Tissue Factor Pathway Inhibitor Activity in Human Plasma
Measurement of Lipoprotein-Associated and Free Forms in Hyperlipidemia
Abstract Tissue factor pathway inhibitor (TFPI), a protease inhibitor that is present in free and lipoprotein-associated forms in plasma and that also occurs as an endothelial cell–associated form, can inhibit the initial reactions of the tissue factor–mediated coagulation pathway. Although a positive correlation between plasma TFPI activity and cholesterol concentration in human plasma has been demonstrated, levels of the various forms of TFPI, ie, the LDL/VLDL-associated form, the HDL-associated form, and the free form, have not yet been completely determined in hyperlipidemia. We therefore established a method for the measurement of each of these forms of TFPI in plasma by gel filtration of plasma in buffer containing 1 mol/L NaCl. The recovery of TFPI activity in the free form was markedly greater as assessed by the new method than the recovery reported when other methods have been used. We employed the new method to analyze TFPI activity in 19 hyperlipidemic patients and compared the results with those for normal control subjects. The level of LDL/VLDL-associated TFPI in hyperlipidemic patients was significantly increased compared with control subjects’ levels (0.383±0.112 versus 0.237±0.077 U/mL), whereas the level of the free form of TFPI in hyperlipidemic patients was significantly decreased (0.381±0.132 versus 0.495±0.106 U/mL), the former being positively correlated with cholesterol level, while the latter was negatively correlated. These results led us to speculate that the decrease in the free form of TFPI in hyperlipidemia was caused by an increase in LDL/VLDL and/or by the inhibition of TFPI synthesis in endothelial cells; such an inhibition may reflect a reduction in the antithrombotic function of vascular endothelial cells. Of note, the decrease in the free form of TFPI was more striking in patients receiving plasmapheresis treatment (0.271±0.089 U/mL); this decrease was due to the repeated removal of LDL/VLDL-associated TFPI and endothelial cell–associated TFPI.
- fluorogenic substrate assay
- gel filtration
- tissue factor pathway inhibitor
- Received August 9, 1994.
- Accepted January 17, 1995.
Tissue factor pathway inhibitor (TFPI), previously referred to as lipoprotein-associated coagulation inhibitor1 or extrinsic pathway inhibitor,2 is a protease inhibitor of the initial steps in the extrinsic blood coagulation pathway. TFPI, which consists of three tandem Kunitz-type inhibitory domains,1 inhibits factor VIIa–tissue factor complex and factor Xa.3 TFPI is mainly synthesized by vascular endothelial cells4 5 and is associated with these cells (50% to 90% of the intravascular pools of TFPI).6 7 8 Plasma contains 10% to 50% of total TFPI, and less than 2.5% is located in platelets.9 Plasma TFPI activity is increased in hyperlipidemic patients10 11 12 but is decreased in patients with abetalipoproteinemia and hypobetalipoproteinemia.13 Positive correlations of plasma TFPI activity with cholesterol10 11 and LDL cholesterol12 levels have also been demonstrated. Gel filtration of plasma has shown that the majority of plasma TFPI activity is associated with LDL/VLDL or with HDL, while the free form of TFPI represents a minor portion.7 14 15 16 We assumed that the measurement of each of the three forms of TFPI in plasma in hyperlipidemia would be more informative than the measurement of plasma TFPI alone as the levels of lipoprotein-associated forms of TFPI vary to a much greater extent than plasma TFPI levels. In fact, LDL/VLDL-associated TFPI in the plasma of crab-eating monkeys increases markedly in response to a high-cholesterol diet.17 The level of the free form of TFPI may also be important, since the endothelial cell–associated form of TFPI is released as the free form by both heparin injection6 7 8 and pathological stimuli.15 In this study we established a method for evaluating the TFPI activity of each of the three forms in plasma by gel filtration. We found that the free form of TFPI accounted for almost 50% of plasma TFPI activity on gel filtration in 1 mol/L NaCl buffer, whereas it accounted for 13% in 0.15 mol/L NaCl. We report the levels of the LDL/VLDL- and HDL-associated and free forms of TFPI in hyperlipidemic patients and the effect of plasmapheresis on these patients. We discuss the relation of the levels of the various forms of TFPI to cholesterol concentration and emphasize the importance of the free form in relation to the anticoagulant activity of vascular endothelial cells.
Human factor X and factor VIIa were kindly supplied by Dr Tomohiro Nakagaki, Chemo-Sera-Therapeutic Research Institute, Kumamoto, Japan. Tissue factor (rabbit brain thromboplastin) was purchased from Kokusai Siyaku. Bovine serum albumin was a product of Boehringer Mannheim. Rabbit anti-human recombinant TFPI (rTFPI; Chemo-Sera-Therapeutic Research Institute) polyclonal antibody was obtained by immunization of a rabbit with human rTFPI and by purification using a protein A cellulofine column (Seikagaku Kogyo). Polybrene was purchased from Aldrich Chemical Co, Inc. Cholesterol concentration was measured with a commercially available kit (Cholesterol E test, Wako Junyaku). Superdex 200 HR (16/60 and 26/60) was a product of Pharmacia LKB Biotechnology.
Blood Sampling and Preparation of Plasma
Blood specimens were obtained from 16 healthy volunteers (aged 30 to 55 years) and 19 hyperlipidemic patients (aged 14 to 65 years) at the National Cardiovascular Center Hospital, Osaka, Japan. Of the 19 hyperlipidemic patients, 13 (aged 32 to 65 years) were untreated. Postheparin plasma was collected 10 minutes after the injection of heparin (30 U/kg). The other 6 patients (aged 14 to 64 years), who had heterozygous familial hyperlipidemia, received plasmapheresis treatment every 2 weeks during which one of two apheresis columns (Evaflux 5A, Kurare, or Liposorber LA-15, Kanegafuchi Kagaku) was used. Blood was collected before the infusion of heparin unless otherwise stated. Heparin (1000 U) was injected before plasmapheresis, and then heparin 1000 U/h was infused during plasmapheresis. The diagnosis of hyperlipidemia was established according to the study criteria of the Japan Atherosclerosis Society (total serum cholesterol concentration, >5.7 mmol/L) (Table 1⇓). Blood was collected in tubes containing trisodium citrate (final concentration, 0.38%), and plasma was separated by centrifugation. Plasma specimens were stored at −80°C until use.
Assay of TFPI Activity
TFPI activity was measured by a two-stage method based on the capacity of TFPI to inhibit factor VIIa–tissue factor activity toward factor X activation. Plasma was diluted 1:10 by 10 mmol/L Tris-HCl buffer containing 0.15 mol/L NaCl, pH 8. In the first step, 50 μL of sample was mixed with 300 μL of a solution containing human factor VIIa (31.5 ng/mL), rabbit brain thromboplastin (80 μg/mL), and CaCl2 (6.67 mmol/L) in 10 mmol/L Tris-HCl buffer, pH 8, containing 0.15 mol/L NaCl, 0.1% bovine serum albumin, and 0.05% NaN3. This mixture was then mixed with 5 μL human factor X (44.1 μg/mL). For the assay of TFPI in fractions after gel filtration in buffer containing 1 mol/L NaCl, undiluted sample was mixed with the solution containing factor VIIa, brain thromboplastin, and CaCl2 in a NaCl-free buffer. After a 15-minute incubation at 37°C, 3 μL of 1 mmol/L Z-Pyr-Gly-Arg–methylcoumarin amide, a fluorogenic peptide substrate for factor Xa, was added as a second step. The increase in the fluorescence intensity of amino methylcoumarin was measured with a centrifugal autoanalyzer (Cobas Fara II, Roche).18 A standard curve was obtained from the serial dilution of human control subject pooled plasma. Relative inhibitory activity was calculated by dividing the rate of release of amino methylcoumarin per minute in the presence of TFPI by this rate in the absence of TFPI. TFPI activity in 1 mL control plasma was defined as 1 U/mL. All plasma samples were heat treated for 15 minutes at 56°C to avoid interference by coagulation factors in plasma. For samples containing heparin, polybrene was added (final concentration, 100 μg/mL). Fig 1⇓ shows the standard curve for the TFPI activity assay, which indicates that TFPI activity in plasma can be measured in a range from 0.01 to 0.1 U/mL. The intra-assay coefficient of variation was estimated to be 3.75% from 10 duplicate plasma samples; the interassay coefficient of variation was estimated to be 5.00% from seven determinations of a single stored plasma.
Plasma (500 μL) was applied to a column (1.6×60 cm) of Superdex 200 HR equilibrated with 10 mmol/L Tris-HCl buffer, pH 8, containing 0.05% NaN3 and either 0.15 or 1 mol/L NaCl. Gel filtration was performed at a flow rate of 1 mL/min at room temperature, and 1-mL fractions were collected. The column was calibrated by using a gel filtration standard obtained from BioRad Laboratories. The TFPI activities of the LDL/VLDL-associated form, the HDL-associated form, and the free form were measured as described above. The recovery of each form was calculated after the gel filtration of plasma in the buffer containing 0.15, 0.3, 0.6, 0.8, and 1 mol/L NaCl. Among the three forms of TFPI, the recovery of the free form of TFPI was markedly dependent on the NaCl concentration in the buffer. When the recovery of free TFPI in 1 mol/L NaCl was assumed to be 100%, the recoveries were 83% in 0.8 mol/L NaCl, 77% in 0.6 mol/L NaCl, 27% in 0.3 mol/L NaCl, and 16% in 0.15 mol/L NaCl, respectively. Thus, levels of each form of TFPI in the patient’s plasma were compared by gel filtration in the buffer containing 1 mol/L NaCl. Interassay coefficients of variation were estimated to be 5.8% for the LDL/VLDL-associated form, 8.0% for the HDL-associated form, and 3.8% for the free form from five determinations each.
Values are expressed as mean±SD. Pearson’s correlation coefficient was calculated for the different variables, and a value of P<.05 was considered significant. The nonparametric Wilcoxon test was employed for statistical comparison, with P<.05 again indicating significance.
Measurement of Plasma TFPI Activity and the Activity of the Lipoprotein-Associated and Free Forms of TFPI
Fig 2⇓ shows the patterns of TFPI activity in normal human pooled plasma after gel filtration in buffer containing 1 or 0.15 mol/L NaCl under the conditions described in “Methods.” In the buffer containing 1 mol/L NaCl, the percentage of each of the three forms of TFPI was 22.6% for the VLDL/LDL-associated form, 22.4% for the HDL-associated form, and 45.6% for the free form. Recovery of total TFPI activity was 90.6%. When gel filtration was performed in the buffer containing 0.15 mol/L NaCl, the percentage of the free form of TFPI was 12.8%, and the recovery of total TFPI activity was 64.8%. However, the recovery of the lipoprotein-associated forms was almost the same as that in 1 mol/L NaCl. When 0.65 U full-length rTFPI or 0.32 U C-terminal part–deleted rTFPI was gel filtered in the buffer containing 1 mol/L NaCl, recovery was 92.8% and 97.2%, respectively, whereas recovery in the buffer containing 0.15 mol/L NaCl was 9.4% and 30.5%, respectively. We then tried to prove directly that the loss of TFPI in the buffer containing 0.15 mol/L NaCl was due to binding of TFPI to the Superdex column by using a shorter column (0.5×5 cm). When 2.5 U rTFPI was applied to the column equilibrated with the buffer containing 0.15 mol/L NaCl and then eluted with the buffer containing 1 mol/L NaCl, 17.5% was found in the fractions eluted with 0.15 mol/L NaCl buffer and 57.1% in the fractions eluted with 1 mol/L NaCl buffer (Fig 3⇓). When 0.1 mL plasma (0.1 U) was applied under the same conditions, 57.6% was found in the 0.15 mol/L NaCl fractions and 3.2% in the 1 mol/L NaCl fractions (data not shown). These results indicate that the low recovery of the free form of TFPI after gel filtration in the buffer containing 0.15 mol/L NaCl was due to the adsorption of this form of TFPI in plasma on the column, mainly due to the interaction of the basic part of TFPI with Superdex 200.
When LDL/VLDL- or HDL-associated TFPI fractions isolated by the gel filtration of normal or hyperlipidemic subjects’ plasma in buffer containing 0.15 mol/L NaCl were subjected to gel filtration in the presence of 1 mol/L NaCl, no free form of TFPI was detected. We concluded that the release of the free form of TFPI from the lipoprotein fractions may not be significant in gel filtration with a buffer of high ionic strength. The recovery of TFPI activity after gel filtration, ie, the percentages of the sum of TFPI activity in the three fractions of TFPI activity in plasma, were 95.4±21.4% for normal individuals (n=16) and 86.1±16.8% for hyperlipidemic patients (n=19). No significant difference was observed in the serial dilution curves of TFPI activity for any of the three fractions. Neither freeze-thawing of plasma nor gel filtration at 4°C affected the recovery of any of the three forms.
TFPI Activity in Hyperlipidemic Patients
Fig 4⇓ shows the significant correlation between plasma TFPI activity and total cholesterol concentration in both normal individuals and hyperlipidemic patients; this correlation agrees with previous findings.10 11 The data for the patients who received repeated plasmapheresis treatment were not included in the calculation since plasma TFPI and endothelial cell–associated TFPI in these patients have been artificially removed as described below.
After gel filtration of the plasma of hyperlipidemic patients, TFPI activity in the lipoprotein-associated and free forms was compared with that found in the plasma of normal individuals (Fig 5⇓). LDL/VLDL-associated TFPI activity in hyperlipidemic patients was significantly higher than that in normal individuals (P<.005); HDL-associated TFPI activity in hyperlipidemic patients was slightly lower than in normal individuals. In hyperlipidemic patients receiving repeated plasmapheresis treatment, this decrease was also not significant. The free form of TFPI activity in untreated hyperlipidemic patients was significantly lower than that in normal individuals (P<.05), while the values in hyperlipidemic patients receiving repeated plasmapheresis were even more markedly decreased (P<.005). Values for TFPI activity in normal individuals and hyperlipidemic patients are summarized in Table 2⇓.
Fig 6⇓ shows the relation between total cholesterol concentration and TFPI activity in the LDL/VLDL-associated and free forms in normal individuals and the hyperlipidemic patients not receiving plasmapheresis treatment. LDL/VLDL-associated TFPI activity was positively correlated and the free form of TFPI activity was negatively correlated with cholesterol concentration (both P<.01). In hyperlipidemic patients receiving repeated plasmapheresis treatment, LDL/VLDL-associated TFPI activity was not positively correlated with cholesterol concentration, whereas the TFPI activity of the free form decreased slightly with increases in cholesterol concentration (data not shown). To determine the effect of plasmapheresis on plasma TFPI levels, we collected blood from the patients immediately before and after plasmapheresis. Fig 7⇓ shows the TFPI activity profiles of plasma from one of six patients after gel filtration. Plasmapheresis treatment produced marked decreases in LDL/VLDL-associated TFPI, while the free form was markedly increased. Similar results were obtained in the plasma of the five other patients. In these patients, LDL/VLDL-associated TFPI decreased to between 20% and 66% of the values before plasmapheresis, and the free form of TFPI increased two- to fivefold. We assumed that the increase in the free form of TFPI was due to the injection of heparin before plasmapheresis. However, the effect of plasmapheresis treatment was not simple. We found the major part of the free form of TFPI in an eluate from the plasmapheresis column (Evaflux 5A) on gel filtration in the buffer containing 1 mol/L NaCl (Fig 7⇓). In contrast, the major part of the free form was associated with lipoproteins when gel filtration was performed in the buffer containing 0.15 mol/L NaCl. Similar observations were obtained on two other patients treated with an Evaflux column. These findings indicate that the repeated plasmapheresis removed not only the LDL-associated form of TFPI but also the endothelial cell–associated TFPI that was released by heparin injection and was weakly associated with lipoproteins in plasma. In another three patients treated with a Liposorber column, we could not recover adequate amounts of TFPI from the column used, possibly because of the tight association of TFPI with dextran sulfate on the column. Therefore, we compared TFPI levels before and after plasmapheresis. The TFPI levels of the plasmapheresis patients did not increase significantly after plasmapheresis (Fig 8⇓). Since the patients were infused with heparin during treatment, TFPI levels in plasma should have markedly increased after plasmapheresis if heparin-releasable TFPI had not been removed by the treatment. In fact, TFPI levels of the patients without plasmapheresis treatment increased markedly after heparin infusion (Fig 8⇓). These findings led us to speculate that heparin-releasable TFPI was removed by the Liposorber column as well as the Evaflux column, possibly by the direct binding with dextran sulfate on the column.
We first established the optimum conditions for the separation of three forms of TFPI, ie, the LDL/VLDL- and HDL-associated and the free forms, by gel filtration of plasma. We then employed this method for the analysis of TFPI activity in the plasma of hyperlipidemic patients. We demonstrated that when gel filtration was performed at a high NaCl concentration, the free form of TFPI accounted for approximately 50% of TFPI in normal plasma; this is consistent with the finding that lipoprotein-associated TFPI separated by ultracentrifugation accounts for 50% to 70% of plasma TFPI.19 20 The low recovery of the free form in the buffer containing the physiological concentration of NaCl was concluded to be due to the retention of the free form on the column by the positive charge of TFPI, because in the buffer containing 0.15 mol/L NaCl the recovery of rTFPI without the carboxy terminal basic part was higher than that of full-length rTFPI. Moreover, the bound TFPI was eluted by the buffer containing 1 mol/L NaCl. The recovery of the free form of TFPI by the gel filtration in buffer containing 0.3 mol/L NaCl is also reported to be low,12 15 as confirmed in the present work. The possibility that the lipoprotein-associated form of TFPI dissociates in high-salt concentrations was excluded, as shown by us and Lesnik et al.19 Our finding of the high recovery of the free form of TFPI corroborates that of Valentin et al,21 who report a significant portion (45%) of low-molecular-weight material corresponding to lipoprotein-free TFPI by gel filtration on a column of Superose 6 with buffer containing 0.5% bovine serum albumin. On the other hand, Novotny et al7 report a low level of the free form of TFPI in plasma after gel filtration on a column of Superose 12 with a buffer containing 1 mol/L NaCl, similar to the buffer in the present experiment that contained 0.15 mol/L NaCl. We speculate that the difference may have been caused by the difference of the column used, ie, Superdex in our work and Superose in theirs.
We did not distinguish the LDL-associated form from the VLDL-associated form. However, we did analyze two forms of TFPI in normal individuals and hyperlipidemic patients by using a Superose 6 column (data not shown). We found that the VLDL-associated form accounted for only 16% to 20% of the LDL/VLDL-associated form in normal individuals; there was no significant increase of the VLDL-associated form of TFPI in hyperlipidemic patients.
We demonstrated that plasma TFPI activity in hyperlipidemic patients not receiving plasmapheresis was higher than that in normal individuals11 12 and that on gel filtration of plasma from hyperlipidemic patients the LDL/VLDL-associated form of TFPI was increased, but the free form of TFPI was decreased. It is generally agreed that TFPI is synthesized mainly in endothelial cells and binds with proteoglycans on the cells and that some portion of the endothelial cell–associated form of TFPI is constantly released into plasma as the free form, some of which then binds with lipoprotein particles, becoming the LDL/VLDL-associated or HDL-associated form, by some as yet unknown mechanisms. Based on this consideration, the decrease in the free form of TFPI in hyperlipidemic patients could indicate that the free form was transferred to the LDL/VLDL fraction due to the increase in LDL, which may cause a reduction in endothelial cell–associated TFPI. Another possibility is that the rate of TFPI synthesis in endothelial cells is reduced or that the rate of degradation of endothelial cell–associated TFPI is accelerated in hyperlipidemic patients, resulting in a decrease of endothelial cell–associated TFPI and, probably, in a decrease of the free form of TFPI in plasma.
We also demonstrated that plasmapheresis caused a pronounced change in TFPI in the hyperlipidemic patients. We presented evidence that the treatment removed LDL/VLDL-associated TFPI together with the free form. Since heparin injection before this treatment releases endothelial cell–associated TFPI into the plasma, this treatment also causes a reduction in the TFPI in endothelial cells. Both the lipoprotein-associated form of TFPI and the free form increase immediately after the injection of heparin into monkeys.17 However, when the plasma was stored for 24 hours at 4°C, the increased level of the LDL/VLDL-associated form decreased to the values seen before the heparin injection. In the present experiments on gel filtration in 0.15 and 1 mol/L NaCl, the eluate from the plasmapheresis column (Evaflux) included the free form of TFPI in loose association with lipoproteins. Although the decrease of TFPI activity in the eluate by the addition of polybrene (data not shown) suggests the presence of heparin in the eluate, we could not confirm that heparin in the column was present as a bound form with TFPI. These findings suggest that the TFPI newly released from endothelial cells is initially weakly associated with lipoproteins and that this association becomes stronger. Therefore, it is evident that in the plasma of hyperlipidemic patients receiving plasmapheresis the lipoprotein-associated form of TFPI includes two types of TFPI, one weakly, the other strongly associated with LDL. However, this was not the case for normal subjects and hyperlipidemic patients not receiving plasmapheresis, since the lipoprotein-associated form of TFPI in the plasma of these subjects did not dissociate in buffer containing 1 mol/L NaCl.
We have no direct evidence that heparin-releasable TFPI was in loose association with lipoproteins. However, the major part of TFPI in the eluate from the Evaflux column was isolated as the free form of TFPI by column chromatography on phenyl-Sepharose, factor Xa–Affigel, and heparin-Sepharose (H. Kato et al, unpublished observations, 1994). When the eluate was applied to a column of phenyl-Sepharose, the major part of TFPI was found in a nonadsorbed fraction, whereas a small portion of lipoprotein-associated forms of TFPI was adsorbed on the column. During the purification procedures, no detergent was included in the buffer. These results support the conclusion that heparin-releasable TFPI in the eluate from the Evaflux column was present in the free form or was weakly associated with lipoproteins but was not present as the tightly associated form.
The free form of TFPI increased two- to fivefold after plasmapheresis of six patients who were infused with heparin during the treatment. However, if heparin-releasable TFPI was not removed by plasmapheresis, the free TFPI level of the patients should have increased sevenfold to 14-fold, ie, to levels shown in normal individuals and the patients not receiving plasmapheresis, immediately after the treatment (data not shown). The increase of TFPI levels of all patients treated with either the Evaflux or Liposorber columns was not significant after plasmapheresis, whereas the TFPI levels of the patients without plasmapheresis markedly increased after heparin infusion. These findings indicate that heparin-releasable TFPI was removed by plasmapheresis regardless of the column used. Heparin-releasable TFPI may have been removed by direct binding with dextran sulfate on the Liposorber column. We speculate that the remarkable decrease in the free form of TFPI was caused by the removal of the free form through repeated plasmapheresis treatment (every 2 weeks) and that the rate of TFPI synthesis in endothelial cells did not overcome this decrease in TFPI. These factors may explain why plasma TFPI activity was not correlated with total cholesterol concentration in the plasmapheresis patients.
Lindahl et al15 have suggested that the free form of TFPI has greater anticoagulant activity relative to amidolytic activity than lipoprotein-associated TFPI. They have also suggested that TFPI loses its heparin-binding capacity when it is associated with lipoproteins, and that therefore it may lose the capacity to bind with proteoglycans on endothelial cells. It has been speculated that TFPI on endothelial cells in vivo plays a role in the regulation of thrombosis.22 Our findings emphasize the important role played by the free form of TFPI in plasma and lead us to speculate that the decrease in free TFPI reflects a decrease in endothelial cell–associated TFPI and subsequently the thrombotic tendency in hyperlipidemia. Thus, the measurement of each form of TFPI in plasma rather than measurement of total plasma TFPI will be more informative in detecting changes in hypercoagulability in hyperlipidemia.
This study was supported in part by a grant-in-aid for Research Promoting Comprehensive Longevity Science and by a Research Grant for Cardiovascular Diseases (5A-2) from the Ministry of Health and Welfare.
Wun TC, Kretzmer KK, Girard TJ, Miletich JP, Broze GJ. Cloning and characterization of a cDNA coding for the lipoprotein-associated coagulation inhibitor shows that it consists of three tandem Kunitz-type inhibitory domains. J Biol Chem. 1988;263:6001-6004.
Rao LVM, Rapaport SI. Studies of a mechanism inhibiting the initiation of the extrinsic pathway of coagulation. Blood. 1987;69:645-651.
Broze GJ, Warren LA, Novotny WF, Higuchi DA, Girard JJ, Miletich JP. The lipoprotein-associated coagulation inhibitor that inhibits the factor VIIa-tissue factor complex, also inhibits factor Xa: insight into its possible mechanism of action. Blood. 1988;71:335-343.
Bajaj MS, Kuppuswamy MN, Saito H, Spitzer SD, Bajaj SP. Cultured normal human hepatocytes do not synthesize lipoprotein-associated coagulation inhibitor: evidence that endothelium is the principal site of its synthesis. Proc Natl Acad Sci U S A. 1990;87:8869-8873.
Enjyoji K, Emi M, Mukai T, Kato H. cDNA cloning and expression of rat tissue factor pathway inhibitor (TFPI). J Biochem. 1992;111:681-687.
Novotny WF, Palmier M, Wun TC, Broze GJ, Miletich JP. Purification and properties of heparin releasable lipoprotein associated inhibitor. Blood. 1991;78:394-400.
Novotny WF, Girard TJ, Meletich JP, Broze GJ. Platelets secrete a coagulation inhibitor functionally and antigenically similar to the lipoprotein associated coagulation inhibitor. Blood. 1988;72:2020-2025.
Sandset PM, Lund H, Norseth J, Abildgaard U, Ose L. Treatment with hydroxymethylglutaryl:coenzyme A reductase inhibitors in hypercholesterolemia induces changes in the components of the extrinsic coagulation system. Arterioscler Thromb. 1991;11:138-145.
Hansen JB, Huseby NE, Sandset PM, Svensson B, Lyngmo V, Nordoy A. Tissue factor pathway inhibitor and lipoproteins: evidence for association with and regulation by LDL in human plasma. Arterioscler Thromb. 1994;14:223-229.
Novotny WF, Brown SG, Miletlich JP, Rader DJ, Broze GJ. Plasma antigen levels of the lipoprotein-associated coagulation inhibitor in patient samples. Blood. 1991;78:387-393.
Novotny WF, Girard TJ, Miletich JP, Broze GJ. Purification and characterization of the lipoprotein-associated coagulation inhibitor from human plasma. J Biol Chem. 1989;31:18832-18837.
Abumiya T, Nakamura S, Takenaka A, Takenaka O, Yoshikuni Y, Miyamoto S, Kimura T, Enjyoji K, Kato H. Response of plasma tissue factor pathway inhibitor (TFPI) to diet-induced hypercholesterolemia in crab-eating monkeys. Arterioscler Thromb. 1994;14:483-488.
Lesnik P, Vonica A, Guerin M, Moreau M, Chapman MJ. Anticoagulation activity of tissue factor pathway inhibitor in human plasma is preferentially associated with dense subspecies of LDL and HDL and with Lp(a). Arterioscler Thromb. 1993;13:1066-1075.
Grabowski EF, Zuckerman DB, Nemerson Y. The functional expression of tissue factor by fibroblasts and endothelial cells under flow conditions. Blood. 1993;81:3265-3270.