A human monoclonal antiprothrombin antibody is thrombogenic in vivo and upregulates expression of tissue factor and E-selectin on endothelial cells


Silvia S. Pierangeli, PhD, Division of Rheumatology, Department of Internal Medicine, University of Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-1165, USA.
E-mail: sspieran@utmb.edu


Prothrombin is now accepted as one of the target antigens recognised by antiphospholipid (aPL) antibodies. However, it is not clear whether anti-prothrombin antibodies are pathogenic in vivo and if so, the possible mechanism(s) involved. Here, we examined the pathogenic effects of the IS6 monoclonal anti-prothrombin antibody isolated from a patient with Antiphospholipid Syndrome (APS). IS6 antibody was purified from hybridoma supernatant. Its pathogenic potentials were studied in an in vivo model of induced thrombosis. The expression of tissue factor (TF) and E-selectin on human umbilical vein endothelial cells (HUVEC) was determined by cyto-enzyme-linked immunosorbent assay. Transcription of TF mRNA was determined by quantitative real time-polymerase chain reaction (RT-PCR). In vivo, the thrombus size increased significantly when treated with IS6 compared with control-treated mice (5388 ± 1035 μm2 vs. 2845 ± 1711 μm2). In vitro, IS6 induced significant expression of TF and E-selectin on HUVEC, when compared with control preparation (3·1- and 5·1-fold increase compared with the control-treated cells). RT-PCR analysis of TF mRNA transcription showed a 2·5-fold increase of IS6-treated cells over the value obtained with control-treated cells. Taken together, these data show that IS6 monoclonal anti-prothrombin antibody promotes thrombosis and this is associated with TF and E-selectin expression.

Thrombosis and/or pregnancy morbidity and positivity for antiphospholipid autoantibodies (aPL) are the main features of antiphospholipid syndrome (APS) (Harris, 1987; Wilson et al, 1999; Miyakis et al, 2006). These antibodies are characterised for either their binding to phospholipids (PLs), generally cardiolipin (CL), in solid phase assays or for their ability to prolong PL-dependent clotting reactions in vitro (thus they are termed lupus anticoagulant (LA) (Love & Santoro, 1990). APL antibodies are a heterogeneous group of antibodies, which recognise various PLs, PL-binding plasma proteins and/or PL-protein complexes and include β2 glycoprotein I (β2GPI), prothrombin, protein C, or protein S (Fleck et al, 1988; Galli et al, 1990; Matsuura et al, 1990; McNeil et al, 1990; Bevers et al, 1993, Oosting et al, 1993, Arvieux et al, 1995). Among these aPL, antibodies to β2GPI and/or its complexes with CL probably account for most positive test results for anti-CL antibodies in APS, while anti-prothromnin and/or anti-β2GPI antibodies are responsible for the majority of the LA activity (Roubey et al, 1992; Bevers et al, 1993; Permpikul et al, 1994; De Groot et al, 1998).

Various mechanisms have been proposed to explain the thrombogenic properties of aPL antibodies in APS patients. First, studies have shown that endothelial cells (EC) expressed significantly higher amounts of adhesion molecules [Intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1) and E-selectin] when incubated with aPL and antibodies against β2 glycoprotein I (β2GPI) in vitro (Simantov et al, 1995; Del Papa et al, 1997). Similarly, the incubation of ECs with antibodies reacting with β2GPI has been shown to induce EC activation with upregulation of adhesion molecules, interleukin (IL)-6 production and alteration in prostaglandin metabolism (Del Papa et al, 1997). Our group has shown, utilising mouse models, that human polyclonal and monoclonal aPL antibodies activate endothelium and enhance thrombus formation in vivo (Pierangeli et al, 1999). Utilising ICAM-1, E-selectin and P-selectin knock-out mice and specific monoclonal antibodies to VCAM-1, we confirmed that the EC-activating properties of aPL are mediated by upregulation of ICAM-1, E-selectin, P-selectin and VCAM-1 in vivo (Pierangeli et al, 2001; Espinola et al, 2003).

Second, aPL may interact with EC and induce a tissue factor (TF)-dependent procoagulant state. TF is a transmembrane protein that is expressed on the surface of a number of cell types, primarily monocytes, vascular ECs and smooth muscle cells. The extracellular domain of TF serves as a receptor for activated factor VII (FVIIa). Under normal circumstances, circulating FVII is not exposed to TF. When the integrity of the vasculature is breached, ECs are induced to express cell surface TF, and TF may then interact with FVIIa and initiate blood coagulation (Nemerson, 1988). Recently, interest has focused on the role of aPL in the induction of TF and procoagulant activity (PCA) in ECs and monocytes. For example, polyclonal and monoclonal aPLs (i.e. TM1G2 and EY1C8 IgM aCL) were found to bind to ECs and induce expression of adhesion molecules (including E-selectin, ICAM-1 and VCAM-1) and monocyte adhesion. Along this line, it was reported that incubation of monocytes with aPL led to surface expression of TF and that TF expression on monocytes was increased in APS patients, particularly in those positive for IgG aCL (Reverter et al, 1996; Amengual et al, 1998; Reverter et al, 1998; Dobado-Barrios et al, 1999; Zhou et al, 2004). Our group recently showed that aPL antibodies induce TF transcription, expression and function on ECs (Vega-Ostertag et al, 2005). Combined, these data suggest that aCL may interact with EC and monocytes and induce a TF-dependent procoagulant state.

Although our understanding of aPL has advanced, it remains unclear whether anti-prothrombin antibodies are actually involved in thrombosis and EC activation. Thus, it is important to obtain monoclonal antibodies from APS patients with various specificities, and to determine their thrombogenic properties and their ability to induce TF and cellular adhesion molecules. This study characterised the in vitro (expression of E-sel and TF on ECs) and in vivo (thrombogenic effects) properties of a human monoclonal antibody with specificity to prothrombin, named IS6 (Zhao et al, 1999). IS6 was obtained from a patient with primary APS, and displayed LA activity. Moreover, IS6 enhanced the binding of prothrombin to damaged endothelial cells, and shorten the EC-based plasma coagulation times (Zhao et al, 1999).

Materials and methods

Monoclonal antibody IS6 was obtained from a female patient with primary APS and anti-prothrombin as described in detail previously (Zhao et al, 1999; Zhu et al, 1999). The patient had a history of venous thrombosis, high titres of anti-CL (aCL) antibodies, anti-β2GPI antibodies and LA. Briefly, in 1992, at 16 years of age, she suffered spontaneous deep venous thrombosis of the calf and was treated for 3 months with heparin/warfarin. At age 18, she developed right-sided hemichorea associated with basal ganglia infarcts visible on magnetic resonance imaging and was treated with prednisone (10 mg/d) and low dose aspirin, which resulted in complete resolution of neurological symptoms. In addition to prothrombin, IS6 also binds to thrombin and a few other serine proteases (including activated protein C and plasmin) that share homologous catalytic domains with thrombin (Yang et al, 2004). Moreover, IS6 displays LA activity, enhances the binding of prothrombin to EC and shortens plasma coagulation times in a human umbilical vein endothelial cells (HUVEC)-based in vitro system (Rand et al, 1997; Zhao et al, 1999).

The IS6 monoclonal antibody was purified from culture supernatants with protein G Sepharose. A commercially available human monoclonal IgG of irrelevant specificity was used as control (control MoAb) and was purchased from EMD Biosciences (Pasadena, CA, USA).

In vitro experiments

Cell surface enzyme-linked immunosorbent assay (ELISA) for the detection of E-selectin.  Human umbilical vein endothelial cells obtained from American Tissue Type Culture Collection (Rockville, MD, USA) were maintained in MCDB110 medium (American Biorganics Inc., Niagara Falls, NY, USA), supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 μmol/l l-glutamine, 30 g/ml H0Neurext, HUVEC monolayers in 96-well plates were treated with IS6 monoclonal at 100 μg/ml or control MoAb in medium for 4 h. Treated HUVEC were paraformaldehyde-fixed and then incubated with phosphate-buffered saline (PBS)/1% bovine serum albumin (BSA) (Pierangeli et al, 1999, 2000, Espinola et al, 2003).

Thereafter, fixed monolayers were incubated sequentially with a saturating concentration of anti-E-selectin (Becton Dickinson) and a peroxidase-conjugated goat anti-mouse IgG (Sigma Aldrich, St Louis, MO, USA) for 1 h. O-phenyldiamine (0·4 mg/ml) – 0·01 % H2O2 in a 0·05–0·025 M phosphate citrate buffer; pH 5·0, was subsequently added to each well. Colour development was stopped with 3 M H2SO4 at 20 min and the optical density (OD) was read at 492 nm wavelength on a Spectramax 250 ELISA plate reader (Molecular Devices). The degree of specific antigen expression was calculated by subtracting non-specific binding of the secondary antibody from all test values.

Cell surface ELISA for the detection of TF.  Human umbilical vein endothelial cells were grown to confluence in collagen-coated 96-well tissue culture plate with complete MCDB110 medium containing 10% heat-inactivated fetal bovine serum (FBS) as described previously (Vega-Ostertag et al, 2005). On the night before the experiments, the medium was replaced with MCDB110 medium containing 1% FBS (Vega-Ostertag et al, 2005). Wells were then treated for 4 h at room temperature or as indicated otherwise for each experiment with 100 μg/ml of IS6 in MCDB110 medium or with 100 μg/ml MoAb Control or with 100 nmol/l phorbol myristate acetate (PMA) (Sigma Aldrich) as positive control. Cell surface TF was measured by ELISA, as described previously (Vega-Ostertag et al, 2005). Briefly, the cells were incubated for 1 h at room temperature with 1:500 dilution of human anti-TF antibody (product 4508; American Diagnostica, Stanford, CT, USA). The antibody solution was then removed, and cells were fixed with 1% formaldehyde for 15 min. Cells were then washed and incubated for 45–60 min with a 1:500 dilution of horseradish peroxidase-conjugate goat anti-mouse IgG (BioRad, Hercules, CA, USA). Then, the cells were washed three to four times with Hanks Balanced Salt Solution and incubated in tetramethylbenzidine solution (Dako, Carpinteria, CA, USA) for 30 min. The colour reaction was stopped by the addition of 8 N sulphuric acid. Optical densities were read at 450 nm in an ELISA reader. In all experiments, a second step background colour was performed in which cells were treated as described in the section above, but without incubating with the first antibody. Experiments were run in duplicate and performed three times.

Analysis of TF mRNA

Human umbilical vein endothelial cells were cultured in a 60 nm culture dish in MCDB110 complete medium containing 10% inactivated FBS. When the cells reached confluence, the medium was replaced with MCDB110 medium containing 1% FBS, and the cells were incubated overnight under those culture conditions (Vega-Osterlag et al, 2005). The cells were then treated with 200 μg/ml of IS6 or control MoAb for 2 h, washed twice with 3 ml of ice-cold PBS and treated with 1 ml of TRI-reagent (Sigma Aldrich) according to the manufacturer's instructions. RNA was then treated with RNAase–free-DNAase for 1 h at room temperature to remove any potential genomic DNA. sDNA (cDNA) was generated using the Superscript II-first-strand cDNA pre-amplification system (Gibco BRL, Rockville, MD, USA), according to the random primer protocol provided by the manufacturer. The indication of mRNA was measured by real time quantitative polymerase chain reaction (RT-PCR) using the GENEAmp 5700 sequence detection system from Perkin Elmer (Emeryville, CA, USA). PCR was performed using Sybir green master mix (Perkin Elmer, Wellesley, MA, USA), with primers optimised as recommended. Primers for the RT-PCR (for TF mRNA), 5′-CAC-CGA-CGA-GAT-TGT-GAA-GGA-T-3′ and 5′-CCC-TGC-CGG-GTA-GGA-GAA-3′; for actin mRNA, 5′CGT-CCA-CCG-CAA-ATG-CTT-3′ and 5′-TCT, GCG, CAA, GTT, AGG, TTT, TGT, C-3′), were synthesised by the Microchemical Facility of the Winship Cancer Center (Emory University School of Medicine, Atlanta, GA, USA). In preliminary experiments, the slopes of the TF and the actin (control) primers were well within 0·1 of each other (1·406 and 1·4875, respectively). The relative level of TF mRNA was calculated by the δδ Ct method, normalising for actin (Vega-Ostertag et al, 2005). Data were expressed as the fold-increase over the non-stimulated media control. Experiments were performed three times.

In vivo experiments

A mouse model of induced thrombosis (‘pinch-induced’ thrombosis) was employed to study the thrombogenic properties of IS6 (Pierangeli et al, 1999, 2000). Mice were housed in the Center for Animal Resources at Morehouse School of Medicine, an approved facility, under the supervision of veterinarians and trained technicians.

Specifically, groups of 7–9 CD1 male mice (Charles River Breeding Laboratories, Bal Harbor, MA, USA), weight 25–30 g, were injected intraperitoneally with IS6 or with control MoAb (500 μg/mouse) at time 0 and 48 h. At that time, each animal was subjected to the surgical procedure described elsewhere (Olee et al, 1996; Pierangeli et al, 1999, 2000, Espinola et al, 2003). Briefly, the right femoral vein of each treated mouse was exposed, resulting in a 1 cm segment of vein free of manipulation and observation. The vein was pinched with a standard pressure of 1500 g/m2 to introduce a thrombogenic injury. Clot formation and dissolution in the transilluminated vein were monitored with a microscope equipped with a closed-circuit video system (including a monitor and recorder). Thrombus sizes (μm2) were measured when the thrombi reached maximum size and the outer margin of the thrombus was traced; the times (min) of disappearance from maximum thrombus size were measured as well.

Statistical analysis

Means were compared among the various treated and control groups using Student's unpaired t-test. P-values ≤ 0·05 were considered statistically significant.


IS6 antibodies are thrombogenic in mice

At the time of the surgical procedures, IS6 treated mice had significant titres of anti-prothrombin antibodies in the serum, as determined by ELISA (0·403 ± 0·086), compared with control mice (0·123 ± 0·098) (P = 0·04). Mice injected with IS6 produced significantly larger thrombi (5388 ± 1035 μm2) when compared with mice treated with a control human monoclonal IgG (2845 ± 1711 μm2, Table I, P = 0·002). In addition, thrombi in mice treated with IS6 persisted longer when compared with mice injected with the corresponding control MoAb (disappearance times and total times) (Table I, P = 0·04 and 0·025).

Table I.   Thrombogenic effects of IS6 in mice.
Group Mice (n)Maximum thrombus size (μm2)Disappearance time (min)Total time (min)
  1. *Statistically significant difference from control.

  2. Disappearance time: the time taken for the thrombus to reduce from maximum size to undetectable.

  3. Total time: the time from the start of thrombus formation to its total disappearance.

IS675388 ± 1035*11·2 ± 2·5*12·3 ± 3·0*
Control MoAb82845 ± 17114·1 ± 2·05·6 ± 2·0

IS6 antibodies induce expression of E-selectin and TF on HUVECs

Surface antigen expression of E-selectin was low under untreated or unstimulated conditions. HUVEC treated with lipopolysaccharide as positive control induced a fourfold increase in E-selectin expression after 4 h incubation (Fig 1). IS6 induced a significant 2·1-fold increase of E-selectin expression over the cells treated with control MoAb (P = 0·002). HUVEC treated with Control MoAb did not express significant E-selectin on their surface.

Figure 1.

 IS6 induces expression of E-selectin on endothelial cells. HUVEC were treated with IS6 or a monoclonal antibody control (control) or with medium alone as indicated in Materials and methods. Expression of E-selectin on HUVEC was determined by cyto-ELISA, and expressed as mean OD units ±SD. *Statistically significant difference from control.

Similarly, IS6 induced a significant 2·5-fold increase in TF expression in HUVEC as determined by cyto-ELISA (P = 0·001) when compared with HUVEC treated with control MoAb. TF expression in HUVEC treated with control MoAb did not differ from the expression observed in HUVEC treated with medium (Fig 2). IS6 also induced a 21·5-fold increase in TF mRNA transcription, as determined by RT-PCR.

Figure 2.

 IS6 induces expression of TF on endothelial cells. HUVEC were treated with IS6 or a monoclonal antibody control (control MoAb) or with medium alone as indicated in Materials and methods. Expression of TF on HUVEC was determined by cyto-ELISA, and expressed as mean OD units ±SD. *Statistically significant difference from control.


To determine the pathological significance of anti-prothrombin antibodies in APS, the IS6 patient-derived IgG monoclonal anti-prothrombin antibody was studied in an in vivo‘pinch-induced’ thrombosis model. The results showed that IS6 induced significantly larger thrombi than the control normal human IgG (Table I) (P = 0·002). In addition, IS6-induced thrombi persisted longer than those induced by the control monoclonal human IgG (disappearance times and total times, Table I, P = 0·04 and 0·025, respectively). Combined, these results indicate that the IS6 anti-prothrombin antibody is prothrombotic.

Subsequently, the in vitro studies showed that IS6 induced expression of E-selectin and TF on HUVEC (Figs 1 and 2). The data suggest that some anti-prothrombin antibodies in APS may activate EC and induce a TF-dependent procoagulant state, similar to aCL or anti-β2GPI antibodies (Simantov et al, 1995; Del Papa et al, 1997; Vega-Ostertag et al, 2005).

Previous studies of IS6 showed that it could enhance prothrombin binding to EC (possibly via cross-linking two prothrombin molecules and thus increasing the avidity of prothrombin to the PL/EC surface) and shorten clotting time on the EC surface (Amengual et al, 1998). Taken together, these data show that IS6 may promote thrombosis by two sequential mechanisms: first activating EC and inducing a TF-dependent procoagulant state; and then increasing the local concentration of prothrombin and speeding up thrombin generation.

In addition to prothrombin, IS6 was found to react with thrombin and other serine proteases (such as activated protein C and plasmin) that share homologous catalytic domains (Hwang et al, 2001, 2003). However, this binding activity did not appear to have any pathological significance. Unlike the CL24 monoclonal aCL, IS6 did not interfere with inactivation of thrombin by antithrombin and, unlike CL15 monoclonal aCL, IS6 did not inhibit the functional activities of activated protein C and plasmin (Hwang et al, 2001, 2003; Yang et al, 2004).

Various mechanisms have been proposed to explain the thrombogenic properties of aPL in APS patients. These include EC activation with upregulation of adhesion molecules and TF, IL-6 production and alteration in prostaglandin metabolism (Reverter et al, 1996; Del Papa et al, 1997; Amengual et al, 1998; Reverter et al, 1998; Dobado-Barrios et al, 1999). In addition, studies have shown that some aPL may inhibit protein C activation and/or the function of activated protein C (Cariou et al, 1988; Borrell et al, 1992; Smirnov et al, 1995; Carson et al, 2000). Moreover, some aPL may bind to heparan sulphate, and interfere with the inactivation of thrombin by antithrombin (Chamley et al, 1993). In this context, the current study showed that IS6 may promotes thrombosis by only two of many proposed mechanisms.

In summary, this study has determined, for the first time, the thrombogenic effects of a human monoclonal anti-prothrombin antibody derived from a patient with APS. These findings may help to understand the pathogenesis of thrombosis in APS.


This study was partially funded by a grant no. G12-RR03034 for Research Center in Minority Institutions and a Minority Biomedical Research Support Grant (GM58268–02) (SP); and AR42506 (PPC) from National Institutes of Health.