The M echanism of Factor VIII Inactivation by H um an Antibodies

ANNALS O F CLINICAL A N D LABORATORY SC IE N C E, Vol. 15, No. 1 Copyright 1985, Institute for Clinical Science, Inc. The M echanism of Factor VIII Inactivation by H um an Antibodies II. The Effect of Factor VIII R: Antigen on the Rate o f Interaction* JOHN LAZARCHICK, M.D. Department of Laboratory Medicine, Medical University of South Carolina, Charleston, SC 29425 ABSTRACT Factor VIII R:Ag by binding to the procoagulant component (VIILC) inhibits the extent and rate of interaction of hum an antifactor VIII antibodies with VIILC. W hen this reaction is examined under ionic strength conditions (0.24M CaCl) which dissociate the two factor VIII components, the extent of the reaction is increased approximately two fold and the initial rate of interaction is increased three to four fold for both intact IgG antibody and its Fab' derivative. W ith isolated procoagulant component, increased ionic strength conditions only influence the rate of interaction. These studies further explain the peculiar tim e-dependence of this in te r action. Introduction It has been dem onstrated in a previous study that the presence of factor VIILR antigen affects the interaction of human antibodies w ith factor V III.3 W hen antibody assays are perform ed under ionic strength conditions which would cause dissociation of the factor VIII molecule, there is an approximate two-fold increase in antibody titer for both intact IgG and Fab' derivative com pared to the same assay being done in physiologic buffer. * Supported in part by Grant HL26044 from the National Heart, Lung and Blood Institute. Since the num ber of antifactor VIII antibody m olecules p re se n t in any p aired solution was identical, it was postulated th a t factor V IIL R antigen by bin d in g V IILC was rate-lim iting in these reactions. Because by convention antibody titers are determ ined at a defined two h o u r e n d p o in t, th ese findings do not allow inference w hether this effect re p resents simply a greater extent of interaction (m ore V IIL C accessible to a n tibody) or w hether the rate of reaction is accelerated (final equilibrium reached sooner) or a com bination of b oth. To approach this question, this interaction has been exam ined at sequential tim e 76 0091-7370/85/0100-0076 $01.20 Institute for Clinical Science, Inc.

INHIBITORY EFFECT OF VIII R: ANTIGEN 77 points enab lin g us to g en e ra te rate of reaction d e te rm in a tio n s w hich p e rm it testing these hypotheses. O ur results confirm that factor VIII :R antigen is rate-lim iting and suggest that th e antibody bin d in g and factor V III inactivation reactions, although closely related, are distinct components of this interaction. M aterials and Methods Pooled normal plasma (PNP) from 16 donors and partially purified factor VIII procoagulant activity, p re p a re d using im m u n o ad so rb an t chrom atography as p reviously d escrib ed, w ere th e factor VIII sources used in all studies.7 The latter m aterial had 118 U per dl factor VIII procoagulant activity (VIII:C), 250 U per dl coagulant antigen (VIII C:Ag), and less than 0.03 U per dl of factor VIII related antigen (VIII R:Ag). Factor V III procoagulant activity was m easured by a onestage m ethod using factor VIII deficient plasma as substrate. Factor VIII C:(Ag) and factor VIII R.Ag were m easured by im m unoradiom etric assays as previously d e s c rib e d.1,4 Pooled norm al hum an plasma (100 U per dl) served as standard for m easurem ent of all factor VIII param ete rs. T he antifactor V III containing serum was o b tained from a p a tie n t in w hom antifactor V III activity arose spontaneously. P ro p e rtie s of this antibody have been described in a previous publication.4 Both IgG and Fab' derivatives w ere prepared as previously described.5 Antifactor VIII activity m easurem ents w ere made under conditions suggested by th e B eth esd a conference w ith one u n it of antifactor V III activity being defined as that which inactivates 50 percent of the procoagulant activity of the control sample at the end of two hours.2 W hen dilutions of antibody were used, the antibody tite r was obtained by m ultiplying th e observed value by the dilution factor. To perm it assessment of the effect of in creased salt (0.24M C ac l2) concentrations on the rate of factor VIII inactivation by antibody and su b sequently the antibody titer, the assay system was modified as follows. The initial m ixtures w ere prep ared by incubating 0.9 ml of pooled normal plasma (PNP) w ith 0.1 ml of e ith e r 2.4M C ac l2 or borate buffered saline (BBS). After a 30 m inute incubation at room tem perature, an 0.5 ml aliquot from each PN P solution was incubated with an equal volume of IgG or Fab' antibody preparation in BBS for two hours at 37 C. Factor VIII :C of th ese m ixtures was th en com p ared to control m ixtures in which BBS was incubated with the PNP solutions instead of the antibody preparations. Standard factor VIII dose-response curve was generated on the control mixtures at 30, 60, and 120 m inutes. Aliquots from the antibody-containing solutions w ere rem oved and residual factor VIII activity at each of th ese tim e points was d e te rm in e d based on th e corresponding paired control mixture dose response curves. Assay resu lts at th re e and 15 m in u tes w ere based on the 30 m inute control mixtures. For purposes of comparison, all antibody titers are expressed in B ethesda units. Studies utilizing partially purified factor VIII procoagulant m aterial as a factor VIII source were perform ed in an identical manner. The procoagulant material was initially diluted with BBS to obtain 100 U per dl of factor VIII procoagulant activity to allow calculation of antibody titers. All antibody containing test solutions w ere assayed at a 1:20 dilution in borate buffered saline. The concentration of CaCl2 used to clot assay mixtures containing 0.24M CaCl2 was modified to ensure a final concentration of 0.0075M CaCl2. Results A n t i b o d y P r e p a r a t io n The antibody preparation used throughout this study had an average B ethesda

78 LAZARCHICK unit titer of 600 units when assayed with BBS buffer but a titer of 1380 U when 0.24M CaCl2 was the buffer condition. Its Fab' derivative assayed under each of th ese conditions was 166 and 280 U, respectively. In each case, the average antibody tite r was increased to approximately two fold. Since these results are reflection of static endpoint analysis (end of two hour incubation), th e n e e d to exam ine early points to d eterm in e the effect of the buffer conditions on rate of th e in teractio n was in v estig ated. The approach utilized is shown in figure 1, w ith residual factor VIII procoagulant activity m easurem ents being done at 3, 15, 30, 60, and 120 m inutes. W ith IgG at 1:300 and 1:600 dilutions, equilibrium is reached by the 60 m inute point in the presence of the 0.24M CaCl2 buffer, i.e., th e re is no fu rth e r loss of factor V III activity at the 120 m inute tim e. An additional feature is that the paired antibody dilutions have totally different slopes up to the 60 m inute tim e point. The final residual activity of the 1:300 dilution of IgG in borate buffered saline is similar to the 1:600 dilution in 0.24M calcium chloride buffer, confirm ing the approxim ate two-fold increase in antibody activity u nder the increased ionic strength buffer condition. W hen F ab' derivatives were used as the antibody source, the results are even more dramatic. In 0.24M calcium ch lo rid e buffer, th e in te ra c tion is essentially com pleted by the 15 to 30 m inute tim e period for both dilutions of antibody used. In contrast, in the presence of borate buffered saline, equilibrium was not reached by 60 m inutes. The INCUBATION TIME (Minutes) INCUBATION TIME (M inutes) F ig u r e 1. The rate of factor VIII inactivation by human antifactor antibodies in physiologic borate buffered saline (BBS) versus increased ionic strength (0.24M calcium chloride) buffers. Section A shows IgG; Section B shows Fab' derivative.

INHIBITORY EFFECT OF VIII R: ANTIGEN 79 effect on th e e x te n t of th e reaction is again n o ted w ith the 1:100 dilution of antibody in 0.24M calcium chloride buffer having th e sam e final residual activity as the 1:50 dilution of the same antibody in borate buffered saline. R a t e o f F a c t o r VIII I n a c t iv a t io n by A n t i- f a c t o r A n t i b o d i e s W hen these data points are examined in term s of rates of interaction, i.e., unit change in factor V III procoagulant activity per unit tim e, the results are striking (table I). W ith IgG at 1:600 dilution in borate buffered saline, no interaction is evident at the three m inute point. At 15 m inutes, the rate has m axim ized with loss of 1.3 U p er m inute. There is a slight d ecrease in th e rate of procoagulant activity loss by the 30 m inute point, and it is not until the 120 m inute point that the rate is significantly changed. In the presence of calcium chloride, the rate is increased several fold at any point com pared to the borate buffered saline with a maximal rate occurring at th re e m inutes. Subsequent points show the rate to decrease by approximately 50 percent of the previous determ ination. The rates of reaction of the Fab' derivatives show a similar pattern with the initial rate markedly accelerated in the presence of calcium chloride com pared to the borate buffered saline. T i m e C o u r s e o f F a c t o r VIII I n a c t iv a t io n W hen this sam e se q u en tial p aired tim e point analysis is utilized to d e te r m ine residual VIII:C activities and antibody titers, the effect of buffer conditions on the interaction is even more apparent (table II). For purposes of com parison only, all antibody titers based on residual factor VIII :C m easurem ents determ ined before the two hour point are expressed in B ethesda units. A significant differ- TABLE I Rate of Factor VIII Inactivation in Borate Buffered Saline and 0.24M CaCl2 Buffers I n c u b a t io n U n i t s o f V I I I : C A n tib o d y T im e I n a c t i v a t e d p e r M in u te S o u r c e ( M in u te s ) BBS 0.2 4 M C a C l2 Ig G 1 :6 0 0 3 0 5. 6 15 1. 3 3. 7 30 1.1 2.1 6 0 0. 8 1. 3 1 2 0 0. 4 0. 7 F a b ' 1 : 1 0 0 3 0 10 15 1. 3 4. 9 30 1.1 2. 7 6 0 0. 7 1. 2 1 2 0 0. 2 0. 6 ence can be seen in antibody titers under each buffer condition at each corresponding tim e point. For IgG, at a 1:600 dilution, antibody activity is already apparent at three m inutes in the presence of th e calcium chloride buffer (152 B e thesda units, representing 11 percent of the final titer). By 15 m inutes, 51 percent of the final titer is achieved; by the 60 m in u te p o in t, th e reaction ap p ears to have reached equilibrium and no further change in the titers is noted at the two h o u r point. In co n trast, no antibody interaction is noted w ith the same antibody dilution in borate buffered saline at the three m inute point and only 30 p ercent of the final titers achieved at 15 m inutes. The effect of the increased ionic stre n g th conditions is even m ore p ro nounced on the Fab' interaction with 20 percent of the final titer being noted at three m inutes and 80 percent by 15 minutes. W ith the univalent antibody derivative th e reaction is essen tially com pleted by 30 m inutes. This is in m arked contrast to th e b o rate b u ffered saline conditions in which zero percent, 20 percent, and 31 percent of the final titer is achieved at each com parative point. D is s o c i a t i o n o f F a c t o r VIII To examine the prem ise that the effect of calcium chloride on this reaction is due

80 LAZARCHICK to its dissociation of factor VIII thus making factor VIII :C m ore available for interaction w ith antibody, additional experim en ts w ere p erfo rm ed w ith F a b ' and isolated factor VIILC procoagulant m aterial. The results are shown in table III. Only the 30, 60, and 120 m inute time points were examined. If the accelerating effect n o ted w ith th e in creased ionic strength solutions w ere due to dissociation of factor V III alone, the antibody reaction w ith isolated procoagulant m aterial u n d er eith er salt condition should show sim ilar rates of in te ractio n and, therefore, sim ilar antibody tite r values should be p re se n t at each tim e point. Again with the calcium chloride solution, the reaction is essentially com pleted at the 30 m inute point similar to the experim ent when PNP was used as a factor V III source. An in te rm e d ia te effect, however, is noted in the presence of the b o rate b u ffered saline. T he two hour antibody tite r is not significantly different than the results in calcium chloride at the same tim e point. However, a distinct difference in titer is evident at 30 m inutes; even at 60 m inutes, the reaction is only 86 percent com pleted. There appears to be retardation interaction in the presence of borate buffered saline com pared to the calcium chloride condition. Similar results w ere noted with TABLE I I Time Course of Factor V III Inactivation by Antibody in Borate Buffered Saline and 0.24M CaCl2 Anti- Residual V I I I :C per Antibody T ite r body Incubation Time (Minutes) Source B u ffer 3 15 30 60 120 IgG BBS 0 / 0 8 0 /1 8 2 6 8 /3 3 0 5 4 /5 4 0 5 0 /6 0 0 1 :6 0 0 (0 %) (30%) (55%) (90%) ( 1 0 0 %) 0.24M C ac l2 8 3 /1 5 2 ( 1 1 %) 4 4 /7 0 8 (51%) 3 2 /9 8 0 (72%) 1 9.6 /1 4 1 6 (103%) 2 0 /1 3 8 0 ( 1 0 0 %) F a b BBS 0 / 0 8 1 /3 0 6 6 /6 0 4 5 /1 1 6 3 5 /1 5 2 1 :1 0 0 ( 0 %) ( 2 0 %) (39%) (76%) ( 1 0 0 %) 0.24M C a d 2 7 0 /5 0 ( 2 1 %) 2 6 /1 9 0 (79%) 1 8 /2 5 0 (104%) 1 9 /2 4 0 ( 1 0 0 %) 1 9 /2 4 0 ( 1 0 0 %) N um bers i n p a r e n t h e s e s a r e p e r c e n t o f e a c h a n tib o d y t i t e r r e l a t i v e t o f i n a l t i t e r d e te r m in e d a t t h e tw o h o u r p o i n t. A n t i bo dy S o u rc e F a b 1 1 :1 0 0 TABLE I I I Time Course of Inactivation of Isolated Procoagulant Factor VIII by Antibody in Borate Buffered Saline and 0.24M CaCl2 BBS B u f f e r 0.24M C ac l2 A n t ib o d y T i t e r (U p e r d l ) I n c u b a t i o n Tim e (M inutes.) 30 6 0 1 2 0 85(56% ) 180(92% ) 130(86% ) 196(100% ) 152(100% ) 195(100% ) N um bers i n p a r e n t h e s e s a r e p e r c e n t o f e a c h a n tib o d y t i t e r r e l a t i v e t o t h e f i n a l t i t e r d e te r m in e d a t tw o h o u r p o i n t. IgG as th e antibody source; therefore, this effect on reaction rates is not due to a property of the valency of the antibody per se. The basis for this difference is uncertain, but several possibilities can be suggested. The increased cation/anion concentrations could have an effect on the rate of the interaction independent of their dissociating effect on factor VIII. It is also conceivable that the isolated procoagulant m aterial behaves in an anom alous m an n e r in th e p resen ce of borate buffered saline with a tendency to polym erize m aking some VIILC inaccessible or small amounts of undetectable factor VIII:R antigen are present causing association to occur in the presence of borate buffered saline. If the latter is the case, the reassociation is not com plete since the rate at 30 m inutes is faster and th e p e rc e n t of final antibody tite r is greater (56 p ercen t versus 39 percent) than when pooled normal plasma is used as a factor VIII source. Discussion T hese stu d ies confirm th e approxim ately tw o-fold increase in apparen t antifactor V III antibody tite r of intact IgG and its F a b ' d eriv ativ e w hen the assays are p e rfo rm ed u n d e r increased ionic stre n g th conditions. Since this effect of salt conditions was abolished w hen isolated procoagulant m aterial

INHIBITORY EFFECT OF VIII R: ANTIGEN 81 devoid of any m easu rab le factor V III antigen was used, it was suggested by us that the increase in titer was probably due to th e in creased m olarity causing dissociation of the factor V III complex w ith th e pro co ag u lan t p o rtio n of the bim olecular factor V III com plex th en being m ore readily accessible to a n tibody. Since this conclusion was based on endpoint determ inations, there were no means of analyzing for an effect of the salt conditions on the rate of the interaction. From the data presented here, it is now apparent that the effect is not only due to dissociation of the factor VIII complex, but also a direct acceleration of the rate of interaction of factor VIII procoagulant antigen and antibody. This contribution of increased salt concentrations to the rate of interaction is a p p aren t w hen a com parison of tim e courses of factor V III inactivation are examined under both buffer conditions and is most evident when initial rates of inactivation are examined. For example, at 15 m in u tes b oth IgG and F a b ' are inactivating factor V III at a rate th ree tim es faster in th e 0.24M calcium chloride buffer than in the borate buffered saline. The comparative acceleration of factor V III inactivation, although lessen ed as th e reactio n p ro ceeds, is still apparent at each assay tim e point during the two hours incubation. The effect of this rate acceleration is to shorten the reaction equilibrium time. Using the two hours endpoint as the reference for the maximum antibody titer and maximum extent of interaction, it is evident that in 0.24M calcium chloride the maximum IgG effect is reached at the 60 m inute and th e F a b ' at th e 30 m inute assay points. W ith the latter antibody even by 15 m inutes, approximately 80 percent of the reaction is com plete versus only 20 percent in the borate buffered saline. That this apparent change in antibody titer is not only due to an effect of the increased salt conditions on the rate of the reaction but also a dissociating effect on the factor V III com plex is e v id en t from the extent of the reaction. At the two hour point, the same dilution of antibody and, therefore, the same num ber of antibody molecules m ore extensively inactivate factor VIII in calcium chloride than in borate buffered saline. Since the num ber of factor VIII molecules is the sam e in both m ixtures, i.e., sam e p ro coagulant activity, the increased salt conditions abolish the protective effect of the factor VIILR antigen on VIII:C and more procoagulant molecules are accessible to inactivation by antibody. The results using isolated procoagulant material support this interpretation. In this case, final two hour titers obtained in e ith er buffer condition w ere essentially the same, implying that the extent of the reaction is influenced only by the physicochem ical properties of the p ro coagulant material. In a dissociated state it is readily accessible to antibody attack; however, when bound, some of the factor VIII:C sites are inaccessible. It is conceivable that if the reaction under borate buffered saline conditions w ere allowed to proceed beyond the two hour point (the limit imposed by the Bethesda Conference), even these sites may become available. W hen the tim e course of inactivation is examined, the secondary effect of salt to increase the rate of reaction is evident. By 30 m inutes, 92 percent of th e F a b ' reaction is co m p leted in th e presence of 0.24M calcium chloride but only 56 percent in the presence of borate buffered saline. P relim in ary data m easuring factor VIII :C antigen with human rad io lab eled F a b ' anti-factor antibody would suggest the rate of antibody binding to the antigen site is similar to the factor VIII procoagulant m olecule either in its associated form with factor VIILR antigen or in its dissociated form. This finding in conjunction w ith th e inactivation data shown here strongly suggest that (1) the antigen binding site on the procoagulant portion is distinct from the procoagulant site, and (2) inactivation of

82 LAZARCHICK the procoagulant site on the intact factor VIII molecule involves an initial binding with antibody and a secondary loss of factor V IILC, perhaps through a steric change in the m olecule which is rate-limited by the presence of factor VIILR antigen. O ur previous d em o n stratio n of unexpectedly small im m une complexes resulting from this interaction are consistent w ith these conclusions.3 This difference in availability of procoagulant sites has also been suggested in a report by M ontgomery and co-workers who found a difference in factor VIII procoagulant activity and factor V IILC antigen in a patient with a variant form of von Willebrand s disease.6 These studies support the concept that factor VIII is a complex of two molecular weight species and that at least in vitro th e p rese n c e of factor V IIL R antigen would appear to interfere with the rapidity that which antibody can react with the procoagulant portion of factor VIII molecule. These studies have fu rth er im plications in that they suggest the conventional m ethod for calculating antibody titers in Bethesda units, although a much needed reference system for comparing antibody titers betw een different laboratories, may be an inadequate indication of antibody concentration which would tend to underestim ate antibody activity in vivo. This has a clinical corollary in that treatm ent of patients who have antibodies w ith am ounts of factor V III replacem ent therapy, based on Bethesda tite r calculations, often resu lts in th e recovery of low er levels of factor VIII activity post infusion than anticipated. References 1. H o v e r, L. W. : Im m unologic studies of antihem ophilicfactor(ahf, factorviii). IV. Radioimmunoassay of A H F antigen. J. Lab. Clin. Med. 80:822-833, 1972. 2. K a s p e r, C. K., A l e d o r t, L. M., C o u n t s, R. B., E d s o n, J. R., F r a n t a n t o n i, J., G r e e n, D., H a m p t o n, J. W., H e l g a r t n e r, M. W., L a z e r - s o n, J., L e v i n e, P. H., M c M i l l a n, C. W., P o o l, J. G., S h a p i r o, S. S., S h u l m a n, N. R., and v a n E y s, J. : A more uniform measurement of factor VIII inhibitors. Throm b. Diath. Haem orrh. 34:869-871, 1975. 3. L a z a r c h i c k, J. : The effect of the physiochemical state of factor VIII on its interaction with human antibodies. Ann. Clin. Lab. Sci. 14:3 75-380, 1984. 4. L a z a r c h i c k, J. and H o y e r, L. W. : Immunoradiometric measurement of the factor VIII procoagulant antigen. J. Clin. Invest. 6 2 :1 0 4 8-1052, 1978. 5. L a z a r c h i c k, J. and H o y e r, L. W. : The properties o f im m une com plexes formed by human an tib od ies to factor VIII. J. C lin. In vest. 60:1070-1079, 1977. 6. M o n t g o m e r y, R. R., H a t h w a y, W. E., J o h n s o n, J., J a c o b s o n, L., and M u n t e a n, W. : A variant of von Willebrand s disease with abnormal expression of factor VIII procoagulant activity. Blood 60:201-2 0 7, 1982. 7. T u d d e n h a m, E. D. G., T r a h o l d, N. C., C o l l i n s, J. A., a n d H o y e r, L. W.: T h e p r o p e r t i e s o f f a c t o r VIII c o a g u l a n t a c t i v i t y p r e p a r e d b y i m m u - n o a d s o r b e n t c h r o m a t o g r a p h y. J. L a b. C l i n. M e d. 93:40-53, 1979.