medium supplemented with 10% heat-inactivated FCS and penicillin/streptomycin. Cells were grown to a concentration of approximately 1 × 10⁷/20 mL media, then washed four times with phosphate-buffered saline, pelleted by centrifugation at 1,000g for 10 minutes, and dissolved in 2 mL of guanidinium thiocyanate extraction buffer. Cells that were adherent to the culture flasks were released by incubation with trypsin/EDTA and then treated as above. Extracts were stored at 4°C pending RNA purification. 293 cells transfected with a wild-type FXI liver cDNA24 25 were grown in DMEM 5% FCS, 2 mmol/L L-glutamine, 1 mmol/L sodium pyruvate, and penicillin/streptomycin/amphotericin B. These cells, which served as a positive control for the RT/PCR process, were processed into guanidinium thiocyanate extraction buffer as above.

Total RNA from platelets and cell lines was isolated using standard guanidinum thiocyantate/phenol:chloroform extraction and precipitated with 70% ethanol.26 Leukocyte, bone marrow, and cell line poly-A RNA were isolated from the guanidinium isothyocyante extractions using a PolyA Tract System 1000 (Promega) according to the manufacturer’s instructions. PolyA RNA was eluted from the system using nuclease-free water. A Ready-To-Go T-Primed first-strand RT kit (Pharmacia) was used for RT reactions. Briefly, 5 μg of total RNA (platelets) or 100 to 1,000 ng of poly A RNA (leukocytes, bone marrow, or cell lines) was mixed with oligo-dT and random hexanucleotides as recommended by the manufacturer. Samples were heated at 65°C for 5 minutes, placed on ice for 2 minutes, and then added to tubes containing the Ready-To-Go beads. After 1 minute of incubation at room temperature, samples were mixed, transferred to a 37°C water bath, and incubated for 2 hours. To show that signals obtained from PCR reactions were RT dependent and, therefore, not due to contamination of RNA preparations with genomic or cDNA, RNA underwent RT in the presence or absence of RT as follows: 2 μg of RNA was added to reaction mixtures containing 1× buffer, 0.5 mmol/L of each deoxyribonucleoside (dNTPs), 0.5 μmol/L of random hexanucleotides, and 10 units of RNase inhibitor, and incubated at 37°C for 2 hours either in the presence or absence of 4 units of Omniscript RT. Reactions were then cooled on ice and 8 to 10 μL was used as template in a 50 μL PCR reaction.

Oligonucleotide primer pairs for PCR (Table1) were designed using OSP software27 and published sequences for human FXI,18,28 glycoprotein IIb (GPIIb),29 and CD-18.30 Three sets of primers were prepared for FXI to amplify sequences represented by exons 3 through 6 of the FXI heavy chain, exons 10-14 from the carboxy-terminus of the heavy chain and catalytic light chain, and exons 3-15 covering the entire sequence encoding the mature protein (Fig 1). An additional pair of primers amplifying parts of exon 10 and 11 were prepared (Table 1) for use in identifying a common polymorphism in exon 1131 known to be present in one of the platelet donors (see below). PCR reactions for GPIIb, CD-18, and exons 3-6, 10-14, and 10-11 of FXI were amplified in reactions containing 1× TaKaRa reaction buffer, 1.5 mmol/L magnesium chloride, 100 μmol/L of each dNTPs, 10 pmol of each primer, 2.5 units of TaKaRa LA Taq polymerase, and 3.0 μL of RT mRNA in a total volume of 50.0 μL. PCR was performed on a Perkin Elmer model 460 DNA Thermal Cycler (Foster City, CA). The following cycling parameters were used: one cycle of denaturation at 94°C for 3 minutes, annealing at 60°C for 40 seconds and extension at 70°C for 1 minute, followed by 38 cycles of denaturation at 94°C for 30 seconds, annealing at 60°C for 30 seconds, and extension at 72°C for 1 minute. Long-range PCR to amplify the 1.6 kb long fragment covering FXI exons 3-15 was performed in a total volume of 50 μL containing 1× TaKaRa reaction buffer, 1.5 mmol/L magnesium chloride, 400 μmol/L of each deoxyribonucleoside (dNTP), 10 pmol of each primer, 2.5 units of TaKaRa LA Taq polymerase, and 3.0 μL of RT mRNA. PCR was performed on a PTC-100 Thermal Cycler (MJ Research) using the following cycling conditions: an initial cycle of denaturation at 94°C for 3 minutes followed by 38 cycles of denaturation at 94°C for 40 seconds and annealing/extension at 66°C for 2 minutes. Small PCR fragments were size fractionated by electrophoresis on 2% NuSieve agarose gels. Long-range PCR fragments were separated on 1% SeaKem LE agarose gels.

The PCR fragments representing FXI exons 3-6 from normal platelets, bone marrow, 293 control cells, MEG-01, HEL, and CHRF-288-11 RNA were subjected to dideoxy-chain termination sequencing using a Thermo Sequenase radiolabeled terminator cycle sequencing kit according to the manufacturers recommendations. Sequencing of the exon 10-11 product from the platelet RNA of an individual known to carry a polymorphism in exon 11 was performed in a similar manner. Fragments were sequenced in both directions using the oligonucleotide primers originally used to produce the PCR fragments. Reaction products were run on standard 7% polyacrylamide gels followed by autoradiography overnight. Sequences were compared with the published sequence for the human FXI cDNA.18 

The human FXI gene contains 15 exons. Exon 1 is not translated, whereas exon 2 encodes a signal peptide that is not present on the mature molecule.28 Exons 3-10 encode the FXI heavy chain and exons 11-15 encode the catalytic light chain.28 Three sets of PCR primers were used to amplify portions of the FXI message from RT RNA (Fig 1). The expected PCR products generated from the primer pairs, based on the published cDNA sequence for FXI18 are: 400 bp for exons 3-6, 552 bp for exons 10-14, and 1,673 bp for exons 3-15. PCR products of the expected sizes were generated when RT RNA from the 293 cell line transfected with a liver FXI cDNA (control cells) was used as template (Fig 2A, lane 1). Human pancreas has previously been showed by Northern blot analysis to express FXI mRNA.25 PCR products similar to those from the control cell line were obtained when a human pancreas cDNA library was used as template for all three sets of PCR primers (data not shown). When the PCR template was RT RNA from normal human platelets, single products were obtained that were identical in size to those from reactions with the 293 control cells (Fig 2A). Note, in Fig 2A the control sample in lane one for the exon 3-15 PCR product appears very heavy. This is due to overexposure of the film to show the PCR products from platelets, which are faint. The exon 3-15 PCR product represents a nearly full-length FXI message. As platelet RNA appears to be partially degraded on ethidium-stained agarose gels (data not shown), it is likely that there are very few full-length FXI messages in platelet RNA, accounting for the weakness of this PCR product. Peripheral blood platelets are anucleate cell fragments lacking the machinery for RNA synthesis (except perhaps for mitochondrial RNA). RNA in platelets, therefore, must be produced by megakaryocytes. We identified FXI message by RT-PCR in RNA from normal human bone marrow (Fig3, lane 1). Whereas the cell of origin of the message is not determined by this experiment, the results are consistent with a megakaryocytic origin for the FXI message identified in platelets. The positive result with bone marrow RNA could have been due to platelets within the bone marrow sample. However, the bone marrow RNA was prepared from a small sample (5 mL), and in our experience there is insufficient platelet RNA in this volume to generate a positive signal by RT-PCR. The DNA sequences of the exon 3-6 PCR fragments from platelets, bone marrow, pancreas, and the 293 control cells were identical to the published sequence for the human FXI cDNA.18 

A recent report describes a truncated FXI message lacking exon 5 in the megakaryocytic leukemia cell line CHRF-288-11.20 The absence of this exon in platelet or bone marrow mRNA should have resulted in PCR products of 241 bp for the exon 3-6 primer pair and 1,514 bp for the exon 3-15 primers. We did not observe either of these smaller species in any fresh platelet preparation (three separate donors, Figs 2A and 3 and data not shown) or in platelet-pheresis products from our blood bank (three separate donors, Figs 2A and 3).

Cell lines MEG-01,21 HEL 92.1.7,22 and CHRF-288-1123 are derived from tissues of patients with acute nonlymphocytic leukemia. These cell lines have been shown to possess features and express proteins typically found in normal megakaryocytes. An FXI cDNA lacking exon 5 and a small part of the 3′-portion of exon 4 has been isolated from a cDNA library prepared from CHRF-288-11 RNA, and the MEG-01 line is reported to express an FXI mRNA of 1.9 kb (the message in liver is 2.1 kb) detectable by Northern blot analysis.20 RNA was isolated from cells and analyzed by RT-PCR with the exon 3-6 primer pair specific for the FXI heavy chain (Fig 2B). PCR products identical to the one from the control cell line were obtained with RNA from all three cell lines. No product that could represent a truncated FXI mRNA was identified in the cell lines. DNA sequences of the exon 3-6 PCR products for all cell lines were identical to the published sequence for the human FXI cDNA from liver.18 

A major concern when isolating RNA from peripheral blood platelets is contamination of the platelet preparation with small numbers of leukocytes that contain relatively large amounts of RNA compared to platelets. Although the platelet isolation technique used in this study produced several platelet preparations with no apparent leukocyte contamination by visual microscopic inspection, some leukocyte contribution to all platelet RNA preparations could not be ruled out. To address this potential problem, RNA was isolated and reverse transcribed from granulocytes and from mononuclear leukocytes (a combination of lymphocytes and monocytes) and along with platelet RNA was analyzed by PCR for leukocyte- and platelet-specific markers, in addition to FXI. The results of a representative experiment are shown in Fig 3. All platelet preparations and bone marrow were positive for GPIIb. As suspected, one of six platelet preparations gave a positive signal for CD-18, a component of the leukocyte marker LFA-1,30 indicating white cell contamination of the platelet preparation (data not shown). Both granulocytes and mononuclear cells were consistently positive for CD18. In addition, some preparations of mononuclear cells were positive for GPIIb, which is typically only found on platelets and megakaryocytes.32This is consistent with the platelet contamination noted in the leukocyte preparation. Most importantly, granulocytes and mononuclear cells did not give a positive signal for FXI. It is our impression that platelets contain very small amounts of FXI mRNA, and the low level of platelet contamination of the mononuclear cells would not provide sufficient platelet RNA to detect the FXI message. Although low levels of leukocyte expression of FXI mRNA can not be ruled out by this analysis, the results show that platelets are the major source of FXI mRNA in peripheral blood.

An obvious concern when using PCR to detect mRNA is the generation of a false-positive signal due to contamination with genomic or cDNA. Analysis by a technique that does not involve amplification steps, such as Northern blot or RNAse protection analysis can overcome this limitation. However, we were unable to show FXI message by these techniques in any of the RNA preparations in this study except for the transfected 293 cells that overexpress the message (data not shown). The amount of FXI message, therefore, is below the level of detection for these types of assays. In our experiments, PCR primer pairs were selected to amplify fragments with more than one exon, eliminating the possibility of genomic DNA being responsible for the signals. Control PCR reactions in which the template was RNA that had not been reverse transcribed were performed for all RNA preparations (Fig4). These reactions did not generate PCR products indicating that RNA preparations were not contaminated with factor XI cDNA, and that PCR products in reactions using RT RNA as template were dependent on the RT reaction. This shows that the template for the PCR signals is, indeed, mRNA.

Finally, several polymorphisms that do not result in amino acid changes have been reported in the coding sequence for the human FXI gene.31 One of these polymorphisms, a T to C change at base pair 1234 within exon 11, was known to be present in one copy of the FXI gene from one of our platelet donors. A PCR primer pair that amplifies portions of exons 10 and 11 (Table 1) was used to generate a 216 bp fragment from the platelet RNA of this donor, and the base pair sequence of the PCR product was determined (Fig5). The sequence shows both a T and a C residue at base pair 1234, consistent with the known heterozygous state for this person. This result could not have been created by genomic DNA contamination of the sample, as the PCR product involves parts of two exons separated by several kilobases of intron sequence in the gene.28 The pattern is highly unlikely to be caused by cDNA contamination because two separate sequences are present in the PCR product and no cDNAs containing the T to C polymorphism have been created in our laboratory. These data also confirm that platelet RNA is the template responsible for the PCR products in our experiments.

The bleeding diathesis associated with FXI deficiency has been perplexing since the first description of FXI-deficient patients in 1953.33 Congenital deficiency of FXI is a relatively rare disorder with a particularly high incidence in people of Ashkenazi Jewish descent.34 Excessive bleeding is typically triggered by trauma or surgical procedures, however, hemorrhage does not correlate well with plasma FXI levels determined by contact activation initiated clotting assays.7,8 In a seminal study, Asakai et al5 presented an explanation for some of the variation when they noted that hemorrhage depends not only on the FXI level in plasma, but also on the tissue involved. Bleeding from the oral cavity and urinary tract are particularly common, likely due to the high levels of fibrinolytic activity in these tissues. Nevertheless, there are severely deficient patients who do not bleed with surgical procedures, suggesting an alternative procoagulant activity may be present in some individuals that substitutes for plasma FXI. It has been proposed that the platelets of some people contain an FXI-like activity that may modify bleeding in the absence of plasma FXI.14 15 

Platelet α-granules, contain numerous proteins involved in hemostasis and fibrinolysis that are also present in plasma.35 The source of granule proteins appears to vary. The vWF 36 and the A chain of factor XIII37 are apparently of megakaryocyte origin, as megakaryocytes express mRNA for these proteins. In contrast, megakaryocytes do not express mRNA for albumin or the α and β chains of fibrinogen, indicating these proteins are taken up into granules from the plasma.38 Our results show FXI mRNA in human platelets and bone marrow, as well as three cell lines with megakaryocyte features. Furthermore, platelets are the major source of FXI mRNA in peripheral blood, because leukocyte RNA does not contain FXI message. The amount of FXI message in platelets and the leukemia cell lines appears to be small. Nevertheless, the data suggest FXI similar to the protein produced by the liver may be present in platelets, and could be released at sites of platelet activation to increase local concentrations of this clotting factor at a wound site. It should be noted that it is difficult to postulate a role for a platelet pool of FXI in the variable bleeding observed in FXI deficiency, because this protein would presumably be affected by the same mutation that produced the plasma deficiency state.

The presence of FXI mRNA in platelets does not provide definitive proof that FXI protein is present. The issue of whether or not platelets contain FXI protein, and the form this protein takes, is a controversial one. Evidence for platelet FXI comes from three major lines of investigation: (1) assays of FXI activity, (2) antibody-based immunologic studies of FXI antigen, and (3) molecular biology investigations of FXI mRNA. Well-washed platelets contain small amounts of activity that shorten the abnormally long clotting time of FXI-deficient plasma in an aPTT assay.13,39,40 The nature of this activity is difficult to determine by this type of assay, however, as complex platelet extracts contain factors that interact with the clotting cascade at many levels. Indeed, Schiffman et al41,42 and Osterud et al43 made this point when they examined platelet extracts for FXI activity and concluded no appreciable activity was present.

Using immunoprecipitation techniques with polyclonal antihuman FXI antibodies, three groups have reported on the partial purification of a 220 kD protein from platelets by immunoprecipitation.14-16The protein appears to be on the platelet surface, because it copurifies with the membrane fraction of sonicated platelets.13 The migration of this protein through SDS-polyacrylamide gels is different from plasma FXI, a 160 kD homodimer comprised of identical disulfide-bond–linked 80 kD polypeptides.17 Under reducing conditions the platelet protein migrates as a single 50 to 55 kD protein implying that the 220 kD protein is a polymer of a smaller subunit. The protein shares some features with plasma FXI, in addition to cross-reactivity with antibodies. Like plasma FXI, the platelet protein is cleaved by factor XIIa, and the cleaved form appears to cleave the synthetic substrate S-2366, indicating it is a protease.15 In contrast, unlike plasma FXI the platelet protein is not activated by trypsin, its proposed catalytic domain is substantially larger than that of factor XIa, and it has not been shown that it has FXIa activity (ie, activates factor IX). Although this protein may eventually be shown to possess FXI activity, it is difficult to invoke it at present as an explanation for variable bleeding in FXI deficiency because it is not known if some people have the protein while others lack it.

Using flow cytometry, Hu et al44 detected binding of anti-FXI antibodies to platelets from normal and FXI-deficient patients, a finding supported by earlier studies.14 These authors propose that the binding site for the antibody is a FXI-like molecule encoded by a modified message from the FXI gene that may circumvent mutations responsible for plasma FXI deficiency. The findings are consistent with this hypothesis, however, other possibilities must be considered. The simplest explanation is that the antibody is cross-reacting with a protein that may have homology to FXI, but is not a product of the FXI gene. In this circumstance it would not be expected to be affected by a mutation causing deficiency of plasma FXI. Alternatively, the antibodies could bind to normal FXI in normal platelets, and to mutant FXI molecules in platelets from FXI-deficient patients. A common point mutation associated with FXI deficiency in Ashkenazi Jews is a phenylalanine to leucine substitution at amino acid 283.5 The mutation interferes with normal dimer formation of FXI, resulting in reduced secretion of the protein into plasma.45 Factor XI associated with platelets may be stored in granules and not be secreted, however, and abnormal proteins could be retained within the platelets. Taken as a whole, the data obtained with FXI antibodies indicate that platelets contain a protein that is a protease and that has some homology to plasma FXI. The range of activities of this protein and its role in hemostasis, if any, remain to be determined.

Using PCR techniques, Hsu et al have recently provided evidence for an alternatively spliced mRNA from the FXI gene,20 and postulate that it codes for the 50 to 55 kD subunit of the platelet protein identified by FXI antibodies. With platelet RNA as template, each exon of the FXI gene could be amplified individually by RT/PCR except for exon 5, which encodes a portion of the apple two domain of the FXI heavy chain. This group also isolated a novel FXI cDNA lacking exon 5 and the 3′-terminal bp of exon 4 from a leukemia cell line cDNA library.20 This cell line, CHFR-288-11, shows megakaryocytic features including expression of platelet factor 4, vWF, and glycoprotein IIb/IIIa, as well as induction of multinucleation and hyperploidy upon stimulation with phorbol ester.23 We were interested to determine if the novel FXI message identified in the cDNA library of this cell line, is present in normal human blood tissues and in freshly prepared RNA from megakaryocyte cell lines.

We did not find evidence for truncated FXI mRNA in normal platelets or in the three megakaryoblast cell lines tested, including CHRF-288-11. All PCR products representing the FXI heavy chain contain normal sequence for exon 5. The results with RT-PCR cannot exclude the presence of low levels of alternatively spliced message, however, the data indicate the predominant message for FXI in platelets and cell lines is identical to the message produced in liver. Generation of FXI mRNA lacking exon 5 that would maintain an open reading frame would require an unusual mRNA splicing event that would ignore nearly perfect consensus donor splice junctions at the 3′-end of exons 4 and 5.28 Our studies failed to identify this species in normal blood cells. It is possible that some individuals may have a polymorphism in the FXI gene that could allow this splicing event to occur, and that we did not identify it in the platelets of the six individuals we tested. However, this possibility would not explain the discrepancy between our data with CHRF-288-11 mRNA and the previously published results of the novel alternatively spliced FXI message from a library prepared from this same cell line.20 

In summary, our results show FXI mRNA in platelets, suggesting platelets may contain FXI of megakaryocytic origin that is similar to plasma FXI produced in the liver. Although platelets appear to contain a membrane protein that cross-reacts with FXI antibodies, our data do not support the premise that this protein is a product of an alternatively spliced message from the FXI gene. This is supported by the failure of mutations in the FXI gene causing plasma FXI deficiency to influence the presence of the platelet membrane protein.44 

The authors are grateful to Dr Jerry Ware (Scripps Inst, LaJolla, CA) for providing cDNA libraries and information regarding HEL and CHRF-288-11 cells, and to Dr Michael Lieberman (University of Cincinnati, Cincinnati, OH) for providing the CHRF-288-11 cell line. We thank Dr Mortimer Poncz (University of Pennsylvania, Philadelphia, PA) for recommendations concerning leukocyte and platelet specific markers, and Dr George J. Broze, Jr (Washington University, St. Louis, MO) for his thoughtful reading of the manuscript. We also wish to thank Mr Mao-Fu Sun for expert technical assistance.