UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Factor VIIa levels in prothrombotic conditions Fong, Whalley K. 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1997-0411.pdf [ 6.36MB ]
JSON: 831-1.0087969.json
JSON-LD: 831-1.0087969-ld.json
RDF/XML (Pretty): 831-1.0087969-rdf.xml
RDF/JSON: 831-1.0087969-rdf.json
Turtle: 831-1.0087969-turtle.txt
N-Triples: 831-1.0087969-rdf-ntriples.txt
Original Record: 831-1.0087969-source.json
Full Text

Full Text

FACTOR Vila LEVELS IN PROTHROMBOTIC CONDITIONS By WHALLEY K. FONG B.Sc, Simon Fraser University, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Biochemistry and Molecular Biology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 1997 © Whalley K. Fong, 1997 In, presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I. further agree that permission for extensive copying of this thesis for scholariy purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^tocAi^'sh^ /^OJICMJ^A ^frv/c^ij The University of British Columbia Vancouver, Canada Date DE-6 (2/88) A B S T R A C T The coagulation cascade is continuously in a state of preparedness to rapidly respond to vascular damage at sites of injury. However, an overactivity of the hemostasis system resulting in a hypercoagulable state can lead to thrombotic disorders such as coronary heart disease (CHD) and stroke. Recent prospective studies have implicated elevated factor VII (FVII) activity as an independent risk factor in incidences of coronary heart disease. Indeed, factor VII and its cofactor, tissue factor (TF), are generally thought to play the principle role in the physiological initiation of clotting which may help explain the association between FVII and the risk of CHD. Recent evidence has indicated that trace levels of the activated form of FVII, called factor Vila (FVlIa), circulates in the blood. This active form is unique among the clotting factors in that FVlIa has a long half life of 2 to 2.5 hours, is highly resistant to inhibitors, and can auto-activate zymogen FVII. However, conventional assays for FVIT activity employing tliromboplastin were not able to separate FVIIa from its zymogen precursor, FVII. This has resulted in a great deal of controversy over whether plasma FVIIa levels per se or total FVII (zymogen and FVIIa) is more closely associated with the risk of CHD. In the present study, a specific assay for factor Vila was developed using a recombinant soluble form of tissue factor (sTF), deleted of the membrane binding region, that is deficient in promoting the conversion of factor VII to factor Vila. However, this soluble tissue factor, in the presence of phospholipid vesicles, retains its ability to clot plasma using the pre-activated factor Vila in the plasma sample. The sTF was engineered by using site-directed mutagenesis to delete the transmembrane domain of the TF gene and the truncated protein was expressed in a eucaryotic expression system employing baby hamster kidney (BHK) cells. Using the unique properties of sTF, an automated clot-based assay was developed and optimized to quantitatively determine the concentration of FVIIa in plasma samples. The assay was employed to measure FVIIa levels in a normal, healthy population and in a group of patients in the late stages of complicated pregnancies. ii The FVUa assay developed in the present study had a sensitivity range of 20 pg/ml to 1000 ng/ml FVHa and required only 5 ul of plasma per assay. The within-run and between-run coefficients of variation were 3.9% and 4.9%, respectively, indicating high reproducibility. The assay also showed excellent correlation (r = 0.998) with a commercial FVUa kit produced by Diagnostica Stago in a patient comparison study over a range of 0.1 ng/ml to 30 ng/ml FVUa. The present study was standardized using an international FVUa standard and therefore FVUa levels could be measured in concentration (ng/ml) or international activity units (mU/ml). Results of the initial population study on FVTJa confirmed the existence of trace amounts of FVJJa in normal healthy subjects. The normal range in the present study was 0.7 to 4.8 ng/ml (11 to 100 mU/ml) and the mean plasma FVUa was 2.08 ± 0.90 ng/ml (42 ± 15 mU/ml). There was a significant correlation between FVTJa levels and increasing age but no significant difference between FVIIa levels in men and women. Factor Vila levels in complicated pregnancy patients were significantly higher than in an age-matched group of healthy women. The mean FVJJa in the patients was 4.1 ± 2.2 ng/ml (85 ± 42 mU/ml) compared to 1.86 ng/ml (37 mU/ml) in healthy women. However, more extensive clinical studies on FVIIa and FVH in other groups at risk of CHD are required to better establish the importance of FVUa levels in prothrombotic conditions. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vii LIST OF FIGURES viii LIST OF ABBREVIATIONS x ACKNOWLEDGMENTS xiii INTRODUCTION 1. Hemostasis 1 2. The Coagulation Cascade 3 3. Characteristics of Tissue Factor 4 4. Characteristics of Factor VH/VIIa 7 5. Thrombosis 8 6. Traditional Assays for FVTJ Activity 10 7. Soluble Tissue Factor and the Factor VTIa-specific Assay 11 8. Previous Work by Others on FVIIa-specific Assays 14 9. Clinical studies using FVIIa Assay 15 10. Objectives of the Present Study 17 MATERIALS 19 METHODS 1. Polymerase Chain Reaction of sTF PCR amplification of sTF 21 Agarose gel electrophoresis 22 GENECLEAN DNA Purification 22 2. Cloning and DNA sequence analysis of PCR fragment Cloning of PCR fragment into pBS 23 Transformation of E.coli cells 23 Mini-prep recovery of plasmid DNA 24 DNA Sequence Analysis 25 3. Transfection of BHK Cells and Expression of sTF Cloning of sTF into pNUT 26 Transfection of BHK cells 27 iv SDS-PAGE 27 Western blot analysis 28 Tissue culture of sTF-BHK cell line 29 4. Purification of sTF from Tissue Culture media Ammonium Sulfate fractionation 30 Anion-exchange Fast Protein Liquid Chromatography 30 Gel-filtration purification 31 5. Characterization of purified sTF 32 6. Development of Factor VTJa Assay Preparation of phospholipid vesicles 32 Optimization of sTF-Reagent 32 Assay for Factor VTIa 33 Factor Vila Standard Curve 33 Comparison of BHK vs. £.c0/i*-derived sTF 34 7. Testing of FVIIa Assay Precision Testing 34 Accuracy Testing 34 Plasma Stability studies 35 Individual variations 35 8. Factor Vila in Normal Populations Ethical approval for blood collection 36 Subjects 36 Phlebotomy 36 Measurement of FVIIa 36 9. Factor Vila in Complicated Pregnancies Ethical approval 36 Subjects 37 Clinical assessment 37 RESULTS 1. Polymerase Chain Reaction of sTF 38 2. DNA sequence analysis of sTF insert 38 3. Expression of sTF in BHK cells 40 4. Ammonium sulfate fractionation 44 5. High Q FPLC purification 46 6. Sephacryl S-100 HR Gel Filtration purification 49 v 7. Characterization of sTF protein 8. Optimization of assay reagents 9. Factor VJJa Standard curve 10. Comparison of BHK-derived sTF and £.co//-derived sTF 11. Testing of FVTJa assay 12. Plasma FVIIa stability testing 13. Intra-individual variations in plasma FVTIa 14. Factor Vila levels in a Normal population 15. Factor Vila levels in Complicated Pregnancy patients DISCUSSION 1. BHK-derived glycosylated sTF 76 2. Factor Vila assay 77 3. Comparison to previous FVTIa assays 79 4. Handling of plasma samples 80 5. Factor Vila in normal populations: effects of age and gender 81 6. Factor Vila in prothrombotic conditions: Complicated pregnancies 83 7. Future work 85 8. Possible mechanisms of FVTJa involvement in thrombotic disorders 87 9. Implications for treatment of thrombotic disorders 88 REFERENCES 53 55 59 62 62 66 69 70 73 vi LIST OF TABLES I. Effects of freezing on plasma FVIIa levels 66 II. Comparison of EDTA vs. citrate as plasma anticoagulant 68 HI. Intra-individual variation in FVIIa levels 69 IV. Age distribution and FVIIa levels in study sub-populations 72 V. Clinical details of complicated pregnancy patients 73 VI. Hemostatic variables of complicated pregnancy samples 74 Via. Hemostatic variables for new complicated pregnancy samples 75 Vii . FVHa levels in normal populations from several studies 81 vii LIST OF FIGURES 1. Coagulation Cascade 2 2. Human Tissue Factor 4 3. X-ray crystal structure of extracellular domain of Tissue Factor 5 4. Dual cofactor function of Tissue Factor 6 5. Human Factor VII 7 6. Factor VII Coagulant (FVILC) Assay 10 7. Factor Vila Assay 12 8. Human TF gene and sTF gene 13 9. Oligonucleotide primers for PCR of sTFi_2i9 21 10. DNA Sequence Analysis Primers 26 11. PCR amplification of sTF gene 39 12. Not I serening of Bluescript-sTF clone 41 13. Not I and EcoRl digestion screening of pNUT-sTF clone 41 14. pNUT expression vector with cloned sTF insert 42 15. Western Blot of tissue culture media from several sTF-BHK cell lines 43 16. 55% Ammonium Sulfate fractionation 45 17. High Q FPLC purification ' 47 18. Calibration curve for Sephacryl S-100 HR Column 50 19. Sephacryl S-100 HR purification 51 20. Purification Summary gel 54 21. Clot time vs. sTF concentration 57 22. Clot time vs. Phospholipid concentration 58 23. FVUa standard curve using Sigma FVII-def plasma 59 24. FVIIa standard curve using George King FVII-def plasma 60 25. Final FVIIa Standard Curve 61 26. Comparison of BHK and E.coii-derived sTF in FVIIa Assay 63 viii 27. Patient comparison study results 64 28. Relation between Factor Vila and Age 71 29. Factor Vila Conversion Curve 79 ix LIST OF ABBREVIATIONS A Adenosine A Angstrom Amp Ampicillin APS Ammonium persulfate API i Activated Partial Thromboplastin Time BHK Baby Hamster Kidney Bis N,N'-methylenebisacrylamide BSA Bovine serum albumin C Cytosine cDNA Complementary DNA CHD Coronary Heart Disease CHO Chinese Hamster Ovary C-terminal Carboxy-terminal C. V. Coefficient of variation DEAE Diethylaminoethyl DHFR Di-hydrofolate reductase DMEM Dulbecco's modified essential medium DMF Dimethylformamide DNA Deoxyribonucleic acid dNTP Deoxyribonucleotide triphosphate EDTA Ethylenediamine tetraacetic acid EGF Epidermal Growth Factor ELISA Enzyme-Linked Immunosorbent Assay FBS Fetal bovine serum FTBR Fibrinogen FPLC Fast Protein Liquid Chromatography FVII Factor VII FVIIa Activated Factor VII or Factor Vila FVUa-TF Factor VHa - Tissue Factor FVII: Ag Factor VII Antigen assay FV1I:C Factor VII Coagulant activity assay FIX Factor DC FX Factor X FXa Activated Factor X or Factor Xa x G Guanosine Gla Gamma carboxylated glutamic acid HEPES N-2-Hydroxyethylpif^ razine-N'-2-ethane sulfonic acid HR High Resolution HRT Hormone replacement therapy IgG Immunoglobulin G IHD Ischaemic Heart Disease IPTG Isopropyl-p^D-Thiogalactopyranoside kb Kilobase kDa KiloDaltons LB Luria-Bertani (media) LUV Large unilamellar vesicles MI Myocardial infarction MTX Methotrexate MT-1 Zinc metallo-protein promoter N/A Not available NPHS Northwick Park Heart Study NPP Normal pooled plasma N-terminal Ammo-terminal OD Optical density PAGE Polyacrylamide gel electrophoresis PBS Phosphate Buffered Saline solution pBS Bluescript (vector) PC Phosphatidylcholine PCR Polymerase chain reaction PE Phosphatidylemanolamine PL Phospholipid PROCAM Prospective Cardiovascular Muenster Study PS Phosphatidylserine psi Pounds per square inch PT Prothrombin Time r Regression coefficient (linear regression analysis) RBC Rabbit Brain Cephalin rpm Revolutions per minute SDS Sodium dodecyl sulfate STA Stago xi sTF Soluble Tissue Factor T Thymine Taq Thermophius aquaticus TAE Tris-base, acetic acid, EDTA buffer TBS Tris-buffered saline solution TBE Tris-base, borate, EDTA buffer TE Tris-base, EDTA buffer TEMED N, N, N', N'-Tetiamethylethylenediamine TF Tissue Factor TFPI Tissue Factor Pathway Inhibitor Tris Tri (hydroxymethyl) aminomethane U Units of enzyme activity USG Ultroser G u v Ultra violet v/v Volume-to-volume ratio w/v Weight-to-volume ratio X-Gal 5-Bromo-4-chloro-3-indolyl-p,-D-galactopyranoside xii A C K N O W L E D G M E N T S I was fortunate enough to work alongside an extraordinarily talented group of researchers without whom the fruits of this work would not have emerged. I owe special thanks to J.P. Heale and Allyson Davey who performed the preliminary stages of the project described in this report. As well, I wish to acknowledge the contributions of Dr. Dana Devine, Dr. Haydn Pritchard, Jeff Hewitt, Alexis Maxwell, Bea Tam, Mark Brown, Leisa Stenberg, Gord Rintoul, Michael Murphy, and Hung Vo for their technical advice and expertise. I wish to thank Dr. Louis Wadsworth and Dr. Cedric Carter for their help with the clinical study, Dr. James Morrissey for the soluble tissue factor and the original FVTIa assay, and Linda Spiller who was Ms. Everything when it came to blood. To my supervisor, Ross MacGillivray, I cannot express my full gratitude for the support and generosity you extended to me both as a scientist and as a friend. I also must thank my mother and father, my brothers, Ian, Glance, and Glenn, and my sister, Karling, for their love and support from the day I was born. Finally, I thank my wife, Naomi Uranishi. You were the light in the darkness that showed me the way. xiii INTRODUCTION 1. Hemostasis Hemostasis, the physiological arrest of bleeding at sites of vascular injury, is one of the most vital protective mechanisms that preserves the internal environment of vertebrates (1). Not only does hemostasis prevent excessive bleeding but it also blocks the portal of entry for invasive organisms (2). The complex mechanism of hemostasis involves the interaction of three events: a transitory constriction of the damaged blood vessels at the site of injury, the formation of a platelet plug, and the coagulation of blood (3). The first two events, comprising the "primary hemostatic" response (4), form a soft, weak mechanical plug that can stem the bleeding from minor injuries (1). However, hemorrhage from more formidable wounds requires the "secondary phase" of hemostasis, blood coagulation. The coagulation or clotting of blood results in the formation of a tough fibrin clot that provides a scaffold for a mechanical plug to prevent blood loss while also allowing for endothelial cell repair and the subsequent release of the fibrin clot by the fibrinolytic system (4). The process of coagulation involves a group of blood-borne clotting proteins that circulate as inactive zymogens. The system responds to tissue damage or other stimuli by initiating a series of proteolytic activation reactions that lead to a cascade or waterfall of activated clotting factors resulting in the rapid formation of a fibrin clot (Figure 1). Under normal conditions, the hemostasis system provides an elegant and efficient system that confers protection against vascular trauma. However, dysfunction of the hemostatic system can lead to several different and potentially life-threatening pathological states. A deficiency in the hemostatic system caused by the absence or decreased activity of just one of the clotting factors can lead to a bleeding disorder with an associated inability of the blood to clot normally. The two most 1 INTRINSIC PATHWAY Factor XII Anionic Surface Kininogen _ Prekalllkrien EXTRINSIC PATHWAY 0 ^ -YTissue Damage I Tissue Factor. Factor VII Factor X COMMON PATHWAY Feedback Loop Fibrin Factor XHIa Cross-linked Fibrin Clot Figure 1. Coagulation Cascade. A simplified diagram illustrating the separation into the two classic pathways, the intrinsic and extrinsic pathways, that converge at factor X into the Common pathway that eventually leads to the formation of a fibrin clot. common of these clotting disorders are hemophilia A and hemophilia B which are caused by deficiencies in factors VIII and LX, respectively. Rare deficiencies in factors V, VII, X, XI, XII, XIH, prothrombin, and fibrinogen as well as multiple factor deficiencies have also been described (5). On the other extreme is an overactivity of the hemostatic system possibly caused by increased levels or overactivity of one or more of the clotting factors. The resulting hypercoagulable state could lead to thrombotic disorders such as coronary heart disease, acute myocardial infarction, and strokes. Recent studies have provided strong evidence for 2 a link between increased levels of certain clotting factors and an increased risk of thrombotic events (6, 7, 8, 9). Thus, the hemostatic system must be tightly regulated to maintain the very delicate balance required for the health of the individual. 2. The Coagulation Cascade As seen in Figure 1, the clotting cascade has been traditionally separated into two systems, the intrinsic and extrinsic pathways, based on the location of the initiating factors in each pathway. The mtrinsic pathway (so named because all of the components are found in plasma) is initiated by contact of blood with anionic molecules such as kininogen or kallikrien. The extrinsic pathway is initiated by the release of tissue factor (TF) from the damaged tissue (located outside of the vasculature and thus extrinsic). The two series of proteolytic activation reactions converge with the activation of factor X (FX) and then share a common pathway leading to the formation of a cross-linked fibrin clot. Calcium ions and phospholipid membranes are essential for the proteolytic activity of the clotting factors (10). In reality, there are many interactions between the two pathways and feedback loops that regulate blood coagulation. For example, thrombin can feedback and activate factor XI as shown in Figure 1. Other studies have shown that the extrinsic factors, factor VII (FVII) and tissue factor (TF), play important roles in the activation of the intrinsic system via activation of factor IX (FIX) in addition to directly activating factor X of the common pathway (11, 12). Thus it is generally thought that the extrinsic pathway is the primary hemostatic response to tissue damage with the factor Vila-tissue factor (FVUa-TF) complex serving as the physiological initiator of blood coagulation (13, 14). Because of this principal role in coagulation, the FVIIa-TF complex has become a prime target for the investigation of thrombotic disorders. 3 3. Characteristics of Tissue Factor Human TF, an essential cofactor of FVIIa, is a ubiquitous membrane bound glycoprotein of apparent molecular weight 45-47 kDa (15,16) on SDS-PAGE. A diagram of TF is shown in Figure 2. Figure 2. Human Tissue Factor. The cofactor of FVIIa consists of a large extracellular domain followed by a single pass transmembrane domain and a short cytoplasmic domain. There are three glycosylation sites and two disulfide bonds in the extracellular domain. TF is expressed as a 295-residue polypeptide which is processed to remove a 32-residue leader peptide. The mature protein, consisting of a 219 amino acid extracellular domain anchored to the cell membrane by a 23-residue transmembrane domain (17), is present on subendothelial components of the vessel wall and has also been localized to^ atherosclerotic plaques (18,19). TF is glycosylated via three N-linked glycosylation sites at Asn-13, Asn-126, and Asn-139 but glycosylation does not appear to be necessary for function (17). 4 Excluding post-translational modification, TF has a predicted molecular weight of 29.5 kDa (15). TF also contains two pairs of cysteine residues on its surface domain that form covalently stable disulfide bonds that are required for co-factor function (17). The extracellular domain of TF (residues 1-219) has been crystallized following partial subtilisin digestion and the X-ray structure solved to 2.4 A (20) resolution (Figure 3). Figure 3. X-ray crystal structure of extracellular domain of Tissue Factor. The extracellular domain consists of two fibronectin-like domains end-to-end forming an extended structure. The normal site of membrane attachment is located at the bottom of the figure. Courtesy of (20). The model shows that the extracellular portion of TF adopts an extended form consisting of two fibronectin-like domains oriented end-to-end (length = 115 A, diameter = 40-50 A) (21). Experiments with TF-gene knockout mice have shown that tissue factor is essential for the development of the vasculature during embryogenesis and thus mutations in the TF-5 gene are lethal (22). The absence of any reported homozygous mutations in the human TF-gene is consistent with TF being essential for life. Under normal conditions, TF is sequestered from the vasculature. However, upon tissue damage, TF is exposed to the blood and forms a complex with FVII. This triggers the activation of FVII to FVIIa and thus initiates the coagulation cascade (Figure 1). TF actually fulfills two essential cofactor functions for FVIIa. First, TF serves as the cofactor during activation of FVII to FVIIa and second, formation of a complex with TF significantly increases (up to 16,000-fold) the serine protease activity of FVIIa on its substrates (FX and FLX) (23) as shown in Figure 4. Figure 4. Dual cofactor function of Tissue Factor. Tissue factor (TF) (a) serves as a cofactor during activation of factor VII (FVII) zymogen to factor Vila (FVIIa) and (b) serves as a cofactor for FVIIa serine protease activity on factor X (FX). PL = Phospholipid surface, Ca 2 + = Calcium ions, FXa = factor Xa. 6 Recently, trace amounts of plasma TF (172±135 pg/ml) have been detected by a very sensitive ELISA method (24) and these low levels of TF may be important for the generation and activity of trace amounts of pre-activated FVIIa in the plasma. 4. Characteristics of Factor VLWIIa Factor Vila belongs to a family of plasma proteins called the Vitamin K-dependent serine proteases which also includes factors FXa, Xa, and Protein Ca. They share a common protein structure composed of an N-terminal Gla-domain, followed by two epidermal growth factor (EGF)-like domains, and a C-terminal serine protease catalytic domain (Figure 5). Figure 5. Human Factor V n . Factor VII, the zymogen precursor to the Vitamin In-dependent serine protease factor Vila, has several characteristic structural motifs including an N-terminal Gla domain, two epidermal growth factor like domains, and a C-terminal serine protease catalytic domain. Cleavage of the Argl52-Ilel53 bond activates factor VII to factor Vila. Courtesy of (90). 7 Factor VII is secreted by the liver as a single chain zymogen of molecular weight 50 kDa and is activated by proteolytic cleavage of the Argl52-Ilel53 bond to form FVIIa. The resulting two polypeptide chains are joined by a disulfide bond between Cys-135 and Cys-262 in activated FVIIa (25). In vitro, this activation can be catalyzed by FXa, FIXa, FXIIa, and thrombin and requires the presence of TF as a cofactor. Recently, studies have shown that FVIIa can also autoactivate itself (14) which is an essential property for an initiation factor. However, it is still not well understood which of the proteases is most responsible for FVU activation in vivo. In addition, factor Vila has several other unique properties that make it a good candidate for the physiological initiator of clotting. Unlike the other serine protease clotting factors, FVIIa is highly resistant to plasma protease inhibitors. As well, once activated most serine proteases have a half-life of a few seconds to minutes, while FVIIa has an in vivo half-life of 2-2.5 hours (14, 26). Recent studies by several groups have confirmed that approximately 0.5 - 1% of factor VII circulates in plasma in the pre-activated FVIIa-form (26, 27) and these trace levels of FVIIa may serve a "priming" function in the coagulation cascade. Small amounts of FVIIa-TF complexes may be involved in the autocatalytic activation of more FVIJ-TF and/or in the generation of an initial burst of FXa that could back-activate TF-bound FVII during coagulation. If FVIIa indeed fulfills this role as the physiological initiator of clotting, then elevated plasma FVIIa levels may contribute to hypercoagulable states leading to thrombotic conditions (14). 5. Thrombosis An overactivity of the hemostasis system can lead to a hypercoagulable state resulting in an increase in risk of thrombotic disorders. The Northwick Park Heart Study (NPHS) (6,7) first implicated factor VII as a potential risk factor for coronary heart disease (CHD). This large scale, long term prospective study found that high levels of FVII 8 coagulant activity (FVII:C) were associated with an increased risk of CHD, especially for events within 5 years of recruitment. During the 5 year period, FVII:C was a better independent predictor of CHD than even cholesterol levels (7). The subsequent follow-up study at Northwick Park found that FVILC was more strongly associated with fatal events of ischemic heart disease (IHD) than non-fatal events. An increase of one standard deviation in FVILC was found to increase the probability of dying of IHD by 44% (28). Several other studies have corroborated this relationship between factor VII coagulant activity and CHD risk. The PROCAM (Prospective Cardiovascular Muenster Study) study re-affirmed FVILC as a strong independent risk factor for fatal coronary events (8). A number of other groups have shown significant correlations between FVILC and serum cholesterol (29, 30), triglyceride levels (31, 32), CHD risk scores (33, 34), diabetes (35, 36), late pregnancy (37), and high fat diets (38, 39, 40, 41). The localization of tissue factor to atherosclerotic plaques (18) provides insight into a possible mechanism for the relationship between extrinsic activation and CHD events. The rupture of an atherosclerotic plaque is known to be an integral event preceding occlusive thrombosis and acute myocardial infarction (MI) (19,42, 43). The exposure of the plaque's necrotic core containing TF to the vasculature following rupture provides the source of thrombogenecity needed for the thrombotic event (18). Ruddock and Meade (28) have suggested that the severity of the occluding thrombus depends on the activation state of the extrinsic factors. Specifically, high levels of FVII activity may facilitate more extensive initiation of clotting following exposure of TF after atherosclerotic plaque rupture (28). 9 6. Traditional Assays for FVII Activity The Northwick Park Heart Study and most other studies on hemostatic variables and CHD risk factors have measured FVII activity using the standard factor VU Coagulant activity (FVILC) assay. This assay employs thromboplastin, a heterogeneous mixture of the organic extracts of emulsified mammalian brain or lung tissue (44), whose active clotting agent is tissue factor. Because of TF's ability to act as both a cofactor for the activation of FVII to FVIIa and as a cofactor for FVIIa proteolytic activity (Figure 4), FVILC actually measures an aggregate of both FVII zymogen and FVIIa levels in the plasma (Figure 6). TF FVIIs Plasma Sample Factor X TF Factor Xa Clotting Time Figure 6. Factor VII Coagulant (FVIItC) Assay. Because of the dual function of tissue factor (TF), the FVILC assay measures an aggregate of both FVU and FVUa levels. In addition, variations in the configuration and reagents used for the FVILC assay between laboratories are thought to affect its relative sensitivity to FVII vs. FVIIa (45). Because of a lack of standardization of the FVILC assay, the results from various studies may implicate different aspects of FVII activity or form. For example, the NPHS FVILC assay was found to be considerably more sensitive to plasma FVIIa levels than the assay used in the PROCAM study (45). This was caused by the virtual absence of protein C in 10 the substrate plasma used in the NPHS assay. As well, differences were found in the source of tissue thromboplastin used. Human brain, rabbit brain, human placental, and bovine brain thromboplastin are commonly used in different laboratories and it has been shown that thromboplastin from bovine sources is most sensitive to FVIIa levels (45). A third problem arises from the practice of measuring FVILC as a relative value with 100% activity defined as the FVILC level for a pooled plasma sample. However, FVILC activity of pooled plasma samples may differ from one population to another (27). There are other assays that measure various aspects of factor VII activity and quantity. The FVII antigen (FVILAg) assay is an ELISA method that measures the total amount of FVII (FVII and FVIIa). The FVII "coupled" amidolytic assay first activates all of the FVII using thromboplastin and the FVIIa, in turn, generates FXa which cleaves a chromogenic substrate (46). A ratio of FVILC and either of these two measures of total FVII was often used as an estimate of the "activation state" of FVII. As well, more general clotting assays such as prothrombin time (PT), which also uses thromboplastin, can be used to measure the activity of the extrinsic system (44). However, because most of these assays also use thromboplastin, they suffer the same problems as the FVILC test. 7. Soluble Tissue Factor and the Factor Vila-specific Assay Because of these innate problems with the standard FVILC assay, there has been a great deal of controversy over the results of the NPHS and other such studies concerning whether plasma FVIIa levels per se or total FVII (zymogen plus FVIIa) is more closely associated with the risk of heart disease. Although providing a relative measure of the amount of pre-activated FVIIa, the FVILC assay cannot separate previously activated FVIIa from FVII activated by thromboplastin in the clotting assay because of the dual function of TF (Figure 6). Thus, the FVILC assay does not provide an accurate quantitative measure of plasma FVIIa levels due to the interference of zymogenic factor VII. An assay that is specific for the measurement of factor Vila and insensitive to zymogenic FVH is required to 11 be able to assess the roles of FVIIa versus total FVII independently as risk factors for thrombotic disorders. Recently Dr. Morrissey at the Oklahoma Medical Research Foundation developed a soluble form of tissue factor (sTF), deleted of the membrane binding region, that is deficient in promoting the conversion of factor VII to factor Vila. However, this truncated TF, in the presence of phospholipid vesicles and calcium ions, retains its ability to clot plasma using the pre-activated factor Vila in the plasma sample (47). This unique partial functionality of sTF can be exploited to develop a clot-based assay with extreme sensitivity for factor Vila (Figure 7). Factor X sTF Factor Xa 1 Clotting Time Figure 7. Factor Vila Assay. Since soluble tissue factor (sTF) cannot activate FVII to FVIIa, only pre-activated plasma FVIIa contributes to the clotting assay. Thus, the FVIIa assay can directly measure FVIIa levels without interference from zymogenic FVII as in the FVII coagulant assay. Such an assay could be used to quantitatively determine the FVIIa levels in plasma without interference from activation of zymogenic FVII. Human TF is encoded by a 12.4 kb gene comprised of 6 exons and 5 introns (Figure 8a) (48) located on chromosome 1 (49). Exon 1 codes for the translation initiation 12 FVIIa Plasma Sample domain and a 32-residue leader sequence that includes a signal peptide which is subsequently cleaved. Exons 2-5 code for the 219-residue extracellular domain of the tissue factor protein and exon 6 codes for the 23-residue transmembrane domain, the 21-residue cytoplasmic domain, and the 3'-untranslated region (17). As shown in Figure 8b, soluble tissue factor was produced by using recombinant DNA techniques to delete exon 6 of the TF gene by adding a stop codon at the end of exon 5 (50). Thus the recombinant sTF protein consisted of only the 219-residue extracellular domain of native TF after processing. (a) T Translation Initiation Domain (123 bp) Y Extracellular Domain (700 bp) Tissue Factor (295 residues) Transmembrane & Cytoplasmic Domain (1278 bp) (b) T Translation Initiation Y Extracellular Domain J Soluble Tissue Factor (219 residues) Figure 8 (a) Human TF gene, (b) sTF gene, (a) Human TF gene is composed of 6 exons: exon 1 codes for translation initiation domain, exons 2-5 code for the extracellular domain, and exon 6 codes for the transmembrane and cytoplasmic domains, (b) Deletion of exon 6 results in soluble tissue factor. Using the unique properties of sTF, Morrissey et al developed a clotting time based assay specific for factor Vila (26,51). The method was a modification of the FVILC assay (Figure 6) with an sTF-reagent replacing thromboplastin as the active clotting agent. Since sTF cannot activate zymogenic FVII, the level of extrinsic pathway initiated clotting was 13 determined only by the endogenous plasma FVIIa level (Figure 7). FVIIa levels could then be quantitated by plotting a standard curve of FVIIa concentrations versus clotting times. Using this assay, it was possible to directly measure the levels of pre-activated FVIIa circulating in the blood without interference from zymogen FVII. 8. Previous Work by Others on FVIIa-specific assays The FVIIa assay developed by Morrissey et al (26) was a clot-based assay dependent on the limited procoagulant properties of sTF. Recently, Kario et al (27) modified this sTF-based technique to develop a fluorogenic assay for FVIIa. Their assay defined the reaction end point as the time elapsed to reach a relative fluorescence of 0.05 during hydrolysis of a fluorogenic thrombin substrate. The fluorogenic assay used less reagents and exhibited high reproducibility but was less sensitive (0.2 - 1000 ng/ml) than Morrissey's assay (0.01 - 10,000 ng/ml) (26). Another assay specific for FVIIa was recently described by Philoppou et al (25). Unlike Morrissey and Kario's functional assays, this new assay avoids using sTF by employing an ELISA method. Philoppou et al were able to prepare a FVIIa-specific antibody raised against the region of FVII exposed following its activation by cleavage at Argl52-Ilel53. The FVIIa ELISA showed excellent correlation with the functional assay over a range of 0.05 - 18 ng/ml FVIIa but measured plasma FVIIa levels differed significantly from the sTF-based assays. Another problem with Philoppou's assay is that ELISAs are innately time-consuming and labor-intensive requiring prolonged blocking and antibody incubation times. An important consideration in the development of any assay is the ability to compare results between laboratories and studies. This was a major obstacle in the interpretation of studies using the FVILC assay. An sTF-based FVIIa assay will inherently have lower variability because the reagents are more strictly defined. It is a quantitative assay that measures the absolute amounts of protein based on the activity of FVIIa. Thus, 14 the determined FVIIa value is completely dependent on the FVIIa standard curve and therefore the FVIIa calibration standards. A potential problem resides in the fact that there are several commercial suppliers of FVIIa (Novo Nordisk, Enzyme Research, etc) as well as various in-house preparations, each of which may vary in their specific activity despite having the same mass/volume concentration. This can be due to differences in the purity of the FVIIa or loss of activity due to instability of the FVIIa. Thus, the National Institute of Biological Standards (in the U.K.) has proposed the adoption of an International factor Vila Standard based on specific activity (52). 9. Clinical studies using FVIIa assay Since the development of the original FVUa assay in 1993 by Dr. Morrissey, there has been a considerable number of clinical studies taking advantage of the new assay to improve our understanding of the role of factor Vila in prothrombotic conditions. Several studies confirmed the presence of circulating FVIIa in plasma at concentrations of 3.6 ng/ml (26), 2.5 ng/ml (27), and 4.1 ng/ml (53) in various populations corresponding to approximately 0.5 - 1% of the total plasma FVII. In all cases, there was considerable variation in FVIIa levels between individuals; "normal" FVUa ranges were as wide as 0.4 -14.3 ng/ml (53). The studies also showed a general trend towards increasing FVIIa levels with age (26, 27, 53). Studies have also examined the relationship between FVIIa levels and other CHD risk factors. There has been mixed results with regards to an association between plasma FVUa and elevated lipid levels. Despite the well known correlation of FVILC and FVII: Ag with serum triglycerides (31, 32), several groups have shown that plasma FVIIa levels were not correlated with serum triglycerides at all (40, 53, 54). However, studies have consistently shown that FVIIa levels are transiently increased in the post-prandial hypertriglyceridemic state following a high fat meal (38, 39,40,41,55). Studies on total cholesterol and FVJJa have provided more conflicting results. Some groups have found no 15 significant correlations with cholesterol (27, 56) while others have shown a significant relationship (53). As well, different studies using various methods to assess CHD risk have come to differing conclusions. One group found that factor Vila levels were not elevated in men who had suffered a myocardial infarction at a young age (56). However, Kario et al (27, 54) did find that elderly patients with arterial cardiovascular diseases showed an increase in plasma FVUa levels. Other groups have found that FVUa increased with a CHD-risk score based on a number of CHD-risk factors in a larger general population (33, 53). In several other potentially prothrombotic conditions, FVIIa levels have also been shown to be elevated. Increased FVIIa levels were associated with non-insulin dependent diabetes (35, 57), late pregnancy (26), post-menopausal women (53), cancer patients (58), oral contraceptive-users (59), and patients suffering from juvenille arthritis (60). It is has been postulated that some of these conditions result in endothelial tissue damage which may activate the extrinsic pathway through release of subendothelial TF at sub-thrombotic levels thereby increasing FVIIa levels. On the other hand, FVIIa levels were generally not found to be associated with smoking (26) or male gender (26, 27), factors that usually predispose CHD. As well, FVUa levels were found to be decreased in patients treated with oral anticoagulant therapy (26) and hormone replacement therapy (53). There also appears to be a genetic link to FVIIa levels. Healthy subjects who are heterozygous for a common polymorphism of the FVII gene called Mspl M2, a single G-to-A substitution changing an Arg to a Gin at residue-353, have approximately 50% lower FVIIa levels than people homozygous for the Mspl M l allele (61). Some evidence suggests that the Mspl M2 allele confers a cardioprotective effect. Thus, there is still great potential to clarify our understanding of the role of factor Vila in coronary heart disease and other potentially thrombotic conditions using the FVUa 16 assay. With such knowledge, more efficacious therapeutics can be developed to treat these diseases which take the lives of nearly half a million North Americans every year. 10. Objectives of the Present Study The purpose of the present study was to develop a fibrin clot-based FVIIa assay using sTF in Vancouver for use in clinical studies. Recombinant soluble tissue factor was produced by using mutagenic PCR oligonucleotide primers to delete the transmembrane and cytoplasmic domains of the TF cDNA, cloning the truncated TF gene into a pNUT expression vector, and expressing the sTF protein in a eucaryotic expression system employing baby hamster kidney (BHK) cells. The FVIIa-specific assay was developed based on the method of Morrissey (26, 51). However, the present study's FVIIa assay differs from Morrissey's assay in several respects. The original assay employed £.co//-derived sTF and rabbit brain cephalin (RBC) as a phospholipid source. The assay in the present study uses a BHK-derived glycosylated form of sTF and a pure phospholipid mixture in place of RBC and studies were performed to optimize the reagents and conditions. These changes could conceivably increase the activity of the sTF-reagent and thus improve the efficiency of the FVIIa assay and provide better accuracy and reproducibility. A comparison study of the two sources of sTF was performed to determine if one source is superior. In addition, the present study's FVIIa assay was automated using an ST A coagulometer from Stago Diagnostica (Paris, France). The robustness of the new assay was assessed using precision and accuracy testing. Precision was judged on within-run and between-run reproducibility and accuracy was judged by comparison of the new FVIIa assay with a commercial FVIIa kit from Diagnostica Stago. To ensure, results from the new assay can be compared between different laboratories, the assay was calibrated using the international FVIIa standard (52) in addition to calibration for measurement in concentration (ng/ml). 17 An initial population study on FVIIa was performed. Factor Vila levels were measured in a normal, healthy population (50 subjects) to determine the mean level and normal range for plasma FVIIa and examine trends in age and gender. As well, a group of women in the late stages of complicated pregnancies was also studied to examine changes in FVIIa levels that accompany pregnancy and determine anomalies caused by the complications. 18 MATERIALS Acrylamide, N,N'-methylbisacrylamide, ammonium persulfate, bromophenol blue, TEMED, and Econo-pac HighQ FPLC cartridges were purchased from Bio-Rad Laboratories. Bactero-tryptone, bactero-agar, and yeast extract were purchased were Difco grade and obtained from the Grand Island Biological Company. IPTG and X-Gal were purchased from 5-prime —> 3-prime. Protran nitrocellulose membranes were purchased from Schleicher & Schuell. Tissue culture flasks and roller bottles were obtained from Corning. Centricon 30 microconcentrators, YM-30 ultrafiltration membranes, and a spiral cartridge concentrator (CH2PRS) fitted with a S10Y10 cartridge were from Amicon. Dialysis tubing was purchased from Spectrum Medical Industries. Ampicillin, ethidium bromide, benzamidine, blue dextran, hemoglobin, ovalbumin, trypsin inhibitor, cytochrome c, and bovine serum albumin were obtained from Sigma Chemical Company. Agarose, goat anti-mouse immunoglobulin G-alkaline phosphatase conjugate, Not I and EcoRl restriction enzymes, React 3 buffer, T4 DNA ligase, ligase buffer, HEPES Buffered Saline, trypsin, Dulbecco's modified essential medium-Ham F12 with phenol red dye, fetal bovine serum, and Ultroser G were from Gibco-BRL. Methotrexate was obtained from David Bull Laboratories, antibody specific for the extracellular domain of tissue factor was from Enzyme Research Laboratories, Coomassie blue dye was from EM Science, and calf intestinal alkaline phosphatase and sephacryl S-100 HR chromatography resin were purchased from Pharmacia-LKB. Deoxy-ribonucleotides were purchased from Perkin-Elmer and Taq polymerase was prepared from a clone in our laboratory. All other chemicals were of reagent grade or higher and were purchased from either Sigma Chemical Company, Fisher Scientific Company, or British Drug Houses Limited. A T7 DNA sequencing kit was obtained from Pharmacia Biotech and two STACLOT VIIa-rTF kits were purchased from Diagnostica Stago. Recombinant factor Vila was purchased from Enzyme Research Labs. Immunodepleted human FVII-deficient 19 plasma was purchased from Sigma Chemical Co.. Congenital human FVII-deficient plasma was purchased from George King Biomedical. According to the suppliers, these deficient plasmas had less than 1% residual FVJJ activity. Pure phosphatidylcholine (PC), phosphatidylserine (PS), and phosphatidylethanolamine (PE) were purchased from Northern Lipids Incorporated. Siliconized Vacutainer tubes containing 3.2% trisodium citrate were obtained from Becton Dickinson. 20 METHODS 1. Polymerase Chain Reaction of sTF PCR amplification of sTF. The polymerase chain reaction was used to amplify the coding region for soluble tissue factor from a human cDNA library. Oligonucleotides primers with engineered NotI restriction sites and termination codons were designed with the aid of "Oligo" software (freeware) and synthesized on an Applied Systems 391 DNA Synthesizer, PCR-MATE, using deoxy-ribonucleotides purchased from Perkin-Elmer. The primers used are shown in Figure 9. a. 5'-primer 5' - ACA GCG GCC GCT AGA CATGGA GAC CCC TGC C - 3' Not I site START b. 3' - primer 5' - ACA GCfi GCC GCT CAT CAT TCT CTG AAT TCC CCT TTC TCC - 3' Not I site 2 STOP Figure 9. Oligonucleotide primers for PCR of sTFi.219. The primers were designed with engineered Not I restriction sites for cloning. An ATG start codon was included in the 5'-primer and two TCA stop codons were included in the 3'-primer. PCR amplification was carried out on a Perkin Elmer Cetus DNA Thermal Cycler 480 under the following conditions: 5 pi of human brain cDNA library (100 ng), 5 pJ of 10X PCR reaction buffer E, 1 pi of 10 mM deoxyribonucleotide triphosphates (dNTP), 1 p:l of 5'-primer (100 ng/ul), 1 ul of 3'-primer (100 ng/ul), 36.5 LLI of H 2 0 , and 0.5 ul of Thermophius aquitus (Taq) DNA Polymerase (1 U). The human brain cDNA library was constructed in the Escherichia coli expression vector pCDNA3 (Clonetech). Buffer E is a 10X PCR reaction buffer composed of 0.67 M Tris, pH 9.0, 0.155 M ammonium sulfate, 21 0.1 M 2-mercaptoethanol, and 25 mM magnesium chloride. The PCR cycle was begun with a hot start at 94°C (4 minutes) and then was set to the following cycle: Denaturation: 94°C, 30 seconds Annealing: 50°C, 30 seconds Extension: 72°C, 60 seconds for a total of 20 cycles followed by a 10 minute final extension at 72°C. Agarose gel electrophoresis. The PCR reaction mixture was checked using agarose gel electrophoresis. The electrophoresis was performed in a Bio-Rad Mini-sub DNA Cell apparatus. The electrode buffer used was IX TAE (diluted from a 50X TAE stock of composition 2 M Tris base, pH 7.5, 1 M glacial acetic acid, 0.1 M EDTA); 2% agarose gels were prepared by combining 1 g agarose with 1 ml 50X TAE in 49 ml of distilled water, heating the mixture to boiling, adding 10 ul ethidium bromide (2.5 mg/ml) pouring into a horizontal slab apparatus with combs, and then allowing to set at room temperature. DNA samples were prepared for electrophoresis by adding 1/10 volume of 10X running dye (40 ul 50X TAE, 500 ul 30% Ficoll-0:5% bromophenol blue, 960 ul water) and mixing with a pipetman. Electrophoresis was carried out at 100 V constant voltage for 20 - 30 minutes using a Pharmacia EPCS 3000/150 power supply. Bands were detected under UV illumination at 260 nm. GENECLEAN DNA Purification. The PCR-amplified DNA was purified from the agarose gel using the GENECLEAN protocol (BIO101 Inc.): the band of interest was cut from the agarose slab and placed into an Eppendorf tube to which 3-equivalents of Nal was added. The agarose was melted by incubating at 55°C for 5 minutes. Five microlitres of vortexed glass milk (BIO 101 Inc.) was added and the mixture incubated again at 55°C for 5 minutes to bind the DNA. The solution was then centrifuged briefly and supernatant removed. The glass milk pellet was then washed three times with 400 ul of New Wash (20 mM Tris-HCl, pH 7.2, 0.2 M NaCl, 2 mM EDTA combined 1:1 with 100% ethanol). The washed pellet was then resuspended in 20 ul of TE-buffer (10 mM 22 Tris-HCl, pH 8.0, 1 mM EDTA) and centrifuged. The supernatant containing the clean DNA was used for ligation reactions. 2. Cloning and DNA sequence analysis of PCR fragment Cloning of PCR fragment into pBS. The purified PCR fragments were cloned into a Bluescript KS- (pBS) vector (Stratagene Cloning Systems) for DNA sequence analysis. To produce sticky ends (6 bp overhangs) for cloning of sTF into the pBS sequencing vector, both the amplified PCR fragment and the vector were cut with the restriction enzyme Not I as follows: Purified PCR fragment (sTF) 3 ul Bluescript KS-(pBS) 1 Hi Not I (10 U) 1 ul Not I (10 U) 2ul lOX-React 3 buffer 2 ul 10-X React 3 buffer 17 ul dH20 14 ul dH20 at 37°C for 1 hour. Following digestion, the cut pBS vector was treated with 1 ul calf intestinal alkaline phosphatase (1 U) and incubated at 37°C for 30 minutes. Digested DNA was purified by the Gene-Clean method. The purified Not I-digested sTF fragment DNA and linearized pBS vector DNA were then ligated together as follows: 2 ul Not I-digested pBS 5 ul Not I-digested sTF fragment 1 ul T4 DNA ligase (1 U) 2 ul 5X ligase buffer at 16°C overnight (16 hours). Transformation of E.coli cells. A colony of E.coli DH5cc was inoculated into 50 ml of YT medium (8 g bactero-tryptone, 5 g yeast extract, 5 g NaCl, pH7 in 1 litre of distilled water, autoclaved). The culture was shaken at 37°C until the OD600 reached 0.4 to 0.6. The culture was then placed on ice for 5 minutes before transferring to a cold sterile 23 centrifuge tube and centrifuged at 6000 rpm in an SS-34 rotor at 4°C for 5 minutes. The supernatant was poured off, the cells gently resuspended in 12.5 ml of ice cold 50 mM calcium chloride, and left on ice for 25 minutes to make them competent (62). The cell suspension was re-centrifuged as before, the supernatant poured off, and the tube placed on ice. The competent cells were resuspended very gently in 2.5 ml of ice cold 50 mM calcium chloride. Five microlitres of the ligation reaction mixture was transformed into 50 ul of competent DH5a E.coli cells. The transformation was placed on ice for 30 minutes then heat shocked at 42°C for 90 seconds. Two hundred microlitres of LB (Luria-Bertani)-media (10 g bactero-tryptone, 5 g NaCl per litre H20, pH 7.4, autoclaved) was added and the mixture incubated at 37°C for 45 minutes with shaking. One hundred microlitres of transformation mixture was plated on a selective LB-agar plate (12 g/1 agar in LB-media) containing 25 mg/1 JPTG, 50 mg/1 X-Gal, 100 mg/ml ampicillin. The remaining 100 ul of transformation mixture was plated on a second LB-IPTG/X-Gal/Amp plate and the plates were incubated upside down at 37°C. Colonies were visible after 14 hours incubation. pBS allowed for colour selection since vectors containing a cloned DNA insert produce white bacterial colonies on the selective plates. Plasmids without inserts appeared blue. Mini-prep recovery of plasmid DNA. White colonies were picked off the plates using a pipette tip and grown in 4 ml of LB media in Falcon tubes for 12 hours at 37°C with shaking. Plasmid DNA (pBS) was isolated by mini-prep as follows: the Falcon tubes were spun in a Sorvall RT6000B swinging basket centrifuge at 3000 g for 15 minutes at 4°C. The supernatant was poured off and the pellet was resuspended in 200 ul of glucose buffer (50 mM glucose, 25 mM Tris-base, 10 mM EDTA, pH8), vortexed, and transferred to a 1.5 ml bullet tube (Eppendorf). Four hundred microlitres of freshly prepared alkaline-SDS solution (0.2 N NaOH, 1% SDS) was added, mixed by inversion, and placed on ice for 5 minutes. Three hundred microlitres of 3 M sodium acetate (pH 4.8) was added and the solution mixed by inversion. A thick precipitate formed. The solution 24 was cooled on ice for 10 minutes and then centrifuged for 5 minutes. Seven hundred and fifty microlitres of supernatant was withdrawn and placed into a new eppendorf. Four hundred and fifty microlitres of cold isopropanol was added, mixed well, and left for 5 minutes at 4°C to allow the DNA to precipitate. The mixture was centrifuged for 5 minutes, the supernatant decanted, the wet pellet briefly re-centrifuged, and the remainder of the fluid removed using a pulled pasteur pipette. The pellet was allowed to air dry and then resuspended in 200 ul of dF^O by vortexing. The pellet was then washed by adding 100 ul of 7 M ammonium acetate (pH 7.5), mixed, and then filling the tube with cold 95% ethanol, and mixed again. The solution was centrifuged for 5 minutes and the supernatant poured off. The pellet was re-washed using 200 ul of 95% ethanol, centrifuged for 2 minutes, and the supernatant removed using a pulled pasteur pipette as before. The washed plasmid DNA pellet was air dried and resuspended in 50 ul of dFkO. The recovered plasmid DNA was screened for an sTF cDNA insert by restriction digest using Not I and electrophoresing on a 2% agarose gel. Clones containing an sTF insert were subject to DNA sequence analysis." DNA Sequence Analysis. Four positive clones (containing sTF insert) were selected to be sequenced using a T7 Sequencing Kit (Pharmacia Biotech). pBS forward and reverse primers were used in addition to two internal primers as shown in Figure 10. Chain termination sequencing reactions (63) and 35S-labelling were performed using the T7 Sequencing kit. Electrophoresis of the DNA sequencing reactions was carried out on a BRL vertical electrophoresis apparatus using 37.5 x 20 cm glass plates and 0.4 mm spacers. The electrode buffer used was 0.5 X TBE (diluted from a 10 X TBE stock of composition 0.89 M Tris, 0.89 M borate, 25 mM EDTA); the sequencing gel was composed of 8% polyacrylamide, 8.3 M urea, and 45 mM Tris borate-1.25 mM EDTA. Polymerization of the gel was initiated by the addition of 0.5 ml 10% (w/v) NH4HS04and 23 ul TEMED. The gel was warmed by pre-running at 50 watts constant power for 30 minutes. The DNA sequencing reaction samples were loading on the gel using Sharkstooth 25 a. pBS primers 5' - GTA AAA CGA CGG CCA GT - 3' FORWARD primer 5' - AAC AGC TAT GAC CAT G - 3' REVERSE primer b. Internal primers 5* - GTC CCG CGC CCC GAG ACC GC - 3' TF100 5' - TTT CTT TCC TGA ACT TGA AG -3' TF594 Figure 10. DNA Sequence Analysis Primers. Two sets of primers, pBS primers (a) specific for vector sequences and internal primers (b) specific for tissue factor gene sequences, were used. combs and electrophoresed at 32 W constant power. After electrophoresis, the gels were dried at 80°C under vacuum on a Bio Rad gel drier for 20-30 minutes and autoradiographed on Kodak XK-1 film overnight at room temperature. The results of sequencing of the four clones was compared to the published sequence of tissue factor (48). A PCR error-free clone was identified and grown in LB medium as above. Plasmid DNA containing the error-free sTF insert was isolated by mini-prep and used for sub-cloning. 3. Transfection of BHK Cells and Expression of sTF Cloning of sTF into pNUT. The sTF insert was cut out of the pBS vector using a Not I restriction digest and ligated into a Not I linearized pNUT eucaryotic expression vector (64) using the same procedure as for cloning into the pBS vector. The ligation mixture was used to transform competent DH5cc E.coli cells as before. The transformed cells were grown on selective LB plates, colonies were picked and cultured, and the pNUT plasmid DNA was isolated by using the mini-prep procedure described 26 previously. Diagnostic restriction digests using Not I and EcoRl of the DNA were performed to ensure the presence of the sTF insert in the pNUT expression vector. Transfection of BHK cells. Baby hamster kidney (BHK) cells were grown in Dulbecco's modified essential medium-Ham F12 (DMEM-F12) with 5% fetal bovine serum (FBS) to approximately 107 cells per 90 mm dish. The isolated pNUT-sTF vector was used to transfect the BHK cells as follows: 7 ug of mini-prepped pNUT-sTF was resuspended in 450 ul of distilled water followed by addition of 50 ul of 2.5 M CaCk and 500 ul of 2X-HEPES buffered Saline (pH6.95). After standing at room temperature for 20 minutes, the mixture was added to 70% confluent BHK cells. After 6 hours, the transfection media was removed and fresh 5% FBS-DMEM media added and the cells allowed to recover for 24 hours at 37°C. The FBS-DMEM media was then replaced with selective 5% FBS-DMEM media containing 500 uM methotrexate (MTX). The surviving BHK colonies were trypsinized and passaged into a 75 cm tissue culture flask, still under MTX selection. After five days, the 75 cm flasks were approaching confluency and a Western blot was performed on the tissue culture media to detect production of sTF by the BHK cells. SDS-PAGE. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, based on the method of Laemmli (65), was performed in a Hoeffer Scientific Mighty Small II SE250 mini-protein gel apparatus using 10 x 8 cm glass plates and 1 mm spacers. The electrode buffer used for electrophoresis was 0.015 M Tris base, pH 8.3,0.192 M glycine, 0.1% SDS (usually prepared by dilution of a 5X electrode buffer). The separating gels used contained 0.375 M Tris-HCl, pH 8.8, 0.1% SDS, and 10% or 12% acrylamide:bis (prepared from a 30% acrylamide (w/v) and 0.8% N,N'-methylenebisacrylamide (w/v) stock solution). Polymerization was initiated by the addition of freshly prepared 10% ammonium persulfate (APS) to a final concentration of 0.05% and TEMED to a final concentration of 0.00033%. The stacking gels were prepared as 0.125 M Tris-HCl, pH 6.8,0.1% SDS, 3% acrylamiderbis (30:0.8). Polymerization was effected by adding fresh 27 10% APS to a final concentration of 0.13% and TEMED to a final concentration of 0.001%. Protein samples were prepared for electrophoresis by addition of 1/5 volume of 5X sample buffer (60 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS, 0.1% bromophenol blue, 14.4 mM 2-mercaptoethanol) and placed in a boiling water bath for 3 minutes. After cooling, the samples were loaded onto the stacking gel and electrophoresed at 50 mA per gel for 30 minutes. Protein gels were stained with Coomassie blue stain (0.25% Coomassie blue, 10% acetic acid, 45% methanol) for 30 minutes to 1 hour and then de-stained using 45% methanol, 10% acetic acid de-stain solution. Western blot analysis. Proteins were separated by SDS-PAGE and electrophoretically transferred (66) to a Protran nitrocellulose membrane using a Hoeffer Scientific Transphor apparatus at 0.1 mA constant current overnight with water cooling. The transfer buffer used was 23.4 mM Tris base, 19.3 mM glycine, and 20% methanol (v/v). The nitrocellulose membrane was transferred to a Sarstedt culture tube and blocked with a blotto/tween solution (5 g skim milk powder, 0.2 ml tween 20, 0.020 g sodium azide in 100 ml phosphate buffered solution (8 g NaCl, 0.2 g KC1, 1.44 g Na2HP04, 0.24 g KH 2 P0 4 , pH 7.4, per litre PBS)) for 1 hour at room temperature with mixing. The membrane was then incubated with a primary antibody specific for the extracellular region of TF (1:3000 dilution in 1/10 blotto/tween) for 2 hours at room temperature with mild agitation. Unbound antibody was removed by washing the membrane twice with phosphate buffer solution and twice with 150 mM NaCl, 50 mM Tris-buffer (pH 7.5). The secondary antibody, an anti-murine IgG-alkaline phosphatase conjugate (1:5000 dilution in 1/10 blotto/tween) was then applied to the membrane and incubated for 1 hour at room temperature with mild agitation. After washing the membrane four times using 150 mM NaCl, 50 mM Tris-buffer (pH7.5), colour detection was carried out by incubating the membrane with 66 ul of nitro-blue tetrazolium (50 mg/ml in 70% DMF) and 33 ul bromochloroindoyl phosphate (50 mg/ml in DMF) substrate in 10 ml alkaline phosphatase buffer (0.1 M Tris-HCl, pH9.5,0.1 M NaCl, 5 mM MgCl2) until bands appeared. 28 Tissue culture of sTF-BHK cell line. Freezer stocks of the sTF-expressing BHK cell line were prepared by trypsinizing the confluent cells and re-suspending them in 2 ml of DMEM with 50% glycerol. Aliquots were stored at -70°C. Tissue culture of the BHK cell lines was carried out on the remaining flasks. The BHK cells were transferred to 250 ml tissue culture flasks and incubated at 37°C in 50 ml of 5% FBS-DMEM media with 500 uM methotrexate (MTX). When the cells became confluent after 3-4 days, they were passed or split as follows: in a sterile fumehood, the old media was suctioned off and 2 ml of trypsin was added to the flask and incubated for 50 - 60 seconds. The trypsin was suctioned off and the treated cells were resuspended in 2 ml of fresh 5% FBS-DMEM-MTX media. Ten to twelve drops of the cell suspension were added to a new 250 ml flask containing 50 ml of fresh 5% FBS-DMEM-MTX media and the fresh flask replaced in the 37°C incubator. After several passes, the sTF-expressing BHK cells were transferred to Roller bottles for protein harvesting. Two 250 ml flasks containing confluent BHK cells were trypsinized and resuspended in 5% FBS-DMEM-MTX as described above for passing of cells. All 4 ml of the cell suspensions was added to a 2 litre roller bottle containing 150 ml of fresh 5% FBS-DMEM-MTX and the roller bottle placed in a 37°C incubator. The cells became confluent after 7 days at which time, the tissue culture media was switched to a minimal 1% Ultroser G (USG)-DMEM media as follows: the 5% FBS-DMEM-MTX media was removed by suction and 150 ml of 1% USG-DMEM media was added to the roller bottle under sterile conditions. The roller bottle was replaced in the 37°C incubator and the cells allowed to express sTF. After 3 days, the 1% USG-DMEM media was reaching its limit of sustaining the cells and the media was collected into a sterile bottle containing 0.2% sodium azide and 25 mM benzamidine. Three hundred milliltres of fresh 1% USG-DMEM was added to the roller bottle and replaced in the 37°C incubator. The media was collected every 3.5 days and replaced with fresh media. The collected media 29 was stored at 4°C. Media was collected until the expressing BHK cells died at which time a new roller bottle was seeded from a freezer stock and the collection cycle repeated. 4. Purification of sTF from Tissue Culture media Soluble tissue factor was purified from the tissue culture media by ammonium sulfate precipitation, anion-exchange FPLC, and gel filtration. Detergent was not required during the purification of sTF. Ammonium Sulfate fractionation. Tissue culture media was concentrated 12-fold using an Amicon CH2PRS spiral cartridge concentrator fitted with an S10Y10 cartridge at 20 psi pressure. To the concentrated crude media, solid ammonium sulfate was slowly added to give a final concentration of 55% (36.4 g ammonium sulfate per 100 ml solution). After precipitation at 4°C overnight, the solution was transferred to centrifuge tubes and spun at 5000 rpm for 30 minutes at 4°C using a Sorvall RC-5B Refrigerated Super-Speed Centrifuge in an SS-34 rotor (Mandel Scientific Company). The supernatant was separated from the precipitated proteins, concentrated 5-fold using a Centricon 30 microconcentrator (Amicon), and then dialyzed against two changes of 4 litres of 20 mM NaCl, 20 mM Tris-HCl, pH 7.5 (buffer A) to lower the salt concentration. The precipitate was checked for sTF by SDS-PAGE and Western blot and discarded. Anion-exchange Fast Protein Liquid Chromatography. The FPLC apparatus consisted of Pharmacia LKB components: LCC-500 Plus Gradient Controller, P-500 pumps, manual valve V-7, UV-MII optical unit with an HR 10 flow cell, REC-481 single channel chart recorder, and a Frac-100 fraction collector. The pumps were primed using low salt buffer A and high salt buffer B (500 mM NaCl, 20 mM Tris-HCl, pH 7.5). Two 5 ml Econo-Pac High Q FPLC cartridges (Bio-Rad) connected in series were used for anion-exchange purification. The cartridges were washed with 2 volumes of buffer B and then equilibrated with buffer A until baseline was reached prior to application of the crude protein sample. 30 Following suction filtration through a 0.45 um filter, the dialyzed 55% ammonium sulfate fraction was loaded onto the High Q FPLC Cartridges using a Superloop (Pharmacia-LKB) at a flow rate of 1.0 ml/min. After washing with buffer A until a stable baseline was re-established, the bound proteins were eluted with a continuous salt gradient of 20 - 500 mM NaCl over 60 minutes at a flow rate of 1.0 ml/min and 1.5 ml fractions were collected. Protein elution was detected by measuring absorbance at 280 nm and the salt gradient was held constant during peak elution. Collected fractions were checked for the presence of sTF using SDS-PAGE and Western blot; fractions containing sTF were pooled and concentrated using Centricon Concentrators 30 with YM-30 ultrafiltration membranes (Amicon) to a volume of 2.5 ml. Gel-filtration Purification. A Sephacryl S-100 HR gel filtration column was prepared as follows: 300 ml of 100 mM NaCl, 20 mM Tris-HCl, pH 7.5 (buffer C) was added to 600 ml of Sephacryl S-100 HR matrix (Pharmacia LKB) and the slurry placed on ice and de-gassed for 10 minutes. At 4°C, the de-gassed slurry was poured slowly into a Bio-Rad 55 x 2.4 cm column to a height of 52 cm giving a bed volume of 300 ml. The gel filtration column was equilibrated with 3 bed volumes of buffer C passaged at 1 ml/min using a Gilson Miniplus HP4 peristaltic pump. Void volume was determined using blue dextran and the column calibrated using 3 ml of a protein standard mix containing 2.5 mg each of hemoglobin, ovalbumin, trypsin inhibitor, and cytochrome c. The concentrated High Q fraction (2.5 ml) was loaded onto the Sephacryl S-100 HR gel filtration column. Proteins were eluted with buffer C at a flow rate of 0.60 ml/min using a peristaltic pump (Gilson Miniplus HP4) and 3.0 ml fractions were collected using a 2112 Redirac Fraction collector (LKB Bromma) Protein elution was determined by measuring absorbance at 280 nm on a Perkin Elmer Lambda 3A UV/Visible Spectrophotometer. Fractions containing protein were checked for sTF using SDS-PAGE and Western blot; fractions containing pure sTF were pooled and concentrated using Centricon Concentrators 30. The sTF concentration was determined spectrophoto-31 metrically using a 1% extinction coefficient at 280 nm of 14.8 (67) and a molecular weight of 39,000. Aliquots of the pure sTF were stored frozen at -70°C. 5. Characterization of purified sTF The purified sTF was characterized using SDS-PAGE on a 12% polyacrylamide gel and Western blotted as described above. Amino-terminal sequence analysis of sTF was performed using an Applied Biosystems 476A Sequencer. The presence of glycosylated residues was also determined using the 476A Sequencer. 6. Development of Factor Vila Assay Preparation of phospholipid vesicles. Phospolipid vesicle mixtures of composition POPS = 75:25 and PC:PS:PE = 64:6:30 were prepared as follows: the phospholipids in the desired molar composition were frozen in liquid nitrogen, lyophilized, and then resuspended in distilled water by vortexing. The phospholipid mixture was then transferred to a cryovial (Simport Plastics) and subjected to 5 freeze-thaw cycles alternating between a liquid nitrogen bath and a 35°C water bath. Large unilamellar vesicles (LUVs) were prepared by extruding the freeze-thawed mixture through a 100 nm pore size polycarbonate filter (Poretics Corporation) using a Lipex Extruder (Lipex Biomembranes Inc.) under 300 psi Nitrogen gas pressure for 10 cycles. Vesicle size was determined using a Nicomp Model 270 Submicron Particle Sizer (Pacific Scientific) and concentration of the phospholipid mixtures was determined using a standard phosphate assay (68). Optimization of sTF-Reagent. The sTF-reagent was composed of purified BHK-derived soluble tissue factor and a phospholipid mixture dissolved in TBS/BSA (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 0.1% bovine serum albumin, 0.02% NaN3). To determine the optimal concentration of sTF, a series of sTF-reagents with sTF-concentrations ranging from 0.01 umol/1 to 1.0 umol/1 was prepared in a 0.27 mM PC:PS:PE (64:6:30) phospholipid, TBS/BSA mixture. As well, an sTF-negative reagent 32 containing no sTF was also prepared. To determine the optimal composition and concentration of the phospholipid mixture, a series of sTF-reagents with varying concentrations of PC:PS:PE (64:6:30) or PC:PS (75:25) phospholipid were prepared in a 0.5 uM sTF, TBS/BSA mixture. All sTF-reagents were stored frozen at -70°C and underwent exactly one freeze-thaw cycle before use. Assay for Factor Vila. The factor Vila assay was performed on an automated STA Coagulometer (Diagnostica Stago) at Vancouver Hospital. The assay was performed on titrated plasma samples thawed at 37°C and held at room temperature until placed in the coagulometer. Sample volumes per assay (plasma, sTF-reagent, and 25 mM CaCl 2 solution) were 50 ul each. Factor VH-deficient plasma, test plasma, and the sTF-reagent were held in a 16°C temperature block and the 25 mM CaCl2 solution was held in a 37°C temperature block prior to use. Assays were carried out in cuvettes containing electromagnetic ball-bearings for mixing. Five microlitres of titrated test plasma was diluted 10-fold in FVJJ-deficient plasma to give a final volume of 50 ul in a cuvette; 50 ul of sTF-reagent was added, the solution mixed, and incubated at 37°C for 120 seconds. Fifty microlitres of pre-warmed CaCl2 was added and the solution electro-magnetically mixed. Clotting times were measured from the addition of calcium to the freezing of the ball-bearing to electromagnetic mixing indicating clot formation. All assays were performed in duplicate and a third assay was run if the duplicates did not agree within 7%. If the third assay did not agree with either of the duplicates, an error was reported. Factor Vila Standard curve. A stock factor Vila concentrate was prepared by diluting pure FVIIa to a concentration of 300 ng/ml in titrated normal pooled plasma (obtained from the Hematology Department, UBC Hospital). Aliquots of this FVIIa stock solution were stored at -70°C. Factor Vila standards of 30, 3.0, 0.3, and 0.03 ng/ml FVUa were prepared by serial dilution of the factor Vila 300 ng/ml stock in FVII-deficient plasma. The clotting times of the four FVIIa standards were measured using the FVIIa assay as described above. Using the calibration mode of the STA Coagulometer, the 33 clotting times were plotted versus the FVIIa concentrations on a log-log scale and a polynomial curve was fit to the line using linear regression analysis. Using this calibration curve, factor Vila concentrations in ng/ml could be generated automatically from the measured clotting times of the test plasma samples. Comparison of BHK vs. E.coli-derived sTF. Soluble tissue factor derived from BHK cells was compared to Zs.co/i'-derived sTF (a generous gift from Dr. James H. Morrissey, Oklahoma Medical Research Foundation) by preparing two sTF-reagents of identical composition 0.5 uM sTF, 1.0 mM PC:PS:PE (64:6:30) in TBS/BSA, except using BHK-derived sTF in one preparation and Rco/i-derived sTF in the other preparation. The four FVIIa standards (30, 3, 0.3, 0.03 ng/ml) and five titrated plasma samples were measured using both sTF-reagents in the FVIIa assay described above. Linear regression analysis of the results was performed to compare the two sources of sTF. 7. Testing of FVIIa Assay Precision Testing. To test the reproducibility of the FVIIa assay, precision testing for within-run and between-run reproducibility was performed. A citrated plasma sample was collected from a single donor and 15 aliquots were frozen at -70°C. One aliquot was measured 10 times (in duplicate) on a single day to establish within-run precision. To establish between-run precision, an aliquot from the donor was measured 4 times per day on 5 additional days . Mean values, standard deviations, and coefficients of variation for within-day and between-day precision were calculated. Accuracy Testing. To test the accuracy of the FVIIa assay, a patient comparison study was performed between the present study's FVIIa assay and a commercial FVIIa kit available from Diagnostica Stago. The STACLOT VHa-rTF kit was set up on the STA Coagulometer according to the manufacturer's instructions. A standard curve using FVIIa international standards (mU/ml), included with the kit, was established for the STACLOT VIIa-rTF assay. Twenty citrated plasma samples ranging from 0.1 ng/ml to 30 ng/ml 34 FVIIa concentration were measured using both the present study's FVIIa assay and the STACLOT VIIa-rTF kit. Linear regression analysis was performed to assess the accuracy of the FVIIa assay in relation to the commercial FVUa kit. A FVIIa conversion curve was plotted to convert FVUa concentrations in ng/ml to international units (mU/ml). Plasma Stability studies. To determine the stability of plasma FVIIa, a series of plasma stability studies was performed. To determine the effects of freezing and frozen storage time on a plasma sample, a freshly collected titrated plasma sample was measured for FVIIa and then aliquots were frozen at -70°C for 15 minutes, 2 days, and 7 days and re-measured. To test the effects of repeated freeze-thaw cycles on plasma FVIIa, three plasma samples were subjected to three freeze-thaw cycles and measured for FVIIa after each cycle. To compare the effects of different anti-coagulants, one titrated and one EDTA-anti-coagulated plasma sample was drawn from a single donor from the same venepuncture. Precision testing, as described above, was performed on both the EDTA and titrated plasma samples and the results compared. To determine the effects of dilution during the FVIIa assay, 10 titrated plasma samples were measured neat and after 10-fold dilution in FVII-deficient plasma and the measured FVIIa compared after correction for dilution. Individual variations. To determine the intra-individual variations in plasma FVIIa levels, 6 titrated plasma samples were drawn from a single donor over a 3 month period (December, 1996 to March, 1997). FVIIa levels were measured for each sample to determine individual fluctuations in plasma FVUa. 35 8. Factor Vi la in Normal Populations Ethical approval for blood collection. Ethical approval for the collection of blood from volunteer donors, who gave informed written consent, was obtained from the University of British Columbia ethical review board. Subjects. Fifty normal healthy volunteers (29 female, 21 male) aged 20-52 years with no history of CHD provided blood samples. No effort was made to control for ethnic origin and samples were collected in the fasting state or within 2 hours after a light breakfast. Phlebotomy. Venous blood (9 volumes) was collected between 9 and 10:30 am sequentially into 2 siliconized Vacutainer tubes (Becton Dickinson) containing 1 volume of 3.2% trisodium citrate. The first tube of blood collected after venepuncture was discarded. Only the second tube was used for analysis. Platelet poor plasma was obtained by centrifugation at 3000 rpm for 20 minutes and aliquots in plastic tubes were frozen at -70°C within one hour of collection. At the time of assay, citrated plasma samples were thawed rapidly at 37°C, mixed gendy, and stored at room temperature until tested. Plasma samples were never subjected to more than one freeze-thaw cycle. Measurement of FVIIa. FVIIa was determined with the automated FVIIa assay employing sTF as described above. The composition of the sTF-reagent used was 0.5 uM sTF, 1.0 mM PC/PS/PE (64:6:30) in TBS/BSA. Clotting times were converted to a FVIIa concentration in ng/ml using the FVIIa standard curve and into FVIIa international activity units (mU/ml) using the FVIIa conversion curve. FVIIa levels in the normal population were examined by age and gender. 9. Factor Vi la in Complicated Pregnancies Ethical approval. Ethical approval for collection of blood samples from hospital patients was obtained through Dr. Cedric Carter of Vancouver Hospital and Dr. Louis Wadsworth of British Columbia Children's Hospital. 36 Subjects. Nineteen pregnant women aged 22-40 years who experienced complications during their pregnancy provided blood samples. Plasma collection was performed as described above. Clinical assessment. Details on the complications and stage of pregnancy of each subject were noted. In addition to the measurement of plasma FVIIa as described above, plasma samples were variously measured for prothrombin time (PT), activated partial thromboplastin time (APTT), plasma fibrinogen (FD3R), and platelet counts at the physicians' discretion. 37 RESULTS 1. Polymerase Chain Reaction of sTF The initial step in the present study was the isolation of a truncated form of tissue factor. Since sTF does not occur naturally, the mutant protein had to be produced using recombinant DNA techniques. Previous studies have generated sTF by isolating restriction fragments from cDNA libraries and then ligating synthetic linker oligonucleotides to the restriction fragments' 3'- and 5'-ends to generate termination codons and cloning sites (50). The present study simplified the process by using polymerase chain reaction with mutagenic primers incorporating engineered stop codons and cloning sites to selectively amplify the coding region for soluble tissue factor. A 785-bp DNA fragment consisting of 752 nucleotides of the TF gene encoding amino acids -32 to 219 as described in (15) flanked by a synthetic Not I restriction site on the 5'-end and two synthetic stop codons followed by a synthetic Not I site on the 3'-end was amplified using PCR. Successful amplification of the sTF fragment by PCR was confirmed using agarose gel electrophoresis. Figure 11a shows the results of PCR amplification of sTF using a human brain cDNA library as the template. The 785 bp band in lane 2 was cut out and used as a template for a second round of PCR. The results shown in Figure lib indicate a stronger 785 bp band confirming specific amplification of the desired fragment. 2. DNA Sequence analysis of sTF insert The PCR fragment with its engineered NotI restriction sites was subsequently cloned into a Bluescript sequencing (pBS) vector and transformed into an E.coli host. pBS allowed for colour selection since vectors containing a cloned DNA insert produced white bacterial colonies on the selective LB-IPTG/X-Gal/Amp plates while plasmids without an 38 Figure 11. PCR amplification of sTF gene. 2% agarose gel electrophoresis, visualized at 260 nm UVillumination of (a) PCR amplification reaction from human cDNA library of sTF gene. A faint band of 785 bp was visible in lane 2. This band was cut out and used for a second round of PCR amplification, (b) Results of PCR. amplification using the excised band as template show two very strong bands of 785 bp in lanes 2 and 3. These bands were excised, punfied by GENECLEAN, and used for DNA sequencing. The large, smudged bands on the bottom of the gels areRNA. Lane 1 shows a standard 1 kb ladder. 3 9 insert appeared blue. Figure 12 shows the results of the Not! digestion screening to confirm the presence of the sTF insert in the isolated pBS from a white colony. Since the Taq polymerase employed in the PCR reactions lacks a 3'-exonuclease for proof-reading (69), errors in the amplification product often occur. To ensure that an error-free clone was selected for sTF protein expression, four of the positive clones were analyzed for their DNA sequence. Of the four clones sequenced, only one was found to be free of PCR errors when compared to the known sequence for TF (48). 3. Expression of sTF in BHK cells The sTF insert from the error-free pBS clone was sub-cloned into a pNUT expression vector again via its Not! restriction sites. The presence of the sTF insert in the pNUT vector was confirmed by Not! and EcoRl digestion screenings. Figure 13 shows the 782 bp Not I band and a 600 bp EcoRl band verifying the presence of the sTF insert. The 600 bp band is a result of two EcoRl restriction sites, one at the 3'-end of the sTF insert and one in the pNUT poly linker as shown in Figure 14. Expression of sTF has previously been carried out in Chinese hamster ovary (CHO) cells (50), S.cerevisiae (70), and E.coli (67). The present study used baby hamster kidney (BHK) cells for expression. The BHK cells provided a eucaryotic expression system without the disadvantages of hyperglycosylation conferred by yeast systems (S.cerevisiae) (70) and the low sTF expression levels of CHO cells (0.5-1.0 mg/1) (50). The pNUT vector serves as a eucaryotic expression vector, with a zinc metallo-protein promoter driving transcription of the cloned sTF insert (Figure 14). The isolated pNUT-sTF vector was used to transfect competent BHK cells and the cells were grown in selective media containing the cytotoxic agent methotrexate (MTX). Since the pNUT vector contains a genetic cassette for di-hydrofolate reductase which breaks down 40 p B S - 3000 bp sTF —782 bp Figure 12. Not I screening of Bluescript-sTF clone. 2% agarose gel electrophoresis, visualized at 260 nrn U V illumination of isolated bluescript -sTF plasmids following digestion with Not I. Lanes 1 to 9 show the results from 9 white colonies. Lane 11 shows a standard 1 kb ladder. The upper (3000 bp) band corresponds to the bluescript vector DNA and the lower (782 bp) band corresponds to the cloned sTF -insert. Thus all clones were positive. MW 1 2 3 4 5 6 7 8 9 10 (<=) MW1 2 3 4 56 7 8 910 - 782 bp •600 bp Figure 13. Not I and EcoRl screening of pNUT-sTF clone. 2% agarose gel electrophoresis, visualized at 260 nm TJV illumination of isolated pBS-sTF plasmid following digestion with (a) Not I or (b) EcoRl. The results indicate a 782 bp band presentin all lanes except 3 and 6 of the Not I screen (a) and a 600 bp band in lanes 2, 4, and 8 of the EcoRl screen (b) confirming the presence of the sTF insert in the pNUT expression vector. A 1 kb ladder is shown in lane MW. 41 •TF cDNA DHFR em—Urn Figure 14. pNUT expression vector with cloned sTF insert. A schematic representation of the pNUT eucaryotic expression vector with the sTF cDNA cloned into its polylinker region. EcoRl and Not I restriction sites are indicated as well as the zinc metallo-protein promoter (MT-1) and the di-hydrofolate reductase (DHFR) cassette that confers resistance to methotrexate. methotrexate, only BHK cells successfully transfected with the pNUT-sTF plasmid survived drug selection. Successful transfection was apparent by day nine of methotrexate selection when discreet surviving colonies were visible after dead cells were removed and fresh selective media re-applied. After transferring to tissue culture flasks, a Western Blot was performed on the tissue culture media to detect production of sTF by the BHK cells. The results of the immunoblot are shown in Figure 15. All of the BHK clones were found to be successfully expressing and secreting soluble tissue factor. It should be noted that the expressed sTF protein contains an endogenous export signal peptide in its 32-residue leader peptide (17) which is 42 1 2 3 4 5 6 7 8 Full-length s T F - * Positive Control >M *%M| •»* M l ^-BHK-expressed sTF (E.coii-derived sTF) Figure 15. Western Blot of tissue culture media from several sTF-BHK cell lines. Samples of tissue culture media were run on a 10% SDS-PAGE and Western blot analysis was performed using an antibody raised against the extracellular portion of tissue factor. A l l 7 clones (lanes 2-8) were positive for expression of sTF protein as indicated by the results of Western blot. Western blot analysis of non-transfectedBHK cells (results not shown) showed no positive clones. Lane 1 shows a positive control containing sTF and full-length TF expressed in E.coli. 43 subsequently cleaved by a signal peptidase. It was via this peptide that the sTF protein was secreted from the BHK cells into the media. (Had such an endogenous signal peptide not existed in the sTF protein structure, a synthetic oligonucleotide coding for such a peptide would have had to be ligated to the sTF fragment or the sTF would have had to be purified from within the BHK cells). Lane 1 shows a positive control of sTF provided by Dr. Morrissey. Since all the BHK clones were positive for sTF expression, they were combined. Tissue culture of the sTF-expressing BHK cell line was carried out in roller bottles at 37°C. Culture media was collected (and replaced with fresh media) when the cells appeared to be stressed. A pH-sensitive phenol red indicator dye in the DMEM-F12 media changed from red to orange-yellow when the media had reached its limit to support the BHK cells. After 5 to 6 changes of media, the sTF-BHK cell line eventually died and new roller bottles were seeded. BHK cells generally express at their highest level when stressed. Approximately 7 litres of tissue culture media from the sTF-expressing BHK cells was collected. 4. Ammonium sulfate fractionation The first step in the purification of soluble tissue factor from the tissue culture media was an ammonium sulfate fractionation. This served two purposes: to concentrate the crude sample to a workable volume and to remove some protein impurities. Prior to addition of ammonium sulfate, the tissue culture media was concentrated 12-fold using an Amicon concentrator to reduce the volume. To determine the optimal ammonium sulfate concentration for fractionation, ammonium sulfate cuts of 30%, 40%, 50%, 60%, 70%, and 80% were performed on samples of the concentrated media. Western blots indicated that sTF began precipitating at 60% ammonium sulfate (results not shown).' Thus, it was decided that a 55% ammonium sulfate fractionation would produce a good initial purification with negligible loss of sTF protein (Figure 16). 44 (a) (b) 1 2 3 4 5 1 2 3 4 5 3 minor < impurities \43 kDa 130 kDa bo kDa M kDa 130 kDa 90 kDa 68 kDa 43 kDa 29 kDa 29 kDa 18 kDa sTF 18 kDa Major 65 kDa Impurity Figure 16. 55% Ammonium Sulfate Fractionation. Supernatant and precipitate following 55% ammonium sulfate precipitation were run on a 12% SDS-PAGE. The protein gel was visualized with Coomassie blue (a) and Western blot analysis (b) was performed using an antibody raised against the extracellular portion of tissue factor. Lanes 1 and 4 are crude protein samples, lane 2 is the precipitate fraction, lane 3 is the supernatant fraction, and lane 5 contains the M W markers. Western blot (b) analysis indicated that no sTF precipitated at 55% ammonium sulfate (lane 2) but the super-natant fraction (lane 3) contained a large amount of sTF of molecular weight 39 kDa. The Coomassie stained gel (a) shows three minor impunty proteins were removed in the 55% precipitate (lane 2) but a major 65 kDa impunty remained in the supernatant fraction with the sTF (lane 3). 45 A photograph of the SDS-PAGE and Western blot of the supernatant and precipitate after 55% ammonium sulfate fractionation of the concentrated tissue culture media is shown in Figure 16. Comparison to low MW standards show that the BHK-derived sTF runs with an apparent MW of 39 kDa indicating a glycosylated form (70) as expected with a eucaryotic expression system. Figure 16a shows only a small amount of purification occurred with the 55% ammonium sulfate cut. Three minor protein impurities (65 kDa, 60 kDa, 25 kDa) were removed in the precipitate but a very large impurity of approximately 65 kDa remained in the supernatant fraction containing sTF. However, the immunoblot in Figure 16b indicated that no sTF precipitated at 55% ammonium sulfate. 5. High Q FPLC Purification In order to remove the major 65 kDa impurity (likely bovine serum albumin (BSA) from the media), FPLC purification was employed. Using the PC-Gene computer application, the isoelectric point of sTF was determined to be pH 4.88. Thus an anion-exchange column was chosen for purification. A MonoQ FPLC column using a phosphate buffer (pH 6.0) and continuous NaCl gradient failed to separate the sTF from the major impurity in initial attempts with anion-exchange FPLC. However, changing to a High Q FPLC cartridge using a Tris-buffer (pH 7.5) and continuous NaCl gradient resulted in excellent resolution between sTF and the major impurity. The purification profile of the High Q FPLC run is shown in Figure 17 along with SDS-PAGE and Western blot analysis of the collected fractions. The salt gradient was held constant during elution of each peak. As can be seen from the results of the SDS-PAGE and Western blot, sTF eluted in the first peak at approximately 100 mM NaCl along with some other minor (but significant) impurities. The major 65 kDa impurity eluted in the very large second peak at 160 mM NaCl; the resolution between the peaks was excellent Fractions containing the semi-purified sTF were pooled. 46 Figure 17. High Q FPLC purification. The ammonium sulfate fraction was purified using a High Q FPLC cartridge and a linear salt gradient of 20 - 500 mM NaCl in 20 mM Tris-HCl, pH 7.5 buffer. The salt gradient was held constant during peak elution; four peaks eluted at 100, 160, 210, and 285 mM NaCl as shown in the elution profile (a). Fractions from peak #1 and #2 were run on a 12% SDS-PAGE and stained with Coomassie blue (b) or Western blotted using an antibody specific for sTF (c). The results indicate that sTF eluted in Peak #1 along with a high MW doublet impurity. However, the major 65 kDa impurity (BSA) was separated and eluted in Peak #2. Lane 1 is the crude ammonium sulfate fraction, lane 2 is the FPLC flow-through, and lane 11 is MW markers. 47 Peak £2 1 160 mM.. I a II /•' I H8&1.1M,. , / • Peats #3 Poak J»4 1 « I I I |< Peak#l >| Pk#2 1 2 3 4 5 6 7 8 9 10 11 12 13 / High M W Doublet |< Peaktfl >| Pk#2 1 2 3 4 5 6 7 8 9 10 11 12 13 ( c) -130kDa -90kDa -68 kDa - 43 kDa - 29 kDa Major 65 kDa Impurity sTF 48 One disadvantage of the High Q FPLC cartridge was that it had a limited capacity for total protein loaded. If overloaded, sTF would elute with the flow-through during loading of the crude sample because of the relatively low affinity of sTF for the column. Thus, during preparative purification of sTF, two High Q cartridges were connected in series and a maximum of 100 mg of total protein was loaded onto the column. Several runs had to be performed to purify all of the protein. A second problem was caused by the pH-indicator dye in the tissue culture media. The phenol red dye bound very strongly to the anion-exchanger during loading. This was one of the main reasons the column was so easily overloaded. As well, since the dye bound so strongly, washing the column with concentrated salt (2 M NaCl) solutions for long periods was necessary to re-generate the anion-exchanger. To avoid the problems caused by the dye, DMEM-F12 media without indicator dye can be purchased from Gibco-BRL and used for tissue culture. Another choice is to perform an initial batch purification of the ammonium sulfate fractionated sample on a DEAE-Sepharose equilibrated in 20 mM Tris-HCl, pH 7.5, 100 mM NaCl (16) to remove the dye prior to loading onto the FPLC column. In this case, the dye would bind to the gel and the sTF would flow through. 6. Sephacryl S-100 HR Gel Filtration Purification Following FPLC, the High Q-purified sample still contained some impurities in the form of a high MW doublet of approximately 70 kDa as seen in Figure 17b (lanes 4, 5, 6). Thus for the final stage of purification, a size exclusion column was used to separate the relatively small (39 kDa) sTF protein from the high MW doublet impurity. A high resolution Sephacryl S-100 HR gel filtration column with a height-to-diameter ratio of 20:1 and bed volume of 300 ml was prepared. Void volume for the column was determined to be 105 ml using blue dextran and calibration using a mixture of protein standards indicated 49 excellent resolution between proteins of 64.5 kDa (hemoglobin), 45 kDa (ovalbumin), and 20 kDa (trypsin inhibitor) molecular weight as shown in Figure 18. 0.4 Hemoglobin -Tryps in Inhibitor E c o CO CM 0.3 H d> o 0.2 H a o w .a o.H < 0.0 0 1 0 20 30 40 Fraction number Figure 18. Calibration curve for Sephacryl S-100 HR Column. Elution profile of protein standards run on Sephacryl S-100 HR size exclusion column using 100 mM NaCl, 20 mM Tris-HCl, pH 7.5. The protein standards were hemoglobin (64.5 kDa), ovalbumin (45 kDa), trypsin inhibitor (20 kDa), and cytochrome c (13 kDa). Flow rate was 0.6 ml/min. Void volume for the column was 105 ml. To ensure the highest resolution, the High Q-purified sample was concentrated to 2.5 ml or 1% of the bed volume prior to loading onto the Sephacryl S-100 HR column and separation was carried out at slow flow rate of 0.6 ml/min. The A 2 8 0 profile of the gel filtration purification run is shown in Figure 19a. Surprisingly, there was some overlap between the elution peaks of the high MW doublet (70 kDa) and sTF (39 kDa). Arrows indicate elution volumes of the protein standards used for calibration. Fortunately, SDS-PAGE and Western blot (Figure 19b and 19c) indicated 50 Figure 19. Sephacryl S-100 HR Purification. The High Q fraction (2.5 ml) was loaded onto the Sephacryl S-100 HR size exclusion column and eluted with 100 mM NaCl, 20 mM Tris-HCl, pH 7.5 at a flow rate of 0.6 ml/min. The elution profile (a) shows that the two peaks were not completely resolved. However fractions from peak #1 and #2 were run on a 12% SDS-PAGE and stained with Coomassie blue (b) or Western blotted using an antibody specific for sTF (c). The results of the electrophoresis indicate that fractions containing pure sTF (lanes 8 - 14) were isolated. The reason for the poor resolution despite the large apparent size difference was due to the extended structure of sTF. 51 (a) 2 Fract ion number (b) , 4 5 6 (0 9 10 11 12 13 14 15 1 2 3 4 J -90 kDa v J68 kDa - 43 kD a 6 7 8 9 10 11 12 13 14 15 29 k D a Pure sTF Pure sTF Low M W sTF 52 that sTF had successfully been separated from the high MW doublet in the collected fractions with nearly no overlap. Thus, fractions containing the pure sTF were pooled. The poor resolution between the sTF and the high MW doublet was surprising since the size exclusion column had successfully resolved hemoglobin and ovalbumin, two proteins with an apparent molecular weight difference smaller than sTF and the high MW doublet. However, consideration of the recently published X-ray crystal structure of sTF (20, 21) revealed the reason for the apparent discrepancy. Soluble tissue factor has an elongated structure, approximately 115 A long and 40-50 A wide (Figure 3). Thus, the apparent Stokes radius of the sTF protein is much larger than a globular protein of similar molecular weight. Since size exclusion chromatography actually separates proteins by Stokes radius (71) and not molecular weight, sTF behaves like a protein much larger (ie. higher MW) than its apparent molecular weight of 39 kDa on SDS-PAGE. This resulted in the poor resolution between sTF and the high MW doublet impurity. A longer, narrower column would improve resolution for future size separations using the same gel filtration matrix. Sephacryl S-100 HR is an excellent size exclusion material that offers high resolution due to its small bead size (25-75 um) and mechanical strength due to its extensive cross-linking. The fractionation range of 1,000 to 100,000 suits the purification requirements well. 7. Characterization of sTF Protein Figure 20 shows an SDS-PAGE at different stages of purification of sTF from the tissue culture media. The crude tissue culture media in lane 1 was dominated by the major BSA impurity (65 kDa) which was removed after the High Q FPLC purification (lane 4). Gel filtration on a Sephacryl S-100 HR column removed the remainder of the impurities (lanes 5-8). 53 1 2 3 4 5 6 7 8 9 . 130 kDa - 90 kD a - 68 kD a - 43 kDa 9 kDa / « / \ Major 65 kDa Impurity H i g h M W doublet sTF L o w M W sTF Figure20. Purification Summary gel. Samples from different stages of purification of sTF were run on a 12% SDS-PAGE and stained with Coomassie blue. Lane 1 shows the crude tissue culture media, lane 2 shows the supernatant of the 55% ammonium sulfate fractionation; lane 3 shows a (trial) batch purification, lane 4 shows the results after HighQ FPLC purifi cation, lanes 5-8 show several fractions of the purified sTF after Sephacryl S-100 HR size exclusion chromatography. Lane 9 shows the molecular weight markers In lane 8, a high M W (39 kDa) and a low M W (35 kDa) form of sTF are clearly visible. 54 The purified sTF consists of a major band of 39 kDa. However, Western blot analysis also detects at least one slightly smaller form of sTF below the major band (Figure 19c, lanes 12-14 and Figure 20, lane 8). This minor band of apparent molecular weight 35 kDa most likely represents a partially glycosylated (or unglycosylated) form of sTF (70). Treatment with endoglycosidase F could confirm this characteristic of the minor band. The minor band behaved identically to the major band when used in the sTF-reagent for the factor Vila assay. Amino-terminal sequence analysis of the purified sTF determined two peptide sequences: Ser-Gly-Thr-Thr-Asn-Thr-Val and Thr-Thr-Asn-Thr-Val-Ala-Ala in equal proportions. The second sequence is identical to the first sequence from the third residue (Thr) on. The two sequences represent two alternate processing sites as described for natural full length tissue factor (15). Thus the N-terminal sequence of the purified protein is in agreement with the expected composition of sTF. The N-terminal sequence analysis also confirmed the purity of the sample to be greater than 98% and the presence of glycosylation in the form of fucose. The concentration of sTF was determined spectrophotometrically using a 1% extinction coefficient at 280 nm of 14.8 for the pure protein (67). Purification of 2.4 litres of tissue culture media yielded 7.36 mg of pure sTF protein. Thus, the overall expression level of the sTF-BHK cell line was approximately 3 mg/litre. This amount represented about 1% of the total protein in the tissue culture media and was a significant improvement over expression levels of CHO cells for sTF (50). Fifty microlitre aliquots of the purified sTF were stored frozen at -70°C prior to use. 8. Optimization of Assay Reagents The first step in the development of the FVIIa assay was to optimize the assay reagents. The assay uses three reagents to measure the FVIIa-level in a plasma sample: the 55 sTF-reagent, CaCl2 solution, and FVII-deficient plasma. The calcium chloride solution is a standard 25 mM aqueous solution and the FVII-deficient plasma is available commercially from several manufacturers. The sTF-reagent is analogous to thromboplastin in a prothrombin time (PT) or FVII coagulant (FVILC) assay . The sTF-reagent contains the soluble tissue factor protein and phospholipid diluted in a TBS/BSA solution. One of the keys to a robust FVIIa assay is the optimization of the sTF-reagent. According to Morrissey's original paper (26), the sensitivity of the FVIIa assay could be manipulated by adjusting the sTF concentration in the sTF-reagent. A "high sensitivity" reagent containing nearly saturating concentrations of sTF was determined to be optimal. Thus to determine the optimal concentration of sTF, a set of sTF-reagents with varying concentrations of sTF was prepared and the clotting time for a normal pooled plasma (NPP) sample was measured using each. A plot of clotting time versus sTF concentration is shown in Figure 21. Clotting times began to plateau at around 0.1 uM sTF concentration and reached a minimum at 0.5 uM sTF. This was lower than the 1.0 uM sTF required in the original assay to achieve saturating levels (26). An sTF-negative reagent failed to produce a clot in 400 seconds confirming the BHK-derived sTF as the active clotting agent in the sTF-reagent. The original assay used rabbit brain cephalin (RBC) as a phospholipid source in the sTF-reagent. Since Sigma no longer sells RBC, a phospholipid replacement was required. Initially a PC/PS (75:25) pure phospholipid mixture was used. However, it was found that this composition was not particularly stable to freezing resulting in variable clotting times after a single freeze-thaw cycle. According to a recent paper by Neuenshwander et al (72), the addition of phosphatidylethanolamine (PE) to a standard PC/PS phospholipid mixture enhanced the activity of the FVIIa-TF complex in its activation of factor X and factor IX. An sTF-reagent containing a PC/PS/PE (64:6:30) mixture was prepared and found to be extremely stable to freeze-thaw. To determine the optimal concentration of PC:PS:PE 56 300 0 H • 1 • 1 • 1 • 1 • 1 • 1 0.0 0.2 0.4 0.6 0.8 1.0 1.2 sTF-conc (uM) Figure 21. Clot time vs. sTF concentration. Clotting times of a neat citrated pooled plasma sample were measured using an sTF-reagent composed of 0.27 mM PC/PS/PE (64:6:30), TBS/BSA with varying concentrations of sTF (0.01 - 1.0 uM). Clotting times began plateauing at 0.1 uM and reached a minimum of aprox. 0.5 uM sTF. (64:6:30) phospholipid mixture, a set of sTF-reagents with varying concentrations of PC/PS/PE was prepared and the clotting time for a normal pooled plasma (NPP) sample was measured using each. A plot of clotting time versus phospholipid concentration is shown in Figure 22. Clotting times reached a plateau at approximately 1.0 mM PC/PS/PE (64:6:30). An updated protocol for the FVIIa assay provided by Dr. Morrissey describes the use of a 0.2 mM PC/PS/PE (40:40:20) mixture as the phospholipid source (51). Comparison of the two PC/PS/PE compositions revealed that the 0.2 mM PC/PS/PE (40:40:20) mixture yielded 30% shorter clotting times than the, 1.0 mM PC/PS/PE 57 (64:6:30) using the same sTF concentration. The shorter clotting times obviously were a result of the higher composition (40%) of phosphatidylserine (PS), a negatively charged phospholipid. The danger of using such a high concentration of negatively charged phospholipids (ie. 60%) is that the negative phospholipids could initiate contact activation of coagulation which could in turn activate FVU to FVIIa via factor XUa (37). This would result in a misleading increase in the measured plasma FVIIa levels. The main objective of optimizing the sTF-reagent composition was to achieve a reproducible assay with convenient clotting times while conserving valuable sTF. An sTF-reagent of composition 0.5 uM sTF, 1.0 mM PC/PS/PE (64:6:30) in TBS/BSA gave convenient clotting times of approximately 60 seconds for neat NPP samples and 120 seconds for 10-fold diluted NPP samples using the automated assay and thus was chosen as the sTF-reagent for the FVIIa assay. 140 T 1 u CD 40 H 1 1 • 1 • 1 0 1 2 3 PC/PS/PE-concentration (mM) Figure 22. Clot time vs. Phospholipid concentration. Clotting times of a neat citrated pooled plasma sample were measured using an sTF-reagent composed of 0.5 uM sTF, TBS/BSA with varying concentrations of PC/PS/PE (64:6:30) phospholipid (0.05 -2.0 mM). Clotting times reached a minimum at approx. 1.0 mM PC/PS/PE. 58 9. Factor Vila Standard Curve In order to quantitate FVIIa levels, a calibration curve to convert clotting times to factor Vila concentration was prepared. Initially, FVIIa standards of 30, 3.0, 0.30, and 0.03 ng/ml were prepared by serial dilution in Sigma immuno-depleted FVII-deficient plasma. However, as shown in Figure 23, the calibration curve began to plateau at FVIIa concentrations below 0.10 ng/ml. 2.2 T : 1 1.4 H • 1 1 1 r • 2 - 1 0 1 2 Log FVIIa coiic Figure 23. FVIIa standard curve using Sigma FVII-def plasma. Measurement of clotting times of FVIIa calibration standards of 30, 3, 0.3, and 0.03 ng/ml FVIIa prepared in Sigma brand immuno-depleted FVII-deficient plasma using the FVIIa assay. A plateau in clotting times is apparent at FVIIa concentrations below 0.10 ng/ml due to the high endogenous levels of FVIIa in the Sigma FVII-deficient plasma. Measurement of the FVIIa level in the FVII-deficient plasma revealed that there was 0.06 -0.07 ng/ml residual FVIIa in the Sigma product. When the FVIIa standard curve was corrected for this endogenous FVIIa, a linear curve resulted. 59 However, the endogenous FVIIa would cause significant deviations in measured plasma FVIIa levels due to the fact that plasma samples would be diluted 10-fold in the Sigma FVII-deficient plasma. For example, a plasma sample containing 2.0 ng/ml FVIIa diluted 10-fold would be expected to give 0.2 ng/ml. However, the endogenous FVIIa in the FVII-deficient plasma used for dilution would increase the apparent FVIIa concentration to 0.27 ng/ml. After correction for dilution, the measured FVIIa would be 2.7 ng/ml, significantly different from the actual FVIIa concentration. Thus, congenital FVII-deficient plasma was purchased from George King Biomedical as an alternative source. A FVIIa calibration curve prepared using FVIIa standards serially diluted in the new FVII-deficient plasma is shown in Figure 24. 2.4 -S 2.2-** fit) .5 2 - ° " "•3 © 5 1.8-O J 1.6-1.4 -- 2 - 1 0 1 2 Log FVIIa cone Figure 24. FVIIa standard curve using George King FVII-def plasma. Measurement of clotting times of FVIIa calibration standards of 30, 3, 0.3, and 0.03 ng/ml FVIIa prepared in George King Biomedical brand congenital FVII-deficient plasma using the FVIIa assay. Only a small plateau in clotting times was apparent at low concentrations of FVIIa due to the very low levels of endogenous FVIIa in the George King plasma. 60 There was still a slight plateau at low FVIIa levels but the George King Biomedical FVTJ-deficient plasma provided a significant improvement over the Sigma product Measurement of the endogenous FVIIa revealed that there was only 0.02 ng/ml FVTJa in the new FVII-deficient plasma. Although not completely negligible, this was judged to be acceptable for the dilution of normal plasma samples. The FVIIa standard curve was corrected for the endogenous FVIIa by plotting the FVIIa standards as 30.02, 3.02, 0.32, and 0.05 ng/ml FVTJa. The final FVUa standard curve used for the assay is shown in Figure 25. CALIBRATION Factor UII Activate! Date 61/27/1937 18:65:26 Control 1 Control 2 Calibrators Rom Data Interpol. ng/al sec. ng/itl 39.82 26.9 28.82 36.62 27.3 27.51 3.62 53.3 3.31 3.62 53.2 3.33 6.32 111.1 6.32 6.32 111.3 6.32 6.65 263.6 6.65 6.65 261.1 6.65 1 r-0.05 0.32 3.02 oancfMrxtion no^ l LogCO - <- 3.1(7 • UgCt» • 5.988 r » - 1.000 Figure 25. Final FVUa Standard Curve. Factor Vila standard curve prepared by STA coagulometer calibration program using corrected FVIIa standard concentrations of 30.02, 3.02, 0.32, and 0.05 ng/ml diluted in George King Biomedical FVII-deficient plasma. This curve was used by the STA coagulometer to convert clotting times to FVUa concentrations in ng/ml. The linear regression coefficient for the standard curve of -1.000 shows a ^-magnitude linear range for the FVTJa assay. Clot times ranged from 26 seconds to 200 seconds for the 61 measurable range with normal 10-fold diluted plasma samples falling in the middle of the range. Plasma FVIIa concentrations as low as 20 pg/ml could be measured indicating an extremely sensitive assay. The FVIIa standard curve shown in Figure 25 was used by the STA Coagulometer to automatically convert clotting times of plasma samples into FVIIa concentrations in ng/ml. 10. Comparison of BHK-derived sTF and E.co/i-derived sTF As previously mentioned, the original FVIIa assay employed a recombinant sTF protein expressed in E.coli. (26). To study the effects of glycosylation on sTF-function, sTF-reagents containing BHK-derived sTF and E.co/i-derived sTF (a gift from Dr. Morrissey) were prepared and several FVIIa standards and plasma samples were measured using each sTF-reagent. Figure 26 shows the results of linear regression analysis of the FVIIa concentrations measured using each sTF-reagent. Measurement of the clotting times for the FVIIa standards (Figure 26a) using each sTF-reagent gave excellent agreement with a linear regression coefficient of r = 1.00. Figure 26b shows the results of measurement of the 5 normal plasma samples. Linear regression analysis shows the correlation between the FVIIa values measured using either sTF-reagent was also extremely good, r = 0.976. The results indicate that BHK-derived sTF behaved identically to the E.co/i-derived form and glycosylation did not appear to affect the co-factor activity of sTF nor its activity in the FVIIa assay. This was in agreement with the findings of Dr. Morrissey (26) and others (50, 70). 11. Testing of FVIIa Assay The next stage of development of the FVIIa assay was to establish the precision and accuracy of the assay. Reproducibility studies established within-run and between-run coefficients of variation of 3.9% and 4.9%, respectively, for the present study's FVIIa 62 Clot times w/ E.coli-sTF (sec) FVIIa w/ E.coli-sTF (ng/ml) Figure 26. Comparison of BHK and E.co/i-derived sTF in FVIIa Assay. Results of linear regression analysis of (a) measurement of clotting times of 4 FVIIa calibrators (30, 3, 0.32, 0.05 ng/ml FVIIa) and (b) measurement of FVIIa concentrations of 5 normal citrated plasma samples using sTF-reagents composed of 1.0 mM PC/PS/PE (64:6:30), TBS/BSA, and either 0.5 uM BHK-derived sTF or 0.5 uM £.co//-derived sTF. assay. This compared very well with Morrissey's results of 2.2% and 8.1% (26), Kario's results of 1.3% and 4.2% (27), and Phillippou's results of 4.5% and 9.8% (25), respectively. A patient comparison study was performed to establish the accuracy of the FVIIa assay. Since no absolute FVIIa standards were available as a comparison, a commercial STACLOT VIIa-rTF kit from Diagnostica Stago was used to define the FVIIa levels of a series of citrated plasma samples. The Stago kit measured FVIIa in international activity units (mU/ml) and included a sample of the FVIIa international standard (52). Since it is the only commercial kit available on the market for measuring FVIIa, the Stago kit was regarded as the standard by which a new FVIIa assay should be tested. Although the kit also used a clot-based method of detection, the assay procedure (which was also automated 63 on the STA Coagulometer) was slightly different. Plasma samples were diluted 10-fold in a physiological buffer and then diluted 1:1 with FVII-deficient plasma prior to measurement. The present study's FVIIa assay diluted plasma samples 10-fold directly in FVII-deficient plasma. However, since the FVIIa standards and test plasma samples were treated identically in each assay, the difference in dilution protocols should not have an effect on the measured FVIIa levels. The results of linear regression analysis of the patient comparison study are shown in Figure 27. 0 200 400 600 800 0 20 40 60 80 100 STAGO FVIIa (mU/ml) STAGO FVIIa (mU/ml) Figure 27. Patient comparison study results. Results of linear regression analysis of measurements of 19 citrated plasma samples containing 0.1 - 30 ng/ml FVIIa using present FVIIa assay and STACLOT VIIa-rTF kit (a). Two spiked high FVIIa samples (15 and 30 ng/ml) were included. The normal human plasma samples gave FVTIa values of 5.0 ng/ml (100 mU/ml) or less. Linear regression analysis of only these 17 normal plasma samples are shown in (b). The plasma samples were chosen to achieve a broad range of FVIIa levels. To assess accuracy at extremely low FVIIa levels, plasma samples were diluted 10-fold in FVII-64 deficient plasma before loading for measurement. To determine accuracy at extremely high FVIIa levels, two plasma samples were spiked with FVIIa to give concentrations of 15 and 30 ng/ml FVIIa. A total of 19 samples were measured using both assays. As shown in Figure 27, over the full range of FVIIa (0.1 - 30 ng/ml), correlation between the two assays was excellent, r = 0.998. When considering only the normal human plasma samples, the results were still extremely good, r = 0.995. It should be noted that the FVIIa calibration curve of the Stago kit had a smaller range than the present study's FVIIa assay. The upper limit of the kit's standard curve was 250 mU/ml which corresponded to approximately 10 ng/ml FVIIa. Thus, the 15 ng/ml FVIIa sample was off-scale and its corresponding international activity level of 400 mU/ml had to be determined manually. The limited FVIIa ceiling was partly due to the faster clotting times of the STACLOT kit. As FVIIa levels increase, the clotting times decrease. For example, the 30 ng/ml FVIIa sample clotted in 26.8 seconds with the STACLOT kit compared to 54.4 seconds using the present FVIIa assay. At extreme FVIIa levels, clot times would reach a lower limit using the STACLOT kit more quickly than with the present study's FVIIa assay. This would limit its use in measurement of extremely high FVIIa levels. The inclusion of the FVIIa international standard in the Stago kit gave the opportunity to establish a conversion curve for FVIIa concentrations. Since the results of the patient comparison study were so accurate, it was concluded that both assays were indeed measuring the same relative level of FVIIa but expressing them in different units. Thus, the linear regression curve of Figure 27 could be used as a FVIIa conversion curve to convert FVIIa concentrations measured in ng/ml to international activity units in mU/ml to allow for better inter-laboratory comparison. 65 12. Plasma FVIIa Stability Testing It was important to clearly establish the method of handling plasma samples to ensure the stability of plasma FVIIa. To determine whether samples that were measured after being freshly drawn (as often is the case in a hospital) could be compared to plasma samples placed in frozen storage, samples were measured immediately after being drawn and again after varying lengths of storage at -70°C. Table la below shows the results of the experiment. Table I. Effects of freezing on plasma' FVIIa levels8, a. Results of fresh vs freeze-thawed Sample Fresh FVIIa Norm lC:pm 3.1 ng/ml Norm lE:pm 3.2 ng/ml Norm lE:pm 3.2 ng/ml Time Frozen 15 minutes 2 days 7 days Freeze-thawed FVIIa 3.1 ng/ml (100%) 3.0 ng/ml (94%) 3.1 ng/ml (97%) b. Results of repeated freeze-thaw cycles. Number of Freeze-Thaw Cycles Sample One Two Three Norm lC:am 2.2 ng/ml 2.0 ng/ml 1.0 ng/ml (91%) (45%) Norm lE:pm 3.0 ng/ml 2.8 ng/ml 3.2 ng/ml (93%) (107%) Norm lE:pm 3.1 ng/ml 3.4 ng/ml 3.1 ng/ml (110%) (100%) a Plasma FVIIa levels measured using present study FVIIa assay after indicated time of frozen storage and/or number of freeze-thaw cycles. Values in brackets refer to relative% FVIIa measured compared to fresh sample (a) or single-freeze-thaw sample (b). 66 The results indicate that plasma FVIIa levels are not affected by a single freeze-thaw cycle nor by the length of frozen storage; the differences between fresh and freeze-thawed FVIIa levels are within the inter-assay variations of 4.9%. To test whether further freeze-thaw cycles would affect plasma FVIIa levels, the samples were subjected to two more freeze-thaw cycles. The results in Table lb show that FVIIa levels were reasonably stable after even two freeze-thaw cycles but the change in FVIIa could not be explained solely by inter-assay variation. However, after the third freeze-thaw cycle, gross changes in FVIIa levels were observed in at least one sample (Norm lC:am). In most assays for hemostatic variables (PT, APTT, FVII:C, etc.), plasma samples are measured fresh or after.at most a single freeze-thaw cycle (73). Thus, plasma samples for the clinical study in this investigation were restricted to a maximum of one freeze-thaw cycle. Another important consideration for the stability of the plasma samples is the anticoagulant used during blood collection. Typically;- plasma used for measurement of hemostatic variables is collected in citrate. To determine whether plasma samples collected in EDTA tubes could be used for measurement of FVIIa levels, a plasma sample from the same, donor was collected in a citrate tube and a second collected in an EDTA tube and both were measured using the FVIIa assay. The results of precision testing on both samples are shown in Table II. The results show very poor agreement between the EDTA and citrated sample in measured FVIIa level. In addition, inter-assay reproducibility of the EDTA sample was very poor with a coefficient of variation of 21.4% versus 4.9% for the citrated sample. EDTA is a much stronger chelator than citrate and thus requires much higher concentrations of calcium to overcome its anti-coagulant effects (74). As well, EDTA can cause clumping of platelets in plasma that may also affect the measured FVUa levels. It was concluded that 67 Table II. bComparison of EDTA vs. Citrate as plasma anticoagulant. Avg FVIIa-conc Intra-assay C V . Inter-assay C V . Citrated plasma 1.65 ± 0.08 ng/ml 3.9% 4.9% EDTA plasma 2.19 ± 0.47 ng/ml 7.1% 21.4% b Measurement of a citrated and EDTA plasma sample collected from a single donor at the same collection using the FVIIa assay. Mean FVIIa with standard deviation and results of intra-assay and inter-assay precision testing for reproducibility are shown. EDTA was a poor choice as an anti-coagulant for FVIIa measurement. Thus, plasma samples were collected only in citrated tubes for the clinical study. Plasma samples were routinely diluted 10-fold in FVII-deficient plasma for measurement of FVIIa and then the measured FVIIa concentration was multiplied by 10 to correct for the dilution. However, the expense of FVII-deficient plasma suggested measurement of plasma samples neat. The results of a comparison between plasma samples measured undiluted versus 10-fold diluted (data not shown) indicated mixed results. In some cases the measured FVIIa levels agreed between undiluted and diluted measurements whereas in other cases undiluted samples gave higher or lower readings for FVIIa when compared to the 10-fold diluted samples (after correction for dilution). i Consideration of the mechanism of coagulation suggested the difference in FVIIa between undiluted and diluted samples was caused by differences in the amounts of clotting factors other than FVIIa (such as fibrinogen, FX, etc.). If the test plasma had different amounts of the other clotting factors than in the FVII-deficient plasma in which the standards were diluted, then these would also influence the clotting time. Samples that gave similar FVIIa levels whether diluted or not probably had levels of the other clotting factors (other than FVII) that were similar to those in the FVII-deficient plasma. Thus, the clotting times of samples that are measured undiluted are determined not only by their 68 endogenous level of FVIIa but also by the levels of other clotting factors. Ten-fold dilution ensures that 90% of the other clotting factors come from the same source, the FVII-deficient plasma. Thus, the differences caused by the other clotting factors (which also vary from individual-to-individual) are diluted out and only differences between FVIIa levels will affect the clotting times and thus the measured FVIIa concentration (26). For this reason all test plasma samples were diluted 10-fold in FVII-deficient plasma. Although formal studies were not carried out, the effects of cold activation and post-prandial hyperlipidemia on FVIIa levels were examined briefly. The results indicated that prolonged storage of plasma samples at 4°C caused an increase in FVIIa levels. As well, plasma samples drawn 3 hours or more after a high fat meal had elevated FVIIa levels compared to the fasting state. However such an increase in FVIIa levels was not observed after light (low fat) meals up to 2 hours before collection. These results are in agreement with several other studies which found that cold activation (26, 27) and post-prandial hyperlipidemia (38, 39, 40, 41, 55) are associated with an increase in FVIIa. 13. Intra-individual variations in plasma FVIIa To study the effects of individual variations in FVIIa, six citrated plasma samples were collected over a 3 month period from the same donor. The results are shown in Table III. Table III. Intra-individual variation in cFVIIa levels. Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 FVIIa 2.2 ng/ml 2.1 ng/ml 2.0 ng/ml 2.3 ng/ml 2.2 ng/ml 2.0 ng/ml c Measurement of FVIIa levels in six citrated plasma samples collected from a single donor over a period of 3 months using the FVIIa assay. Values are from duplicate measurements agreeing within 7%. Average FVIIa level is 2.13 ± 0.12 ng/ml (CV. = 5.6%). 69 The results indicate that FVTIa levels are quite stable within the same donor. The variations are not within inter-assay variability and thus there is a small amount of real fluctuation in FVUa levels within an individual donor. However, generally it appeared that FVIIa levels were very characteristic of an individual. This was also in agreement with previous studies (26, 27). 14. Factor Vila Levels in a Normal Population The first stage of the clinical study on FVIIa was to establish the normal range for FVIIa and to determine the effects of age and gender on FVIIa levels. A population study was undertaken in which plasma samples were collected from 50 healthy donors aged 20 -52 years. The overall mean FVIIa level was 2.08 ng/ml (42 mU/ml) with a standard deviation of 0.90 ng/ml (15 mU/ml). The range of FVIIa in the study group was 0.7 ng/ml (11 mU/ml) to 4.8 ng/ml (100 mU/ml) corresponding to 33.7% to 231% of the mean value. Previous studies by others have determined mean FVIIa levels to be 3.48 ng/ml (26), 2.5 ng/ml (27), and 4.1 ng/ml (53). The lower mean FVIIa level in the present study could be a result of the smaller population size surveyed or simply characteristic of a different population than in the other studies. However as Figure 28 shows, FVIIa levels increased with age. The average age of the donors in the present study was 35.4 ± 8.9 years which was significantly lower than the studies by Kario (57.9 years) and Scarabin (45.5 years) but similar to Morrissey's group (36.2 years). However, the present study did not include any elderly patients (over 55) who would be expected to have higher FVIIa levels as indicated by the trend in Figure 28 and previous studies (26, 27, 53). Thus the absence of elderly subjects could have resulted in the lower mean FVIIa value recorded in the present study. As Figure 28 shows, a strong correlation between increasing age and FVIIa levels existed (r = 0.443 for males, r = 0.390 for females). 70 • Males (n=21) • Females (n=29) R = 0.443 for Males R = 0.390 for Females 10 20 30 40 50 60 Age (years) Figure 28. Relationship between Factor Vila and Age. Plasma FVIIa levels in a normal population of 50 healthy subjects (21 male, 29 female, average age = 35.4 year, age range = 20 - 52 years) as a function of age. Linear regression coefficients (R) are indicated separately for males and females. Mean factor Vila of the normal population was 2.08 ± 0.90 ng/ml. The association between increasing age and FVIIa had very similar patterns in both males and females as shown by the two lines in Figure 28. In addition, there was no significant differences between the FVIIa levels of men and women in terms of mean FVIIa. Statistics on sub-populations (by age and gender) of the study subjects are summarized in Table IV. 71 Table IV. Age Distribution and FVIIa levels in Study Sub-populations'1 Sub-population Mean FVIIa Average Age Group size All Normals 2.08 ± 0.90 ng/ml 35.4 ± 8.9 years n = 50 Normals, <40 yrs 1.82 ±0.76 ng/ml 29.8 ± 5.2 years n = 32 Normals, >40 yrs 2.55 ± 0.97 ng/ml 45.3 ± 3.9 years n = 18 All Females 2.10 ±0.90 ng/ml 34.7 ± 8.4 years n = 29 All Males 2.07 ± 0.94 ng/ml 36.3 ± 9.6 years n = 21 All Females, <40 yrs 1.86 ± 0.75 ng/ml 30.1± 5.2 years n = 20 All Females, >40 yrs 2.62 ± 1.00 ng/ml 45.0 ± 3.3 years n = = 9 All Males, <40 yrs 1.76 ± 0.80 ng/ml 29.3 ± 5.3 years n = 12 All Males, >40 yrs 2.48 ±0.99 ng/ml 45.7 ± 4.6 years n = 9 d Mean ± Standard deviation of plasma FVIIa levels for various sub-populations of the study group. Mean ± SD for age and number of subjects per sub-population are also indicated. No significant difference between males and females was observed in any of the sub-populations. In both genders, FVIIa levels increased with age at approximately the same rate. Corresponding male and female sub-populations in Table IV were closely age-matched. Thus, gender did not seem to affect FVIIa levels in the study population. 72 15. Factor Vi la Levels in Complicated Pregnancy Patients Clinical details on nine patients who experienced complications during pregnancy are summarized in Table V. Table V. Clinical details of Complicated pregnancy patients Sample H-26 H-77 H-39 H-16 H-106 H-61 H-134 H^tl H-179 Comments On Heparin Had anti-cardiolipin antibody 22+ weeks Post-partum Hemorrhage New-bom had hemolytic disease Age 40 years 32 years 26 years 35 years 22 years 32 years 36 years 36 years 28 years Clinical Stage Ante-partum Ante-partum Ante-partum Post-partum Post-partum Post-partum Post-partum Post-partum Post-partum Platelet count N/A 311 N/A N/A 450 170 126 191 267 The patients were either full-term and plasma samples were drawn following a complicated delivery (post-partum) or mid- to late-term and experienced complications during their pregnancy (ante-partum) when the plasma samples were taken. Normal hemostatic variables measured may have included platelet counts, prothrombin time (PT), activated partial thromboplastin time (APTT), and plasma fibrinogen (FIBR) at the physician's discretion. FVUa levels were measured in all subjects. The results are shown in Table VI. 73 Table VI. Hemostatic variables of complicated pregnancy samples Sample PT (10.4-12.2) APTT (23.3-32.8) FIBR (3.75-6.3) FVIIa (0.7-4.8) H-26 12.1 sec 31.5 sec N/A 0.9 ng/ml H-77 10.9 sec 27.4 sec N/A 3.9 ng/ml H-39 10.6 sec N/A N/A 1.8 ng/ml H-16 10.7 sec 24.9 sec . 6.7 6.2 ng/ml H-106 10.5 sec 31.9 sec N/A 4.8 ng/ml H-61 11.1 sec 22.0 sec N/A 2.2 ng/ml H-134 N/A 27.0 sec 4.16 5.2 ng/ml H^U 9.5 sec 27.5 sec 7.5 6.0 ng/ml H-179 10.6 sec 31.2 sec N/A 6.0 ng/ml Overall, for the 9 pregnancy samples the mean FVIIa was 4.11+2.0 ng/ml (85±40 mU/ml) which was significantly higher than the mean FVIIa for normal subjects. This agrees with previous studies by Morrissey on FVIIa in pregnancy (26) and supports the general increase in hemostatic activity accompanying pregnancy; factor VII levels normally increase to 150 - 250% of normal levels during pregnancy (75) as a protective mechanism. However, there appeared to be a few obvious outliers among the pregnancy samples. Specifically, samples H-26 (0.9 ng/ml), H-61 (2.2 ng/ml), and H-39 (1.8 ng/ml) exhibited significantly lower FVIIa levels than the other pregnancy samples. Closer inspection of the patient details in Table V provided explanations for the outliers. Patient H-26 was on heparin therapy, an anticoagulant. Factor Vila levels have been previously shown to be lowered in patients on anticoagulants (26). The PT and APTT times were also extended indicating the anti-clotting effects of heparin. Patient H-61 suffered from post-partum hemorrhage indicating failure of the protective hemostatic mechanisms that normally stops excessive bleeding after delivery. The normal level of FVIIa in the patients blood was an indication of the unprepared state of the hemostatic system to the massive blood loss accompanying parturition. Patient H-39 was only 22 weeks or 5 months into her pregnancy and thus activation of the hemostatic system had not reached its peak. 74 However, other undocumented complications in the patient may also have influenced the low plasma FVIIa concentration. After removing the 3 outliers, the mean FVIIa concentration for the remaining pregnancy samples was 5.35 ± 0.89 ng/ml (115 ± 15 mU/ml) which was the same mean FVIIa level measured by Morrissey (26) in pregnant subjects. This was a very significant increase over the mean FVIIa of 1.86 ± 0.75 ng/ml (37 ± 12 mU/ml) in an age-matched group of non-pregnant female subjects. There did not appear to be any significant correlation between FVIIa levels and any of the other hemostatic variables measured in Table VI. An additional 10 plasma samples from complicated pregnancy patients were received several weeks after the initial study. Unfortunately, complete clinical details (as in Table V) were riot available for the new patients at the time of publication. However, FVIIa concentration, PT, and APTT were measured for each sample. The results are presented in Table Via. Table Via. Hemostatic variables for new complicated pregnancy samples Sample Age PT APTT FVIIa 10 22 10.8 s N/A 2.6 ng/ml 12 30 11.1 s 26.2 s 3.0 ng/ml 13 33 11.0 s 30.6 s 5.1 ng/ml 14 31 10.7 s 25.5 s 4.6 ng/ml 15 33 N/A 44.4 s 0.8 ng/ml 16 36 10.5 s 35.3 s 4.2 ng/ml 17 30 10.1 s 28.8 s 1.9 ng/ml 18 29 11.5 s 30.2 s 1.7 ng/ml 19 34 11.5 s 27.0 s 8.7 ng/ml 20 29 11.7 s 32.5 s 7.7 ng/ml The overall mean FVIIa level for all nineteen complicated pregnancy samples was 4.07 ± 2.2 ng/ml (84 ± 45 mU/ml), similar to the mean FVIIa for the first nine samples. 75 DISCUSSION Recent studies implicating factor VII as an independent risk factor for coronary heart disease (7, 8, 28, 33) have necessitated improved techniques to characterize plasma levels of the serine protease. Conventional assays for the measurement of FVII have typically relied on thromboplastin to target the extrinsic factors. However, recent studies exposing the heterogeneity of such assays (23, 45) as well as the postulation of the existence of trace amounts of pre-activated factor Vila in plasma (76), which may be important in the initiation of clotting, have suggested the need for an assay specific for factor Vila. Until recendy, there was no way to separate the activated form of FVIIa from its zymogen precursor, FVII, using a functional assay. The development of a soluble mutant form of tissue factor (47, 50) has provided an avenue for the design of a clot-based assay specific for factor Vila. The FVIIa assay in the present study was developed based on the original FVIIa assay employing sTF reported by Morrissey et al (26). The present study's assay modified the original procedure in several important ways. First, a glycosylated form of sTF was uniquely expressed in BHK cells rather than E.coli as in the original assay. Second, a new phospholipid source had to be found to replace rabbit brain cephalin (RBC) which was used in the original paper but has since been discontinued by the manufacturer (Sigma). Third, the present study's assay was run on an automated STA Coagulometer. Finally, the present assay was calibrated to determine plasma FVIIa activity in international standard units (mU/ml) as well as in concentration (ng/ml). 1. BHK-derived glycosylated sTF Soluble tissue factor protein was successfully expressed in baby hamster kidney cells at a level of 3 mg per litre and purified using anion-exchange FPLC and gel filtration chromatography. Previous studies have used E.coli, yeast, and Chinese hamster ovary 76 cells for expression. The advantage of using a mammalian eucaryotic expression system such as BHK or CHO is that they confer post-translational modifications to the recombinant proteins similar to native human forms of the protein. The BHK-expressed form of sTF was glycosylated with a molecular weight of 39 kDa which was similar to the protein produced in CHO cells (50) and the minor form of sTF produced in yeast (70). A minor lower molecular weight form of the sTF protein was also detected by immunoblot in BHK expression. Its apparent molecular weight of 35 kDa on SDS-PAGE matched the CHO- and yeast expressed forms of sTF following treatment with endoglycosidase F (50, 70) suggesting that the minor species was a less glycosylated form of sTF. Glycosylation of sTF occurs at three N-linked sites: Asn-11, Asn-124, and Asn-137 (70). Analysis of the BHK-derived protein confirmed the presence of fucosyl residues in the sTF. Comparison with a 25 kDa unglycosylated form of sTF expressed in E.coli, however, revealed that the BHK-derived sTF behaved identically to the £.co//-derived form in the clotting assay. The 39 kDa and 35 kDa forms of the BHK-derived proteins were also functionally identical. Therefore, despite the increased molecular mass, glycosylation had no effect on the cofactor activity of sTF. This was consistent with previous studies (26, 70). However, there were other advantages of producing sTF in BHK cells over bacteria. The signal peptide included in the cloned sTF cDNA sequence ensured that the protein was secreted by the BHK cells into the growth medium. This allowed for milder purification conditions since harsh methods of homogenization were not required. Previous studies had used affinity chromatography to purify sTF (16, 67, 77). The present study showed that the high expense of affinity columns could be avoided by purifying sTF using general commercially available columns and matrices. 2. Factor V i l a Assay The FVIIa assay developed for the present study was optimized to run on an STA coagulometer using an sTF-reagent composition of 0.5 uM sTF, 1.0 mM PC/PS/PE 77 (64:6:30) in TBS/BSA. The method required only 5 ul of test plasma and 45 ul of FVII-deficient plasma per assay and as many as 100 samples could be measured simultaneously using the STA coagulometer. In addition, only 50 ul of sTF-reagent, which contains just 0.8 ug sTF, was used per assay; thus 1.0 mg of sTF would be good for over one thousand assays. The assay was sensitive to as little as 20 pg/ml FVIIa and samples up to 300 ng/ml FVIIa could be measured, giving the assay a tremendous range. Furthermore, the upper limit could easily be extended to 1000 ng/ml with no modifications. However, most plasma samples would be expected to be well within the established range. Precision testing of the assay revealed high reproducibility; within-run and between-run coefficients of variation were 3.9% and 4.9%, respectively. The FVIIa assay also proved to be extremely accurate when compared to a commercial FVIIa kit produced by Diagnostica Stago. The two assays agreed over a range of 0.1 ng/ml to 30 ng/mPFVTIa with a correlation coefficient of r = 0.998. All plasma samples measured fell within this range of FVIIa concentrations. The inclusion of an international FVIIa standard in the Stago kit allowed for the standardization of the present study's FVIIa assay. Thus samples measured for FVIIa in ng/ml could be converted to international activity units (mU/ml) using a FVUa conversion curve (Figure 29). Use of standardized units of FVIIa activity would facilitate comparison of results between different laboratories. This was a major shortcoming of previous assays for FVII coagulant activity using thromboplastin. The use of FVIIa standards from different sources for calibration of the FVIIa assay would produce problems since different methods of assigning mass concentration (eg. optical density or ELISA antigen) are often used in different labs (52). Results from studies using the present investigation's FVIIa assay can be compared directly with other FVIIa assays that have been standardized using the international FVIIa assay (including the STACLOT VUa-rTF kit). 78 40 FVIIa (mU/ml) Figure 29. Factor Vila Conversion Curve. Factor Vila values measured in ng/ml using the present study's FVIIa assay can be converted to international units (mU/ml) using this conversion curve to improve between-laboratory comparison. 3. Comparison to previous FVIIa assays Previous assays described by Morrissey (26), Kario (27), and Phillopou (25) have not made use of the international FVIIa standard. Generally, the FVIIa assay in the present study is on par with or improves upon these previous assays. Testing revealed that the present study's FVIIa assay (referred to simply as the "present assay" for this comparison) was at least as robust as the other assays. The inter-assay reproducibility of 4.9% for the present assay was significantly better than both Morrissey's and Phillopou's assays (8.1% and 9.8%, respectively). Use of the STA coagulometer ensured reproducibility because of its excellent automated duplicate run program which repeated an assay if the duplicate results did not agree within 7%. Another advantage the present assay had over the Morrissey assay was that the present assay used an sTF-concentration of 0.5 uM compared to 1.0 uM sTF recommended in the Morrissey assay. Thus, using the present assay, twice as many samples can be measured using the same amount of soluble tissue factor. 79 The present assay's sensitivity down to 20 pg/ml FVIIa is another strong point that sets it apart, especially from Kario's assay which only measures to 200 pg/ml. Another disadvantage of Kario's assay is that the fluorescent endpoint typically take 200 - 500 seconds (3-8 minutes) to be reached after addition of the initiating reagent. This results in a slow assay. The present assay's reagents were optimized to achieve convenient clotting times for the FVIIa standards (26 - 200 s) and the normal plasma FVIIa range (90 - 180 s). Typically, a single measurement required about 4 minutes total assay time (including a 2 minute incubation period). One problem of the extreme sensitivity of the present assay is that it is easily influenced by contamination or carry-over of thromboplastin from other assays. Since thromboplastin is a very sticky substance, it tends to be easily carried over on syringes or other pipetting devices despite rinsing. A small amount of carryover can activate zymogen FVII in a plasma sample and thus cause an increased FVIIa reading when measured using the present assay. Since FVIIa values are multiplied by 10 to correct for dilution, the error in FVIIa due to thromboplastin carryover is also compounded. Error of this type was apparent when plasma samples were measured for FVIIa and FVII:C or PT, simultaneously, using the STA coagulometer. To avoid the problem, samples had to be measured separately when both FVIIa and a thromboplastin-based assay were used and generally, it was better to measure FVIIa levels first to avoid thromboplastin carryover. 4. Handling of plasma samples The results of plasma FVIIa stability testing emphasized some guidelines for the collection and handling of plasma samples for the FVIIa assay. Samples should be collected only in siliconized tubes using trisodium citrate as the anti-coagulant. Two tubes of blood were sequentially collected from the same venepuncture. The first tube of blood was collected to remove the tissue plug that might result in in vivo tissue factor mediated activation of FVII. Only the second tube was used for FVIIa measurement. However, 80 studies by other groups have suggested this precaution is unnecessary (26, 27). Consumption of a light (uncooked, low fat) breakfast within 2 hours of blood collection did not appear to affect FVIIa levels. However, due to the difficulty of controlling meal content intake, fasting (12 hours) samples are highly recommended. Once collected, the plasma samples may be measured immediately or after at most a single freeze-thaw cycle. Frozen samples should be thawed at 37°C and held at room temperature until assay. Samples should never be stored at 4°C for extended periods to avoid cold activation nor should they have any contact with untreated glass to avoid contact activation. 5 . Factor Vi la in a normal population: effects of age and gender The present study confirmed the existence of trace amounts of FVIIa in normal healthy subjects as previously reported (26, 27). The normal range in the present study was 0.7 to 4.8 ng/ml and the mean plasma FVIIa was 2.08 ± 0.90 ng/ml. Previous studies have determined mean FVIIa levels to be 3.48 ng/ml by Morrissey et al (26), 2.5 ng/ml by Kario et al (27), and 4.1 ng/ml by Scarabin et al (53) with ranges of 0.5 - 8.4 ng/ml, 1.2 -5.1 ng/ml, and 0.4 - 14.3 ng/ml, respectively (Table VII), using larger populations. Table VII. FVIIa Levels in Normal Populations from Several Studies Study Mean FVIIa FVIIa Range Study size Age of Group Morrissey 3.48 ng/ml 0.5 - 8.4 ng/ml n= 188 36.2 yrs (18-72) Kario 2.5 ng/ml 1.2 - 5.1 ng/ml n= 110 57.9 yrs (21-84) Scarabin 4.1 ng/ml 0.4 - 14.3 ng/ml n = 684 41.1 yrs (26-64) Present 2.08 ng/ml 0.7 - 4.8 ng/ml n = 50 35.4 yrs (20-52) As in previous studies, FVIIa levels varied substantially between individuals in the normal population. In fact, the 7-fold variation of FVIIa levels in the present study was small 81 compared to other studies which found FVIIa levels to vary up to 36-fold (53). Normal FVII antigen levels typically vary from 130 to 1000 ng/ml, a 7.7-fold range (78). Thus, differences in FVII activity (ie. FVII:C) associated with CHD risk could be due to these large variations in FVIIa levels observed in other studies. Assuming an average FVII antigen concentration of 470 ng/ml, the mean FVIIa would represent 0.5% to 0.9% of the total factor W protein in plasma. The reason for the smaller range of FVIIa in the present study was partially due to the smaller population surveyed (n=50) and partly due to the limited age-range of the individuals. The oldest subject in the present study was 52 years old and the mean age of the population was 35.4 years which was significantly lower than the studies by Kario (57.9 years) and Scarabin (45.5 years) but similar to Morrissey's group (36.2 years). However, as Table VII indicates the present study did not include any elderly patients (over 55) who would be expected to have higher FVIIa levels as indicated by the trends observed in Figure 28 and the previous studies (26, 27, 53). Thus the mean FVIIa value and the overall FVIIa range in the present study was skewed lower due to the absence of elderly subjects. The significant correlation between FVIIa levels and increasing age in both men and women was also consistent with previous studies (26, 27). The results of the present study indicated that there was no apparent difference in FVIIa levels between men and women in any age group. However, some studies have shown that FVIIa levels increase significantly in post-menopausal women (53). Menopausal status was not considered in the present study. Interestingly, the general trend of increasing FVIIa levels with advancing age and menopause parallels the increase in CHD risk with age and onset of menopause. Of course, many other changes in hemostatic status and other CHD risk factors such as lipid levels also accompany age and menopause. However, since FVII coagulant activity has been implicated in fatal incidences of CHD (28) and FVIIa plays an essential role in the 82 initiation of coagulation (13), it is conceivable that FVIIa is an important factor in thrombotic conditions. 6. Factor Vi la in prothrombotic conditions: Complicated pregnancies A comprehensive prospective clinical study is the only way to confirm an association between risk of coronary heart disease and FVIIa levels. Although such a study is clearly beyond the scope of the present investigation, an initial study to examine potential correlations between FVIIa levels and certain prothrombotic conditions was undertaken. A group of 9 patients at B.C. Children's Hospital were included in a study of FVIIa levels in women experiencing complicated pregnancies. A significant increase in FVIIa levels in pregnant women was observed; the corrected mean FVIIa concentration during pregnancy was 5.35 ± 0.89 ng/ml. This was substantially higher than the mean FVIIa in an age-matched group of non-pregnant females (1.86 ± 0.75 ng/ml) as well as the population average of 2.08 ng/ml. The overall effect of pregnancy is to increase the activity of the coagulation system as a protective mechanism against post-partum hemorrhage. However, these changes can put the patient at increased risk of venous thrombosis. Total factor VII levels reach 150 to 250% of normal levels during pregnancy (75) and this has been shown to be reflected in a substantially higher FVII coagulant activity (79). The present study found FVIIa levels during pregnancy to be on average 287% of normal levels. Although this high level of FVIIa activity may have been in part due to the complications experienced by the patients during their pregnancy, it does seem to indicate that FVIIa increases proportionally as much or more than the total FVII mass during pregnancy. To better understand the interaction between FVIIa and total FVII levels, FVII antigen levels must be measured (by ELISA) and the ratio of FVIIa-to-FVII: Ag examined. Although a complete study was not performed to determine the effects of anti-coagulants on FVIIa, the results of one of the complicated pregnancy subjects provided 83 some insight into the cardio-protective effects conferred by heparin. The patient, a 40 year old woman, was placed on heparin therapy because she was considered to be a high-risk pregnancy. Because of her age, the risk of thrombotic complications during pregnancy was greater than average. Previous results would suggest that her age and her pregnancy would combine to produce extremely elevated FVIIa levels. However, her plasma FVIIa was 0.9 ng/ml, below average even for a non-pregnant subject. This was likely a direct result of the heparin therapy. Thus, heparin's cardio-protective effect is in part mediated through its ability to directly decrease FVIIa levels or FVIIa activity in the plasma. Although this may not be surprising since heparin is known to target serine proteases (80), the 500% decrease in FVIIa activity/concentration is very large. The only complicated pregnancy patient with lower than expected FVIIa levels (2.2 ng/ml) after delivery suffered from post-partum hemorrhaging. Interestingly, the. patient's PT and APTT times were normal. Hence the patient's risk of post-partum bleeding could not have been predicted using the standard tests. However, the FVIIa assay was able to detect a problem that would have alerted medical personnel to the potential danger. Unfortunately, plasma fibrinogen was not measured for this subject as a comparison to determine if other clotting factors were also depressed; however, the fact that FVIIa was critically low (for a full term pregnancy) but both PT, a measure of extrinsic activity, and APTT, a measure of the intrinsic pathway (44), were normal could be a clue that depressed FVIIa levels are sufficient to cause bleeding problems post-partum. The remaining 10 complicated pregnancy samples (Table Via) were not included in the above discussion because clinical details about the patients' conditions were not available at the time of publication. However, as with the first nine patients, the overall mean FVIIa level was significantly elevated compared to normal, healthy donors. This group included patients with extremely high FVIIa levels (samples 19, 20) as well as a few patients with lower than expected FVIIa levels (samples 15, 17, 18). However, more information about the patients is required for a full discussion. 84 7. Future Work Clearly more work must be done to establish FVIIa as an independent risk factor in prothrombotic states. The clinical investigation undertaken in the present study had many deficiencies. The healthy population used to establish the normal range of FVIIa was small (n=50) and comprised a narrow age range (20-52 years). As well, even though subjects were healthy with no known history of CHD, it would have been useful to better characterize them by measuring BMI, blood pressure, total cholesterol, total triglycerides, and hemostatic variables such as FVILC, FVILAg, and fibrinogen. Thus, correlations between these different variables and FVIIa levels could have been determined. However, the most indispensable test not included in the present study was for factor VH antigen. Measurement of FVH:Ag would allow the determination of the ratio of FVIIa-to-FVII: Ag, a more useful value representing the actual proportion of FVII in the activated form. For example, an increase in FVIIa could represent either an overall increase in the total amount of FVII protein with the same proportion of activated FVIIa or it could represent an actual increase in the proportion of total FVII in the activated FVIIa form. Without data on FVILAg, it was impossible to differentiate specific increases in activated FVIIa from general increases in total FVII mass. Thus, the significance of the observed increase in FVHa levels during pregnancy was not completely clear. The study of FVIIa in prothrombotic states is presently incomplete. To this point, only 19 samples from complicated pregnancies have been tested. More samples from various clinically defined groups are absolutely required before a comprehensive study can be achieved. It would be interesting to study FVIIa levels in patients with dyslipidemia to determine if there are associations between FVIIa levels and various lipid components. Studies by other groups have shown that a common polymorphism of the FVII gene is associated with lower FVIIa levels. It would be interesting to follow such individuals to determine if they suffered from a lower incidence of CHD. Recent studies have found a 85 close link between factor VII activity and fatal incidences of CHD (28); this suggests a study that would separate subjects by the outcome of an MI event and measure each group's mean FVIIa levels at the time of the event The most meaningful study would be a long term, large scale prospective study such as the Northwick Park Heart Study (7): The best way to accomplish such an extensive study would be to examine several hemostatic variables together in a well defined group of subjects who could be recruited and followed according to a well defined study plan over at least a 10 year period. The best groups to study would be those that fit into traditionally high risk categories but with no prior history of heart disease such as middle-aged to elderly men or post-menopausal women. Besides research on CHD risk, the FVIIa assay developed in the present study would be useful for routine clinical work such as monitoring of patients receiving recombinant FVIIa therapy or anti-coagulant therapy in which FVE levels must be carefully controlled. As well, the assay could help in biochemical studies of cold and contact activation of clotting factors, the mechanism of the tissue factor pathway, and determination of the importance of pre-activated FVHa in initiation of normal clotting. Since FVII can be activated by several different factors (thrombin, FXa, FXIIa, FVIIa) in vitro (81, 82), the new assay could be used to elucidate which of these factors is most important in FVII activation in vivo by performing studies in patients with various clotting factor deficiencies. Another use of the FVIIa assay could be in studying individuals with FVII-deficiencies. Since individuals deficient for FVII exhibit a broad range of clinical symptoms (some suffer from severe bleeding disorders while others have no apparent problems (83)), measurement of such subjects' FVIIa levels may provide some insight into the apparent discrepancies. For example, a study could be done on two patients who each have only 20% of normal FVII antigen levels; however one suffers a severe bleeding disorder while the other does not. Measurement of their FVIIa levels could reveal that the 86 patient who does not bleed maintains normal levels of FVIIa despite the decreased total FVII mass. 8. Possible mechanisms of FVIIa involvement in thrombotic disorders Many recent studies have improved the understanding of the significance of FVIIa levels in various pathological conditions. The results of many of the clinical studies on FVIIa were described in the introduction. Other studies have provided insight into the possible mechanism of FVIIa and tissue factor in thrombotic states Here, speculation into possible mechanisms of FVIIa's role as a prothrombotic risk factor in atherosclerosis, diabetes, and menopause are discussed. Histological evidence has confirmed the presence of TF in atherosclerotic plaques (18). The rupture of a plaque is often the critical event in acute MI (19,42). The exposure of TF from the plaque core to the vasculature initiates an explosive thrombus formation near the site of plaque rupture. The size of the thrombus formed and thus the outcome of the MI event may be determined by the plasma FVIIa level. High levels of circulating FVIIa would produce a stronger and faster activation of extrinsic coagulation initiated by the released TF and thus a more severe occlusion at the site of injury resulting in a greater chance of the event ending fatally. Thus, FVIIa levels may not be a direct indicator of the risk of CHD but may be important in determining the severity of the event after it has occurred. This is consistent with the studies of Northwick Park where they found FVII:C levels were better predictors of fatal incidences of CHD than non-fatal events (28). Several studies have also determined that plasma FVIIa levels are elevated in patients with non-insulin dependent diabetes mellitus (35, 57). This is a pathological state that has been associated with endothelial damage. Studies have show that the greater the extent of endothelial damage, the higher the associated FVIIa levels (57). Endothelial damage results in the release or exposure of TF, a component of the subendothelium. Thus, increased plasma TF would result in activation of FVII. However, the increased 87 extrinsic activity would remain sub-thrombotic and be characterized only by elevated FVIIa levels. One of diabetes' most devastating side effects is an increased risk of CHD and sudden cardiac death (84). Thus, the elevated FVIIa level mediated by diabetes-related endothelial damage could be the cause of the CHD risk associated with diabetes. Hormone replacement therapy (HRT) in post-menopausal women has been shown in several studies to offer cardio-protective benefits (85). HRTs seems to confer these cardio-protective benefits through their effect on the vascular wall and specifically the endothelial layer. Thus, HRT may in fact be reversing the endothelial damage that releases TF and activates FVII resulting in lower risk of CHD. It has been shown by Scarabin et al (53) that HRT decreases FVIIa levels in post-menopausal women and such women suffer from fewer incidences of CHD than post-menopausal women not on such therapy (85). Since factor W activity and possibly FVHa may be closely linked to the outcome of a coronary event, measures of FVII activity (FVILC, FVIIa, FVII: Ag) could be useful in determining the short-term risk of CHD. If FVII(a) levels were found to increase quickly over a short period of time, it could be an indication of plaque rupture or excessive endothelial damage which might lead to an acute thrombotic event. Thus, careful monitoring of FVIIa levels could be important for the survival of patients who are already at known risk of CHD such as patients suffering from atherosclerosis, diabetes, or post-menopausal women. 9. Implications for treatment of thrombotic disorders. Because of the growing amount of epidemiological evidence implicating factor VII and more recently factor Vila to the risk of CHD and cardiovascular death, the TF-FVHa complex has become a target for the development of new anticoagulants for the treatment of thrombotic disorders. Several groups have begun work on therapeutics which specifically focus on the inhibition of factor Vila and/or the extrinsic pathway. 88 Some groups are exploring the use of recombinant forms of natural inhibitors of factor VII. Yamanobe et al (86) have studied the use of recombinant tissue factor Pathway Inhibitor (TFPI) as a therapeutic anticoagulant. They found that TFPI injected intravenously into rats initially binds to endothelial cells and hepatocytes removing it from circulation. However, upon injection of heparin, the TFPI was released and improved the anti-coagulant effects of the heparin. Another group, Sorensen et al (87) are working on the use of an active site-inhibited FVIIa as a potential antithrombotic agent. The group treated FVIIa with D-Phe-L-Phe-L-Arg-chloromethyl-ketone to inhibit its active site. Early work has shown that, the inhibitor alters FVIIa's affinity for tissue factor thus inhibiting TF-initiated coagulation. Other groups are working on recombinant proteins and synthetic polypeptides as FVfla-targeting anticoagulants. Kelley et al (88) have developed a mutant form of soluble tissue factor which binds FVIIa with the same affinity as the native form but shows a 34-fold reduction in catalytic efficiency for FX activation. The sTF mutant was shown to successfully prolong prothrombin times. Finally, Ronning et al (89) have studied the possibility of using synthetic peptide analogs of TF and FVII to inhibit FXa formation by the TF-FVIIa complex. They prepared several synthetic polypeptides corresponding to regions of TF and FVII which have been suggested to be important in binding between TF and FVUa or the TF-FVIIa complex and FX to determine the potential inhibitory effects on coagulation. Early results were encouraging. . Thus, there is much excitement over this new potential area for the development of anti-thrombotic agents targeting factor Vila and the tissue factor pathway. Further understanding of the specific role and mechanism that FVIIa plays in thrombotic disorders will allow better targeting of the potential therapeutics. Therefore, continued studies using the FVIIa assay developed in the present investigation are essential to the development of more efficacious anti-thrombotic drugs targeting the extrinsic pathway of coagulation. 89 REFERENCES 1. Ratnoff, O.D. Hemostasis and coagulation. In: Physiology, edited by Beme, R.M. and Levy, M.N. St. Louis: The C V . Mosby Company, 1988, p. 371-391. 2. Kay, L.A. Whole blood Coagulation in Haemostasis. In: Essentials of Haemostasis and Thrombosis, edited by Kay, L.A. Edinburgh: Churchill Living Stone, 1988, p. 40-53. 3. Jones, P. Bleeding and Clotting. In: Living with Haemophilia, edited by Jones, P. Oxford: Oxford University Press, 1995, p. 39-58. 4. Saito, H. Normal Hemostatic Mechanisms. In: Disorders of Hemostasis, edited by Ratnoff, O.D. and Forbes, C D . Orlando: Grune and Stratton, Inc., 1984, p. 23-42. 5. Bowie, E.J.W. and Owen, C A . The Clinical and Laboratory Diagnosis of Hemorrhagic Disorders. In: Disorders of Hemostasis, edited by Ratnoff, O.D. and Forbes, C D . Orlando: Grune and Stratton, Inc., 1984, p. 43-72. 6. Meade, T.W. Factor VII and Ischaemic Heart Disease: Epidemiological Evidence. Haemostas 13:178-185, 1983. 7. Meade, T.W., Brozovic, M. , Chakrabarti, R.R., Haines, A.P., Imeson, J.D., Mellows, S., Miller, G.J., North, W.R.S., Stirling, Y. and Thompson, S.G. Hemostatic Function and Ischaemic Heart Disease: Principal Results of the Northwick Park Heart Study. Lancet 533-537, 1986. 8. Heinrich, J., Balleisen, L., Schulte, H., Assman, G. and van de Loo, J. Fibrinogen and Factor VII in the Prediction of Coronary Risk. Results from the PROCAM Study in Healthy Men. Arterioscler Thrombos. 14:54-59, 1994. 9. Hultin, M.B. Fibrinogen and Factor VII as Risk Factors in Vascular Disease. Prog Hemostas Thrombos 10:215-241, 1991. 10. Ratnoff, O.D. The Evolution of Knowledge about Hemostasis. In: Disorders of Hemostasis, edited by Ratnoff, O.D. and Forbes, C D . Orlando: Grune and Stratton, 1984, p. 1-22. 11. Osteoid, B. and Rapaport, S.I. Activation of factor LX by the reaction product of tissue factor and factor VII: additional pathway for initiating blood coagulation. Proc Natl Acad Sci USA 74:5260-5264, 1977. 12. Zur, M., Shastri, K. and Nemerson, Y. Kinetics of the tissue factor pathway of blood coagulation: activation of factor DC. Blood 52:198,1978. 13. Nemerson, Y. Regulation of the Initiation of Coagulation of Factor VII. Haemostas 13:150-155, 1983. 14. Morrissey, J.H. Tissue Factor Modulation of Factor Vila Activity: Use in Measuring Trace Levels of Factor Vila in Plasma. Thrombos Haemostas 74:185-188,1995. 15. Morrissey, J.H., Fakhrai, H. and Edgington, T.S. Molecular Cloning of the cDNA for Tissue Factor, the Cellular Receptor for the Initiation of the Coagulation Protease Cascade. Cell 50:129-135, 1987. 90 16. Paborsky, L.P., Tate, K.M., Harris, R.J., Yansura, D.G., Band, L., McCray, G., Gorman, C.M., O'Brien, D.P., Chang, J.Y., Swartz, J.R., Fung, V.P., Thomas, J.N. and Vehar, G.A. Purification of Recombinant Human Tissue Factor. Biochemistry 28:8072-8077, 1989. 17. Edgington, T.S., Mackman, N., Brand, K. and Ruf, W. The Structural Biology of Expression and Function of Tissue Factor. Thrombos Haemostas 66:67-79,1991. 18. Wilcox, J.N., Smith, K.M., Schwartz, S.M. and Gordon, D. Localization of tissue factor in the normal vessel wall and in the atherosclerotic plaque. Proc Natl Acad Sci USA 86:2839-2843, 1989. 19. Falk, E. Plaque rupture with severe pre-existing stenosis precipitating coronary thrombosis. Characteristics of coronary atherosclerotic plaques underlying fatal occlusive thrombi. Br Heart J 50:127-134,1983. 20. Muller, Y.A., Ultsch, M.H., Kelley, R.F. and de Vos, A.M. Structure of the Extracellular Domain of Human Tissue Factor: Location of the Factor Vila Binding Site. Biochemistry 33:10864-10870, 1994. 21. Banner, D.W., D'Arcy, A., Chene, C , Winkler, F.K., Guha, A., Konigsberg, W.H., Nemerson, Y. and Kirchhofer, D. The crystal structure of the complex of blood coagulation factor Vila with soluble tissue factor. Nature 380:41-46,1996. 22. Carmeliet, P., Mackman, N., Moons, L., Luther, T., Gressens, P., Van Vlaenderen, I., Demunck, H., Kasper, M., Breier, G., Evrard, P., Muller, M., Risau, W., Edgington, T.S. and Collen, D. Role of tissue factor in embryonic blood vessel development. Nature 383:73-75, 1996. 23. Hayes, E.Y., Pike, J. and Tracy, R.P. Factor VII Assays. Arch Pathol Lab Med 117:52-57, 1993. 24. Albrecht, S., Kotzsch, M., Siegert, G., Luther, T., Grobmann, H., Grober, M. and Muller, M. Detection of Circulating Tissue Factor and Factor VII in a Normal Population. Thrombos Haemostas 75:772-777', 1996. 25. Philoppou, H., Adami, A., Amersey, R.A., Stubbs, PJ. and Lane, D.A. A Novel Specific Immunoassay for Plasma Two-Chain Factor Vila: Investigation of FVUa Levels in Normal Individuals and in Patients with Acute Coronary Syndromes. Blood 89:767-775, 1997. 26. Morrissey, J.H., Macik, B.G., Neuenschwander, P.F. and Comp, P.C. Quantitation of Activated Factor VII Levels in Plasma Using a Tissue Factor Mutant Selectively Deficient in Promoting Factor VII Activation. Blood 81:734-744,1993. 27. Kario, K., Miyata, T., Sakata, T., Matsuo, T. and Kato, H. Fluorogenic Assay of Activated Factor VII. Plasma Factor Vila Levels in Relation to Arterial Cardiovascular Diseases in Japanese. Arterioscler Thrombos. 14:265-274,1994. 28. Ruddock, V. and Meade, T.W. Factor VII activity and ischaemic heart disease: fatal and non-fatal events. QJMed 87:403-406,1994. 91 29. Hoffman, C.J., Miller, R.H. and Hultin, M.B. Correlation of Factor VII Activity and Antigen with Cholesterol and Triglycerides in Healthy Young Adults. Arteriol Thrombos 12:267-270, 1992. 30. Miller, G.J., Walter, S.J., Stirling, Y., Thompson, S.G., Esnouf, M.P. and Meade, T.W. Assay of factor VII activity by two techniques: evidence for increased conversion of VII to Vila in hyperlipidaemia with possible implications for ischaemic heart disease. Br J Haemostas 59:249-258, 1985. 31. Mukherjee, M., Dawson, G., Sembhi, K. and Kakkar, V.V. Triglyceride Dependence of Factor VII Coagulant Activity in Deep Venous Thrombosis. Thrombos Haemostas 76:500-501, 1996. 32. Folsom, A.R., Ma, J., Eckfeldt, J.H., Shahar, E. and Wu, K.K. Plasma phospholipid fatty acid composition and factor VII coagulant activity. Atheroscler 111:199-207,1994. 33. Miller, G.J., Bauer, K.A., Barzegar, S., Cooper, J.A. and Rosenberg, R.D. Increased Activation of the Haemostatic System in Men at High Risk of Fatal Coronary Heart Disease. Thrombos Haemostas 75:767-771, 1996. 34. Pankow, J.S., Folsom, A.R., Province, M.A., Rao, D.C., Eckfeldt, J.H., Heiss, G., Shahar, E. and Wu, K.K. Family History of Coronary Heart Disease and Hemostatic Variables in Middle-aged Adults. Thrombos Haemostas 77:87-93, 1997. 35. Kario, K., Sakata, T., Matsuo, T. and Miyata, T. Factor VII in non-insulin-dependent diabetic patients with microalbuminuria. Lancet 342:1552,1993. 36. Fuller, J.H., Keen, H., Jarret, R.J. and et al., Haemostatic variables asociated with diabetes mellitus and its complications. Br Med J 2:964-966,1979. 37. Mitropoulos, K.A., Martin, J.C., Reeves, B.E.A. and Esnouf, M.P. The Activation of the Contact Phase of Coagulation by Phhysiologic Surfaces in Plasma: The Effect of Large Negatively Charged Liposomal Vesicles. Blood 73:1525-1533,1989. 38. Mennen, L.I., Schouten, E.G., Grobbee, D.E. and Kluft, C. Coagulation Factor VII, Dietary Fat and Blood Lipids: A Review. Thrombos Haemostas 76:492-499, 1996. 39. Sanders, T.A.B., Miller, G.J., de Grassi, T. and Yahia, N. Postprandial Activation of Coagulant Factor VII by Long-chain Dietary Fatty Acids. Thrombos Haemostas 76:369-371, 1996. 40. Green, D., Chamberlain, M.A., Ruth, K.J., Folsom, A.R. and Liu, K. Factor VII, Cholesterol, and Triglycerides: The CARDIA Study. Arterioscler Thrombos Vase Biol. 17:51-55, 1997. 41. Miller, G.J., Martin, J.C., Mitropoulos, K.A., Esnouf, M.P., Cooper, J.A., Morrissey, J.H., Howarth, D.J. and Tuddenham, E.G.D. Activation of Factor VII during Alimentary Lipemia occurs in Healthy Adults and Patients with Congenital Factor XII or Factor XI Deficiency, But not in Patients with Factor IX Deficiency. Blood 87:4187-4196, 1996. 42. Ridolfi, R.L. and Hutchins, G.M. The relationship between coronary artery lesions and myocardial infarct: Ulceration of atherosclerotic plaques precipitating coronary thrombosis. Am Heart J 93:468-486, 1977. 92 43. Levin, D.C. and Gardiner, G.A. Coronary Atheoslerotic Plaque Morphology and its Consequences. In: Heart Disease: A textbook of Cardiovascular Medicine, edited by Braunwald, E. Philadelphia: W.B. Saunders Company, 1988, p. 295-348. 44. Kay, L.A. The Coagulation Laboratory. In: Essentials of Haemostasis and Thrombosis, edited by Kay, L.A. Edinburgh: Churchill Livingstone, 1988, p. 228-276. 45. Miller, G.J., Stirling, Y., Esnouf, M.P., Heinrich, J., van de Loo, J., Kienast, J., Wu, K.K., Morrissey, J.H., Meade, T.W., Martin, J.C., Imeson, J.D., Cooper, J.A. and Finch, A. Factor VD-Deficient Substrate Plasma Depleted of Protein C Raise the Sensitivity of the Factor VII Bio-Assay to Activated Factor VII: an Internation Study. Thrombos Haemostas 71:38-48, 1994. 46. Seligsohn, U., Osterud, B. and Rapaport, S.I. Coupled Amidolytic Assay for Factor VII: ITs use with a Clotting Assay fo Determine the Activity State of Factor VII. Blood 52:978-988, 1978. 47. Neuenschwander, P.F. and Morrissey, J.H. Deletion of the Membrane Anchoring Region of Tissue Factor Abolishes Autoactivation of Factor VII but not Cofactor Function. JBiol Chem 267:14477-14482, 1992. 48. Mackman, N., Morrissey, J.H., Fowler, B. and Edgington, T.S. Complete Sequence of the Human Tissue Factor Gene, a Highly Regulated Cellular Receptor That Initiates the Coagulation Protease Cascade. Biochemistry 28:1755-1762, 1989. 49. Carson, S.D., Henry, W.M. and Shows, T.B. Tissue Factor Gene localized to Chromosome 1. Science 229:991-993, 1985. 50. Ruf, W., Rehemtulla, A. and Edgington, T.S. Phospholipid-independent and -depandent Interactions Required for Tissue Factor Receptor and Cofactor Function. J Biol Chem 266:2158-2166, 1991. 51. Morrissey, J.H. Clotting Assay for Measuring Plasma Factor Vila Based on Soluble Mutant Tissue Factor (sTF). Personal Communication , 1996. 52. Hubbard, A.R. and Barrowcliffe, T.W. Measurement of Activated Factor VII Using Soluble Mutant Tissue Factor - Proposal for Standardization. Thrombos Haemostas 72:649-650, 1994. 53. Scarabin, P.Y., Vissac, A.M., Kirzin, J.M., Bourgeat, P., Amiral, J., Agher, R. and Guize, L. Population Correlates of Coagulation Factor VII: Importance of Age, Sex, and Menopausal Status as Determinants of Activated Factor VII. Arterioscler Thrombos Vase Biol. 16:1170-1176, 1996. 54. Kario, K., Matsuo, T., Sakata, T. and Miyata, T. Factor VII hyperactivity and ischaemic heart disease. Lancet 343:233,1994. 55. Silveira, A., Karpe, F., Blomback, M., Steiner, G., Walldius, G. and Hamsten, A. Activation of Coagulation Factor VII During Alimentary Lipemia. Arterioscler Thrombos. 14:60-69, 1994. 93 56. Moor, E., Silveira, A., van't Hooft, F., Suontaka, A.M., Eriksson, P., Blomback, M. and Hamsten, A. Coagulation Factor VII Mass and Activity in Young Men With Myocardial Infarcation at a Young Age. Role of Plasma Lipoproteins and Factor VII Genotype. Arterioscler Thrombos Vase Biol. 15:655-664, 1995. 57. Kario, K., Matsuo, T., Kobayashi, FL, Sakata, T. and Miyata, T. Activation of Tissue Factor-Induced Coagulation and Endothelial Cell Dysfunction in Non-Insulin-Dependent Diabetic Patients With Microalbuminuria. Arterioscler Thrombos VascBiol. 15:1114-1120, 1995. 58. Kakkar, A.K., DeRuvo, N., Chinswangwatanakui, V., Tebbut, S. and Williamson, R.C.N. Extrinsic-Pathway activation in cancer with high factor Vila and tissue factor. Lancet 346:1004-1005, 1995. 59. Quehenberger, P., Loner, U., Kapiotis, S., Handler, S., Schneider, B., Huber, J. and Speiser, W. Increased Levels of Activated Factor VII and Decreased Plasma Protein S Activity and Circulating Thrombomodulin during Use of Oral Contraceptives. Thrombos Haemostas 76:729-734, 1996. 60. Inamo, Y., Pemberton, S., Tuddenham, E.G.D. and Woo, P. Increase of Activated Factor Vila and Haemostatic Molecular Markers in Juvenile Chronic Arthritis. Br J Rheum 34:466-469, 1995. 61. Silveira, A., Green, F., Karpe, F., Blomback, M., Humphries, S. and Hamsten, A. Elevated Levels of Factor VII Activity in the Postprandial State: Effect of the Factor VII Arg-Gln Polymorphism. Thrombos Haemostas 72:734-739, 1994. 62. Dagert, M. and Ehrlich, S.D. Prolonged Incubation in Calcium Chloride Improves the Competence of Escherichia coli Cells. Gene 6:23-28,1979. 63. Sanger, F., Nicklen, S. and Coulsen, A.R. DNA Sequencing with Chain Terminating Inhibitors. Proc Natl Acad Sci USA 74:5463-5467, 1977. 64. Palmiter, R.D., Behringer, R.R., Quaife, C.J., Maxwell, F., Maxwell, I.H. and Brinster, R.L. Cell lineage ablation in transgenic mice by cell-specific expressed toxic gene. Cell 50:435-443, 1987. 65. Laemmli, U.K. Cleavage of Structural Proteins during the assembly of the head of Bacteriophage T4. Nature 227:680-685, 1970. 66. Towbin, H., Staehelin, T. and Gordon, J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc Natl Acad Sci USA 76:4350-4354, 1979. 67. Reziae, A.R., Fiore, M.M., Neuenschwander, P.F., Esmon, C.T. and Morrissey, J.H. Expression and purification of a soluble tissue factor fusion protein with an epitope for an unusual calcium dependent antibody. Prot Express Purif 3:453,1992. 68. Fiske, C H . and Subbarow, Y. The colorimetric determination of phosphorus. J Biol Chem 66:375-379, 1925. 69. Saiki, R.K., Glefand, D.H., Ttoffel, S.J., Schart, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. Primer-directed Enzymatic Amplification of DNA with a thermostable DNA Polymerase. Science 239:487-491,1988. 94 70. Shigematsu, Y., Miyata, T., Higashi, S., Miki, T., Sadler, J.E. and Iwanaga, S. Expression of Human Soluble Tissue Factor in Yeast and Enzymatic Properties of Its Complex with Factor Vila. J Biol Chem 267:21329-21337, 1992. 71. Porath, J. and Flodin, P. Gel Filtration: a method for desalting and group separation. Nature 183:1657-1659,1959. 72. Neuenschwander, P.F., Bianco-Fisher, E., Rezaie, A.R. and Morrissey, J.H. Phosphatidylethanolamine (PE) augments Factor Vila-Tissue Factor Activity: Enhancement of Sensitivity to Phosphatidylserine. Biochemistry 34:13988-13993,1995. 73. Williams, C.E. Critical Factors in Haemostasis, ChichestenEllis Horwood Ltd, 1988. pp. 1-11. 74. Williams, C.E. Global screening test in coagulation. In: Critical Factors in Haemostasis, edited by Williams, C.E., Short, P.E., George, A.J. and Entwistle, M.B.P. Cambridge: Ellis Horwood, 1988, p. 1-13. 75. Bonnar, J. Haemostasis and coagulation disorders in pregnancy. In: Haemostasis and Thrombosis, edited by Bloom, A.L. and Thomas, D.P. Edinburgh: Churchill Livingston, 1981, p. 454-469. 76. Miller, B.C., Hultin, M.B. and Jesty, J. Altered factor VII activity in hemophilia. Blood 65:845, 1985. 77. Morrissey, J.H., Revak, D., Tejada, D., Fair, D.S. and Edgington, T.S. Resolution of Monomeric and Heterodimeric Forms of Tissue Factor, The High-Affinity Cellular Receptor for Factor VII. Thrombos Res 50:481-493, 1988. 78. Zur, M. and Nemerson, Y. Tissue factor pathways of blood coagulation. In: Haemostasis and Thrombosis, edited by Bloom, A.L. and Thomas, D.P. Edinburgh: Churchill Livingstone, 1981, p. 124-139. 79. Hubbard, A.R., Parr, L.J. and Baines, M.G. Pregnancy and Factor VII. Br J Haemostas 80:265, 1992. 80. Barrowcliffe, T.W. and Thomas, D.P. Antithrombin III and heparin. In: Haemostasis and Thrombosis, edited by Bloom, A.L. and Thomas, D.P. Edinburgh: Churchill Livingstone, 1981, p. 712-724. 81. Prydz, H. The Role of Factor VII in the Intrinsic Pathway: A Brief Review. Haemostas 13:156-160, 1983. 82. Nakagaki, T., Foster, D C , Berkner, K.L. and Kisiel, W. Initiation of the Extrinsic Pathway of Blood Coagulation: Evidence for the Tissue Factor Dependent Autoactivation of Human Coagulation Factor VII. Biochemistry 30:10819-10824, 1991. 83. Roberts, H.R. and Lefkowitz, J.B. Inherited Disorders of Prothrombotic Conversion. In: Haemostasis and Thrombosis: Basic Principles and Clinical Practice, edited by Colman, R.W., Hirsh, J., Mardy, V.J. and Salzman, E.W. Philadelphia: J.B. Lippincott Company, 1994, 95 84. Winocour, P.D. and Richardson, M. Thrombosis and Atherogenesis in Diabetes. In: Diabetes and Atherosclerosis: Molecular basis and Clinical aspects, edited by Draznin, B. and Eckel, R.H. New York: Elsevier Science Publishing Company, 1993, p. 213-254. 85. Sitruk-Ware, R. Benefits and risks of hormonal replacement therapy in postmenopausal women. In: Medical-Economic Aspects of Hormone Replacement Therapy, edited by Cossery, J.M. Basel, Switzerland: 1993, p. 1-16. 86. Yamanobe, F., Mochida, S., Ohno, A., Ishikawa, K. and Fujiwara, K. Recombinant Human Tissue Factor Pathway Inhibitor as a possible Anticoagulant targeting hepatic sinusoidal walls. Thrombos Res 85:493-501, 1997. 87. Sorensen, B.B., Persson, E., Freskgard, P.O., Kjalke, M., Ezban, M., Williams, T. and Rao, L.V.M. Incorporation of an active site inhibitor in Factor Vila alters the affinity for Tissue Factor. J Biol Chem 272:11863-11868, 1997. 88. Kelley, R.F., Refino, C.J., O'Connel, M.P., Modi, N., Sehl, P., Lowe, D., Pater, C. and Bunting, S. A Soluble Tissue Factor Mutant is a Selective Anticoagulant and Antithrombotic agent Blood 89:3219-3227, 1997. 89. Ronning, H.F., Risoen, U.C., Orning, L., letten, K. and Sakariassen, K.S. Synthetic peptide analogs of Tissue Factor and Factor VII which inhibit Factor Xa formation by the Tissue Factor/Factor Vila complex. Thrombos Res 84:73-81,1997. 90. Hagen, F.S., Gray, C.L., O'Hara, P., Grant, F.J., Saari, G.C., Woodbury, R.G., Hart, C.E., Insley, M., Kisiel, W., Kurachi, K. and Davie, E.W. Characterization of a cDNA coding for human factor VII. Proc Natl Acad Sci USA 83:2412-2416, 1986. 96 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items