Alternatively spliced human tissue factor: a circulating ... - Nature

F-nm842.qxd

3/13/03

1:42 PM

Page 458

ARTICLES

Alternatively spliced human tissue factor: a circulating, soluble, thrombogenic protein VLADIMIR Y. BOGDANOV1, VIJI BALASUBRAMANIAN1, JAMES HATHCOCK1, OANA VELE1, MARK LIEB2 & YALE NEMERSON1

© 2003 Nature Publishing Group http://www.nature.com/naturemedicine

1

Division of Thrombosis Research, Department of Medicine and 2The Cardiovascular Institute, Mount Sinai School of Medicine, New York, New York, USA V.Y.B., V.B. and J.H. contributed equally to this work. Correspondence should be addressed to Y.N.; e-mail: [email protected]

Published online 10 March 2003; doi:10.1038/nm841

Tissue factor (TF) is an essential enzyme activator that forms a catalytic complex with FVIIa and initiates coagulation by activating FIX and FX, ultimately resulting in thrombin formation1. TF is found in adventitia of blood vessels2 and the lipid core of atherosclerotic plaques3. In unstable coronary syndromes, plaque rupture initiates coagulation by exposing TF to blood4,5. Biologically active TF has been detected in vessel walls and circulating blood6. Elevated intravascular TF has been reported in diverse pro-thrombotic syndromes such as myocardial infarction, sepsis, anti-phospholipid syndrome and sickle-cell disease7–10. It is unclear how TF circulates, although it may be present in pro-coagulant microparticles11,12. We now report identification of a form of human TF generated by alternative splicing. Our studies indicate that alternatively spliced human tissue factor (asHTF) contains most of the extracellular domain of TF but lacks a transmembrane domain and terminates with a unique peptide sequence. asHTF is soluble, circulates in blood, exhibits pro-coagulant activity when exposed to phospholipids, and is incorporated into thrombi. We propose that binding of asHTF to the edge of thrombi contributes to thrombus growth by creating a surface that both initiates and propagates coagulation. To examine the expression of F3, the gene encoding TF, during granulocyte differentiation, HL-60 cells were exposed to dimethyl sulfoxide (DMSO) and granulocyte colony-stimulating factor (G-CSF) to induce differentiation. RT-PCR was conducted using primers that span exons 3 and 6. In addition to the expected PCR product, a second product of lower molecular weight was noted (Fig. 1a). Subsequent RT-PCR using primers designed to amplify the entire F3 open reading frame (ORF) confirmed the existence of two distinct mRNAs (Fig. 1b). The smaller PCR product was subcloned and sequenced, revealing a previously unknown F3 ORF in which exon 5 is absent and exon 4 is spliced directly to exon 6. Because exon 6 begins with an incomplete codon and exon 4 terminates with a complete codon, such a fusion creates an ORF frameshift. The new ORF encodes alternatively spliced human TF (asHTF), whose mature peptide comprises 206 amino acids. Residues 1–166 are identical to the extracellular domain of TF, and residues 167–206 correspond to a unique C terminus. We note that the 165–166 lysine doublet involved in FVIIa binding13 is maintained in asHTF. Primary structures of TF and asHTF are shown in Fig. 1c. 458

To determine whether human tissues generate alternatively spliced F3 mRNA, we used the Basic Alignment Sequence Tool to search expressed sequence tag (EST) libraries for clones containing the last 12 bases of exon 4 fused to the first 12 bases of exon 6 (AGTTCAGGAAAG|AAATATTCTACA). Four ESTs were identified: one derived from human keratinocytes (BF149254) and three from human lung libraries (Fig. 1c). To investigate the structure of the alternatively spliced F3 mRNA 3′ end, the lung ESTs were obtained from the I.M.A.G.E. Consortium (http://image.llnl.gov) and sequenced. The analysis showed that, similar to that of F3 mRNA, the 3′ untranslated region of alternatively spliced F3 mRNA is entirely encoded by exon 6 (Fig. 1c). To further investigate the presence of alternatively spliced F3 mRNA in human tissues, a multiple-tissue cDNA panel was subjected to PCR analysis using a pair of F3-specific primers. A pair of GAPD-specific primers served as an internal control. This analysis detected alternatively spliced F3 mRNA in lung, placenta, and pancreas (Fig. 1d). Whereas levels of F3 mRNA in these three tissues were similar, levels of alternatively spliced F3 mRNA were markedly lower in placenta and pancreas than in lung tissue. Lung tissue is heavily infiltrated with monocytes derived from circulating blood14. To assess whether circulating mononuclear cells produce alternatively spliced F3 mRNA, a multiple-tissue cDNA panel for human blood fractions was subjected to PCR analysis as described above. The results showed that only CD14+ monocytes had detectable levels of F3 and alternatively spliced F3 mRNA (Fig. 1e). Further studies exploring the co-localization of asHTF and CD14 in lung tissue will be necessary to determine whether monocytes are the primary source of asHTF in lung. To evaluate plasma TF activity not associated with cells or microparticles, diluted human plasma was centrifuged at 260,000g for 4 h. This yielded an upper lipid fraction, a pellet and two distinct particulate-free fractions: a clear aqueous fraction and a lower pigmented fraction. TF in each fraction was immunocaptured using a polyclonal antibody against soluble TF6, relipidated and assayed by the conversion of FX to FXa in the presence of FVIIa and CaCl2. TF activity (per equal volume) associated with the starting material was 158 pM FXa per min (n = 5). TF activity associated with the two particulate-free fractions was 139 pM FX a per min for the clear aqueous fraction (n = 5) and 555 pM FXa per min for the NATURE MEDICINE • VOLUME 9 • NUMBER 4 • APRIL 2003

F-nm842.qxd

3/13/03

1:42 PM

Page 459


ARTICLES

a

b

d

e

c

f

Fig. 1 Identification of asHTF. a, Gel electrophoresis of RT-PCR products amplified using primers from exons 3 and 6 of human F3. RNA was isolated from undifferentiated HL-60 cells (lane 1), cells treated for 5 d with DMSO alone (lane 2) or in combination with G-CSF (lane 3), and cells treated for 8 d with DMSO alone (lane 4) or in combination with G-CSF (lane 5). Lane 6, control (no template). Lettered arrows indicate the amplified products (A, GAPD; B, F3; C, alternatively spliced F3). b, Gel analysis of RT-PCR using primers spanning the entire F3 ORF. Lane 1, HL-60 RNA (top band, F3; lower band, alternatively spliced F3); lane 2, control (no template). M, 100-bp DNA ladder in a, b, d, e. c, Schematic representation of the F3 splice variants (top), cloned alternatively spliced F3 (‘as-F3’ and ‘ORF’; GenBank accession number AF487337) and inserts from 3 alternatively spliced ESTs identified by their accession numbers. Red, leader sequence; green, extracellular domain; blue, transmembrane domain; yellow, cytoplasmic domain; orange, C-terminal region unique to asHTF; white, untranslated mRNA regions. d, Gel electrophoresis of PCR products amplified from a multiple-tissue cDNA panel using a pair of F3-specific primers. Top band, GAPD; middle band, F3; bottom band, alternatively spliced F3. C, con-

trol (no template). Lane 1, heart; lane 2, brain; lane 3, placenta; lane 4, lung; lane 5, liver; lane 6, skeletal muscle; lane 7, kidney; lane 8, pancreas. e, Gel electrophoresis of PCR products amplified from a multiple-tissue cDNA panel for human blood fractions, using a pair of F3-specific primers. Top band, GAPD; middle band, F3; bottom band, alternatively spliced F3. C, control (no template; d and e); lane 1, mononuclear cells (B and T cells and monocytes); lane 2, resting CD8 + cells (T suppressor and T cytotoxic cells); lane 3, resting CD4+ cells (T helper cells); lane 4, resting CD14+ cells (monocytes); lane 5, resting CD19+ cells (B cells); lane 6, activated CD19+ cells; lane 7, activated mononuclear cells; lane 8, activated CD4+ cells; lane 9, activated CD8+ cells. f, Characterization of asHTF-specific polyclonal antibody by western blotting. Human brain TF (lane 1), soluble TF (lane 2) and recombinant asHTF (lane 3) were probed with a monoclonal antibody against human brain TF. Recombinant asHTF (lane 6), soluble TF (lane 7) and human brain TF (lane 8) were also probed with asHTF-specific polyclonal antibody. Lane 5, empty; lane 4, molecular weight markers.

lower pigmented fraction (n = 3). TF activity was 46 pM FXa per min in the lipid fraction (n = 3) and 40 pM FXa per min in the resuspended pellet (n = 1). When platelet-poor plasma was recalcified in the presence of recombinant asHTF (1.3 µg/ml) and phospholipids (30:70 phosphatidylserine/phosphatidylcholine ratio; 15 µM), the clotting time was shortened to 150 ± 4 s from 233 ± 4 s in the presence of phospholipids alone. This result shows that asHTF is a biologically active protein. asHTF does not require actual incorporation into a phospholipid bilayer to be active, as simply combining recombinant asHTF with phospholipids generates coagulant activity. To detect asHTF antigen in human plasma specimens, we used ELISA with a commercially available monoclonal antibody against TF (capture antibody) and a digoxigenin-labeled Fab fragment of a polyclonal antibody against asHTF (detection antibody; used in western blot in Fig. 1f). The results were 18 ± 16 pg/ml from six specimens and 400 pg/ml from one specimen. The values for these specimens obtained with the commercially available TF antibody were 61 ± 59 and 672 pg/ml, respectively. These data indicate that asHTF constitutes a substantial fraction of total TF in plasma. Human thrombi were immunostained for TF and asHTF using antibodies specific for asHTF, membrane-bound TF

(C28)15 and soluble TF (TF1–218), a domain that includes epitopes common to both forms of TF. All three antibodies stained the thrombi (Fig. 2a), indicating the presence of TF and asHTF. To assess whether both TF variants are also incorporated into ex vivo thrombi, we immunostained thrombi formed by shearing whole human blood over arterial segments16. Similar to spontaneous thrombi, all three antibodies stained the ex vivo thrombi on arterial segments (Fig. 2b). To assess whether asHTF co-localizes with platelets, citratetreated whole blood was labeled with mepacrine (a platelet marker) and asHTF-specific polyclonal antibody coupled to Alexa-568, and was then recalcified and perfused over collagen-coated glass slides. Fluorescent images of thrombi on these slides showed extensive co-localization of the platelet signal (green) and the asHTF signal (red) (Fig. 3). We have identified a variant of TF, asHTF, that results from alternative splicing of the primary RNA transcript. asHTF is soluble, circulates in plasma and is biologically active. The mature asHTF peptide comprises the first 166 amino acids of TF and a unique 40-residue C-terminal domain. No such procoagulant TF variant had been previously described, although a splice variant of F3 mRNA lacking exon 5 had been observed17. The alternatively spliced F3 mRNA is only 160 bases shorter than the 2,153-base wild-type F3 mRNA, making these

NATURE MEDICINE • VOLUME 9 • NUMBER 4 • APRIL 2003

459

F-nm842.qxd

3/13/03

1:42 PM

Page 460

ARTICLES two species difficult to distinguish by northern blotting. Furthermore, recent reports have emphasized the importance of TF-bearing microparticles in circulation18. Commercially available ELISA assays designed to detect circulating TF use antibodies that may or may not react with asHTF, thus generating results that are likely to encompass different forms of blood-borne TF. This study shows that asHTF co-localizes with platelets and has pro-thrombotic activity when exposed to phospholipids. In addition, under wall shear rates well within the rates encountered in diseased coronary arteries, both TF and asHTF are recruited from whole blood to thrombi formed on arterial segments. This study also indicates that alternatively spliced F3 mRNA is present in human tissues. Alternative TF splicing is not essential for mouse development, as knockout mice lacking F3 gene and expressing low levels of the membranebound form of human TF develop to term19. However, these rescued mice have substantially shorter life spans than wild-type mice and show deposition of hemosiderin and replacement fibrosis in the heart20. These abnormalities may result from faulty hemostasis, but further study is necessary to establish their cause. Coronary artery thrombi are macroscopic structures that can occlude vessels several millimeters in diameter. In addition to its role in the initiation of thrombus formation, vesselwall TF is also assumed to be responsible for Fig. 2 Immunohistochemical analysis of spontaneous and ex vivo human thrombi. a and thrombus propagation. During arterial thromb, Specimens of spontaneously formed human thrombi (a) and pig media previously perbosis, however, exposed vessel-wall TF is fused with whole human blood (b) were processed using asHTF-specific polyclonal antibody rapidly covered by adherent platelets and fib- (pAb-asHTF), soluble TF–specific polyclonal antibody (pAb-sTF) and TF-specific monoclonal 5 rin and is thus physically separated from cir- antibody C28 (mAb-TF). Growing edge of thrombus is shown. Upper rows of a and b, pAbculating blood. A typical clotting enzyme of 50 asHTF (left), pAb-sTF–specific polyclonal antibody (middle) and non-immune rabbit IgG kD takes, on average, 3 h to diffuse 1 mm in (right) were used as primary antibodies. Lower rows of a and b, mAb-TF (left) and non-imwater at 20 °C21. Because coronary thrombi mune mouse IgG (right) were used as primary antibodies. Brown color indicates positive imhave been noted to grow much faster than munostaining. Magnification ×10. this22, it would appear that vessel-wall TF is rapidly separated from the circulating FVII, FVIIa, FIX, and FX needed to sustain thrombus growth. Thus, propagation in this model, asHTF could be considered a poit is unlikely that vessel-wall TF alone is responsible for the tential therapeutic target. In particular, reagents recognizing propagation of macroscopic thrombi. Our data suggest this its unique C terminus might provide a means to inhibit scenario for thrombus formation: when an atherosclerotic thrombus propagation under various pathological condiplaque ruptures, exposed vessel-wall TF forms a catalytic tions. In light of this potentially important therapeutic applicomplex with circulating FVII and FVII a, thereby initiating cation, further efforts to study the precise role of asHTF in coagulation. Although platelet deposition physically sepa- thrombosis are warranted. Specifically, techniques should be rates this complex from circulating blood, binding of asHTF- developed to evaluate the relative contribution of asHTF and and TF-bearing microparticles to platelets provides a renew- TF to thrombus propagation. able pro-thrombotic interface between a growing thrombus and circulating blood. When localized on platelet surfaces, Methods asHTF activity facilitates incorporation of newly formed FX a Cells. HL-60 cells (no. CCL-240, American Type Culture Collection, into prothrombinase complexes. Increases in shear rate, asso- Manassas, Virginia) were cultured in RPMI 1640 medium supplemented ciated with vessel narrowing, favor this process by promoting with 10% fetal bovine serum (Invitrogen, Grand Island, New York). Differentiation was induced by 1.3% DMSO and 50 ng/ml G-CSF (Intergen, asHTF recruitment to platelets and enhancing substrate delivPurchase, New York). ery. Ultimately, higher shear rates might limit this process, 23 just as they inhibit fibrin formation , probably by removing RT-PCR and PCR. RNA was extracted from HL-60 cells using the RNeasy kit intermediate products of the coagulation cascade. Because in- (Qiagen, Valencia, California) and analyzed by RT-PCR (Qiagen) with F3 hibition of asHTF would be expected to disrupt thrombus primers (bases 506–529, 892–913, 102–122 and 1013–1032 of F3 cDNA;


a

b

460


F-nm842.qxd

3/13/03

1:42 PM

Page 461

ARTICLES clonal antibody followed by digoxigenin-specific horseradish peroxidase–conjugated mouse antibody (BioRad). To verify specificity, we also processed each plasma sample using a non-immune rabbit IgG molecule (Jackson ImmunoResearch, West Grove, Pennsylvania). Recombinant asHTF and TF1–218 (extracellular domain of TF) served as standards.

a


b

Fig. 3 Co-localization of asHTF and platelet aggregates. Fluorescent images of platelets (green) and asHTF (red) in ex vivo thrombi generated by shearing labeled whole human blood in a parallel-plate chamber. a, asHTF-specific polyclonal antibody was used to detect asHTF protein in thrombi and b, non-immune rabbit IgG was used as a negative control to verify specificity of the signal generated by asHTF-specific polyclonal antibody. Left, phase-contrast image; middle, platelets; right, asHTF-specific polyclonal antibody. Magnification ×20.

NCBI no. NM 001993). GAPD primers served as controls. Products were separated by electrophoresis on 2% agarose, visualized with ethidium bromide under ultraviolet light, subcloned (pCR4-TOPO; Invitrogen) and sequenced. For PCR analysis of multiple-tissue cDNA panels (Clontech Laboratories, Palo Alto, California), a pair of F3 primers (bases 221–244 and 1013–1036) was used; products were separated by electrophoresis and visualized. Centrifugation of plasma. Citrate-treated blood was centrifuged at 1,800g for 15 min. The supernatant was diluted 1:1 with Tris-buffered saline and centrifuged at 260,000g for 4 h. Four layers were observed: upper lipid fraction (17% of initial volume), clear aqueous fraction (52%), lower yellow-pigmented fraction (23%) and pellet (3%). Samples of starting material and lipid, aqueous and pigmented fractions were prepared as follows. A 480-µl aliquot of sample was incubated for 1 h with 20 µl of Triton X-100 (4%), then overnight with soluble TF–specific polyclonal antibody immobilized on Affigel-10 beads (Bio-Rad Laboratories, Hercules, California). The beads were washed and eluted with 6 M guanidine HCl. Phospholipids (30:70 phosphatidylserine/phosphatidylcholine ratio; 70 µM; Avanti Polar Lipids, Alabaster, Alabama) and n-octyl-β-D-glucopyranoside (100 mM; Calbiochem, La Jolla, California) were added and samples were dialyzed. TF activity assay. Samples were assayed in a reaction mixture containing FX (100 nM), FVIIa (10 nM) and CaCl2 (5 mM). Aliquots of the resulting mixture were combined with EDTA (25 mM). FXa substrate (Spectrozyme Xa; 0.5 mM; American Diagnostica, Greenwich, Connecticut) was added at intervals and absorption was measured at 405 nm. TF inhibitors FVIIai (10 nM) and polyclonal antibody against asHTF (40 µg/ml) were used to verify specificity of the measured activity. Recombinant asHTF. Alternatively spliced F3 ORF was amplified by RTPCR, cloned into pTriEx-3 Neo expression vector (Novagen, Milwaukee, Wisconsin), sequenced and expressed in Escherichia coli. Recombinant asHTF, with a (His)8 tag sequence at the C terminus, was purified from the soluble fraction of E. coli lysates and verified by western blotting and Coomassie staining (data not shown). Protein concentration was determined by Bradford microassay (Bio-Rad, Hercules, California). ELISA. TF antigen was captured using a pre-coated 96-well plate (American Diagnostica) and probed with digoxigenin-labeled asHTF-specific polyNATURE MEDICINE • VOLUME 9 • NUMBER 4 • APRIL 2003

Antibodies. Monoclonal antibody specific for the cytoplasmic domain of TF (C28)15 and polyclonal antibody specific for the extracellular domain of TF24 are well characterized and have been used to identify TF in numerous experiments. The polyclonal antibody specific for asHTF was generated in rabbits using a peptide comprising the last 11 amino acids of asHTF, with an N-terminal Cys (CEWGRAGRRTPH) conjugated to keyhole limpet hemocyanin (KLH), and purified using a protein-G column followed by peptide–Affigel-10 column. On western blots, the asHTF-specific polyclonal antibody recognized recombinant asHTF but not soluble or membrane-bound TF (Fig. 1f). In addition, recombinant asHTF completely blocked immunostaining by the asHTF-specific polyclonal antibody but not by C28 or the soluble TF–specific antibody (data not shown).

Immunostaining. Specimens of spontaneous human thrombi were processed as previously described24. Paraffin-embedded, paraformaldehyde-fixed specimens were deparaffinized, washed with PBS, incubated in 4% H2O2, and immunostained with C28 (140 µg/ml) and polyclonal antibodies against soluble TF (140 µg/ml) and asHTF (70 µg/ml). Mouse IgG2B against KLH (140 µg/ml, R&D Systems, Minneapolis, Minnesota) and whole-molecule rabbit IgG (140 µg/ml, Jackson ImmunoResearch) served as negative controls. After incubation with primary antibodies, specimens were washed, incubated with appropriate biotinylated secondary antibodies and peroxidase-streptavidin (BioGenex, San Ramon, California), developed with diaminobenzidine (Dako Corporation, Carpinteria, California) and counterstained with hematoxylin. To ensure the consistency of obtained results, every specimen type was immunostained as described in at least 4 separate experiments. Institutional Review Board approval for the use of human specimens was granted.

Labeling of blood and perfusion experiments. Mepacrine (10 µM, Sigma, St. Louis, Missouri) was used as a platelet marker. Alexa-568 (Molecular Probes, Eugene, Oregon) was used to label the asHTF-specific polyclonal antibody (20 µg/ml). ChromPure whole-molecule rabbit IgG (20 µg/ml, Jackson ImmunoResearch) served as a negative control. Citratetreated blood was incubated with markers for 30 min at 37 °C and recalcified with 5 mM CaCl2. Collagen-coated cover slips were placed in a parallel-plate Grabowski chamber25. Labeled, re-calcified blood was perfused at 100 s-1 (wall shear rate) until the chamber was occluded.

Acknowledgments We thank M.B. Taubman, M.C. Moschella, and J.T. Fallon for discussions and advice, S. Carson for monoclonal antibody against membrane-bound TF, E. Grabowski for parallel-plate chamber, J.J. Badimon for specimens of ex vivo thrombi, and H. Vaananen for technical assistance. V.Y.B., V.B., J.H., O.V., M.L. and Y.N. are partially supported by the National Institutes of Health. Competing interests statement The authors declare that they have no competing financial interests.

RECEIVED 24 SEPTEMBER 2002; ACCEPTED 14 FEBRUARY 2003 1. Bach, R.R. Initiation of coagulation by tissue factor. CRC Crit. Rev. Biochem. 23, 339–368 (1988). 2. Nemerson, Y. & Gentry, R. An ordered addition, essential activation model of the tissue factor pathway of coagulation: evidence for a conformational cage.

461

F-nm842.qxd

3/13/03

1:42 PM

Page 462


ARTICLES Biochemistry 25, 4020–4033 (1986). 3. Zaman, A.G., Helft, G., Worthley, S.G. & Badimon, J.J. The role of plaque rupture and thrombosis in coronary artery disease. Atherosclerosis 149, 251–266 (2000). 4. Zeldis, S.M., Nemerson, Y., Pitlick, F.A. & Lentz, T.L. Tissue factor (thromboplastin): localization to plasma membranes by peroxidase-conjugated antibodies. Science 175, 766–768 (1972). 5. Baumgartner, H.R., Turitto, V.T. & Weiss, H.J. Effects of shear rate on platelet interaction with subendothelium in citrated and native blood. II. Relationships among platelet adhesion, thrombus dimensions, and fibrin formation. J. Lab. Clin. Med. 95, 208–221 (1980). 6. Giesen, P.L. et al. Blood-borne tissue factor: another view of thrombosis. Proc. Natl. Acad. Sci. USA 96, 2311–2315 (1999). 7. Suefuji, H. et al. Increased plasma tissue factor levels in acute myocardial infarction. Am. Heart J. 134, 253–259 (1997). 8. Gando, S., Nanzaki, S., Sasaki, S., Aoi, K. & Kemmotsu, O. Activation of the extrinsic coagulation pathway in patients with severe sepsis and septic shock. Crit. Care Med. 26, 2005–2009 (1998). 9. Amengual, O., Atsumi, T., Khamashta, M.A. & Hughes, G.R. The role of the tissue factor pathway in the hypercoagulable state in patients with the antiphospholipid syndrome. Thromb. Haemost. 79, 276–281 (1998). 10. Key, N.S. et al. Whole blood tissue factor procoagulant activity is elevated in patients with sickle cell disease. Blood 91, 4216–4223 (1998). 11. Mallat, Z. et al. Shed membrane microparticles with procoagulant potential in human atherosclerotic plaques: a role for apoptosis in plaque thrombogenicity. Circulation 99, 348–353 (1999). 12. Mallat, Z. et al. Elevated levels of shed membrane microparticles with procoagulant potential in the peripheral circulating blood of patients with acute coronary syndromes. Circulation 101, 841–843 (2000). 13. Kelley, R.F., Costas, K.E., O’Connell, M.P. & Lazarus, R.A. Analysis of the factor VIIa binding site on human tissue factor: effects of tissue factor mutations on the kinetics and thermodynamics of binding. Biochemistry 34, 10383–10392 (1995). 14. Rosseau, S. et al. Monocyte migration through the alveolar epithelial barrier: adhe-

462

sion molecule mechanisms and impact of chemokines. J. Immunol. 164, 427–435 (2000). 15. Carson, S. & Yoder, S. Monoclonal antibodies against the C-terminal peptide of human tissue factor for studies of the cytoplasmic domain. Blood Coagul. Fibrinolysis 3, 779–787 (1992). 16. Badimon, L., Badimon, J.J., Galvez, A., Chesebro, J.H. & Fuster, V. Influence of arterial damage and wall shear rate on platelet deposition. Ex vivo study in a swine model. Arteriosclerosis 6, 312–320 (1986). 17. Guo, W. et al. Effect of all-trans retinoic acid and arsenic trioxide on tissue factor expression in acute promyelocytic leukemia cells. Chin. Med. J. (Engl.) 114, 30–34 (2001). 18. Berckmans, R.J. et al. Cell-derived microparticles circulate in healthy humans and support low grade thrombin generation. Thromb. Haemost. 85, 639–646 (2001). 19. Parry, G.C., Erlich, J.H., Carmeliet, P., Luther, T. & Mackman, N. Low levels of tissue factor are compatible with development and hemostasis in mice. J. Clin. Invest. 101, 560–569 (1998). 20. Pawlinski, R. et al. Tissue factor deficiency causes cardiac fibrosis and left ventricular dysfunction. Proc. Natl. Acad. Sci. USA 99, 15333–15338 (2002). 21. Berg, H.C. Diffusion to capture. In Random Walks in Biology Chap. 3, 37–47 (Princeton University Press, Princeton, New Jersey, 1993). 22. Ambrose, J.A., Almeida, O.D., Sharma, S.K., Dangas, G. & Ratner, D.E. Angiographic evolution of intracoronary thrombus and dissection following percutaneous transluminal coronary angioplasty (the Thrombolysis and Angioplasty in Unstable Angina [TAUSA] trial). Am. J. Cardiol. 79, 559–563 (1997). 23. Turitto, V.T. & Hall, C.L. Mechanical factors affecting hemostasis and thrombosis. Thromb. Res. 92, S25–S31 (1998). 24. Thiruvikraman, S.V. et al. In situ localization of tissue factor in human atherosclerotic plaques by binding of digoxigenin-labeled factors VIIa and X. Lab. Invest. 75, 451–461 (1996). 25. Grabowski, E.F. Platelet aggregation in flowing blood at a site of injury to an endothelial cell monolayer: quantitation and real-time imaging with the TAB monoclonal antibody. Blood 75, 390–398 (1990).
