Enhancement of the enzymatic activity of activated coagulation factor IX by anti-factor IX antibodies

Abstract

SummaryBackground: Factor VIIIa (FVIIIa) binds to activated FIX and enhances the activation of FX by several orders of magnitude. Deficiency of FVIII causes the bleeding disorder hemophilia A and is treated by i.v. infusion of FVIII concentrates. Objectives: To explore whether or not FVIII activity can be supplied by alternative molecules, e.g. molecules with FIXa-binding activity. Methods: Conventional hybdridoma technology was used to discover antibodies exhibiting FVIII-like activity. Results: We identified a series of antibodies specific for human FIX that mimicked the stimulatory effect of FVIIIa on FIXa. Upon binding to human FIXa, these antibodies enhanced the protease activity of FIXa towards its natural substrate FX about tenfold. A similar enhancement was also achieved with 5 pm FVIIIa (i.e. 16 mU mL−1 or 1.6% activated FVIII). Procoagulant activity of these anti-FIXa antibodies was observed in model systems containing purified proteins as well as in plasma. Conclusion: Our findings show that FVIII can, at least partially, be replaced by an unrelated molecule. Procoagulant antibodies might potentially aid the development of an FVIII substitute for hemophilia A treatment. Background: Factor VIIIa (FVIIIa) binds to activated FIX and enhances the activation of FX by several orders of magnitude. Deficiency of FVIII causes the bleeding disorder hemophilia A and is treated by i.v. infusion of FVIII concentrates. Objectives: To explore whether or not FVIII activity can be supplied by alternative molecules, e.g. molecules with FIXa-binding activity. Methods: Conventional hybdridoma technology was used to discover antibodies exhibiting FVIII-like activity. Results: We identified a series of antibodies specific for human FIX that mimicked the stimulatory effect of FVIIIa on FIXa. Upon binding to human FIXa, these antibodies enhanced the protease activity of FIXa towards its natural substrate FX about tenfold. A similar enhancement was also achieved with 5 pm FVIIIa (i.e. 16 mU mL−1 or 1.6% activated FVIII). Procoagulant activity of these anti-FIXa antibodies was observed in model systems containing purified proteins as well as in plasma. Conclusion: Our findings show that FVIII can, at least partially, be replaced by an unrelated molecule. Procoagulant antibodies might potentially aid the development of an FVIII substitute for hemophilia A treatment. One of the key events during hemostasis is the conversion of the zymogen factor X (FX) into its enzymatically active form FXa, subsequently leading to prothrombin activation and stable clot formation. In vivo, FX activation is initially achieved through the tissue factor (TF)–FVIIa pathway, a process that is readily inhibited by the TF pathway inhibitor. Activation of most FX occurs via the intrinsic coagulation pathway and requires formation of the intrinsic FX-activating complex. This complex consists of the activated coagulation factors IX (FIXa) and VIII (FVIIIa), assembled on a phospholipid surface in the presence of calcium ions. The intrinsic FX-activating complex generates sufficient FXa to enable the formation of a stable clot [1Mann K.G. Biochemistry and physiology of blood coagulation.Thromb Haemost. 1999; 82: 165-74Crossref PubMed Google Scholar]. FVIIIa functions as an activator of FIXa, which increases the rate of FIXa-catalyzed FXa formation approximately two thousandfold [2Van Dieijen G. Tans G. Rosing J. Hemker H.C. The role of phospholipid and factor VIIIa in the activation of bovine factor X.J Biol Chem. 1981; 256: 3433-42Abstract Full Text PDF PubMed Google Scholar]. The exact mechanism by which FVIIIa enhances the catalytic activity of FIXa towards FX is not yet fully understood [3Fay P.J. Activation of factor VIII and mechanisms of cofactor action.Blood Rev. 2004; 18: 1-15Crossref PubMed Scopus (160) Google Scholar]. Naturally occurring mutants and site-directed mutagenesis [4Fay P.J. Jenkins P.V. Mutating factor VIII: lessons from structure to function.Blood Rev. 2005; 19: 15-27Crossref PubMed Scopus (44) Google Scholar], as well as homology modeling of FIXa and FVIIIa [5Autin L. Miteva M.A. Lee W.H. Mertens K. Radtke K.P. Villoutreix B.O. Molecular models of the procoagulant factor VIIIa–factor IXa complex.J Thromb Haemost. 2005; 3: 2044-56Crossref PubMed Scopus (0) Google Scholar], suggest that there are multiple interaction sites between FIXa, FVIIIa and FX that help to enhance the enzymatic activity of FIXa [4Fay P.J. Jenkins P.V. Mutating factor VIII: lessons from structure to function.Blood Rev. 2005; 19: 15-27Crossref PubMed Scopus (44) Google Scholar]. FVIII is generally believed to have several functions within the intrinsic FX-activating complex: (i) FVIIIa stabilizes a conformation of FIXa that has increased protease activity towards FX [6Kolkman J.A. Mertens K. Insertion loop 256–268 in coagulation factor IX restricts enzymatic activity in the absence but not in the presence of factor VIII.Biochemistry. 2000; 39: 7398-405Crossref PubMed Scopus (0) Google Scholar], and orients the cleavage sites in FX towards the active site of FIXa [7Lapan K.A. Fay P.J. Localization of a factor X interactive site in the A1 subunit of factor VIIIa.J Biol Chem. 1997; 272: 2082-8Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar], which together result in an increase of the kcat of FX activation; (ii) FVIIIa acts as a receptor for FIXa on phospholipid vesicles [8Van Dieijen G. Van Rijn J.L. Govers-Riemslag J.W. Hemker H.C. Rosing J. Assembly of the intrinsic factor X activating complex – interactions between factor IXa, factor VIIIa and phospholipid.Thromb Haemost. 1985; 53: 396-400Crossref PubMed Scopus (0) Google Scholar] and on activated platelets [9Ahmad S.S. London F.S. Walsh P.N. Binding studies of the enzyme (factor IXa) with the cofactor (factor VIIIa) in the assembly of factor-X activating complex on the activated platelet surface.J Thromb Haemost. 2003; 1: 2348-55Crossref PubMed Scopus (0) Google Scholar], which in vivo provide the procoagulant phospholipid surface on which intrinsic FX activation occurs [10Rosing J. Van Rijn J.L. Bevers E.M. Van Dieijen G. Comfurius P. Zwaal R.F. The role of activated human platelets in prothrombin and factor X activation.Blood. 1985; 65: 319-32Crossref PubMed Google Scholar]; and (iii) FVIIIa reduces the Km for the substrate FX [11Fay P.J. Koshibu K. The A2 subunit of factor VIIIa modulates the active site of factor IXa.J Biol Chem. 1998; 273: 19049-54Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar]. The crucial role of FVIII in hemostasis is demonstrated by hemophilia A, a severe X-chromosome-linked recessive disorder that is characterized by the absence of FVIII activity and a pronounced bleeding tendency. Treatment of patients with hemophilia A is achieved through substitution of FVIII via i.v. injection of either plasma-derived or recombinant FVIII. Today, hemophilia A therapy is safe and highly efficient, but it still has several limitations, e.g. the short half-life of FVIII and the route of administration. In particular, the high molecular weight of FVIII and its low bioavailability after non-i.v. application confines the substitution mode to i.v. infusion, which is a difficult route, especially in young children. Furthermore, approximately 30% of severely affected patients develop inhibitory antibodies against FVIII, leading to a serious and life-threatening complication [12Brettler D.B. Inhibitors in congenital haemophilia.Baillieres Clin Haematol. 1996; 9: 319-29Abstract Full Text PDF PubMed Scopus (41) Google Scholar, 13Penner J.A. Haemophilic patients with inhibitors to factor VIII or IX: variables affecting treatment response.Haemophilia. 2001; 7: 103-8PubMed Google Scholar]. We started our search for a functional substitute for FVIII with a low molecular weight compound by assessing the possibility of replacing FVIII with an unrelated molecule. Conceptually, we viewed FVIIIa as a ligand of FIXa capable of inducing a conformational change in this protease, leading to an increased catalytic activity in FX activation. We assumed that it would be possible to replace FVIII with a completely different FIXa-specific ligand that induces a similar conformational change leading to enhanced FX-cleaving activity. Such a molecule should be safe and well tolerated, and, because it would not be FVIII, should not be affected by FVIII inhibitors, thereby providing a therapeutic alternative for patients with FVIII inhibitors. In addition, design of an FIXa ligand with FVIII-like properties allows the mode of administration to be addressed. Antibodies are versatile binding proteins capable of mimicking a variety of protein–protein interactions [14Monfardini C. Kieber-Emmons T. VonFeldt J.M. Godillot A.P. Voet D. Weiner D.B. Williams W.V. Rational design, analysis, and potential utility of GM-CSF antagonists.Proc Assoc Am Physicians. 1996; 108: 420-31PubMed Google Scholar, 15Magliani W. Conti S. Salati A. Vaccari S. Ravanetti L. Maffei D.L. Polonelli L. Therapeutic potential of yeast killer toxin-like antibodies and mimotopes.FEMS Yeast Res. 2004; 5: 11-8Crossref PubMed Scopus (34) Google Scholar]. Recently, we described antibodies capable of amplifying the activity of FVIII by increasing the affinity of FVIIIa for FIXa and by increasing the kcat of the FIXa–FVIIIa complex [16Kerschbaumer R.J. Riedrich K. Kral M. Varadi K. Dorner F. Rosing J. Scheiflinger F. An antibody specific for coagulation factor IX enhances the activity of the intrinsic factor X-activating complex.J Biol Chem. 2004; 279: 40445-50Abstract Full Text Full Text PDF PubMed Scopus (9) Google Scholar, 17Valentino L.A. Scheiflinger F. Future aspects of hemophilia research and care.Semin Thromb Hemost. 2006; 32: 32-8Crossref PubMed Scopus (4) Google Scholar]. Here, we aimed to evaluate the potential of anti-human FIXa antibodies to function as cofactors in the intrinsic FX-activating complex in the absence of FVIII. We report the isolation and characterization of anti-FIX antibodies that have FVIII-like activity in vitro. We show that an antibody against FIX can at least partially replace FVIII and exhibits procoagulant activity. Human FIX, FIXa, FX, FXa and α-thrombin were purchased from Enzyme Research Laboratories (South Bend, IN, USA). Recombinant human FVIII (rFVIII) was prepared by Baxter BioScience (Thousand Oaks, CA, USA). Bovine serum albumin (BSA) was purchased from Calbiochem (San Diego, CA, USA). Human FVIII-deficient patient plasma and human pooled normal plasma were purchased from George King, Bio-Medical (Overland Park, KS, USA). Human FVIII inhibitor plasma was prepared from patient plasma by Technoclone (Vienna, Austria). Non-specific polyclonal mouse IgG was purchased from Sigma ( St Louis, MO, USA). Pathromtin SL was purchased from Dade Behring (Deerfied, IL, USA). FEIBA and Dapptin were obtained from Baxter BioScience (Vienna, Austria). Fluorogenic TGA substrate (2 mm Z-GGR-AMC, 30 mm CaCl2) and TGA reagent C (micelles of phospholipids, 3.2 μm, containing 71.6 pm recombinant human tissue factor) were purchased from Technoclone. Balb/c mice were immunized four times with human FIXa at intervals of 1 week. Per immunization, 100 μg of antigen was applied using Al(OH)3 as adjuvant. Spleen cells were obtained 3 days after the last immunization, and hybridoma cell lines were generated according to standard procedures [18Kohler G. Milstein C. Continuous cultures of fused cells secreting antibody of predefined specificity.Nature. 1975; 256: 495-7Crossref PubMed Google Scholar]. Per mouse, 880 hybridoma cell lines were grown in 96-well plates, first in HAT selection medium and later in normal growth medium (RPMI-1640 supplemented with 5% fetal bovine serum; Invitrogen, Carlsbad, CA, USA). After 4–5 weeks, supernatants were screened for FVIII-like activity in a screening assay (see below). Cell lines expressing antibodies that were identified as displaying FVIII-like activity were subcloned four to six times to ensure that the cell line was monoclonal and stable. The commercially available test kit for FVIII activity, the COATEST VIII:C/4 (Chromogenix, Milan, Italy) was used to screen hybridoma supernatants for antibodies that display FVIII-like activity. The assay was performed essentially as described by the manufacturer, except that the sample and reagent volumes were reduced to a 96-well format and that the reactions were not stopped after 5 min, but after 3 h. Furthermore, 20 ng of human FIXa was added to each well to increase the sensitivity of the assay. The COATEST VIII:C/4 is an all-in-one assay, and cleavage of FX to FXa by FIXa, as well as cleavage of the FXa-specific chromogenic substrate Bz-lle-Glu-Gly-Arg-p-nitroaniline (pNA) (S-2222), takes place in the same reaction well. Stimulation of catalytic activity of FIXa by a specific antibody that functions as a cofactor of FIXa results in enhanced FXa generation and thus in an increase in the rate of cleavage of the chromogenic substrate. Released pNA was measured at 405 nm in a 96-well microplate reader (iEMS-Reader, Labsystems, Helsinki, Finland). Monoclonal hybridoma cell lines were grown in RPMI-1640 medium supplemented with 5% fetal bovine serum (Invitrogen) for 2–3 weeks. IgGs were purified over protein G–Sepharose 4 Fast Flow (Amersham Biosciences, Uppsala, Sweden) according to standard procedures [19Jungbauer A. Tauer C. Reiter M. Purtscher M. Wenisch E. Steindl F. Buchacher A. Katinger H. Comparison of protein A, protein G and copolymerized hydroxyapatite for the purification of human monoclonal antibodies.J Chromatogr. 1989; 476: 257-68Crossref PubMed Scopus (0) Google Scholar]. Reactions were performed in PPN tubes (Micronic, Lelystad, the Netherlands) in a water bath at 37 °C as follows: 220 μL of HNaBSA5 buffer (25 mm HEPES, 175 mm NaCl, 5 mg mL−1 BSA, pH 7.35) containing 12.8 μm phospholipids and 5.9 mm calcium ions were prewarmed to 37 °C. Twenty microliters of FX, 20 μL of FIXa and 40 μL of the respective cofactor (antibody or thrombin-activated rFVIII) were added yielding, a reaction mixture that contained 10.7 μm phospholipid, 5 mm CaCl2, and 0.5 mg mL−1 BSA. Phospholipid vesicles (60% dioleoyl-phosphatidylcholine, 40% dioleoyl-phosphatidylserine) were prepared by Baxter BioSciences (Vienna, Austria) from synthetic phospholipids (Avanti Polar Lipids, Alabaster, AL, USA). The final concentrations of FIXa, rFVIII, FX and antibody were dependent on the kind of analysis, and are described in Results. After varying time intervals, aliquots (10 μL) taken from this reaction mix were transferred into 500 μL of ice-cold EDTA buffer (50 mm Tris, pH 8.3, 9 mm EDTA, 428 mm NaCl) to stop FXa formation. The amount of FXa generated was determined by mixing 210 μL of the diluted aliquot with 40 μL of a substrate-αNAPAP mixture [5 mm Pefachrome FXa (Pefa-5523) + 6 μmαNAPAP; Pentapharm, Basel, Switzerland] in a 96-well microplate and measuring the rate of chromogenic substrate cleavage (ΔOD/min) at 405 nm (ΔOD/min) at 37 °C in a microplate reader. The FXa concentration was calculated for each time point from a standard calibration curve made with known amounts of FXa. Thrombin generation using a contact trigger was performed by adding 10 μL of antibody dilution to 40 μL of FVIII-deficient plasma or FVIII inhibitor plasma in a black 96-well microplate (Cliniplate, Thermo Labsystems, Milford, MA, USA). After addition of 25 μL of Pathromtin SL, wells were incubated for 5 min at 37 °C on a plate incubator. Thrombin generation was initiated by adding 25 μL of a TGA substrate-CaCl2 mixture. The plate was immediately transferred to a preheated (37 °C) microplate fluorescence reader (Tecan Safire 2, excitation 360 nm/emission 460 nm), and conversion of the fluorogenic substrate by thrombin was followed by kinetically reading the plate every 30 s. The first derivative of the original fluorescence data divided by a calibration factor yielded the amount of generated thrombin as a function of time [16Kerschbaumer R.J. Riedrich K. Kral M. Varadi K. Dorner F. Rosing J. Scheiflinger F. An antibody specific for coagulation factor IX enhances the activity of the intrinsic factor X-activating complex.J Biol Chem. 2004; 279: 40445-50Abstract Full Text Full Text PDF PubMed Scopus (9) Google Scholar]. The FVIII activity of human pooled normal plasma is inhibited by incubating (2 h, 37 °C) 40 μL of human normal plasma with 1.5 μL of heat-inactivated anti-human FVIII plasma (3880 BU mL−1) raised in goat. TGA substrate (40 μL) was premixed with 10 μL of TGA reagent C (containing phospholipid vesicles consisting of 80% dioleoyl-phosphatidylcholine, and 20% dioleoyl-phosphatidylserine) and 10 μL of antibody or FEIBA reference standard dilutions, and thrombin generation was initiated by adding 40 μL of the FVIII-inhibited plasma. The plate was immediately transferred to a microplate fluorescence reader and analyzed as outlined above. We isolated a series of murine monoclonal antibodies (mAbs) specific for human FIXa that have FVIIIa-like activity in the following screening procedure. Balb/c mice were immunized with FIXa, and hybridoma cell lines were obtained after immortalization of spleen cells. Individual supernatants from hybridoma cells were screened for FVIII-like activity as described in Methods. Briefly, a mixture of the enzyme FIXa, the substrate FX, phospholipids and calcium ions was incubated with hybridoma supernatants. We assumed that agonistic antibodies, i.e. FIXa-binding antibodies that are able to augment FIXa protease activity, would accelerate FIXa-catalyzed FX activation. Generation of FXa, indicative of the presence of agonistic antibodies, was monitored with an FXa-specific chromogenic substrate. Supernatants of hybridoma cell lines expressing non-specific mouse IgG were used as negative controls. In parallel, hybridoma supernatants were tested for FIXa-binding antibodies by enzyme-linked immunosorbent assay (ELISA). With this rather simple screening procedure, we identified 88 antibodies with different degrees of FIXa-agonistic activity out of 5280 hybridoma supernatants tested, whereas approximately 60% of the hybridoma cell lines produced FIXa-binding antibodies. All mAbs displaying FVIII-like activity were purified and further characterized in FXa generation assays. Three mAbs, termed 198A1, 198B3, and 224F3, were identified as having the most pronounced FVIII-like activity. The in-depth characterization of these three antibodies is described further. Antibodies 198A1 and 198B3 were found to be strictly specific for human FIX and FIXa, because ELISA and western blot analysis did not show binding to bovine or murine FIX. Furthermore, incubation of bovine FIXa with these antibodies did not enhance FXa generation. However, antibody 224F3 bound to human and bovine, but not to murine, FIX and FIXa. A competition ELISA assay revealed that the three mAbs interact with the same binding site on human FIX, but do not compete with FVIII for binding to FIXa (data not shown). Proteolytic conversion of FX into its enzymatically active form FXa is achieved by the intrinsic FX-activating complex. This complex consists of the protease FIXa bound to its cofactor FVIIIa on a negatively charged phospholipid surface in the presence of calcium ions. To quantify the effect of the procoagulant antibodies on the catalytic activity of FIXa, FXa generation assays were performed as described in Methods. In our first set of experiments, 11 nm FIXa, 150 nm FX and 10 μm phospholipid vesicles were incubated with varying amounts of antibody (1–30 nm) in a reaction mixture containing 5 mm CaCl2, and FXa generation was followed with an FXa-specific chromogenic substrate. Antibodies 198A1, 198B3 and 224F3 gave similar results, and as an example, FXa generation curves obtained with antibody 224F3 are shown. As compared with the control (buffer or non-specific polyclonal mouse antibodies), a tenfold rate enhancement of FXa generation was observed at ∼10 nm antibody, which turned out to be the optimal antibody concentration (Fig. 1A). Higher antibody concentrations (up to 30 nm) did not yield improved results (data not shown). For comparison, FXa generation assays were performed with 15 pm FVIIIa (50 mU mL−1). The curve obtained with FVIIIa was linear for approximately 5 min, after which FXa generation stopped, presumably because of inactivation of FVIIIa through dissociation of its A2 domain [20Lollar P. Parker E.T. Structural basis for the decreased procoagulant activity of human factor VIII compared to the porcine homolog.J Biol Chem. 1991; 266: 12481-6Abstract Full Text PDF PubMed Google Scholar]. In contrast, FXa generation by FIXa in the presence of antibodies was linear for at least 30 min (Fig. 1A). Antibody titrations (1–200 nm) further revealed a dose-dependent effect on the rate of FXa formation (Fig. 1B). At 11 nm FIXa, a linear dose response was obtained up to a mAb concentration of ∼5 nm, i.e. up to the presence of one antigen-binding site per FIXa molecule. Between antibody concentrations of 10 and 30 nm, the rate of FX activation reached a maximum, and it decreased at higher antibody concentrations (Fig. 1B). In the case of 50 nm FIXa, linearity up to 25 nm antibody was obtained (data not shown). Thus, a linear dose response can be seen until FIXa is saturated with antibody. In control experiments, in which FIXa, FX, phospholipids or calcium ions were omitted from the reaction mixture, no detectable amounts of FXa were formed (data not shown). However, it is theoretically possible that in our assays the substrate FX was activated by traces of FXa via an autocatalytic reaction. To rule out this possibility, we performed FXa generation assays with 60 μm Pefablock Xa (Pentapharm, Switzerland), a competitive FXa inhibitor that completely inhibits FXa activity at this concentration. After subsampling into EDTA buffer (see Methods), Pefablock Xa was diluted to 2 μm, a concentration that allowed FXa to be quantified with the chromogenic substrate. The same FXa generation rates were obtained with and without Pefablock Xa (data not shown). This indicates that FXa is also formed when the FXa present or generated in the reaction mixture is inhibited, which excludes the possibility that autocatalysis contributed to antibody-mediated FXa generation. In summary, the experiments described in this paragraph show that the antibodies do not catalyze or induce a bypass reaction by themselves, but rather function as specific cofactors for proteolytic activation of FX by FIXa. The FIXa-activating antibodies 198A1, 198B3 and 224F3 were further characterized by determining the kinetic parameters of human FX (hFX) activation by the FIXa–antibody complexes. Rates of hFX activation were measured at different substrate, i.e. hFX, concentrations in a reaction mixture that contained 11 nm FIXa, 25 nm respective antibody, and FX concentrations between 0 nm and 150 nm. The rate of hFX activation (nm min−1) was plotted as a function of the hFX concentration (Fig. 2), and the Michaelis–Menten constants (Vmax and Km) were obtained by fitting the curve to a hyperbola using the solver function of Windows Excel. For comparison, Michaelis–Menten hyperbolas were also determined in the presence of FVIIIa. FIXa alone as well as FIXa with 25 nm non-specific polyclonal mouse IgG gave identical hyperbolas, and allowed the kinetic parameters of FIXa without effector to be calculated (Table 1). The Vmax values obtained with antibodies 198B3 and 224F3 were increased some 8 to tenfold, and were similar to the Vmax determined with 5 pm FVIII, an amount of FVIII that corresponds to an FVIII activity of 16 mU mL−1, or 1.6% of the plasma concentration. The turnover number, kcat, was calculated by dividing Vmax by the enzyme–complex concentration. The kcat of FX activation by FIXa was increased approximately tenfold by antibodies 198B3 and 224F3, and 8-fold by antibody 198A1. Furthermore, FVIII decreased the Km of hFX for FIXa ∼2.5-fold, independently of the FVIIIa concentration. The antibodies did not influence the Km of FX for FIXa, with the exception of antibody 198B3, which caused a 2-fold increase of the Km. However, in all cases, the Km was far below the concentration of FX in human plasma (∼140 nm). Further experiments revealed that the Km was not affected by variation of the antibody concentration (data not shown).Table 1Kinetic parameters of factor F (FX) activation by FIXa, FIXa–FVIIIa complex and FIXa–antibody complexVmax (nm min−1)Km (nm)kcat* (min−1)FIXa (11 nm) without effector0.06334.20.0057FIXa–FVIIIa complex5 pm FVIIIa0.6414.212815 pm FVIIIa2.113.413645 pm FVIIIa5.014.8111FIXa–198B3 complex (11 nm FIXa, 25 nm IgG)0.6684.30.060FIXa–198A1 complex (11 nm FIXa, 25 nm IgG)0.4545.00.041FIXa–224F3 complex (11 nm FIXa, 25 nm IgG)0.5948.00.053*kcat values were calculated by dividing Vmax by the concentration of FIXa (11 nm), FIXa–FVIIIa complex (5 pm, 15 pm, and 45 pm, respectively), or FIXa–antibody complex (11 nm). Open table in a new tab *kcat values were calculated by dividing Vmax by the concentration of FIXa (11 nm), FIXa–FVIIIa complex (5 pm, 15 pm, and 45 pm, respectively), or FIXa–antibody complex (11 nm). The experiments described so far show that the antibodies 198A1, 198B3 and 224F3 display an FVIII-like effect in a well-defined model system containing purified components. Thrombin generation experiments were conducted in FVIII-deficient plasma to investigate whether the antibodies also have FVIII-like activity in human plasma [21Hemker H.C. Beguin S. Thrombin generation in plasma: its assessment via the endogenous thrombin potential.Thromb Haemost. 1995; 74: 134-8Crossref PubMed Scopus (261) Google Scholar]. In this assay system, the whole intrinsic coagulation cascade, from FXII activation to thrombin formation, as well as the inactivation of the coagulation factors by the plasma protease inhibitors, is probed. A time course of thrombin generation is characterized by a lag phase, which reflects the initiation of coagulation followed by a burst of thrombin formation. Thrombin is subsequently inhibited by anti-thrombin and α2-macroglobulin. The effect of FIXa-activating antibodies on thrombin generation was examined in FVIII-deficient plasma (Fig. 3). As a negative control, antibody-purified mouse IgG antibodies were used in the same concentration range as FIX-specific mAbs. Addition of FVIII activity to FVIII-deficient plasma has two effects on thrombin generation in FVIII-deficient plasma: (i) the thrombin burst occurs earlier, which is indicative of a shortening of the lag phase; and (ii) the amount of thrombin that is generated increases [16Kerschbaumer R.J. Riedrich K. Kral M. Varadi K. Dorner F. Rosing J. Scheiflinger F. An antibody specific for coagulation factor IX enhances the activity of the intrinsic factor X-activating complex.J Biol Chem. 2004; 279: 40445-50Abstract Full Text Full Text PDF PubMed Scopus (9) Google Scholar]. Both effects were shown by antibodies 198A1, 198B3 and 224F3 (Fig. 3), although to different extents. The peak time was shortened from 30 min (control antibody or no antibody) to approximately 15 min, and the maximum thrombin concentration increased from 50 nm (controls) to 150 nm. As in model systems, the effect of the antibodies on thrombin generation was dependent on the antibody concentration. Additional experiments revealed that all three antibodies gave an optimal effect on thrombin generation at plasma antibody concentrations of 40–60 nm, which is approximately half the FIX plasma concentration (90 nm). To investigate whether the activity of the procoagulant antibodies is affected by FVIII inhibitors, thrombin generation experiments were repeated in FVIII inhibitor plasma. The presence of inhibitors makes it impossible to compare the activity of the procoagulant antibodies with that of FVIII. We therefore compared the procoagulant activity of the anti-FIX antibodies with that of FEIBA, an activated prothrombin complex therapeutic agent that is frequently used to treat bleeding episodes in patients with inhibitors [22Turecek P.L. Varadi K. Gritsch H. Auer W. Pichler L. Eder G. Schwarz H.P. Factor Xa and prothrombin: mechanism of action of FEIBA.Vox Sang. 1999; 77: 72-9Crossref PubMed Scopus (0) Google Scholar]. In FVIII inhibitor plasma, increasing amounts of FEIBA had the same effect on thrombin generation as FVIII had in FVIII-deficient plasma. When the procoagulant antibodies were applied as FIXa-stimulating agents, their effects in FVIII inhibitor plasma were similar to those observed in FVIII-deficient plasma, although the peak thrombin levels reached were somewhat lower. Fig. 4 shows the time course of thrombin generation obtained with antibody 224F3. Under optimal conditions, thrombin formation occurs 10 min earlier (thrombin peak after 15 min) than without antibody (thrombin peak after 25 min). As in the experiments with FVIII-deficient plasma, the optimal effect was obtained at 40–60 nm antibody in plasma. For comparison, thrombin generation experiments were also conducted at a therapeutic FEIBA concentration (0.5 U mL−1). Initiation with tissue factor is an alternative method for triggering thrombin formation in plasma. The extrinsic coagulation pathway resembles the natural activation of clot formation. Thrombin formation experiments were performed using the TF–FVII pathway, as described in Methods. The influence of each mAb on thrombin generation in FVIII-inhibited plasma is followed continuously via cleavage of the thrombin-specific fluorogenic substrate Z-GGR-AMC. Antibody 198B3 contributes to higher thrombin generation also after triggering of the extrinsic pathway with recombinant human TF on the phospholipid surface. The activity of 22.5 nm 198B3 corresponds to 30 mU FEIBA mL−1 (Fig. 5). Hemophilia A, a hereditary disease that affects men, is characterized by either the absence of FVIII or the presence of non-functional FVIII. FVIII plays a pivotal role in blood coagulation. Therefore, a lack of FVIII causes serious and often life-threatening bleeding episodes in patients, unless FVIII activity is substituted for. Hemophilia care was revolutionized in the late 1950s, when FVIII substitution therapy became available. Since then, plasma-derived FVIIIs with varying degrees of purity have been widely used in the treatment of hemophilia A. A new era began in the early 1990s, when recombinant FVIII products [23Kaufman R.J. Wasley L.C. Dorner A.J. Synthesis, processing, and secretion of recombinant human factor VIII expressed in mammalian cells.J Biol Chem. 1988; 263: 6352-62Abstract Full Text PDF PubMed Google Scholar] entered the market. The availability of such products has resulted in a great improvement in the quality of life and life-expectancy of affected patients. Unfortunately, FVIII substituted into patients may be inactivated by FVIII-inhibiting alloantibodies, leading to a serious and life-threatening complication. Activated prothrombin complex concentrates, such as FEIBA [22Turecek P.L. Varadi K. Gritsch H. Auer W. Pichler L. Eder G. Schwarz H.P. Factor Xa and prothrombin: mechanism of action of FEIBA.Vox Sang. 1999; 77: 72-9Crossref PubMed Scopus (0) Google Scholar] or recombinant FVIIa [24Hedner U. Bjoern S. Bernvil S.S. Tengborn L. Stigendahl L. Clinical experience with human plasma-derived factor VIIa in patients with hemophilia A and high titer inhibitors.Haemostasis. 1989; 19: 335-43PubMed Google Scholar], are commonly used to arrest bleeding episodes in patients with inhibitors. In this article, we describe the stimulation of FIXa by FIXa-specific antibodies. This finding has the potential to open up new strategies for treating patients with hemophilia A and FVIII inhibitors in which FVIII is substituted by a molecule that does not induce and is not affected by FVIII inhibitor antibodies. Furthermore, it may then be possible to develop compounds that can be delivered via a non-i.v. route. The interaction between FVIIIa and the serine protease FIXa is a well-known example of an effector function. Binding of the cofactor FVIIIa to FIXa converts the otherwise very poor enzyme to a highly active protease by inducing a conformational change in FIXa that leads to enhanced FX-cleaving activity. To screen specific antibodies that induce a similar conformational change in the protease FIXa, we adapted a commercially available photometric FVIII assay that is widely used to measure FVIII levels in plasma [25Barrowcliffe T.W. Raut S. Sands D. Hubbard A.R. Coagulation and chromogenic assays of factor VIII activity: general aspects, standardization, and recommendations.Semin Thromb Hemost. 2002; 28: 247-56Crossref PubMed Scopus (83) Google Scholar]. In the search for alternative molecules capable of binding to FIXa, we started out with murine anti-FIX antibodies. Although 60% of the hybridoma cell lines screened expressed an FIX-binding antibody, only 1.6% of the hybridoma supernatants showed FVIII-like activity in our screening assay. Three of these antibodies were selected for further investigation. Kinetic experiments revealed that the antibodies enhanced the catalytic activity (kcat) of FIXa approximately tenfold, whereas the Km of hFX for hFIXa was hardly affected. Also, FVIIIa mainly increased the kcat of FIXa and, as in the case of the antibody–FIXa complex, the effect of FVIIIa on FIXa was strictly dependent on the presence of calcium ions and phospholipids. We conclude that the conformational changes induced by FVIIIa and our antibodies are similar. Interestingly, the procoagulant antibodies analyzed did not compete with FVIIIa for FIXa binding, but competed with each other. This indicates that a few ‘hot spots’ are responsible for a correct conformational change. The partial inhibition of FXa formation at higher antibody concentrations (Fig. 1B) might be due to a non-specific concentration effect, i.e. the formation of large antibody–FIXa complexes that might sterically interfere with the FIXa–substrate interaction. However, the exact structure of the active antibody–FIXa complex remains to be resolved by crystallographic studies. The effector function of the anti-FIX antibodies was confirmed in plasma assays. In contrast to FXa generation assays, in model systems that contain purified proteins, plasma assays probe the entire intrinsic or extrinsic coagulation pathway, depending on the trigger applied in the experiment. The antibodies were effective in plasma assays in a dose-dependent manner until FIX was saturated with antibody. Their effectiveness was lower than that of FVIII, which is in line with the kinetic data obtained in the model systems. The antibodies affected neither the other procoagulant reactions nor the inactivation of the activated coagulation factors by plasma protease inhibitors. This can be concluded because thrombin generation curves showed an identical burst and inactivation of thrombin with the antibodies and with FVIII. FVIII-like activity of antibodies was observed in FVIII-deficient and in FVIII inhibitor plasma, which demonstrates that inhibitory antibodies directed against FVIII do not affect the activity of the procoagulant antibodies. An interesting option for developing a potential drug from these antibodies with which to treat hemophilia A would be to design a small molecule from the antibody-binding site. We assume that the effector function of the antibody is achieved by binding the antibody that induces conformational changes in FIXa. This can theoretically also be achieved by a small proteomimetic or peptidomimetic that corresponds to the antibody paratope. However, in order to make such a concept clinically feasible, it would be necessary to find a compound that binds exclusively to activated FIX, thereby conferring FVIII-like activity at the site of injury where FIXa is generated, and avoiding side effects such as FIX structural changes or FIX depletion. According to Gao et al. [26Gao C. Mao S. Lo C.H. Wirsching P. Lerner R.A. Janda K.D. Making artificial antibodies: a format for phage display of combinatorial heterodimeric arrays.Proc Natl Acad Sci USA. 1999; 96: 6025-30Crossref PubMed Scopus (0) Google Scholar], the antibody molecule is a biological device for presenting a combinatorial array of peptide elements in three-dimensional space. Thus, it could be possible to develop a compound (an FIXa ligand) exhibiting FVIII-like activity that can be administered via a non-i.v. route and that is also active in plasma that contains FVIII inhibitors. The authors state that they have no conflict of interest.