Background
The blood coagulation system is triggered by either the intrinsic or extrinsic pathway and involves sequential enzymatic activations of serine protease zymogens enhanced by non-enzymatic cofactors, factors Va and VIIIa, resulting in generation of thrombin [
1,
2]. Thrombin generation is central to the regulation of hemostasis and thrombosis and to the pathogenesis of cardiovascular disease and venous thrombosis. The intrinsic pathway is triggered by contact activation. Recent studies using mouse thrombosis models suggest that the intrinsic pathway is essential for pathological thrombus formation in both the arterial [
3‐
7] and venous systems [
7]. The “contact activation” of plasma involves 3 zymogens, factor XII, factor XI and prekallikrein, plus a non-enzymatic cofactor, high molecular weight kininogen, that participate in a set of interrelated proteolytic activation reactions [
4,
8‐
15]. During the initial stages of contact activation, factor XII and prekallikrein participate in reciprocal proteolysis in which factor XIIa activates prekallikrein to kallikrein, which in turn converts factor XII into factor XIIa. Activation of factor XII can be enhanced by autoactivation of zymogen factor XII by the active form α-factor XIIa. These reactions are accelerated by negatively charged surfaces (
e.g., kaolin, sulfatides, and dextran sulfate).
Plasma lipids and lipoproteins can influence both procoagulant and anticoagulant reactions [
16,
17]. Furthermore, dyslipidemia and dyslipoproteinemia are associated with hypercoagulability and venous thromboembolism (VTE) [
16,
17]. The lipid transfer protein, cholesteryl ester transfer protein (CETP), which carries and transfers lipids to modulate plasma lipoprotein levels, can exert procoagulant activity by enhancing prothrombinase activity in purified systems [
18]. Furthermore, plasma CETP mass level was correlated with relative hypercoagulability of plasma independent of high density lipoprotein (HDL) levels. Although molecular mechanisms for CETP procoagulant activity remain unclear, we hypothesized that other lipid transfer proteins can also affect blood coagulation.
Plasma phospholipid transfer protein (PLTP), a homolog of CETP, circulates in plasma and facilitates the transfer of phospholipids and cholesterol among lipoproteins [
19‐
27]. It can mediate the conversion of HDL into larger and smaller particles [
28,
29] and generate pre-β HDL in the process [
30]. PLTP also transfers phospholipids from very low density lipoprotein (VLDL) to HDL by PLTP shuttling from HDL to VLDL particles [
31]. In clinical studies, high PLTP activity was reported to be associated with the increased risk of coronary artery disease [
32,
33], notably in statin-treated patients [
34]. In contrast, low PLTP activity was reported to be associated with peripheral atherosclerosis [
35], suggesting that the relationship of plasma levels of PLTP activity to cardiovascular risk is controversial [
32‐
38]. No study has reported an assessment of the effects of PLTP on blood clotting reactions, such as reflected in thrombin generation assays, or on the association of VTE risk with PLTP activity.
Here we report that PLTP can inhibit sulfatide-induced contact activation of thrombin generation, that factor XII binds directly to PLTP, and that low plasma PLTP activity levels may be associated with VTE risk.
Methods
Materials
Fatty acid-free bovine serum albumin (BSA) was purchased from Calbiochem (San Diego, CA). Human factor XIa and corn trypsin inhibitor were from Hematologic Technologies Inc. (Essex Junction, VT). Human factor XII, prekallikrein and kallikrein were from Enzyme Research Laboratories (South Bend, IN). Human α-factor XIIa was from Aniara (Mason, OH). VLDL was purchased from Intracel (Frederic, MD). Sulfatide was obtained from Matreya (Pleasant Gap, PA). Kaolin was from Fisher Scientific Inc. (Pittsburgh, PA) and dextran sulfate was from GE Healthcare (Parsippany NJ). Innovin was from DADE (BioMerioeux). Chromogenic substrate S2302 and fluorogenic substrate I-1140 were obtained from Chromogenix (Franklin, OH) and Bachem Bioscience Inc. (King of Prussia, PA), respectively. Normal human pooled plasma was prepared using blood obtained from 20 adult healthy donors (10 males and 10 females) by routine venipuncture from the Scripps General Clinical Research Center’s (GCRC) blood donation program after an overnight fast. Blood was mixed with 0.129 M sodium citrate at 1:9 ratios. Plasma was prepared by centrifugation at 2,000 × g for 20 min at room temperature from each donor and pooled. The pooled normal human plasma was stored at −80 °C. Blood from VTE patients was collected in the GCRC at least 3 months after VTE diagnosis and after a 12 h fast.
Recombinant (r) PLTP
Recombinant wild-type PLTP was made and characterized as full length molecules as reported [
29,
39]. Briefly, rPLTP was isolated by Ni
2+-nitrilotriacetic acid resin column chromatography from serum-free conditioned culture medium collected from baby hamster kidney cells transfected with a His-tagged human rPLTP cDNA using methotrexate as selection agent. The isolated rPLTP fractions were assayed for phospholipid transfer activity and evaluated for purity by SDS-PAGE. The concentration of rPLTP was determined by the absorbance at 280 nm. rPLTP was stabilized by adding phosphatidylcholine vesicles in Tris-buffered saline (TBS) containing 0.05 M Tris, 0.15 M NaCl, pH 7.4. The same buffer containing the same concentration of phosphatidylcholine vesicles without rPLTP was used as control for activity assays.
Preparation of sulfatide vesicles
Sulfatide vesicles were prepared by sonication and stored up to two days. Briefly, 1 mg/ml bovine sulfatide in TBS was sonicated 5 times for 30 s with 1 min intervals using the ultrasonic processor XL (Heat System, Inc., Farmingdale, NY) under the flow of N2.
Thrombin generation assay in plasma
Plasma thrombin generation assays were performed as described with some modifications [
40]. Pooled normal human plasma (30 μl) was incubated with various rPLTP concentrations for 15 min at 37 °C. Then, tissue factor (Innovin, final 4 pM) or sulfatide vesicles containing 30 mM CaCl
2 and fluorogenic thrombin substrate solution (I-1140) was added to the plasma mixture (total 110 μl) to initiate coagulation activation. In factor XIa (0.13 nM, final) -induced thrombin generation assays, corn trypsin inhibitor (50 μg/ml, final) was also pre-incubated with plasma. Thrombin generation was followed continuously using SPECTRAmax GEMINI XS fluorometer (Molecular Devices, Sunnyvale, CA) with excitation and emission wavelengths set at 360 and 460 nm, respectively. The first derivative of fluorescence versus time was used to produce thrombin generation curves.
Kallikrein generation assay in plasma
The generation of plasma kallikrein in the presence of sulfatide was determined as described with some modifications [
8]. Briefly, plasma (1:100, final dilution) was incubated with PLTP for 15 min at room temperature, followed by the addition of sulfatide (0.5 μM, final) and then incubated at various times. Kallikrein amidolytic activity was measured by hydrolysis of the kallikrein amidolytic substrate, S2302 (0.4 mM, final). Soy bean trypsin inhibitor (500 μg/ml, final), which inhibits kallikrein activity, inhibited S2302 hydrolysis in this assay system by approximately 95 %, whereas corn trypsin inhibitor, which specifically inhibits factor XIIa, inhibited S2302 hydrolysis by 5 % (data not shown). These showed that the S2302 hydrolysis by contact phase activation in this plasma assay system mainly reflects kallikrein generation rather factor XIIa generation due to the different reactivities of kallikrein and factor XIIa with the substrate [
13].
Factor XII activation in purified protein system
For factor XII autoactivation assays, factor XII (0.1 μM, final) was pre-incubated with rPLTP in the presence or absence of VLDL (25 μg protein/mL, final) for 15 min at room temperature, and then sulfatide vesicles (0.5 μM, final) were added to start autoactivation of factor XII. S2302 (0.4 mM, final) was added at various time points and the amidolytic activity of factor XIIa was measured [
8,
9].
Activation of factor XII by kallikrein was measured as described [
10]. The reaction mixture containing factor XII (0.1 μM, final) with or without rPLTP (5 μg/mL, final) in the presence or absence of VLDL (25 μg protein/mL, final) was pre-incubated for 15 min at 37 °C, followed by the addition of kallikrein (0.4 nM, final) and incubated for optimal time. The reaction was stopped by adding soybean trypsin inhibitor (500 μg/ml final) to block kallikrein. The generated factor XIIa activity was measured as amidolytic activity for S2302 (0.4 mM, final).
Prekallikrein activation by factor XIIa
Activation of prekallikrein by factor XIIa was measured as described [
12]. The reaction mixture containing prekallikrein (0.25 μM, final) with or without rPLTP (5 μg/mL, final) in the assay buffer was pre-incubated in the presence or absence of VLDL (25 μg protein/mL, final) for 15 min at 37 °C, followed by the addition of factor XIIa (0.1 nM, final) and incubated for optimal times. The kallikrein activity generated was measured as amidolytic activity for S2302 (0.4 mM, final).
Factor XII binding to PLTP
Binding was assessed by surface plasmon resonance (SPR) analysis using a BIAcore 3000 biosensor system. An anti-His tag monoclonal antibody was covalently immobilized on the carboxymethylated dextran (CM5) sensor chip (BIAcore) using amine coupling chemistry according to the manufacturer’s instructions. A nonreactive mouse IgG was used as a control for nonspecific binding. rPLTP with a C-terminal His-tag (100 μg/ml) was diluted in 50 mM Hepes, 150 mM NaCl, and 5 mM CaCl2 (pH 7.4) and injected at a flow rate of 10 μL/min with 10 min contact time generating a response. Then, each concentration of factor XII was injected in this buffer for 1.5 min at a flow rate of 5 μL/min. After each Factor XII sensorgram was obtained, the His tag-antibody surface was regenerated with 10 mM glycine/HCl, pH 2.5, and a new injection of His-tag PLTP was used to regenerate the surface. Any influence of mass transport effects was discounted from the results of binding and dissociation at different flow rates. Data analysis was performed by using BIAevaluation software 3.0 (BIAcore). The association and dissociation phases of all ‘sensorgrams’ were fitted globally. Rate constants for association (ka) and dissociation (kd) were obtained by globally fitting the data from five to six injections of factor XII (0–750 nM) by using the BIAevaluation software version 3.2, using the simple Langmuir binding model.
VTE study group
The Scripps Venous Thrombosis Registry is an ongoing case–control study of risk factors for VTE as described [
41]. Inclusion criteria for this study included age at thrombosis < 55 years old, >3 months since diagnosis of acute thrombosis, a life expectancy of at least 3 years and no lipid lowering medications or cancer. Age matched (±2 years) healthy male controls were recruited through the GCRC blood donation program at Scripps. The protocol was approved by the Institutional Review Board of Scripps Clinic and subjects provided written informed consent. Forty of 49 VTE patients (82 %) presented with idiopathic VTE, defined as events that did not occur within 90 days after surgery, trauma, or major immobilization. In this study, male idiopathic VTE patients (
n = 40) and controls (
N = 40) were analyzed for PLTP activity and mass (Additional file
1: Table S1).
Clinical analytes
Serum lipid profile data were obtained from the routine clinical lab using standard techniques.
Lipoprotein subclass particle concentrations were measured by Nuclear Magnetic Resonance Spectroscopy (NMR) at LipoScience, Inc (Raleigh,NC) [
41]. PLTP activity (total plasma activity) and mass were determined as described [
23,
26].
Statistical analysis
Data for VTE patients and controls was compared for median values using the Mann–Whitney test using Prism™ 4.0 software (Graph Pad Software Inc., San Diego, CA). Odds ratios for VTE were determined with a logistic regression model using Minitab software. The difference was considered significant when p was < 0.05.
Discussion
In this study we showed that a recombinant plasma protein, rPLTP, can inhibit sulfatide-induced thrombin generation and contact activation in plasma, implying that it may function as an anticoagulant factor where contact activation is involved. This activity of rPLTP was not apparent when thrombin generation in plasma was triggered by factor XIa or tissue factor, implying a direct effect of rPLTP on contact activation reactions. Factor XII is the key enzyme for contact activation [
1‐
7] and, indeed, we found that factor XII binds directly to rPLTP, thereby providing a starting point for mechanistic studies for the actions of rPLTP.
Interestingly, rPLTP did not inhibit contact activation in plasma stimulated by two other negatively charged activators of the contact system, dextran sulfate and kaolin, further implying that the anticoagulant actions of rPLTP were not due to a nonspecific blocking of negative charges. These findings might be supported by reports that mechanisms responsible for surface activation triggered by different negatively charged molecules (
e.g., sulfatide and kaolin) differ from each other as regards the molecular interaction with the contact factors [
11,
14]. However, the details for the mechanistic relations of PLTP to surface materials for contact activation of coagulation including other naturally occurring negatively charged surfaces effect (
e.g., misfolded proteins, and polyphosphates) need to be evaluated in future studies.
The requirement of VLDL, which also could provide surface for contact pathway of coagulation system [
42], for the inhibition of sulfatide-induced autoactivation of factor XII by rPLTP in purified protein reaction mixtures may reflect direct or indirect effects of VLDL on factor XII and/or rPLTP. Recent work shows that a significant fraction of PLTP is bound to VLDL in plasma [
39], so the ability of rPLTP to inhibit sulfatide-induced contact activation in plasma or factor XII autoactivation likely reflects actions of a PLTP●VLDL complex in plasma. Extensive and detailed mechanistic studies using rPLTP, factor XII and α-factor XIIa plus VLDL and many proteins that associate with VLDL [
43] would be needed for studies to clarify details and elucidate the structural basis for the effects of these molecules on contact activation or factor XII autoactivation. Further, other lipoproteins (
i.e., HDL and LDL), which were not tested here, could also possibly contribute to the anticoagulant activity of PLTP.
Binding studies using purified proteins show direct interactions between factor XII and PLTP. For this protein-protein interaction, the apparent Kd value of 190 nM is below the plasma level of factor XII which is 300 nM [
44] but above the plasma level of PLTP at 25 nM [
19]. Perhaps reflecting this modest value of apparent Kd, PLTP alone does not inhibit sulfatide-induced factor XII autoactivation in purified reaction systems. However, when VLDL was added to the purified system, PLTP does inhibit Factor XII autoactivation, showing additional plasma components are needed. Thus, because VLDL particles carry many associated proteins in addition to PLTP [
43], we speculate that the whole VLDL particles or one or more proteins in the VLDL interactome enhances affinity of PLTP for factor XIIa which at very low levels is responsible for triggering factor XII autoactivation. Additionally, factor XII conformational changes induced by sulfatide or zinc ions [
14] might influence the affinity of factor XIIa for PLTP. However, mechanisms for inhibition of factor XII autoactivation by PLTP in plasma are not completely clear and they need further clarifications including the potential influences of VLDL, sulfatide, or zinc ions on factor XIIa-PLTP binding interactions.
Factor XII is not required for normal hemostasis in man or various animals. However, studies of thrombosis injury models in mice genetically deficient in factor XII or XI suggest that factor XII-dependent thrombin generation via the intrinsic coagulation pathway can significantly contribute
in vivo to fibrin formation and thrombosis including a pulmonary embolism model [
3‐
7]. This concept plus the ability of PLTP to inhibit thrombin generation in plasma led us to assay PLTP activity and mass in plasmas of a young adult male VTE cohort that we have previously described [
23,
41]. The initial analysis of our data without any adjustments for variables failed to show any association of plasma PLTP activity and mass levels with VTE risk. However, a very significant association between low PLTP activity and VTE became apparent after making adjustments for various lipoprotein levels (Table
1, models IV and V) or by analyzing separately the subgroup of normolipidemic patients (Additional file
1: Figure S4).
It appears that levels of plasma lipoproteins, which are indeed related to PLTP-mediated phospholipid transfer activity [
23], can mask the association of PLTP activity with VTE risk. Overall, our finding here is consistent with the concept that active PLTP molecules, putatively in VLDL●PLTP complexes [
43], contribute to the multiple, highly varied antithrombotic activities of plasma. One notes that major limitations of this pilot VTE study include the low number of subjects, the age of subjects under 55 years old, and the absence of female cohorts.
Factor XII-dependent contact activation not only might contribute
in vivo to fibrin formation and excessive thrombosis including pulmonary embolism model [
3‐
7] but also might contribute to inflammation and pathologies via bradykinin formation and complement activation [
5,
6,
45]. Thus, inhibition of contact activation mediated by PLTP might inhibit not only thrombosis but also inflammatory processes stimulated or supported by contact activation. In this regard, it is interesting that PLTP was shown to exert anti-inflammatory activities [
46,
47], although no information relating PLTP’s anti-inflammatory actions to any pro-inflammatory actions due to contact activation have been described.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JHG, HD and JJA participated in the conception of the study. GW and MCC were responsible for rPLTP preparation and quantification of plasma levels of PLTP mass and activity. HD was responsible for performing experiments with PLTP activity and statistical analysis. DJE was responsible for consenting the VTE patients and controls and for obtaining blood specimens. YB and JAF were responsible for performing SPR experiments. HD, DJE and JHG were responsible for writing the manuscript. All authors read and approved the final manuscript.