Role of anti-idiotypic antibodies in immune tolerance induction.

Replacement therapy using factor VIII (FVIII) elicits FVIII-specific antibodies (abs) in about 25% of the patients. A majority of such abs are directed towards specific FVIII regions in which major epitopes have been identified (C-terminal end of the C2 domain, the N-terminal end of the A2 domain and C1 domain in cases of mild/moderate haemophilia A). We derived five human monoclonal abs (mabs) that react with high affinity to the FVIII C1, C2 or A2 domains respectively and are representative of most of the specific inhibitors observed in haemophilia A patients. We generated mouse anti-idiotypic mabs (anti-Ids) against the paratope of each of the inhibitors. We demonstrated that a combination of these anti-Ids (anti-anti-A2, -C1, -C2) had the ability to neutralize the inhibitory properties of human polyclonal abs in plasma. In 16 of the 18 plasmas tested, the inhibiting FVIII activity was neutralized up to 100% by the anti-Ids mixture with restoration of full FVIII activity. These data allow us to conclude that polyclonal high-affinity FVIII inhibitors could be neutralized with an anti-Ids mixture and that only a limited number of anti-Ids were required for inhibitor neutralization in 90% of the patients. We also demonstrated that anti-Id Abs bound to anti-FVIII human B cell line produced the corresponding anti-FVIII Ab and that this binding was followed by surface capping of complexes. Data obtained in vitro at monoclonal and polyclonal level, confirmed by in vivo assays, and the preliminary results obtained at BCR level, indicate that anti-id mixture made of only a limited number of anti-Ids could be useful in the restoration of haemostasis in haemophilia patients with inhibitor.

Factor VIII Inhibitors. Natural autoantibodies and anti-idiotypes.

Natural antibodies to factor VIII are present in the normal antibody repertoire as other self-reactive antibodies to soluble proteins. The question as to whether they represent just a chance occurrence linked to the huge diversification of the antibody repertoire or whether these antibodies have an actual physiological relevance is not entirely settled. Evidence is in favor of a role in the maintenance of immune homeostasis, however, namely self-reactive antibodies are required to maintain the capacity of the immune system to distinguish self from nonself. Anti-factor VIII antibodies pose an interesting case in point because they exhibit the capacity to inhibit the function of factor VIII. Such a property is neutralized at least in part by the production of corresponding anti-idiotypic antibodies. Normal homeostasis can therefore be viewed as a network of interacting molecules, idiotypes, and anti-idiotypes; disruption of this equilibrium leads to the development of autoimmunity. A question that remains open for the time being is whether this network of interactions can be modulated in a defined way for the treatment of autoimmune reactions. This would mean either passive administration of anti-idiotypic antibodies or active immunization with idiotypes. The former has proved to be efficient, and the latter has still to be demonstrated. Further, and probably most importantly, is the question of the possible application of the idiotypic network concept to the treatment of hemophilia patients producing inhibitors. This essentially requires that an analysis of the anti-factor VIII immune response be carried out at the clonal level. Such work is ongoing in our laboratory.

Humanized severe combined immunodeficient mice as a potential model for the study of tolerance to factor VIII.

A Severe Combined Immunodeficient (SCID) mouse model has been established to evaluate experimental conditions leading to the production of factor VIII (FVIII) autoantibodies. To this end, we humanized 10 groups of 7 mice with peripheral blood mononuclear cells of 10 unrelated healthy blood donors (15 x 10(6) cells/mouse). Mice were injected with saline or immunized i.p. with 50 IU of a plasma derived human FVIII 24 h after reconstitution. Further immunization was made with 25 IU of FVIII every fortnight during 6 weeks and animals were sacrificed after 8 weeks. All reconstituted mice showed a spontaneous production of anti-FVIII antibodies in the absence of immunization with the corresponding antigen. However, no differences were observed regarding the quantity or the quality of these antibodies produced in the immunized or the saline group, indicating that tolerance to FVIII had been transferred with cell reconstitution. Affinity purified FVIII specific antibodies were capable of inhibiting FVIII activity and preventing the binding of FVIII to phospholipids in a dose-dependent manner. Immunoprecipitation experiments showed that the antibodies recognized only the C1 and C2 light chain domains. Since antibodies of interest can be found in the SCID mouse model and, moreover, since they are qualitatively comparable with the source donor's antibodies, this model provides a tool to study the regulation of tolerance against self antigens in normal subjects and in acquired haemophilia patients.

A novel cause of mild/moderate hemophilia A: mutations scattered in the factor VIII C1 domain reduce factor VIII binding to von Willebrand factor.

The mechanisms responsible for the low factor VIII (fVIII) activity in the plasma of patients with mild/moderate hemophilia A are poorly understood. In such patients, we have identified a series of fVIII mutations (Ile2098Ser, Ser2119Tyr, Asn2129Ser, Arg2150His, and Pro2153Gln) clustered in the C1 domain and associated with reduced binding of fVIII to von Willebrand factor (vWf). For each patient plasma, the specific activity of mutated fVIII was close to that of normal fVIII. Scatchard analysis showed that the affinity for vWf of recombinant Ile2098Ser, Ser2119Tyr, and Arg2150His fVIII mutants was reduced 8-fold, 80-fold, and 3-fold, respectively, when compared with normal fVIII. Given the importance of vWf for the stability of fVIII in plasma, these findings suggested that the reduction of fVIII binding to vWf resulting from the above-mentioned mutations could contribute to patients' low fVIII plasma levels. We, therefore, analyzed the effect of vWf on fVIII production by Chinese hamster ovary (CHO) cells transfected with expression vectors for recombinant B domain-deleted normal, Ile2098Ser, Ser2119Tyr, and Arg2150His fVIII. These 3 mutations impaired the vWf-dependent accumulation of functional fVIII in culture medium. Analysis of fVIII production by transiently transfected CHO cells indicated that, in addition to the impaired stabilization by vWf, the secretion of functional Ile2098Ser and Arg2150His fVIII was reduced about 2-fold and 6-fold, respectively, by comparison to Ser2119Tyr and normal fVIII. These findings indicate that C1-domain mutations resulting in reduced fVIII binding to vWf are an important cause of mild/moderate hemophilia A.

A human antibody directed to the factor VIII C1 domain inhibits factor VIII cofactor activity and binding to von Willebrand factor.

The occurrence of factor VIII (fVIII) inhibitory antibodies is a rare complication of fVIII substitution therapy in mild/moderate hemophilia A patients. fVIII mutations in certain regions such as the C1 domain are, however, more frequently associated with inhibitor, for reasons which remain unclear. To determine whether inhibitors could map to the mutation site, we analyzed at the clonal level the immune response of such a patient with an inhibitor to wild-type but not self-fVIII and an Arg2150His substitution in the C1 domain. Immortalization of the patient B lymphocytes provided a cell line producing an anti-fVIII IgG4kappa antibody, LE2E9, that inhibited fVIII cofactor activity, following type 2 kinetics and prevented fVIII binding to von Willebrand factor. Epitope mapping with recombinant fVIII fragments indicated that LE2E9 recognized the fVIII C1 domain, but not the Arg2150His-substituted C1 domain. Accordingly, LE2E9 did not inhibit Arg2150His fVIII activity. These observations identify C1 as a novel target for fVIII inhibitors and demonstrate that Arg2150His substitution alters a B-cell epitope in the C1 domain, which may contribute to the higher inhibitor incidence in patients carrying such substitution. (Blood. 2000; 95:156-163)

Anti-idiotypic antibodies: from regulation to therapy of factor VIII inhibitors.

Evidence has recently accumulated showing that anti-idiotypic antibodies specific to anti-FVIII antibodies are present in the plasma of healthy individuals and of haemophilia A patients with or without inhibitors, where they can neutralise the FVIII inhibitory activity. Additionally, patients successfully desensitised towards FVIII have an increased production of anti-idiotypic antibodies with no significant reduction in anti-FVIII antibodies. We review here possible strategies for modulating the anti-FVIII immune response by idiotypic interactions.

Some factor VIII (FVIII) inhibitors recognise a FVIII epitope(s) that is present only on FVIII-vWF complexes.

A mild haemophilia A patient (LE) with an Arg2150His mutation in the C1 domain of the factor VIII (FVIII) light chain was shown to have anti-FVIII antibodies inhibiting wild type but not self FVIII. Polyclonal anti-FVIII antibodies of this patient were purified by affinity adsorption using recombinant FVIII (rFVIII) and/or plasma-derived FVIII-von Willebrand factor (vWF) complexes. A distinct population of antibodies was obtained that bound to FVIII-vWF complexes but not to rFVIII, indicating that an epitope was created by the association of FVIII to vWF. Such antibodies belonged to the IgG2 isotype, but the FVIII epitopes to which they bind could not be mapped with precision due to vWF dependency. Depletion experiments showed that anti-FVIII antibodies recognising FVIII-vWF complex also distinguished wildtype from mutated self FVIII, indicating that the Arg2150His mutation alters the B cell epitope formed by the association of FVIII to vWF. To determine whether the Arg2150His substitution also alters the formation of the FVIII-vWF complex, the interaction between mutated or normal FVIII with vWF was evaluated in plasma. The dissociation rate of mutated FVIII from vWF was found to be significantly increased. The presence of an Arg2150His mutation therefore results in the disappearance of a FVIII B cell epitope generated by the association of FVIII with vWF. Patients carrying such an Arg2150His mutation and receiving infusion of wild-type FVIII may therefore be at risk of developing inhibitors to allogeneic FVIII only.

Antifactor VIII antibody inhibiting allogeneic but not autologous factor VIII in patients with mild hemophilia A.

Two unrelated patients with the same Arg2150His mutation in the factor VIII (FVIII) C1 domain, a residual FVIII activity of 0.09 IU/mL, and inhibitor titres of 300 and 6 Bethesda Units, respectively, were studied. Further analysis of patient LE, with the highest inhibitor titer, showed that (1) plasma or polyclonal IgG antibodies prepared from LE plasma inhibited the activity of allogeneic (wild-type) but not of self FVIII; (2) the presence of von Willebrand factor (vWF) increased by over 10-fold the inhibitory activity on wild-type FVIII; (3) the kinetics of FVIII inhibition followed a type II pattern, but in contrast to previously described type II inhibitors, LE IgG was potentiated by the presence of vWF instead of being in competition with it; (4) polyclonal LE IgG recognized the FVIII light chain in enzyme-linked immunosorbent assay and the recombinant A3-C1 domains in an immunoprecipitation assay, indicating that at least part of LE antibodies reacted with the FVIII domain encompassing the mutation site; and (5) LE IgG inhibited FVIII activity by decreasing the rate of FVIIIa release from vWF, but LE IgG recognized an epitope distinct from ESH8, a murine monoclonal antibody exhibiting the same property. We conclude that the present inhibitors are unique in that they clearly distinguish wild-type from self, mutated FVIII. The inhibition of wild-type FVIII by LE antibody is enhanced by vWF and is associated with an antibody-dependent reduced rate of FVIIIa release from vWF.

Mechanism and kinetics of factor VIII inactivation: study with an IgG4 monoclonal antibody derived from a hemophilia A patient with inhibitor.

The development of an immune response towards factor VIII (fVIII) remains a major complication for hemophilia A patients receiving fVIII infusions. The design of a specific therapy to restore unresponsiveness to fVIII has been hampered by the diversity of the anti-fVIII antibody. Molecular analysis of the specific immune response is therefore required. To this end, we have characterized an fVIII-specific human IgG4kappa monoclonal antibody (BO2C11) produced by a cell line derived from the memory B-cell repertoire of a hemophilia A patient with inhibitor. BO2C11 recognizes the C2 domain of fVIII and inhibits its binding to both von Willebrand factor (vWF) and phospholipids. It completely inhibits the procoagulant activity of native and activated fVIII, with a specific activity of approximately 7,000 Bethesda units/mg. vWF reduces the rate of fVIII inactivation by BO2C11. The antibody-fVIII association rate constant (kass approximately 7.4 x 10(5) M-1 s-1) is eightfold lower than that for vWF-fVIII association, whereas its dissociation rate constant (kdiss < or = 1 x 10(-5) s-1) is 100-fold lower than that for the vWF-fVIII complex, which suggests that BO2C11 almost irreversibly neutralizes fVIII after its dissociation from vWF. BO2C11 is the first human monoclonal anti-fVIII IgG antibody that has been isolated and allows the study of fVIII inactivation at the molecular level.

Restricted epitope specificity of anti-FVIII antibodies that appeared during a recent outbreak of inhibitors.

We recently described an outbreak of anti-factor VIII (FVIII) antibodies in a population of haemophilia A patients non-responsive to FVIII (1). To find out what part of the FVIII molecule had been altered, we purified specific anti-FVIII antibodies from the plasma of the five patients showing high titres of inhibitors. An average of 100 micrograms antibodies per ml of initial plasma was recovered by immunoadsorption on insolubilised FVIII. The antibodies followed the normal isotypic distribution, including the presence of specific IgG2 antibodies; the relative increase in IgG4 that is usually observed in patients with long-standing inhibitors, was not present. The regions of FVIII to which human antibodies bound were determined by a competition assay using a panel of murine monoclonal antibodies: two major regions were identified, one located in the A2 heavy chain domain, and the other made of determinants of both the A3 and C2 light chain domains. Affinity-purified antibodies inhibited the function of FVIII as determined in a chromogenic assay. However, variations existed in the affinities with which antibodies bound to soluble FVIII. This study shows that the immunogenicity of two particular regions of FVIII has been altered. A screening for alterations located in these two regions should possibly be included in the preclinical evaluation of FVIII concentrates.

Heating lyophilised factor VIII does not alter its recognition by specific antibodies.

BACKGROUND AND OBJECTIVES: Alterations of factor VIII (FVIII) during preparation procedures can potentially affect its immunogenicity. One method evaluating such alterations could be by determining the reactivity of FVIII with specific antibodies. MATERIALS AND METHODS: Since heat treatment is currently used to reduce the risk of viral transmission, we evaluated the immunoreactivity of plasma-derived FVIII before and after heating at different temperatures and for different periods. Freeze-dried FVIII was used for these experiments as part of the validation procedure of a novel FVIII preparation. RESULTS: Heating FVIII for up to 72 h at 80 degrees C does not alter its reactivity with specific rabbit antibodies or mouse monoclonal antibodies, although some loss of FVIII activity occurred after 72 h. After heating for 2 h at 100 degrees C, a procedure that reduced FVIII activity by about 50%, there were still no significant effects on FVIII reactivity with monoclonal antibodies. CONCLUSIONS: Freeze-dried preparations of plasma-derived FVIII seem to be resistant to heat-induced structural denaturation.

Factor VIII inhibitors in previously treated haemophilia A patients with a double virus-inactivated plasma derived factor VIII concentrate.

Antibodies to factor VIII (inhibitors) are usually produced at the beginning of treatment with factor VIII and are rare in multitransfused patients. Such antibodies are deemed to be patient-related, as supported by the description of a number of associated risk factors. However, a second category of inhibitors has recently been identified, namely antibodies occurring in multitransfused patients as a result of exposure to a particular factor VIII concentrate. A first outbreak of product-related inhibitors was recently described. The present paper describes the second well-documented occurrence of such inhibitors. Eight out of 140 multitransfused patients with severe haemophilia A developed an inhibitor to factor VIII shortly after changing treatment to a double-virus inactivated plasma-derived factor VIII concentrate. In addition to solvent-detergent treatment, this concentrate was pasteurised at 63 degrees C for 10 hours. Exposure to the pasteurised product before inhibitor detection ranged from 9 to 45 days. Inhibitor titers varied between 2.2 and 60 Bethesda Units and recovery of transfused factor VIII ranged from 0.21 to 0.68 (expressed as i.u./dl factor VIII rise per i.u./kg administered). In contrast to usual inhibitors in haemophilia A patients, these product-related inhibitors showed complex inhibition kinetics. They were found specific for the factor VIII light chain. The inhibitors gradually declined when exposure to the pasteurised product was stopped, despite further treatment with other factor VIII concentrates. The present data stress the importance of carefully monitored clinical studies, both in previously treated and previously untreated patients, before introduction of a new or modified clotting factor concentrate.

Neutralizing antiidiotypic antibodies to factor VIII inhibitors after desensitization in patients with hemophilia A.

Hemophilia A patients producing antibodies towards FVIII are usually treated with infusions of high doses of FVIII in an attempt to "desensitize" them. To examine the mechanisms by which such desensitization operates, sequential plasma samples of two unrelated inhibitor patients were analyzed for anti-FVIII and antiidiotypic antibodies before and during infusions of high doses of FVIII. Anti-FVIII antibodies were separated from antiidiotypic antibodies by immunoaffinity chromatography before analysis. We show in the present study that the concentration of anti-FVIII antibodies did not change during a successful desensitization and that antibodies maintained their capacity to inhibit the procoagulant function of FVIII, even though the number of Bethesda units in plasma was reduced to undetectable levels. Using a competition assay with mAbs, we further show that the specificity of human antibodies did not vary significantly during therapy. Finally, we show that the treatment elicited antiidiotypic antibodies, which neutralized the inhibitory capacity of anti-FVIII antibodies. Inhibitor antibodies can therefore not be accurately evaluated in plasma, as their function appears to be neutralized by antiidiotypic antibodies. These findings could have implications for the design of new therapies for hemophilia A patients with inhibitors.

Strategy for pre-clinical evaluation of factor VIII concentrates.

A series of methods and assay systems was designed, using both mouse monoclonal antibodies and purified human polyclonal antibodies, by which alterations in the antigenic properties, and potentially therefore in the immunogenic properties, of FVIII concentrates could be identified. Those methods could be applied to the pre-clinical evaluation of FVIII concentrates. It has become evident that very subtle alterations in FVIII can have dramatic effects on its antigenicity and immunogenicity and it is suggested that an analysis of FVIII preparations, using reagents such as mouse monoclonal antibodies, should be carried out as a pre-clinical evaluation. It is also thought that any haemophiliac patient is at risk of developing anti-FVIII antibodies, even though he had been considered as tolerant to the infusion of FVIII. The diversity of FVIII concentrates now available on the market has multiplied the chances of encountering a product that would give an immune response in some patients and this further stresses the need for a careful check of FVIII concentrates.

Healthy subjects produce both anti-factor VIII and specific anti-idiotypic antibodies.

Anti-Factor VIII (FVIII) antibodies were prepared by a combination of salt precipitation, gel filtration chromatography, and specific adsorption over insolubilized FVIII from the serum of 10 healthy subjects with normal levels of FVIII. Antibody specificity was confirmed by the capacity to recognize soluble and insolubilized FVIII and to neutralize FVIII cofactor activity in FX activation. Epitope mapping was carried out using a competition ELISA in which affinity-purified human antibodies inhibited the binding of labeled monoclonal antibodies. In most cases, a single region of the A3 domain of the FVIII light chain was recognized by the antibodies, while the reactivity toward heavy chain epitopes differed from one antibody preparation to the other. Sera or IgG fractions of the serum before immunoadsorption over insolubilized FVIII did not bind to FVIII. The IgG fraction that was not retained on the FVIII immunosorbent contained IgG that bound to the variable part of anti-FVIII mouse monoclonal antibodies and inhibited the binding of labeled FVIII; in addition, the IgG fraction inhibited the binding of affinity-purified human antibodies to FVIII, thereby strongly suggesting the presence of anti-idiotypic antibodies. These findings indicate that the presence of anti-FVIII antibodies is a more universal phenomenon than previously thought and that anti-idiotypic antibodies capable of inhibiting the binding of anti-FVIII antibodies to FVIII are produced spontaneously.

Anti-factor VIII antibodies of hemophiliac patients are frequently directed towards nonfunctional determinants and do not exhibit isotypic restriction.

A significant proportion of hemophilia A patients receiving transfusions of factor VIII (FVIII) develop a specific antibody response towards FVIII. These antibodies are usually detected by assays in which they inhibit the function of the molecule, such as the Bethesda clotting test. We have prepared anti-FVIII antibodies by specific immunoadsorption from the plasma of four hemophiliacs with stable inhibitor levels. The isotypic distribution of such antibodies was determined and their capacity to bind to insolubilized FVIII was compared with their inhibitory activity in two functional assays, namely, the Bethesda assay and a chromogenic assay. In addition, the FVIII epitope specificity was determined by competition with monoclonal antibodies for the binding to insolubilized FVIII. We show here that (1) anti-FVIII antibodies are not isotypically restricted; thus, a significant proportion of specific IgG2 was found; (2) antibodies are frequently directed towards epitopes of FVIII that are not directly involved in the function of the molecule and therefore escape detection in the Bethesda method or chromogenic assay; and (3) each patient shows a unique pattern of FVIII epitope recognition. We conclude that evaluation of anti-FVIII antibodies by a functional method does not provide an accurate evaluation of the specific antibody response. These findings have important implications for the comparison of the immunogenicity of FVIII molecules produced by different technologies and for the development of methods to control anti-FVIII antibody production.

A higher than expected incidence of factor VIII inhibitors in multitransfused haemophilia A patients treated with an intermediate purity pasteurized factor VIII concentrate.

In May 1990, 218 patients with haemophilia A regularly attending the Leuven Haemophilia Center were randomly assigned to a group receiving either of two newly introduced factor VIII concentrates: factor VIII-P, an intermediate purity pasteurized concentrate, or factor VIII-SD, a high purity concentrate treated with solvent-detergent for viral inactivation. Patients were followed from May 1990 until October 1991. Between August 1991 and October 1991 a clinically important factor VIII inhibitor was detected in five out of the 109 patients receiving factor VIII-P while none of the 109 patients receiving factor VIII-SD developed such antibodies. All patients acquiring an inhibitor had previously been clinically tolerant to transfused factor VIII with 200 to more than 1,000 days of exposure to factor VIII prior to May 1990. Patients with inhibitors were transfused daily with 30 U factor VIII-SD per kg body weight, which was associated with a gradual decline of the inhibitor level. In all patients the antibodies were relatively slow-acting and predominantly directed towards the light chain of factor VIII. This study demonstrates a higher than expected incidence of factor VIII inhibitors associated with the use of a specific factor VIII concentrate in multitransfused haemophilia A patients. It indicates the usefulness of evaluating newly introduced concentrates in prospective, randomized trials.