SEARCH

SEARCH BY CITATION

Keywords:

  • antibodies to factor XII;
  • antiphospholipid antibodies;
  • antiphospholipid syndrome;
  • factor XII;
  • MultipinTM peptide synthesis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Summary.  Phospholipid binding proteins, including factor XII (FXII), are known to be targeted by antiphospholipid antibodies (aPA). Factor XII antibodies (FXIIab) have been described in some patients with the antiphospholipid syndrome (APS) and have been shown to lead to reduced levels of FXII. The antigenic binding site(s) and the pathophysiological effects of FXIIab are unknown. In an attempt to elucidate the binding site of these antibodies, immobilized plasma kallikrein was used to cleave FXII into its 52-kDa heavy-chain (HCFXII) and 28-kDa light-chain (LCFXII) components. Plasma samples from 12 female patients with definite APS and FXIIab were investigated for the presence of antibodies to FXII, HCFXII and LCFXII. All but one patient's plasma reacted to FXII, HCFXII and LCFXII in a similar manner. One patient gave markedly reduced positivity to HCFXII and LCFXII, suggesting that the FXIIab in this patient had a higher affinity for the intact FXII molecule. To further investigate the antigenic binding site(s) of FXII, 150 biotinylated peptides of the known FXII sequence were synthesized using a MultipinTM peptide synthesis procedure. The IgG and IgM fractions of the 12 patients’ plasma were purified by affinity chromatography. The synthesized peptides were captured on streptavidin plates and individual patients’ purified FXIIab assayed against the peptides in a modified enzyme-linked immunosorbent assay (ELISA). Two regions were identified as possible antigenic binding site(s) for FXIIab: one in the growth factor domain and the other in the catalytic domain.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Although the blood plasma protein factor XII (FXII) was originally identified as a coagulation defect, by in vitro coagulation assays [1], it is now assumed that FXII has a limited role in blood coagulation. Recent research has shown that this protein is intimately linked with the other defence systems of blood (kallikrein–kinin, complement and fibrinolysis) and affects the activities of platelets, polymorphonuclear lymphocytes and other cells [2,3].

Factor XII is a serine protease zymogen, produced by the liver. It has a single peptide chain (596 amino acids), with a molecular weight of 80–90 kDa and an isoelectric point between 6.1 and 6.5 [4]. Factor XII circulates as an inactive zymogen at a concentration of approximately 30 μg mL−1 in plasma [5].

The process of activation of FXII is still being debated; however, it is known that in vitro FXII is adsorbed onto negatively charged surfaces and spontaneously changes to an active form, FXIIa. FXIIa converts prekallikrein (PK) to kallikrein (KK) and KK then activates more FXII by cleaving it at specific sites.

Factor XII affects fibrinolysis either directly by the activation of plasminogen [6], or indirectly, either through the release of tissue-type plasminogen activator (t-PA) [7] or by converting prourokinase to urokinase via activation of plasma kallikrein [8].

Previous investigations in patients with the antiphospholipid syndrome (APS) revealed the presence of antibodies to factor XII (FXIIab) in approximately half of the subjects. FXIIab were identified using an enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance [9,10]. Further studies showed that the presence of these antibodies led to reduced levels of FXII and showed a statistically significant association with multiple fetal loss [11,12]. The significance of FXIIab on the physiological activities of FXII is as yet unexplained. However, the identification of the antigenic binding site for FXIIab may give an insight into the possible role(s) of FXII.

In the present study, attempts were made to identify antigenic binding sites for FXIIab on the FXII molecule. These studies were effected by the use of purified heavy chain (HCFXII) and light chain (LCFXII) of FXII and synthesized dodeca-peptides of the FXII sequence.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

Patients and normal blood donors

Twenty one patients with well-characterized APS and 64 lupus anticoagulant-negative normal blood donors were enrolled in the study. Informed consent was obtained from all patients and normal donors and the study approved by the ethical committee of East Kent Hospitals NHS Trust. Blood samples were collected and plasma samples prepared and stored as described previously [9]. Twelve of the patients with well-characterized APS had tested positive for FXIIab on several occasions. Patients and donors were all screened for aPA as described previously [11].

Factor XII purification

Factor XII was purified as previously described [11]. The resulting FXII was >99% pure, giving a single band on sodium dodecyl sulfate (SDS) gel electrophoresis followed by silver staining and densitometric scanning. FXII showed <1% activation when added to a chromogenic peptide substrate for FXII (Unitrate FXIITM; Technoclone Ltd, Dorking, Surrey, UK).

PK purification

The procedure for PK purification was based on the method of Gallimore et al. [13]. Blood from a normal donor was collected into trisodium citrate containing benzamidine and polybrene and the plasma PK isolated using anion and cation exchange chromatography (Q Sepharose and SP Sepharose fast-flow, respectively) (Amersham Biotech, Buckinghamshire, UK). This was followed by Sephadex G100, Sephacryl, and protein G affinity chromatography (Amersham Biotech), to remove any remaining contaminants. The resultant PK gave a typical electrophoresis pattern with one band of approximately 88 kDa on SDS polyacrylamide electrophoresis followed by silver staining (Fig. 1).

image

Figure 1. Purity of prekallikrein (PK) determined by SDS gel electrophoresis followed by silver staining. Lanes 1 and 4 contain molecular weight markers, lanes 2 and 3, contain 0.5 μg of ‘in-house’ purified PK.

Download figure to PowerPoint

Activation of PK to KK

Prekallikrein was converted to KK using CONT-ACT PKTM (Channel Diagnostics, Walmer, Kent, UK) and the KK isolated using DEAE Sephadex ion-exchange chromatography (Amersham Biotech).

FXII cleavage

Purified FXII was incubated with immobilized KK coupled to cyanogen bromide (CN-Br)-activated Sepharose (Amersham Biotech) for 24 h at 37 °C. The gel mixture was then centrifuged at 800 g for 10 min and the supernatant removed.

Separation of the light and heavy chains of FXII

FXII cleavage products were separated on a Superdex 200 chromatography column, (Amersham Biotech).

Chromogenic peptide substrate assay

Factor XIIa activity was measured using a modification of the Chromogenic peptide substrate (CS) assay kit method for the determination of FXII (Unitest FXIITM; Technoclone Ltd), substituting the FXII activator in the kit with kit buffer.

Identification of LCFXII and HCFXII by ELISA

Mouse monoclonal antibodies to LCFXII and HCFXII (Biodesign International, Maine, USA) were used to identify LCFXII and HCFXII adsorbed to an ELISA plate, followed by a goat antimouse IgG alkaline phosphatase conjugate (Dako Ltd, Cambridge, UK).

Inactivation of LCFXII

The molarity of the LCFXII was increased by dialyzing against 0.05 m Tris/0.5 m NaCl for 4 h. A 1-mm solution of 4-(2-aminoethyl) benzene sulfonyl fluoride (AEBSF) (Sigma Aldrich, Poole, Dorset, UK) in 0.05 m Tris/0.5 m NaCl buffer was then added and the mixture left overnight at 4 °C on a rotor mixer. The treated LCFXII was then dialyzed against 0.05 m Tris containing 0.15 m NaCl pH 7.3 (TBS).

Antibodies to FXII, LCFXII and HCFXII by ELISA

Antibodies to FXII were determined by ELISA as previously described [9]. All samples were tested in duplicate. Antibodies to LCFXII and HCFXII were identified in an identical manner by replacing FXII with the relevant FXII fraction. The coefficients of variation for the ELISAs used are shown in Table 1.

Table 1.  Intra- and inter-assay batch variability for the determination of antibodies to FXII, FXIIHC and FXIILC
 Intra-batch CV%Inter-batch CV%
IgGIgMIgGIgM
FXII2.652.723.975.98
FXIIHC2.342.515.267.23
FXIILC2.123.26.27.95

Chiron MultipinTM peptide synthesis

Peptides were synthesized in two separate blocks and to ensure that no potential epitope binding sites were missed, a three-peptide overlap was added to the second block of peptides. Block one therefore contained 87 peptides, and block two 63 peptides, a total of 150 peptides in all for FXII (Fig. 2). Two control peptides of 12 contiguous histidine residues were added to the end of each block schedule as a control making a total of 154 peptides in all. Peptide synthesis was accomplished by repetitive cycles of Fmoc-deprotection, washing and amino acid coupling, adding one amino acid residue per cycle. After completing the synthesis of the desired peptides, a ‘spacer’ sequence of SGSG (seryl–glycl–seryl–glycl) was added at the N-terminus of each of the freshly synthesized peptides to avoid any obscuring of the peptide binding sequence on the addition of a biotin moiety. The final Fmoc-protecting group was removed and the terminal amino group was biotinylated for use in streptavidin binding to ELISA plates. Peptides generated were assessed for purity by reverse-phase high-performance liquid chromatography (HPLC) and peptide content quantitated by amino acid analysis. Positive and negative non-cleavable controls were assessed by standard ELISA techniques.

image

Figure 2. Schematic diagram of FXII showing overlapping peptides and to which, domain the peptide is associated. Peptides (150 in total) were synthesized in two blocks. Block 1, consisting of peptides 1–87; block 2, consisting of peptides 88–150.

Download figure to PowerPoint

Antibodies to FXII against streptavidin-fixed peptides

Affinity-purified IgG and IgM fractions from patients A, C, D, H, J, K and L were prepared as previously described [10]. The biotinylated synthetic peptides (5 μg ml−1) were bound to a microtiter plate coated with streptavidin. After incubation for 1 h at room temperature and non-bound peptides were removed by washing three times with TBS. Antibodies to FXII present in the plasma, bound to peptides already bound to the plate, and any non-bound antibodies were removed by washing with TBS. Dilutions of either antihuman IgG or IgM horse radish peroxidase (HRP) (Dako) conjugates were added to the plate to detect antibodies bound to the peptide. After incubation, non-bound antihuman IgG or IgM was removed by washing with TBS. A K blue substrate for peroxidase (Skybio Ltd, Bedfordshire, UK) was then added, generating color proportional to the amount of human IgG or IgM bound to the peptide. Results were calculated as optical density (OD) values at 450 nm, read on a Spectramax 340 spectrophotometer (Molecular Devices, Wokingham, UK) for either IgG or IgM.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

FXII cleavage

Plasma KK immobilized on a CN-Br activated sepharose matrix, cleaved FXII into its 52-kDa HCFXII and 28-kDa LCFXII components. The resultant mixture was then separated from the gel matrix by centrifugation and run on an SDS electrophoresis gel and the protein bands visualized by silver staining (Fig. 3). Two bands were clearly visible at molecular weights of 52 and 28 kDa. No band was visible at 80 kDa, the molecular weight of intact FXII.

image

Figure 3. Cleavage products of FXII (HCFXII and LCFXII) identified by SDS electrophoresis followed by silver staining. Lanes 1 and 3 contain molecular weight markers and lane 2 contains 0.5 μg of purified ‘in-house’ FXII cleaved with kallikrein.

Download figure to PowerPoint

Separation of the light and heavy chains of FXII

The two FXII fragments were purified on a Superdex 200 chromatography column. Each fraction from the column was read for protein at 280 nm and for enzyme activity using a CS assay (Fig. 4A). Two distinct peaks are clearly visible on the optical density chart recording. These peaks when calculated as a ratio of Ve/Vo against protein standards of known molecular weight, which corresponded to molecular weights of approximately 52 kDa and 28 kDa. The CS assay detected LCFXII containing the catalytic active site, and the assay peak corresponded to a molecular weight of 28 kDa. No activity was detected by the CS assay with HCFXII. Fractions containing either HCFXII (fraction numbers 94–104) or LCFXII (fraction numbers 122–130) were pooled separately and each showed a single band of >95% purity on a SDS electrophoresis gel followed by silver staining and densitometric scanning (Fig. 4B).

image

Figure 4. (A) Elution profile of HCFXII and LCFXII from the Superdex 200 column. (B) Silver-stained SDS electrophoresis gel of separated HCFXII and LCFXII. Lanes 1 and 6 contain molecular weight markers and lanes 2 and 3 contains 0.5 and 10 μg of LCFXII and lanes 4 and 5 contains 0.5 and 10 μg of HCFXII, respectively.

Download figure to PowerPoint

Antibodies to FXII, HCFXII and LCFXII

The integrity of our purified products was verified by a modified ELISA. A mouse monoclonal HCFXII bound to FXII and HCFXII but not LCFXII. A mouse monoclonal anti-LCFXII bound to FXII and LCFXII but not HCFXII (Table 2). The catalytic activity of LCFXII was inactivated by active site titration with AEBSF to ensure that the catalytic domain did not bind to proteins other than FXIIab. The inactivated LCFXII was then used in the subsequent FXIIab ELISA assays. Plasma were then investigated for the presence of antibodies to FXII, HCFXII and LCFXII at protein concentrations of 5 μg ml−1 by a modification of the previously published ELISA method for the detection of FXIIab [9]. The normal range was determined as the mean OD unit ±2 standard deviations (SD) of plasma samples obtained from 64 blood donors (32 male and 32 female). Antibody positivity was determined as a qualitative evaluation of optical density units not as an absolute value. Positivity for patients was determined when OD values were: FXII IgG >0.11, IgM >0.233, HCFXII IgG >0.23, IgM >0.260, LCFXII IgG >0.20, IgM >0.31. The positive patients’ values ranged from 0.137 to 0.871 (see Table 3). Plasma samples from 12 female patients with APS already identified as having either IgG or IgM FXIIab were investigated. The 12 selected patient plasmas were tested against FXII, HCFXII and LCFXII using the modified ELISA. Results are shown in Table 3. Ten of the patients tested positive for IgG FXIIab. Of these, nine patients tested positive for HCFXII and the same nine patients tested positive for LCFXII. Five of 12 patients tested positive for IgM antibodies to FXII. All five of these patients tested positive for HCFXII and for LCFXII.

Table 2.  Confirmation of FXII, HCFXII and LCFXII binding to the microtitre plate surface using polyclonal antibodies to FXII
FXII speciesProtein concentration added to plate OD 450 nm (μg mL−1)Blank OD 450 nmRabbit antihuman FXII IgG OD 450 nmMouse monoclonal to FXII heavy chain OD 450 nmMouse monoclonal to FXII light chain OD 450 nm
FXII50.0591.31.361.23
HCFXII50.0580.6080.9320.08
LCFXII50.0620.4830.0730.819
Table 3.  Patient IgG and IgM antibody binding to FXII whole molecule, HCFXII and LCFXII
 IgGIgM
FXII antibodiesHCFXII antibodiesLCFXII antibodiesFXII antibodiesHCFXII antibodiesLCFXII antibodies
  1. *Positive; Negative.

Patient A0.171*0.286*0.245*0.1980.2450.302
Patient B0.137*0.332*0.312*0.1540.2230.216
Patient C0.161*0.292*0.205*0.2190.2650.261
Patient D0.183*0.374*0.23*0.356*0.324*0.345*
Patient E0.15*0.262*0.256*0.2240.2310.265
Patient F0.221*0.1860.1690.1450.1560.123
Patient G0.110.1620.1890.254*0.268*0.327*
Patient H0.155*0.422*0.295*0.329*0.326*0.356*
Patient I0.146*0.353*0.292*0.345*0.312*0.451*
Patient J0.1020.2250.1350.255*0.332*0.462*
Patient K0.871*0.268*0.354*0.2240.2530.31
Patient L0.244*0.351*0.398*0.1560.1680.215

Enzyme-linked immunosorbant assay for the detection of FXII antibody epitope-binding sites using generated peptides

Seven patients positive for FXIIab (IgG or IgM) were chosen because of availability of plasma and their positivity for FXIIab by ELISA. IgG or IgM preparations from three FXIIab-negative patient plasma were used as controls. For each patient, two controls were used: a peptide containing 12 histidine residues (detected using a monoclonal mouse anti-6-histidine (diluted 1 : 500) and a biotinylated whole FXII protein (detected using a rabbit antihuman FXII-HRP conjugate 1 : 200). Each patient's purified IgG or IgM was tested against each peptide in duplicate and the average result represented as a bar chart. All three of the patient's plasma that tested negative for FXIIab, did not bind to any of the synthesized peptides. Plasma from only one of the seven patients who tested positive for antibodies to FXII showed clear binding to the synthetic peptides (Fig. 5). In this patient (K), peptides 43, 44 and 45 showed a classical general net binding skewed bell curve, e.g. binding strength increasing to a peak and then trailing off as the epitope moves out of the sequential peptide sequence. The peptide sequences that showed binding for FXIIab in patient K were as follows: 43: CLEVEGHRLCHC, 44: EGHRLCHCPVGY and 45: LCHCPVGYTGPF. All three peptides share the tetramer sequence LCHC. This tetramer sequence occurs only once in the FXII protein sequence and is found in the second growth factor domain of the molecule. Additional peptide sequences that showed binding for FXIIab were seen with peptides 93–98 (see Figs 2 and 5).

image

Figure 5. Patient's purified plasma IgG used as probes against biotinylated peptides (1–150) of the known FXII sequence.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References

The APS is a complex and hetereogenous condition where the presence of aPA are associated with recurrent fetal loss and/or venous or arterial thrombosis [14]. Antiphospholipid antibodies are generally not targeted toward phospholipids but rather to proteins that bind to phospholipids including beta-2-glycoprotein 1, prothrombin, FXI and PK associated with high-molecular-weight kininogen (HK) [15]. The role of these antibodies is still unclear.

Recent studies have now identified the presence of antibodies to FXII in patients with APS [9]. The FXII molecule consists of several domains (negatively charged surface binding, fibronection finger, epidermal growth factor, kringle and catalytic), that may have importance for different physiological functions.

It has been suggested that there is an association between the presence of FXIIab and multiple fetal loss in women with APS [12]. We were interested to find whether these antibodies were targeted to any particular domain of the FXII molecule. Primarily, we hoped to identify FXIIab binding selectively to either the heavy or light chain of FXII. We therefore prepared LCFXII and HCFXII from purified FXII and studied the binding of the various patients IgG/IgM to these preparations. Most of the patient plasmas studied (IgG/IgM) reacted to FXII and both the HCFXII and LCFXII in a similar manner. One patient (K) gave a markedly reduced positivity to HCFXII and LCFXII and patient F showed binding only to the whole molecule, suggesting that the FXIIab in these patients had a higher affinity for the intact FXII molecule in its native form (Table 3).

Although FXIIab appeared to be polyclonal in nature, it was still possible that specific linear epitopes of FXII might be particularly favored for binding by FXIIab. Specific binding to these linear epitopes might lead to particular clinical manifestations. Using the MultipinTM peptide synthesis technique, 154 linear peptides were successfully generated. One patient showed binding to amino acid sequences found in the second growth factor domain and also in the catalytic domain of FXII (Fig. 5). It is interesting to note that this patient (K) had the highest titre of FXIIab, which also displayed the strongest avidity for FXII. This patient had APS and a clinical history of both multiple fetal loss and venous thrombosis.

The MultipinTM peptide synthesis procedure allows the production of peptides through the linkage of amino acids to an amide bond by the process of ‘solid-phase’ chemical synthesis [16]. The carboxy-terminal amino acid is first attached to a solid phase, and then by the addition of amino acids one at a time to the amino-terminal end, long chains can be built up. This concept is opposite to the way peptides and proteins are made in nature. Natural peptide biosynthesis occurs from the amino-terminal amino acid, adding an amino acid to the carboxy-terminus. All the amino acids used in the synthesis are chemically ‘protected’ because of the chemical reactivity of their side chains. This protection is removed once the peptides are synthesized. Antibodies to FXII fulfill the criteria for successful epitope mapping by the use of MultipinTM synthesized peptides. First, the FXII amino acid sequence was fully established by Edman degradation [17] and confirmed by using full-length cDNA clones [18]. Secondly, FXIIab in lupus anticoagulant-positive patients (LA+) do not dissociate significantly under the conditions of an immunoassay as they have been identified by ELISA and through surface plasmon resonance [10]. MultipinTM synthesis will only generate the former linear ‘continuous’ peptides and it has been calculated that only 5–10% of native antibodies will bind to linear epitopes [16]. Therefore, it was not surprising that only one of the seven patients tested in the assay showed binding to linear peptides in our study. This suggests that FXIIab are likely to be directed against epitopes formed by the conformation of the FXII molecule.

It is of interest that the peptides, which bound to this patient's FXIIab, are found in the epidermal growth factor (EGF) domain of FXII and the KK cleavage site. The presence of the EGF-like domain in many different proteins with biological functions unrelated to cell proliferation, such as blood clotting (and also cell adhesion) has been suggested to be due to the evolution of these proteins from molecules with different biological functions. However, FXII has been shown to enhance human hepatoma cell proliferation [19] and to have a mitogenic effect on several other EGF-sensitive target cells by activating signal transduction pathways [20]. If FXII does have a significant role in cellular development the presence of FXIIab may have a deleterious effect on this process.

The binding of FXIIab to peptides found in a sequence in the FXII molecule around the KK cleavage site is also of interest. This site is essential for the activation of surface-bound FXII by plasma KK. This observation, together with the binding of FXIIab to the light chain of FXII (containing the catalytic domain) suggests that FXIIab could have a pronounced effect on the activation of FXII and plasma KK in vivo.

These findings raise the intriguing possibility that FXIIab in patients with APS may in someway inhibit the physiological function of FXII and thus contribute to the clinical symptoms suffered by this patient group.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  • 1
    Ratnoff OD, Colopy JE. A familial hemorrhagic trait associated with a deficiency of a clot promoting fraction of plasma. J Clin Invest 1955; 34: 60213.
  • 2
    Colman RW. Biologic activities of the contact factors in vivo. Potentiation of hypotension, inflammation, and fibrinolysis and inhibition of cell adhesion and thrombosis. Thromb Haemost 1999; 82: 156877.
  • 3
    Schmaier AH. Plasma kallikrein/kinin system: a revised hypothesis for its activation and its physiologic contributions. Curr Opin Hematol 2000; 7: 2615.
  • 4
    Griffin JH, Cochrane CG. Human factor XII (Hageman factor). In: LorandL, ed. Methods in Enzymology XLV: Proteolytic Enzymes B. New York: Academic Press, 1976: 5665.
  • 5
    Saito H, Ratnoff OD, Pensky J. Radioimmunoassay of human Hageman factor (factor XII). J Lab Clin Med 1976; 88: 50614.
  • 6
    Hauert J, Nicoloso G, Schleuning WD, Bachmann F, Schapira M. Plasminogen activators in dextran sulfate-activated euglobulin fractions: a molecular analysis of factor XII- and prekallikrein-dependent fibrinolysis. Blood 1989; 73: 9949.
  • 7
    Smith O, Gilbert M, Owen WG. Tissue plasminogen activator release in vivo in response to vasoactive agents. Blood 1985; 66: 8359.
  • 8
    Ichinose A, Fujikawa K, Suyama T. The activation of pro-urokinase by plasma kallikrein and its inactivation by thrombin. J Biol Chem 1986; 261: 34869.
  • 9
    Jones DW, Gallimore MJ, Harris SL, Winter M. Antibodies to factor XII associated with lupus anticoagulant. Thromb Haemost 1999; 81: 387390.
  • 10
    Jones DW, Nicholls P, Donohue S, Gallimore MJ, Winter M. Antibodies to factor XII are distinct from antibodies to prothrombin in patients with the anti-phospholipid syndrome. Thromb Haemost 2002; 87: 426430.
  • 11
    Jones DW, Gallimore MJ, Mackie IJ, Harris SL, Winter M. Reduced factor XII levels in patients with the antiphospholipid syndrome are associated with antibodies to factor XII. Br J Haematol 2000; 110: 7216.
  • 12
    Jones DW, Gallimore MJ, Winter M. Antibodies to factor XII: A possible predictive marker for recurrent foetal loss. Immunobiology 2003; 207: 4346.
  • 13
    Gallimore M.J, Fareid E, Stormorken H. The purification of a human plasma kallikrein with weak plasminogen activator activity. Thromb Res 1978; 12: 409420.
  • 14
    Wilson WA, Gharavi AE, Koike T, Lockshin MD, Branch DW, Piette JC, Brey R, Derksen R, Harris EN, Hughes GRV, Triplett DA, Khamashta MA. International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome. Arthritis Rheum 1999; 42: 130911.
  • 15
    Mcintyre JA, Wagenknecht DR. Anti-phosphatidylethanolamine (aPE) antibodies: a survey. J Autoimmun 2000; 15: 18593.
  • 16
    Geysen HM, Rodda SJ, Mason TJ, Tribbick G, Schoofs PG. Strategies for epitope analysis using peptide synthesis. J Immunol Methods 1987; 102: 25974.
  • 17
    McMullen BA, Fujikawa K. Amino acid sequence of the heavy chain of human alpha-factor XIIa (activated Hageman factor). J Biol Chem 1985; 260: 532841.
  • 18
    Cool DE, Edgell C-JS, Louie GV, Zoller MJ, Brayer GD, McGillivray RTA. Characterization of human blood coagulation factor XII cDNA. Prediction of the primary structure of factor XII and the tertiary structure of beta-factor XIIa. J Biol Chem 1985; 260: 1366676.
  • 19
    Schmeidler-Sapiro KT, Ratnoff OD, Gordon EM. Mitogenic effects of coagulation factor XII and factor XIIa on HepG2 cells. Proc Natl Acad Sci USA 1991; 15: 43825.
  • 20
    Gordon EM, Venkatsen N, Salazar R, Tang H, Schmeidler-Sapiro K, Buckley S, Warburton D, Hall FL. Factor XII-induced mitogenesis is mediated via a distinct signal transduction pathway that activates a mitogen-activated protein kinase. Proc Natl Acad Sci USA 1996; 93: 21749.