Levels of Factor VIIc Associated With Decreased Tissue Factor Pathway Inhibitor and Increased Plasminogen Activator Inhibitor-1 in Dyslipidemias
Abstract Tissue factor pathway inhibitor (TFPI), a kunitz-type inhibitor of the extrinsic coagulation pathway, factor VII coagulant (FVIIc), FVIIa, and the fibrinolytic factors plasminogen activator inhibitor-1 (PAI-1) and tissue plasminogen activator (TPA) have been studied in various hyperlipidemias. Compared with a normal lipidic group, mean TFPI activity was 70% higher (P<.001) and 36% higher (P<.001) in type IIa and IIb hyperlipidemias, respectively, and was lower by 13% in type IV hyperlipidemia (P=.05). TFPI was correlated with LDL cholesterol (P<.001), total cholesterol (P<.001), HDL cholesterol (P<.01), apolipoproteins (apo) AI (P<.001) and B (P<.001) and lipoprotein a (P<.01). TFPI was negatively correlated with the triglyceride level (P<.05); the correlation was dependent on LDL cholesterol and HDL cholesterol levels, which were decreased in type IV hyperlipidemia. FVIIc activity (P<.001) was increased by 30% in both type IV and type IIb hyperlipidemia and was correlated with triglyceride levels. FVIIa was not significantly increased in any group compared with control group. FVIIc was correlated with triglyceride level (P<.001), while FVIIa was not. Interestingly, FVIIa was correlated with FVIIc (r=.5, P<.001) in the control group as well as in the hyperlipidemic groups (r=.32, P<.01). These results favor the hypothesis that higher FVIIc concentrations in hyperlipidemic patients are likely due to enhancement of synthesis of FVII and that a part of this FVII circulates in an activated chemical form. Compared with the control group, PAI-1 activity was twofold higher (P<.08) in type IIa hyperlipidemia, threefold higher (P<.001) in type IIb hyperlipidemia, and fourfold higher in type IV hyperlipidemia (P<.001). PAI-1 activity correlated with triglyceride levels (P<.001), apoB levels (P<.001) and total cholesterol levels (P<.05). These correlations were dependent on apoB and probably reflect the correlation between PAI-1 and VLDL. In contrast, TPA level was normal in the different hyperlipidemias. No correlation was found between TFPI, FVIIc, and PAI-1. Variation of TFPI activity appears to be related to the variations of its main lipoprotein carriers: LDL, HDL, and Lp (a). The association in hypertriglycemic patients of hypercoagulability (increased FVIIc and decreased TFPI) and hypofibrinolysis (increased PAI-1) may explain thrombosis predisposition of some of these patients. However, it would be interesting to study the increased levels of endothelium-derived TFPI in plasma induced by the injection of heparin.
- tissue factor pathway inhibitor
- factor VIIa
- plasminogen activator inhibitor type 1
- tissue-type plasminogen activator
- Received November 11, 1994.
- Accepted September 19, 1995.
The importance of hyperlipidemia in the development of cardiovascular disease is well established. Dyslipidemias are involved in the atherogenic process and also may adversely affect the hemostatic system, resulting in a prethrombotic state. Several reports point to the influence of hyperlipidemia in the extrinsic coagulation system.1 2 Plasma FVIIc activity is positively correlated with serum triglyceride levels and is increased in type IIb and type IV hyperlipidemic subjects.3 4 The correlation between FVIIc and lipids may be explained by the binding of FVII with triglyceride-rich lipoprotein particles.4 5 6
FVII must be associated with its cofactor TF in an FVIIa-TF complex to activate factor IX and factor X.7 In arterial disease, the disruption of atherosclerotic plaques where TF is prevalent8 may expose TF to circulating blood and trigger blood coagulation. The FVIIa-TF complex is regulated by a serine protease inhibitor, TFPI. In the presence of its cofactor, factor Xa, TFPI can inhibit the FVIIa-TF complex. Plasma TFPI is mostly associated with lipoproteins, ie, LDL and HDL. Different proportions of TFPI associated with LDL and HDL have been mentioned in the literature. TFPI also is associated with Lp (a).12 Surprisingly, in subjects with cardiovascular disease both FVIIc and TFPI are increased.11 The role of the TFPI-lipoprotein complex in hyperlipidemic hypercoagulability has been studied by few authors. It has been shown that in type IIa hypercholesterolemia, TFPI is increased while FVIIc remains in the normal range. Treatment with a hypolipidemic agent (ie, simvastatin) decreases LDL and TFPI but affects FVIIc only slightly.13 FVIIa measured by coagulation and fluorogenic techniques does not seem to be involved in the enhancement of FVII activity in hypertriglyceridemia,14 arterial cardiovascular disease,15 or premature coronary heart disease16 ; this fact suggests that the enhancement of FVIIc is likely related to increased FVII synthesis. Lipoproteins may also affect the fibrinolytic process at different levels. PAI-1 is associated with the triglyceride level.17 18 19 Lp (a) has a structure similar to that of plasminogen20 and could inhibit the fibrinolytic activity.21 22 23 24 25 26 Together these data emphasize the relation between dyslipidemia and the alteration of coagulation and fibrinolytic pathways.
The aim of this prospective study was to evaluate the influence of various hyperlipidemic states on the expression of TFPI activity. We explored the relationship between serum lipids and lipoproteins and several important hemostatic and fibrinolytic factors: TFPI, FVIIc, FVIIa, TPA, and PAI-1 in patients with type IIa, IIb, and IV hyperlipidemias.
Sixty-two consecutive outpatients attending the Department of Nutrition in Hotel-Dieu Hospital, Paris, France, for various types of dyslipidemia were selected. These patients presented with a primary hyperlipidemia and had no general endocrinological or known cardiovascular disease. Patients with secondary hyperlipidemia (resulting from hypothyroidism, cholestasis, nephrotic syndrome, diabetes, etc) and those undergoing drug treatment including contraceptive steroids were excluded. Hyperlipidemic patients (32 women and 30 men) were separated into three groups: type IIa (n=22, age, 46±4 years, BMI, 25±4); type IIb (n=24, age, 45±11 years, BMI, 26±4); and type IV (n=14, age, 44±3 years, BMI, 26±6) (mean±SD) hyperlipidemias. These subjects were compared with a control group of 37 individuals matched for age and sex (21 women and 16 men; age, 39±11 years, BMI, 22±3) (mean±SD) (Table 1⇓).
Samples were collected in vacuum tubes at 8 am after an overnight fast. For determination of plasma TFPI activity, FVIIc activity, FVIIa activity, Lp (a), and PAI-1, blood was collected in siliconized glass tubes containing 1:10 vol sodium citrate 3.8% (Becton Dikinson). For TPA activity assay, samples were collected in acidic citrate tubes (Stabilyte tubes, Biopool). Plasma was prepared by centrifugation at 3000g for 30 minutes at 8°C. Aliquots of plasma were stored within 2 hours in plastic tubes at −70°C until analysis.
Samples were collected in EDTA tubes (Becton Dikinson) for determination of serum triglyceride and cholesterol levels, in sodium oxalate tubes (Becton Dikinson) for serum glucose determination, and in dry tubes for serum apolipoprotein assays. Blood was centrifuged 15 minutes at 7000g and was tested the same day.
Lyophilized FVII-deficient human plasma was purchased from Diamed; calcium-containing rabbit brain thromboplastin (Simplastin Plus) was purchased from General Diagnostics; Substrat CBS 3139, human FVII (2 U/μg), human factor X (0.1 U/μg), and rabbit thromboplastin without calcium and polybrene (10 mg/mL) were purchased from Diagnostica Stago.
TFPI assay buffer was composed of 0.05 mol/L Tris-HCl (Prolabo), 0.01 mol/L trisodium citrate (Prolabo), 0.2% NaN3 (Sigma Chemical Co), 2% bovine serum albumin (Sigma), and 2 μg/mL Polybrene (Sigma). Reference plama was pooled plasma from 20 healthy donors. rFVIIa was a kind gift from Dr U. Hedner, Novo Nordisk, Bagsvaerd, Denmark. Truncated recombinant thromboplastin (TF1-218) was a gift from Prof Y. Nemerson, Mt Sinai Medical School, New York, NY.
Serum total cholesterol and triglyceride concentrations were determined by conventional enzymatic methods with the cholesterol kit from Labinter and with the kit from Bio Merieux, respectively. Serum HDL-C level was determined after precipitation with antiserum against VLDL and LDL from Behring. LDL-C was calculated by using the Friedewald formula27 : apoAI and B were determined after precipitation with specific antibodies against apoAI and apoB from Behring. Plasma Lp (a) antigen level was determined by an ELISA technique with the Tint-Elise kit from Biopool. Serum glucose was measured with a kit from Beckman.
Coagulation and Fibrinolysis Activators and Inhibitors Assay
FVII clotting activity was determined by an automatic method on Coag-a-mate X2 from Organon Technica Corp. Human FVII–deficient plasma (100 μL) was mixed with 100 μL plasma diluted 1:40 with Owren buffer; 200 μL calcium-containing tissue thromboplastin was added, and the clotting times were recorded. Standard curves were established by the dilution of the hemostasis reference plasma from Biopool calibrated by conventional one-stage clotting assay against the first international standard 84/665 from the National Institute for Biological Standards and Control. FVIIa was determined with a newly developed one-stage clotting assay as described by Wildgoose et al28 using an ACL-300 R automated coagulation instrument from the Instrumentation Laboratory with clotting times determined by spectrophotometry (threshold set at 6.5% of baseline). Briefly, the procedure was as follows: plasma test samples were diluted fivefold in 0.1 mol/L NaCl, 0.05 mol/L Tris-HCl, 0.1% bovine serum albumin (m/v), pH 7.4. Equal volumes of test samples of bovine phospholipids (Thrombofax-Ortho Diagnostic Systems) and of immunodepleted FVII-deficient human plasma (Diagnostica Stago) were mixed (90 μL total volume) and incubated for 300 seconds; coagulation was then initiated by addition of a 60-μL aliquot of 12 nmol/L recombinant thromboplastin (TF 1-218) truncated to interact only with FVIIa.
A standard curve was performed by using rFVIIa at increasing concentrations of 0.01, 0.02, 0.05, 0.1, 0.5, 1, 3, and 10 U · mL−1 instead of diluted plasma test samples. One unit of rFVIIa is equivalent to 30 ng of protein and to 1 U of the international standard FVIIa concentrate (89/688) from The National Institute of Biological Standards and Control calibrated by conventional one-stage clotting assay against the first international standard (84/665).29
PAI-1 and TPA activities were determined with a Spectrolyse/P1 kit and Spectrolyse/Fibrin kit, respectively, from Biopool.
Statistical analyses were performed at the INSERM U258 laboratory by using sas software.31 ANOVA and ANCOVA were performed with the General Linear Model procedure to compare the different mean values. Pearson’s coefficient of correlation was determined to study the correlations and to determine the independence of the correlations. Statistical significance was accepted when the probability of occurrence by chance was <.05.
Serum Lipids and Apo Levels in Hyperlipidemias Compared With Control Group
Table 1⇑ displays the mean values of serum lipids and apolipoproteins in the different groups. The mean serum triglyceride level was fivefold higher and twofold higher in type IV and IIb hyperlipidemias, respectively, than in the control group. The total serum cholesterol mean values were 53% (P<.001) and 44% (P<0.001) higher in type IIa and IIb hyperlipidemias, respectively, than in the control group. This increase was due to the increase of LDL-C, which was 75% and 50%, respectively, higher (P<.001) than the control group. Mean HDL-C was moderately higher in type IIa hyperlipidemia (13%, P<.05) but not in type IIb hyperlipidemia; it was lower in type IV hyperlipidemia (28%, P<.001). ApoAI, present in HDL, was also lower in type IV hyperlipidemia (28%, P<.001); apoB, present in VLDL and LDL, was higher in all types of hyperlipidemia: 55% (P<.001), 67% (P<.001), and 70% (P<.001) in type IIa, IV, and IIb hyperlipidemia, respectively, compared with the control group. Lp (a) was higher in IIa hyperlipidemia (P<.02) than in the control group.
Variation of Serum Glucose Level and BMI in Different Hyperlipidemias Compared With Control Group
Subjects with hyperlipidemia had significantly higher BMIs than the control group: 12% (P<.02), 23% (P<.001), and 18% (P<.001) in type IIa, IV, and IIb hyperlipidemia, respectively (Table 1⇑).
The mean fasting glucose level was significantly higher in type IV hyperlipidemia (35%, P<.001) and in type IIb hyperlipidemia (13%, P<.05).
Comparison of Hemostatic and Fibrinolytic Activators and Inhibitors in the Three Groups of Hyperlipidemia and in the Control Group
Compared with the control group, FVIIc activity was not significantly higher in type IIa hypercholesterolemia but was significantly higher in both type IV and IIb hyperlipidemia (30%, P<.001) (Table 2⇓). FVIIa was not significantly different in any group compared with the control group.
TFPI activity was significantly higher (70% and 36%, P<.001) in type IIa and type IIb hyperlipidemia, respectively, compared with the control group and with type IV hyperlipidemia. It was slightly decreased in type IV hyperlipidemia compared with the control group (13%, P=.05) (Table 2⇑). This lower TFPI activity is dependent on the LDL-C level, which is also lower in that group (Table 2⇑).
Compared with the control group, PAI-1 activity was twofold higher (P<.02) in type IIa hyperlipidemia and threefold higher (P<.001) in type IIb and IV hyperlipidemia.
In contrast, TPA activity was not significantly different between groups (Table 3⇓).
Correlation Between TFPI, FVIIc, FVIIa, and Blood Lipids and Apolipoproteins
TFPI was positively correlated with total cholesterol (r=.774, P<.001), LDL-C (r=.800, P<.001), HDL-C (r=.251, P<.01), apoB (r=.406, P<.001), and apoA-I (r=.306, P=.001). It was also correlated with Lp (a) (Table 4⇓). Statistical analysis demonstrated the correlation dependent on LDL-C and HDL-C levels.
FVIIc activity but not FVIIa correlated with the triglyceride level (r=.444, P<.001). That correlation was dependent on apoB level (marker of LDL and VLDL) but was independent of LDL-C (marker of LDL). FVIIc was also correlated with total cholesterol but not with LDL-C and HDL-C. FVIIc was correlated with age, but the correlation between FVIIc and lipids is independent of age (Table 4⇑). FVIIa was correlated with FVIIc in the control group (r=.5, P<.001) as in hyperlipidemic groups (r=.32, P<.01).
There was no significant correlation between FVIIc and TFPI.
Correlation Between PAI-1 Activity, Lipids, and Glucose Level
PAI-1 activity was correlated with triglycerides (r=.307, P<.01), total cholesterol (r=.217, P<.05), and apoB (r=.370, P<.001) but was not correlated with LDL-C (Table 4⇑). The correlation between PAI-1 and triglycerides and PAI-1 and cholesterol was apoB dependent.
PAI-1 was also correlated with the glucose level (r=.404, P<.001) and with BMI (r=.518, P<.001) (Table 4⇑). The correlation between PAI-1 and triglycerides was glucose level dependent.
TPA activity was not correlated with any parameter measured (not shown).
In this prospective study of hyperlipidemic subjects, marked differences in TFPI levels were observed according to the type of dyslipidemia. TFPI was significantly higher in subjects with type IIa and IIb hypercholesterolemia and slightly lower in subjects with type IV hypertriglyceridemia compared with control subjects. TFPI activity was correlated with lipid and protein components of LDL (LDL-C and apoB) and of HDL (HDL-C and apoAI) in agreement with other studies.13 15 16 Interestingly, we found a correlation in addition between Lp (a) and TFPI. These correlations may be explained by the presence of TFPI complexed with HDL, LDL, and Lp (a). In hypercholesterolemia, in accordance with the observations of Sandset et al, increased TFPI activity is associated with close-to-normal FVIIc activity. However, the techniques measuring the TFPI using very long incubation times as used in the present study do not take into account the kinetics of TFPI activity. Previous reports had established the existence of a correlation between higher levels, in hyperlipidemia, of dense small LDL (lipoprotein species most prone to oxidation in vitro32 ) and atherosclerosis.33 Lesnik et al12 suggested that, during oxidation, TFPI complexed to dense LDL in the subendothelium may be inactivated and therefore be unable to inhibit TF expressed on macrophages that generate atherosclerosis. Another hypothesis is that increased circulating LDL may attract TFPI from the endothelium, making the endothelial surface more susceptible to thrombosis.34 In contrast to Sandset et al,13 we found a negative correlation between TFPI activity and triglyceride levels. The lower level of TFPI activity in type IV hyperlipidemic subjects compared with other groups may be related to several factors. First, the decrease of lipoprotein lipase activity in hypertriglyceridemia may lead to a decreased degradation of VLDL and subsequently to a decrease in free TFPI, known to have a higher anticoagulant activity than bound TFPI. Second, a decrease in the amount of TFPI may also be induced by TFPI consumption after binding to excess FVIIa/TF or by the decrease of LDL and HDL levels, its main carrier lipoproteins. This last hypothesis is the most likely explanation, as the negative correlation between triglycerides and TFPI is dependent on LDL-C level. FVIIc was significantly increased in type IIb and IV hyperlipidemia and correlated with triglycerides while FVIIa was not. Our results using a new specific method for the evaluation of FVIIa are in accord with the previous published works.15 16 These results favor the hypothesis that higher FVIIc concentrations in hyperlipidemic patients are due partly to enhancement of synthesis of FVIIc and that a part of this FVII circulates in an activated chemical form.
Fibrinolysis is also modified in hyperlipidemia (especially in hypertriglyceridemia), and the correlation between increased PAI-1 activity, triglyceridemia, and cholesterolemia may be explained by an increase of PAI-1 synthesis in endothelial cells, induced by a high level of VLDL.35 The high PAI-1 level in hyperlipidemic subjects may also be explained by PAI-1 correlation with glycemia and BMI, which also are both higher than in the control group.
The correlation of PAI-1 with blood glucose is consistent with the correlation between PAI-1 and insulin observed by other authors36 and with the increased synthesis of PAI-1 induced by insulin.37 These observations are in accord with the hypothesis of a possible link between increased PAI-1 levels and insulin resistance.20
We did not find any correlation between TFPI, TPA, and PAI-1. In contrast, FVIIc is correlated with PAI-1, and this correlation is dependent on triglycerides. The association in hypertriglyceridemic patients of hypercoagulability (increased FVIIc and the decreased TFPI activity) and hypofibrinolysis (increased PAI-1) may explain the predisposition to thrombosis of some of these patients.22
It appears that plasma TFPI activity does not substantially regulate the hyperlipidemia hypercoagulability. However, the slight decrease of TFPI activity observed in this preliminary study in hypertriglyceridemic patients should be taken with caution as a risk factor for thrombosis and should be confirmed by further clinical studies. TFPI bound to endothelial cell may play a more important role in vivo. It would be interesting to study the increased levels of endothelium-derived TFPI in plasma induced by the injection of heparin.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|FVIIc||=||coagulant factor VII|
|Lp (a)||=||lipoprotein (a)|
|PAI-1||=||plasminogen activator inhibitor-1|
|TFPI||=||tissue factor pathway inhibitor|
|TPA||=||tissue plasminogen activator|
We thank Prof B. Guy-Grand and coworkers from the Department of Nutrition at Hôtel-Dieu Hospital, Paris, France, and M.F. Bloch from Laboratoire de Thrombose Expérimentale, Université Pierre et Marie Curie-Paris VI for their excellent technical assistance, and Dr P.Y. Scarabin and V. Nicaud from Unité INSERM 258 Hopital Broussais, Paris, for their help in statistical analysis.
De Sousa JC, Soria C, Ayrault-Jarrier M, Pastier D, Bruckert E, Amiral J, Bereziat G, Caen JP. Association between coagulation factors VII and X with triglyceride rich lipoproteins. J Clin Pathol. 1988;41:940-944.
Nemerson Y. Tissue factor and haemostasis. Blood. 1988;71:1-8.
Wilcox JN, Smith KM, Schwartz SM, Gordon D. Localization of tissue factor in normal vessel and in the atherosclerotic plaque. Proc Natl Acad Sci U S A, 1989;86:2839-2843.
Novotny WF, Girard TJ, Miletich JP, Broze GJ Jr. Purification and characterization of the lipoprotein-associated coagulation inhibitor from human plasma. J Biol Chem. 1989;264:18832-18837.
Lesnik P, Vonica A, Guerin M, Moreau M, Chapman MJ. Anticoagulant 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.
Sandset PM, Lund H, Norseth J, Abildgaard U, Ose L. Treatment with hydroxymethylglutarylcoenzyme A reductase inhibitors in hypercholesterolemia induces changes in components of the extrinsic coagulation system. Arterioscler Thromb. 1991;11:138-145.
Bara L, Zitoun D, Bloch MF, Basdevant A, Samama M. Study of coagulation fibrinolysis balance in human hyperlipidemia=contribution of factor VIIa in extrinsic hypercoagulability. XIIth International Congress on Fibrinolysis, Leuven, Belgium, September 18, 1994. Fibrinolysis. 1994;8(suppl):I-14. Abstract 39.
Kario K, Miyata T, Sakata T, Matsuo T, Kato H. Fluorogenic assay of activated factor VII plasma factor VIIa levels in relation to arterial cardiovascular diseases in Japanese. Arterioscler Thromb. 1994;14:265-274.
Edelberg JM, Pizzo SV. Lipoprotein (a): the link between impaired fibrinolysis and atherosclerosis. Fibrinolysis. 1991;5:135-143.
Schneiderman J, Sawdey MS, Keeton MR, Bordin GM, Bernstein EF, Dilley RB, Loskutoff DJ. Increased type 1 plasminogen activator inhibitor gene expression in atherosclerotic human arteries. Proc Natl Acad Sci U S A. 1992;89:6998-7002.
Rath M, Niendorf A, Reblin T, Dietel M, Krebber HJ, Beisiegel U. Detection and quantification of lipoprotein(a) in the arterial wall of 107 coronary bypass patients. Arteriosclerosis. 1989;9:579-592.
Cushing GL, Gaubatz JW, Nava ML, Burdick BJ, Bocan TMA, Guyton JR, Weilbaecher D, DeBakey ME, Lawrie GM, Morrisett JD. Quantitation and localization of apolipoproteins [a] and B in coronary artery bypass vein grafts resected at re-operation. Arteriosclerosis. 1989;9:593-603.
Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499-502.
Wildgoose P, Nemerson Y, Hansen LL, Nielsen FE, Glazer S, Hedner U. Measurement of basal levels of factor VIIa in hemophilia A and B patients. Blood. 1992;80:25-28.
Hubbard AR, Barrowcliffe TW. Measurement of activated factor VII using soluble mutant tissue factor—proposal for standardization. Thromb Haemost. 1994;72-4:649-650. Letter.
sas (Statistical Analysis System) procedure. Cary, NC: SAS Institute Inc; 1993.
Dejager S, Bruckert E, Chapman MJ. Dense low-density lipoprotein subspecies with diminished oxidative resistance predominate in combined hyperlipidemia. J Lipid Res. 1993;34:295-308.
Hansen J-B, Huseby N-E, Sandset PM, Svensson B, Lyngmo V, Nordøy A. Tissue-factor pathway inhibitor and lipoproteins:evidence for association with and regulation by LDL in human plasma. Arterioscler Thromb. 1994;14:223-229.
Stiko-Rahm A, Wiman B, Hamsten A, Nilsson J. Secretion of plasminogen activator inhibitor-1 from cultured human umbilical vein endothelial cells is induced by very low density lipoprotein. Arteriosclerosis. 1990;10:1067-1073.
Juhan-Vague I, Alessi MC, Joly P, Thirion X, Vague P, DeClerck PJ, Seradimigni A, Collen D. Plasma plasminogen activator inhibitor-1 in angina pectoris: influence of plasma insulin and acute-phase response. Arteriosclerosis. 1989;9:362-367.