2003; baxter - Supplements - Haematologica

haematologicahinformation for authors, readers and subscribersHaematologica (print edition, ISSN 0390-6078) publishes peer-reviewed papers across all areas of experimental and clinicalhematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and the way it serves the scientificcommunity is detailed online: http://www.haematologica.org/main.htm (journal’s policy).Papers should be submitted online: http://www.haematologica.org/submission. For the time being the journal considers alsopapers submitted via surface mail (Editorial Office, Haematologica, Strada Nuova 134, 27100 Pavia, Italy) or as attachments toemail messages (office@haematologica.org). However, these submission modalities are discouraged and will be abolishedshortly.Haematologica publishes editorials, research papers, decision making & problem solving papers, review articles and scientificletters. Manuscripts should be prepared according to the Uniform Requirements for Manuscripts Submitted to BiomedicalJournals, prepared by the International Committee of Medical Journal Editors (ICMJE) and fully available online(http://www.icmje.org). Additional information is available online: http://www.haematologica.org/instructions.htm (instructionsto authors).Additional papers may be considered for the purely online journal (Haematologica on Internet, ISSN 1592-8721). Becausethere are no space constraints online, Haematologica on Internet will publish several items deemed by peer review to be scientificallysound and mainly useful as educational papers. These will include case reports, irreplaceable images, educational materialfrom scientific meetings, meeting abstracts, and letters to the Editor.Galley Proofs and Reprints. Galley proofs should be corrected and returned by email, fax or express delivery within 72 hours.Minor corrections or reasonable additions are permitted; however, excessive alterations will require editorial re-evaluation andwill be possibly charged to the authors. Papers accepted for publication will be printed without cost. The cost of printing colorfigures will be communicated upon request. Preprints may be ordered at cost by returning the appropriate form sent by thePublisher.Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articlesto the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of publishedpapers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercialintent will require written permission and payment of royalties.Haematologica is published in two printed editions: International (worldwide except Spain, Portugal and Latin Americas) andSpanish (in Spain, Portugal and Latin Americas). Detailed information about subscriptions is available online:http://www.haematologica.org/subscribe.htm (subscriptions). While access to the online journal is free, online access to additionalitems of the website http://www.haematologica.org/ will require either institutional or personal subscription.Rates of the International edition for the year 2003 are as following:InstitutionalPersonalPrint edition and full access to the online journal plus additional items of haematologica.org Euro 350 Euro 150Full access to the online journal plus additional items of haematologica.org Euro 350 Euro 75To subscribe to the International edition, please visit our web site http://www.haematologica.org/subscribe.htm or contact:Haematologica Journal Office, Strada Nuova 134, 27100 Pavia, Italy (phone +39.0382.531182, fax +39.0382.27721, E-mailoffice@haematologica.org). To subscribe to the Spanish print edition, please contact: Ediciones Doyma SA, Travesera de Gracia,17-21, 08021 Barcelona, Spain (phone +34.3.4145706, fax +34.3.414-4911, E-mail: doyma@doyma.es).Advertisments. Contact the Advertising Manager, Haematologica Journal Office, Strada Nuova 134, 27100 Pavia, Italy (phone+39.0382.531182, fax +39.0382.27721, E-mail: mikimos@haematologica.org).Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data,opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles oradvertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorialboard and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurateor misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities arepresented accurately, readers are advised that new methods and techniques involving drug usage, and described within thisjournal, should only be followed in conjunction with the drug manufacturer’s own published literature.Associated with USPI, Unione Stampa Periodica Italiana.Premiato per l’alto valore culturale dal Ministero dei Beni Culturali ed Ambientali

haematologicahIV International Workshop onImmune Tolerance in Hemophilia and the Treatmentof Hemophiliacs with an InhibitorAugust 30-September 1, 2001,Bonn-Königswinter, GermanyGuest Editor: Hans-Hermann BrackmannForeword ...................................................................................................................1Molecular GeneticsGenetic determinants of inhibitor formation in patients with hemophilia AAnne Goodeve......................................................................................................22003; vol. 88; supplement no. 12 - September 2003(indexed by Current Contents/Life Sciences and in Faxon Finder and Faxon XPRESS, also available on diskette with abstracts)http://www.haematologica.org/Genetics of inhibitor development in hemophilia BRolf C.R. Ljung....................................................................................................4Immunological AspectsIdentifying B-cell epitopes on factor VIII: what for?Jean-Marie R. Saint-Remy ....................................................................................8Molecular characterization of the immune response to factor VIIIPete Lollar.........................................................................................................11Analysis of factor VIII inhibitors using phage displayJan Voorberg, Wendy S. Bril, Edward N. van den Brink.........................................18General Treatment of Bleeding EpisodesThe Feiba NovoSeven Comparative Study (FENOC)Erik Berntorp, Jan Astermark ..............................................................................26Round Table: Discussion on Immune Tolerance TreatmentAntibody reactivity to factor VIIa may impede the effect of by-passingagents in patients with hemophilia BJan Astermark, Erik Berntorp and the MIBS study group...........................................29Twenty years of experience with low dose immune toleranceEvelien P. Mauser-Bunschoten, H. Marijike van den Berg, Goris Roosendaal ...........34Possible advantages of protocols with a waiting phaseto treat inhibitors against factor VIIIR. Kobelt ...........................................................................................................40High infection rate in ports used for immune tolerance. The Spanish RegistryJ. Tusell, A. Villar, MF. Lopez-Fernandez, S. Haya, D. Brito, R. Sosa,F. Rodriguez-Martorell ......................................................................................42The Malmö International Brother Study (MIBS). An updateJan Astermark, Erik Berntorp..................................................................................45

haematologicahImmunobiology of Tolerance InductionCatalytic antibodies to factor VIIISébastien Lacroix-Desmazes, Michel D. Kazatchkine, Srini Kaveri...........................48Idiotypic control of inhibitorsJean Guy Gilles..................................................................................................52Murine models for the study of factor VIII inhibitorsBirgit M. Reipert, Maria Sasgary, Christina Hausl, Elisabeth Maier,Rafi U. Ahmad, Peter L. Turecek, Hans P. Schwarz ...............................................55T-lymphocytes in the anti-factor VIII immune responseMarc G. Jacquemin, Jean-Marie R. Saint-Remy ....................................................64Prospectives of immonotherapy for inhibitor patientsJean-Marie R. Saint-Remy ..................................................................................66New Aspects in Treatment of Hemophilia B Patients2003; vol. 88; supplement no. 12 - September 2003(indexed by Current Contents/Life Sciences and in Faxon Finder and Faxon XPRESS, also available on diskette with abstracts)http://www.haematologica.org/Immune tolerance induction in hemophilia B. International immune tolerance registryI. Warrier .........................................................................................................69The Malmö immune tolerance experience in hemophilia BErik Berntorp.....................................................................................................71Factor IX mutations and inhibitor development in hemophilia BDavid Lillicrap...................................................................................................75Acquired Inhibitors in Non-hemophiliacsModified Bonn-Malmö Protocol (MBM-P)L. Hess, H. Zeitler, Ch. Unkrig, W. Nettekoven, T. Albert, R. Schwaab,W. Effenberger, J. Oldenburg, H. Vetter, P. Hanfland, H.H. Brackmann .................78Management of severe hemorrhage and inhibitor-elimination in acquired hemophilia:the modified Heidelberg-Malmö protocolA. Huth-Kühne, P. Lages, H. Hampel, R. Zimmermann ........................................86Acquired factor VIII and factor IX inhibitors: a survey of Italian Hemophilia CentersFrancesco Baudo, Gianni Mostarda, Francesco de Cataldo .....................................93Immunological responsiveness of maternal T cells to self antigens during pregnancy:pregnancy as a model to study peripheral T cell tolerance and the role of co-stimulatorymolecules in tolerance inductionMelanie S. Vacchio, Richard J. Hodes .................................................................100Ten years experience with immune tolerance induction therapy in acquired hemophiliaLaszlo Nemes, Ervin Pitlik ................................................................................106Gene TherapyThe host immune response and risk of inhibitor development following adenoviral genetherapy for hemophiliaDavid Lillicrap.................................................................................................111Gene therapy for hemophilia A: immune consequences of viral-vector mediated factorVIII gene transferThierry VandenDriessche, Desire Collen, Marinee K.L. Chuah .............................115Immunologic sequelae and potential for inhibitor development in adeno-associatedviral gene therapy for hemophilia BRoland W. Herzog............................................................................................122Index of authors

eview paperForewordHANS-HERMANN BRACKMANNhaematologica 2003; 88(suppl. n. 12):1http://www.haematologica.org/free/immunotolerance2001.pdfThe 4 th International Workshop on Immune Tolerancewas held in August 2001 in Bonn-Königswinter, Germany and attracted approximately160 participants to share and discuss themost burning and unresolved issues related toinhibitor treatment in hemophilia. To host all participantsin Königswinter was a particular honourbecause the same year the Hemophilia Center inBonn celebrated its 30 th anniversary.Looking back into the history of hemophiliatreatment we have to recognize that it took a verylong time to understand the pathophysiology ofhemophilia and to develop a diagnostic methodsuch as a simple clotting test.With the increasing knowledge in coagulation,treatment could rapidly be developed. Nowadayssafe products are available and treatment scheduleshave been established; however, development ofinhibitors remains a major challenge.We know that genetic defect and factor exposuredays play a role in the development ofinhibitors, we know how to induce immune tolerancein some of our patients using specific treatmentprotocols. Nevertheless many questions stillremain to be answered, in particular as regards ourunderstanding of the mechanism of inhibitordevelopment. We should therefore focus ourefforts on the most important questions, forinstance the question why some patients developinhibitors, why others do not, why immune toleranceis successful in some patients and in somenot. Researchers and hemophilia treaters all overthe world are working hard to find those answers.The agenda of this meeting focused on immunologicaltopics in association with hemophilia andinhibitors in order to provide an understanding ofcurrent concepts and new approaches.The 4th International Workshop was placedunder the same motto that is inscribed at the baseof the Statue of Liberty, welcoming people from allover the world with the promise of freedom, whichwe would also like to give to all our patients: «Giveme your tired, your poor, your huddled masses, yearningto be free».Hans-Hermann BrackmannAcknowledgmentsI am grateful to all speakers and authors whohave contributed to this supplement which providesvaluable information for researchers, healthcareprofessionals and caregivers dedicated tohemophilia issues and hemophilia patients.I wish also to thank all my co-workers who havedone their best to make this workshop successful.Last but not least I would like to thank my friendGuglielmo Mariani for the partnership in thisworkshop and his exceptional friendship.This publication was made possible by an educationalgrant provided by Baxter BioScience.haematologica vol. 88(supplement n. 12):september 2003

[Molecular Genetics]review paperGenetic determinants of inhibitorformation in patients with hemophiliaAhaematologica 2003; 88(suppl. n. 12):2-3http://www.haematologica.org/free/immunotolerance2001.pdfANNE GOODEVEDivision of Genomic Medicine, Royal Hallamshire Hospital,Glossop Road, Sheffield, UKThere are several genetic influences uponinhibitor formation in hemophilia A. Inaddition to HLA type and race these includefactor VIII (FVIII) mutation type and mutationlocation within the FVIII gene.This article will consider two sources of evidencefor the genetic influence on inhibitor formation;the Recombinate PUP mutation study 1and an analysis of the hemophilia A mutationdatabase; HAMSTeRS:(http://europium.csc.mrc.ac.uk/usr/WWW/WebPages/main.dir/main.htm).The first evidence that mutation type determinesrisk of inhibitor formation in hemophiliaA was presented by Schwaab and colleagues in1995. 2 A study of locally recruited patients, plusthose then reported to the international hemophiliaA mutation database showed that largedeletions, FVIII inversions and stop mutationswere each associated with an inhibitor incidenceof around 35%, whilst the incidence in patientshaving missense mutations and short deletionswas only around 5%. The patients included inthe study had been treated with a variety of differentFVIII products on a variety of differenttreatment regimes. Therefore a study was undertakento examine the relationship betweeninhibitor incidence and mutation type in acohort of previously untreated hemophilia Apatients, each having the same severity of hemophiliaA; severe disease, treated with the sameproduct; Recombinate, and each examined forinhibitor development following the same protocol,with the testing being performed at a singlecenter. 1 DNA analysis was performed on 55of the 73 patients enrolled in the study. Eachwas examined for the presence of the FVIII geneinversion mutation using Southern blotting.Where the mutation was absent, patients wereCorrespondence: Dr. Anne Goodeve, Division of Genomic Medicine,Royal Hallamshire Hospital, Glossop Road, Sheffield, S102JF, UK. Phone: international +44.114.2712679. Fax: international+44.114. 2721104. E-mail: a.goodeve@sheffield.ac.ukscreened for mutations elsewhere in their FVIIIgene using conformation sensitive gel electrophoresisfollowed by direct DNA sequencingand mutations were characterized in 51 of the55 patients. As found previously by Schwaab etal., 33% of patients with a FVIII inversion developedan inhibitor and of the other four patientswith a stop codon or large deletion mutation,half developed an inhibitor. The HAMSTeRSdatabase (2001) details 117 patients with stopmutations, and overall 35% of them have developedan inhibitor. The distribution of the stopmutations throughout the FVIII gene identifiedin inhibitor patients is very uneven; only 13% ofpatients with stop mutations in the A1, a1, A2and B domains developed an inhibitor, whilst69% of patients with stop mutations in A3, C1and C2 domains did so.Interestingly, in the Recombinate PUP study,none of the 11 patients with a small deletion orinsertion of one or a few nucleotides developedan inhibitory antibody. Five of these patients hadan insertion or deletion of an A nucleotide intoa run of As. A possible mechanism whereby suchpatients may produce very small quantities ofFVIII mRNA and protein, thus avoiding inhibitorformation, has been presented. 3 These findingswere compared with patient data available onthe HAMSTeRS database. Of patients with aninsertion or deletion of an A nucleotide into arun of As, 3 of 29 (10.3%) had developed aninhibitor, compared with 14 of 64 (21.9%) withinsertions or deletions into other sequences.Development of inhibitors in patients withinsertions or deletions into runs of A nucleotidestherefore appears relatively rare.Virtually all cases of moderate and mild hemophiliaA result from missense mutations and veryfew cases develop inhibitory antibodies. TheHAMSTeRS database details 29/606 patientswith moderate and mild hemophilia A and aninhibitor (4.8%). Again, the distribution of themutations resulting in inhibitors is uneventhroughout the FVIII gene with the majority(72%) of the mutations resulting in inhibitorshaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 3residing in the C1 and C2 domains.These observations show that both the mutationtype and its location within the FVIII geneinfluence the propensity of patients to developinhibitory antibodies. The observations ofSchwaab and colleagues 2 were replicated to anextent by the Recombinate PUP study, althoughinhibitor and mutation data from further suchstudies will help to enhance understanding of themutation-inhibitor relationship. The HAMSTeRSdatabase, although the inhibitor informationavailable is incomplete, also provides a usefulsource of information on previous patients’ experience.Newly diagnosed patients with hemophiliaA can now be analysed for their mutationtype and location within the FVIII gene and thisinformation can be used to predict inhibitor risk.In the future, this may enable treatment alterationin an attempt to reduce inhibitor formation.References1. Goodeve AC, Williams I, Bray GL, Peake IR, for theRecombinate® PUP Study Group Relationship betweenfactor VIII mutation type and inhibitor development ina cohort of previously untreated patients treated withrecombinant factor VIII (Recombinate®). ThrombHaemost 2000; 83:844-8.2. Schwaab R, Brackmann HH, Meyer C, Seehafer J,Kirchgesser M, Haack A, Olek K, Tuddenham EGD, OldenburgJ. Haemophilia A: Mutation type determines riskof inhibitor formation. Thromb Haemost 1995; 74:1402-6.3. Young M, Inaba H, Hoyer LW, Higuchi M, Kazazian HH,Antonarakis SE. Partial correction of a severe moleculardefect in hemophilia A, because of errors during expressionof the factor VIII gene. Am J Hum Genetics 1997;60:565-73.haematologica vol. 88(supplement n. 12):september 2003

[Molecular Genetics]review paperGenetics of inhibitordevelopment in hemophilia Bhaematologica 2003; 88(suppl. n. 12):4-7http://www.haematologica.org/free/immunotolerance2001.pdfROLF C.R. LJUNGDepartments of Pediatrics and Coagulation Disorders, LundUniversity, University Hospital, Malmö, SwedenThe prevalence of inhibitors to FIX in hemophilia Bpatients varies from 1.5% to 23% in the literature,although in most series around 4% of the severecases. The vast majority of inhibitors are of the highrespondingtype and of the IgG4 subclass type. Thenature of the mutation in the FIX gene is an importantfactor in determining whether or not a patientwith hemophilia B will develop an inhibitor. Differencesin the frequency of inhibitors between populationscan be due to differences in the spectrum ofmutations. Up to 30% of patients with with grossdeletions, nonsense and frameshift/stop codonmutations develop inhibitors in contrast to virtuallynone of those with missense mutations. Genetic factorsother than the type of mutation are of importancefor inhibitors as suggested by the concordancebetween brothers with hemophilia.©2003, Ferrata Storti FoundationKey words: hemophilia B, inhibitors, factor IX,epidemiology.Correspondence: Rolf Ljung, Pediatric Clinic, University Hospital,SE-205 02 Malmö, Sweden. Phone: international+46.40.331639 Fax: international +46.40.336226.E-mail: rolf.ljung@pediatrik.mas.lu.seThe prevalence of inhibitors to FIX in hemophiliaB patients varies from 1,5% to 23% inthe literature, 1-5 but most studies show a frequencyof approximately 4% for patients withsevere hemophilia B. This is considerably lowerthan the corresponding figures for hemophilia A.The reason for this is not known. One couldspeculate that factors of importance could be thesize of the gene/protein, the concentration ofthe protein in plasma, the distribution in thebody of the protein etc. Table 1 shows the differencesand similarities in inhibitor developmentbetween hemophilia A and B. Inhibitors inhemophilia B affect all ethnic and racial groups,the type of mutation is known to be of importance,modulators of immune response may be ofimportance and certainly genetic factors otherthan type of mutation and immune response areof importance.The factor IX geneThe FIX gene is located in the distal part of thelong arm of the X-chromosome (Xq27.1).Yoshitake and coworkers have elucidated theentire nucleotide sequence, which is 33.5 kblong, has 8 exons (a-h) and manifests stronghomology with other vitamin K-dependent coagulationfactors. 6 The factor IX protein is a serineprotease synthesized in the liver via a vitamin K-dependent process and occurs in plasma as a singleglycoprotein of 415 amino acids and a molecularweight of 57,000 Daltons. 7 The normalFIX plasma concentration is 5 mg/mL which isconsiderably higher than the normal FVIII plasmaconcentration of 0.1 mg/mL.Hemophilia B is caused by a wide variation ofmutations distributed over the entire FIX gene.Since 1990, an annually updated database hasbeen published of characterised point mutations,small deletions and insertions. 8,9 In the 2001update of the database, a total of 2,421 mutationsare listed, of which approximately a thirdare unique molecular events while the remainderare repeats. 9The factor IX inhibitorInhibitors to FIX in hemophilia B may be of thehigh responding type or low responding type. Ahigh responder manifests a marked increase ininhibitor titer 4-10 days after exposure to FIX,that normally takes 6-12 months to return to itsoriginal level. During this period it is not possi-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 5Table 1. Differences between inhibitors in Hemophilia A andB.Hemophilia AHemophilia Bhigher prevalence (15-52%) lower prevalence (1.5-23%)high and low respondersmainly high responderstype of mutation predisposing factor type of mutation predisposing factorbut less than in hemophilia B more than in hemophilia Ano anaphylactoid reactionsgood response to ITIno nephrotic syndrome during ITIanaphylactoid reactionspoor response to ITInephrotic syndrome during ITIble to saturate the inhibitor by increasing dosesof FIX. A low responding patient manifests onlya slight increase in inhibitor titer after exposureto FIX and it is possible to neutralise the inhibitorby giving higher doses of FIX. By definition thepeak titer is 10 BU for a high responder(some authors use 10 BU). Some patients mayhave very strongly neutralising inhibitors with>1000 BU. Most inhibitors in hemophilia B areof the high responding type, 82% in the survey bySultan, 1 58% in the survey in North America byKatz 5 and all patients in a survey in Japan. 10The inhibitors (allo-antibodies) to Factor IX areof the IgG type and mainly of the subclass IgG4.In the series of Sawamoto (n=10) all sampleswere positive for IgG4, however, three were alsopositive for IgG2 antibodies. 11 In an early work,Örstavik and Miller also found all inhibitors tobe of the IgG4 subclass, although some patientshad a heterogenous nature of the inhibitorsincluding also IgG1 and IgG2 subclasses. 12 TheIgG4 subclass is specific since it does not bindcomplement.The epitopes of the inhibitor have been mappedin a few cases in a few studies and were found tobe amino acid 168 to amino acid 182. 13 It hasalso been shown that synthetic peptides includingthis epitope and sorrounding regions arecapable of neutralising some FIX inhibitors. 13Non-neutralising inhibitors to FIX has also beendiscussed. 14FIX inhibitors and geneticsBefore the genes for FVIII and IX were characterised,it was suggested by several studies thatthere may exist a genetic predisposition forinhibitor development. 15 It was the fact thatinhibitors developed mainly in severe hemophiliaA and B and the observed concordancebetween brother pairs that led to this conclusion.In 1983 Giannelli and coworkers found thatthree out of four hemophilia B patients withinhibitors had gross deletions of the FIX gene. 16In subsequent larger studies it has been clearlyshown that the nature of the mutation in the FIXgene is an important factor in determiningwhether a patient with hemophilia B will developan inhibitor. 17-18 In a recent overview of theInternational database of hemophilia B mutationsby Peter Green, UK at the ISTH congress2001, 39 patients with inhibitors were reportedworldwide in whom the type of mutation wascharacterised. All but 2 had mutations abolishingprotein synthesis. There were only two reportedmissense mutations, Q191K and S365G, inHemophilia B patients with inhibitors. It is reasonableto assume that exogenous FIX are moreimmunogenic if the patient is totally devoid ofthe protein: i.e. carrying a deletion, nonsense,frameshift/stop codon mutation. Most patientswith missense mutations have a circulating, butin many cases, functionally less active protein.There were 76 reported patients with large deletions(Aug. 2001) of whom 20 (26%) had developedinhibitors. However, one has to take intoconsideration that the material in the Internationaldatabase may be biassed since some centersreport all cases and many centers do notreport at all.Table 2 shows details of the patients reportedwho had point mutations and developedinhibitors. As can be seen there is a cluster ofinhibitors for the mutation C 6460 T,Arg29→stop although this a frequent mutationper se. Furthermore, 4 of the other point mutationsare in the same part of the gene, i.e. nc.6383-6402.The International database is not representativefor a certain population and may be biassed dueto the mode of inclusion of patients (random ona volontary basis from different centers/countries).The most representative population basedfigures on record are from UK and Sweden.In UK the mutation has been characterised in421 families which represent 75% of the totalnumber of families (P Green, personal communicationat the ISTH Congress 2001). Only 8patients developed inhibitor of whom 3 had totaldeletions, 1 partial deletion, 2 frame-shift stopcodon and 2 Arg 29→stop. Thus the frequency ofinhibitors in hemophilia B in UK is very low.In Sweden the mutations have been characterisedin all 77 families with hemophilia B. 19 Tenof the mutations recurred in 1-6 other families.Haplotype analysis by study of polymorphismsrevealed that at least 65 families had uniquemutation, i.e. almost all families had their ownunique mutation. In this population based study,11/48 (23%) patients with severe hemophilia Bdeveloped inhibitors, and all of them had deletionsor nonsense mutations. If one looked at thepatients with severe hemophilia B due to deletion/nonsensemutation, 11/37 (30%) developedinhibitors, as compared to 0/11 of thepatients 19 with missense mutations. One mayhaematologica vol. 88(supplement n. 12):september 2003

6R.C.R. LjungTable 2. Int database - point mutations in patients withinhibitor (n=22) (Aug. 2001).Number (n) Nucleotide change(s) Amino acid change7 C 6460 T 29 R → stop2 C 20551 A, C 20551 T 191 Q → K (missense)2 C 30863 T 2481 TA 31213 CG 365 S → G (missense)10 92, 6383, 6392, 6401, −−−6402, 6680, 20561,30820, 30821, 30950speculate on whether the high incidence ofinhibitors in hemophilia B in Sweden comparedto other studies on record, is accidental, due tothe limited number of cases, or is real and due to,for example, the choice of concentrate or mode ofdelivery of concentrate. One important factor isprobably the fact that the frequency of large deletions(>30 bp) seems to be higher in Sweden (8%of the families, but 13% of the total number ofpatients with severe hemophilia B) than elsewhere(2-5% of reported cases) 17, 20, 21 bearing inmind that we do not have reliable data on the frequencyof large deletions in the Internationaldatabase since they have only been included duringthe last years. As mentioned, in the Britishseries, large deletions were only found in 7/412families (1.7%). 22Prophylactic treatment, started at an early ageand used since the beginning of the ‘70s inhemophilia B in Sweden, does not result in ahigher frequency of inhibitors. 23 Furthermore,the age distribution of the inhibitor patients doesnot indicate a higher frequency since high purityconcentrates became available. 24 It is reasonableto assume population specific differences inthe frequency of various types of mutationswhich may have clinical implications for thatpopulation. A similar finding was recently reportedin a small series of sporadic cases in Mexico. 25On the other hand in our series one family witha total deletion had three affected members ofwhom none developed inhibitor. 26 The questionwhy these patients have not developed inhibitorscan only be answered through the understandingthat the etiology of inhibitor formation is multifactorial.Genetic factors other than the type of mutationAlthough it is obvious that the type of mutationis an important genetic predisposing factor,other genetic factors must be involved. A reasonablefactor are the modulators of immuneresponse and markers in the major histocompatibilitycomplex have been investigated in severalstudies 27 in hemophilia A. In a study of 176patients, Hay and coworkers found an increasedfrequency of HLA class II antigen in hemophiliaA inhibitor patients, but only HLA DQA1*0102reached statistical significance. 28 This finding wasnot supported in other studies. 29, 30 There are, tomy knowledge, no studies on record studyingexclusively HLA and inhibitors in hemophilia B.However, unique features of the immune systemin hemophilia B patients have been found sincemany patients with inhibitors have developedanaphylaxis or severe allergic reactions to FIX andmany of these patients have total deletions.Another unexplained feature that differs hemophiliaB from hemophilia A is the poor responseto immune tolerance induction - these topics arebeyond the scope of this review.Genetic factors other than expressed by thetype of mutation and the immune response, maybe found in studies like the Malmö InternationalBrother Study (MIBS), which was started in1996. The registry includes siblings with hemophiliain Europe and North America, with anwithout a history of inhibitors. Data has beenpublished comprising 460 brother pairs of whom72 had hemophilia B. 31 It was found in the wholematerial that the risk for a sibling to develop aninhibitor was three-fold higher in families witha previous known inhibitor. However, the specificsituation in hemophilia B can not be elucidatedsince only 3/29 with severe hemophilia Bin the series had developed inhibitor. As pointedout in the study it is clear that genetic predispositionexists but the mechanisms are not yetknown.To sum up, inhibitors are less common inhemophilia B than hemophilia A although differencesdo exist between different populations.The nature of the mutation is an important predisposingfactor for inhibitor development butother still unknown genetic factors contribute.The type of mutation is of importance for predisposingof anaphylactic reactions after FIXinfusion, but does not seem to be of importancefor type of inhibitor (high/low) or the outcomeof immune tolerance induction.FundingResults presented in this review was supported bygrants from the Swedish Medical Research Council(no. 13493), the University of Lund and funds ofMalmö University Hospital.References1. Sultan Y. Prevalence of inhibitors in a population of3435 hemophilia patients in France. French HemophiliaStudy Group. Thromb Haemost 1992;67:600-2.2. Shapiro S. Antibodies to blood coagulation factors. ClinHaematol 1979;8:207-14.3. Ljung R, Petrini P, Lindgren AC, Tengborn LNilsson IM.Factor VIII and factor IX inhibitors in hemophiliacs.Lancet 1992;339:1550.4. Lusher J. Inhibitors in young boys with haemophilia B.Ballieres Clin Haematol 2000;13:457-68.5. Katz J. Prevalence of Factor IX inhibitors among patientshaematologica vol. 88(supplement n. 12):september 2003

12P. Lollaragainst a limited number of immunodominantepitopes. For example, the immune response tothe Mr 76,000 influenza A hemagglutinin heterodimeris directed against four immunodominantB-cell epitopes 5 that cover only a fraction ofthe protein surface.The mechanisms that produce this restrictionof the immune response are not known. Oneview is that T-cells orchestrate the process andthe structure of the antigen itself is relativelyunimportant. However, it is possible that thepotential antibody repertoire is biased to recognizeendemic pathogens and to avoid self-reactivespecificities. 6 When there is a loss of tolerance,intrinsic structural properties of the antigen mayinfluence immunodominance. For example, theautoantibody response to cytochrome c found insystemic lupus erythematosus (SLE) and otherdisorders is directed to a limited number of siteswhich are similar to those recognized by mice thatare immunized with human cytochrome C. 7Analysis of the polyclonal response within asingle immunodominant epitope reveals considerableheterogeneity when the component monoclonalantibodies are studied. An exhaustiveanalysis of the number of anti-hemagglutininmurine B-cell hybridomas produced in responseto infection with an influenza A strain producedan estimate of 1,500 different antibodies in therepertoire directed against the four immunodominantepitopes. 8 Thus, an immunodominantepitope could be viewed as an area underneathan antibody footprint, or set of overlappingfootprints, in which there is considerable variationat the clonal level in the atomic contacts thatdetermine the strength of the interaction.Structural analysis of antigen-antibody complexesby X-ray crystallography has demonstratedthat typically 20-25 amino acid residues of eachcomponent are buried. 9 However, most of thestructural epitope does not contribute significantlyto the binding energy because most of theresidues in contact with antibody can be mutatedwithout loss of binding energy. 10 However,mutation of one of a few key residues typicallyresults in a major decrease in affinity. The set ofthese residues has been defined as the functionalepitope. 2,10FVIII structure and functionWith this background, we can discuss factorVIII (FVIII) as a potential target of the immunesystem. FVIII is a large, complex plasma glycoprotein.It contains 2,332 amino acids and ispost-translationally glycosylated and sulfated. Themolecular weight of the mature protein is approximately300,000. FVIII circulates bound noncovalentlyto von Willebrand factor (vWf). vWfis a huge, multimeric protein that weights severalmillion Daltons and has a hydrodynamicradius in the submicron range. 11 Thus, the FVIIIvWfcomplex may look something like a bacteriumor virus to the immune system. Disruption ofthis interaction, for example in severe von Willebrand’sdisease or naturally occurring vWf mutations,results in rapid clearance of FVIII and a secondarydeficiency of FVIII.FVIII contains a sequence of domains designatedA1-A2-B-ap-A3-C1-C2, where ap is an activationpeptide. FVIII is cleaved intracellularly withinthe B domain by a protease known asPACE/furin. 12 This results in fragmentation ofthe B domain and production of a heterodimer,which still requires proteolytic activation in thecoagulation cascade. During the activation ofFVIII by thrombin, cleavages occur between A1and A2 at Arg 372 , between A2 and B at Arg 740 , andbetween ap and A3 at Arg 1689 . Consequently, theB domain fragments and the acidic 41-residue appeptide are released, producing an A1/A2/A3-C1-C2 activated FVIII (FVIIIa) heterotrimer. 13 FVIIIalso can be activated by factor Xa, 14 which producesa more complex cleavage pattern. 15The only known function of FVIII is to participate,in activated form, as a cofactor for factor IXaduring intrinsic pathway factor X activation. Theenzymatic complex consists of factor IXa, FVIIIaand factor X on a phospholipid membrane surface.FVIII is assayed in coagulation or chromogenicassays under conditions in which it islimiting relative to factor IXa, factor X and phospholipid.Inhibitory antibodies are most commonlymeasured using the Bethesda assay 16 or theNijmegen modification of this assay. 17 Theseassays measure the loss of FVIII activity in a coagulationassay and return a single number — thedilution of antibody sample that produces 50%inhibition — as a global measurement of inhibitoryantibody activity.FVIII inhibitorsThe FVIII molecule represents the most commonlytargeted plasma protein for the developmentof inhibitory antibodies. Hemophilia A isthe most common severe hereditary bleeding disorderand produces a clinical setting in whichpatients are infused with a potentially immunogenicprotein. Most patients with severe hemophiliahave gene deletions, insertions, inversionsor nonsense mutations, all of which could possiblylead to complete lack of circulating FVIII. Inthis setting FVIII would represent a foreign proteinand would be fair game for the immune system.Alloantibodies arise in approximately 25%of patients with hemophilia A 18 and it is perhapssurprising that the incidence is not higher.Patients with severe hemophilia usually do nothave detectable levels of FVIII antigen. However,it is possible that in some of these patients, thereis a small amount of antigen that is sufficient toproduce tolerance.It is not possible to predict whether a previouslyuntreated patient with hemophilia A will developan inhibitor. However, patients with missensemutations and circulating levels of FVIII antigenare less likely to develop inhibitors, either becauseof a state of immune tolerance or because themissense mutation does not produce severehaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 13hemophilia and the frequency of FVIII infusionsis less common. Occasionally, patients with missensemutations develop inhibitors in responseto increased FVIII usage.Additionally, FVIII is the most likely coagulationprotein to be targeted in autoimmune reactions,producing a condition termed acquiredhemophilia or acquired hemophilia A. The incidenceof this condition is very low, being approximately1 in 5 million people. It usually manifestsas a serious bleeding disorder, which anecdotally,and for unknown reasons, is more severethan when inhibitory antibodies develop inpatients with hemophilia A. FVIII autoantibodiesdevelop in a variety of clinical settings, includingthe postpartum period, systemic lupus erythematosus,and chronic lymphocytic leukemia. Inapproximately 50% of the cases, acquired hemophiliais idiopathic. 19In both hemophilia A and acquired hemophilia,FVIII inhibitors are polyclonal IgG populationsthat are usually directed toward more than oneepitope. For unknown reasons, the IgG4 subclassis more common in plasma of inhibitor patientsthan in the normal population. Inhibitors areclassified into two types. 20,21 Type I inhibitorsinactivate FVIII completely with second-orderkinetics, which would be expected for a simplebimolecular antigen-antibody reaction. Type IIinhibitors inactivate FVIII incompletely and displaymore complex kinetics of inhibition. Mostautoantibody patients have type II inhibitors. Inmost cases, type II inhibitors behave like type Iinhibitors when they are tested against FVIII inthe absence of vWf, indicating that inhibition bythe type II antibodies is blocked due to competitionfor binding by vWf. 22FVIII domains recognized by inhibitoryantibodiesA comprehensive analysis of the antibodyresponse to FVIII in either human disease or inanimal models does not yet exist. A hierarchicalanalysis would answer whether: 1) the humoralimmune response is restricted to only some of theFVIII domains; 2) antibody binding to epitopesin different domains is independent or co-operative;3) there are single or multiple immunodominantepitopes within a domain; 4) andwhether there is variation in the amino acidswithin a single FVIII epitope (corresponding toan antibody footprint) that contribute significantlyto antibody binding energy.SDS-polyacrylamide gel electrophoretic analysisof FVIII reveals bands corresponding to FVIIIheavy chain (A1-A2-B) and light chain (ap-A3-C1-C2). After exposure to thrombin, the threeheterotrimer bands, A1, A2, and A3-C1-C2, areapparent. Initial epitope mapping studies byWestern blotting of purified FVIII revealed thatmost patients’ antibodies bound to the A2 andA3-C1-C2 domains. 23 The relative lack of anti-A1 reactivity was an early indicator that the B-cell response to FVIII was restricted. Subsequently,the C2 domain was identified as the mostcommon target for anti-light chain antibodies byWestern blotting of defined domain fragments ofrecombinant FVIII produced in E. coli. 24,25 Anti-A3antibodies were also detected, although less frequently.Some inhibitory antibodies do not binddenatured FVIII in Western blotting experimentsand can be detected by immunoprecipitationusing soluble FVIII fragments. 26An important complementary method to Westernblotting analysis is antibody neutralization, inwhich recombinant or plasma-derived fragmentsof FVIII are used to block inhibitory antibodies.25,27,28 The sum of neutralization of inhibitoryactivity by the A2 domain and FVIII light chain(ap-A3-C1-C2) fragment approaches 100% inmost cases, 28 indicating that antibodies to the A1domain do not contribute significantly to mostinhibitor titers. By comparing antibody neutralizationby the ap-A3-C1-C2 fragment to that bythe C2 fragment, the relative contribution of anti-C2 antibodies can be assessed. Most anti-lightchain activity is against the C2 domain, althoughin some plasmas there is significant activityagainst ap-A3-C1. The dominant inhibitors inmost autoantibody plasmas are directed onlyagainst C2 or A2, but not both. In contrast, mosthemophilia A inhibitor plasmas recognize multipleepitopes. However, the similarity betweenalloantibodies and autoantibodies is perhaps morestriking than the difference, because the A2 andC2 domains are immunodominant in both settings,despite the different immunologic backgroundsfrom which they arise.Mechanism of action of FVIII inhibitorsFVIIIa binds factor IXa, 29 phospholipid 30 andfactor X 31 in the intrinsic FXase complex. Thus,disruptions of any of these interactions by anantibody could be inhibitory and all these possibilitieshave been observed. Additionally,inhibitors could function by binding to sites thatare recognized by thrombin or factor Xa. However,this mechanism of action has not been convincinglydemonstrated.Anti-A2 inhibitors inhibit intrinsic FXase noncompetitively,i.e., most likely by blocking thebinding of FVIIIa to factor X. 31 However, inhibitionof binding of FVIIIa to factor IXa by anti-A2inhibitors also has been observed. 32 Many, if notall, anti-C2 antibodies inhibit the binding of FVIIIto phospholipid. 33 However, the C2 domain alsobinds vWf, 34,35 at or near the phospholipid bindingsite. Thus, vWF could interfere with the bindingof anti-C2 antibodies, producing the type IIantibody pattern described above. Consistentwith this, polyclonal patients’ antibodies that recognizeonly the C2 domain are much more commonin autoantibody patients, 28 in whom mosttype II antibodies are found. 20,22Thus, another possible mechanism of action ofinhibitors could be to block the binding of FVIIIto vWf. In contrast to disruption of the intrinsicFXase complex or inhibition of proteolytic acti-haematologica vol. 88(supplement n. 12):september 2003

14P. Lollarvation, inhibition of vWf binding would not bedetected by the Bethesda assay. Inhibitors havebeen described that block the binding of FVIII tovWf, and this may be a common property of anti-C2 inhibitors. 36 However, the relative contributionto the bleeding diathesis of inhibition ofbinding of FVIII to vWf versus binding to phospholipidis unknown.High resolution mapping of FVIII inhibitorepitopesHomolog scanning mutagenesis using hybridhuman/porcine, human/canine, or human/murineFVIII molecules has been successful inobtaining higher resolution maps of the B-cellepitopes. 37,38 This strategy is based on the observationthat inhibitors usually have limited crossreactivitywith non-human FVIII. A hybrid FVIIImolecule that is less reactive with an inhibitor tohuman FVIII localizes the inhibitor to the regionin which the non-human substitution has beenmade. Because the hybrids are active, reduction inantigenicity is unlikely to be due to improper foldingof the protein. In contrast, epitope localizationby deletion mapping is complicated by thepossibility that loss of antigenicity can occur bydenaturation of the polypeptide chain.Using this method, the A2 epitope recognizedby most A2-specific antibodies has been localizedto residues Arg 484 -Ile 508 . Analysis of several hybridscontaining A2 substitutions indicates that the A2epitope appears to be confined with the Arg 484 -Ile 508 linear segment. This is consistent with thefact that this segment protrudes as a large loopbetween β-strands in a homology model of the A2domain based on the X-ray structure of ceruloplasmin.The A2 epitope has been mapped further by alanine-scanningmutagenesis. Human A2-specificpolyclonal antibodies recognize several aminoacids in this segment, including Tyr 487 and to alesser extent Ser 488 , Arg 489 , Pro 492 , Val 495 , Phe 501 ,and Ile 508 . 39 There is variability between the antibodies,indicating that although the antibodyfootprint may be relatively invariant, individualamino acid residues recognized by the antibodiesdiffer.The region recognized by several C2-specificinhibitors has been mapped using a panel ofhybrid human/porcine FVIII molecules to residuesGlu 2181 -Val 2243 . 37 Recently, an X-ray structure ofthe human FVIII C2 domain has been solved, providinga more detailed approach to the analysis ofC2 inhibitors. The structure reveals a putativehydrophobic three-prong phospholipid membrane-bindingsite consisting of Met 2199 /Phe 2200 ,Val 2223 , and Leu 2251 /Leu 2252 . 40 Additionally, severalbasic residues contribute to a ring of positivelycharged residues that may contribute to an electrostaticinteraction of FVIII with negativelycharged phosphatidylserine. Several active humanFVIII mutants were constructed based on differencesbetween human, porcine, murine, andcanine FVIII at proposed phospholipid-bindingsites. The antigenicity of the mutants toward C2-specific polyclonal human antibodies, a humanmonoclonal anti-C2 antibody, BO2C11, 41 and amurine C2-specific monoclonal antibody, NMCVIII-5, 42 was measured. The results suggest thatC2 inhibitors frequently target the Met 2199 /Phe 2200and Leu 2251 /Leu 2252 β-hairpins, which is consistentwith the hypothesis that these residues participatein binding to phospholipid membranes. Incontrast, mutations at Val 2223 increased antibodybinding, indicating that Val 2223 may oppose antibodybinding sterically or through stabilization ofa low-affinity membrane-binding conformationof the C2 domain. Subsequently, the X-ray structureof BO2C11 in complex with the FVIII C2domain was solved and revealed contact with theMet 2199 /Phe 2200 and Leu 2251 /Leu 2252 . 43Delineation of the boundaries of an epitopeusing hybrid FVIII molecules can be limited bysequence identity. For example, human andporcine FVIII are identical from residues 2243(the end of the sequence mapped using human/porcinehybrids) through 2252. Thus, identificationof the epitope extending to residue 2252was not possible using hybrid human/porcineFVIII.Subsequently, we identified four C2-specificplasmas that do not appear to recognize the phospholipid-bindingregion (Barrow RT and Lollar P,personal observations). The functional propertiesof these antibodies have not been characterized.Overall, the analysis of C2 inhibitors appearsconsiderably more complex than that of A2inhibitors both with respect to epitope structureand mechanism of inhibition.Regions outside the A2 and C2 domains appearto be minor targets for inhibitory antibodiesbecause the inhibitory activity in most plasmascan be neutralized by the A2 and C2 domains. 28However, three inhibitor IgGs have been identifiedthat are neutralized by a synthetic peptidecorresponding to residues 1804-1819 in the A3domain. 44 This region encompasses a regiondefined by residues 1811-1819 that constitutespart of the factor IXa binding site. 45 Additionally,several inhibitor plasmas have been identified inwhich substitution of the human ap region withthe corresponding porcine segment results in areduction of antigenicity. This indicates that theap segment also contains an inhibitory epitope. 46Unusual FVIII inhibitorsInhibitory antibodies have been described thatrecognize epitopes outside the A2, C2, A3 and apregions. A human monoclonal antibody, LE2E9,is an IgG4k antibody that it is directed against theC1 domain and was derived from a patient withmild hemophilia A who has an Arg 2150 His mutation.47 The antibody recognizes wild-type FVIIIbut not endogenous FVIII, indicating that theantibody recognizes Arg 2150 . The antibody is a type2 inhibitor and blocks FVIII binding to vWf. Similarly,an inhibitor from a patient with mildhemophilia A due to an Arg 593 Cys mutation inhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 15the A2 domain has been described that recognizeswild-type FVIII but not the endogenous FVIII A2domain. 48An inhibitor with specificity confined to the A1domain was identified in a severe hemophilia Apatient who had a deletion of the A1 domain, butnevertheless produced detectable circulating FVIIIantigen. 49 Three type II inhibitory autoantibodieshave been identified that recognize the A3domain, and not the C2 domain, which typicallyis the target of type II antibodies. 50 These antibodiescirculate as immune complexes with FVIII,which has not been previously described.It is unusual for previously treated patients withhemophilia A to develop an inhibitor if they havenot developed one within the first year of treatment.18,51 Outbreaks of C2-specific inhibitorshave been described in previously treated hemophiliaA patients receiving specific lots of plasmaderivedFVIII. 52,53 The C2-specificity is remarkablebecause it is unusual in the hemophilia A population.The results suggest that exposure to denaturedforms of FVIII or generation of a neoepitopemay be immunogenic.The immunogenicity of FVIII in murinehemophilia AThe analysis of FVIII inhibitors is man is limitedby the inaccessibility of lymphoid tissue andthe difficulty associated with cloning antigen-specificB-cells. The polyclonal nature of the inhibitorantibody population makes functional and structuralanalysis of B-cell epitopes difficult. Only twohuman monoclonal antibodies have been produced(vide supra), in contrast to the potentialease of generation of murine monoclonal antibodies.Thus, the analysis of the immune responseto FVIII in mice holds considerable promise as amodel. There are two strains of FVIII knockoutmice, 54 which were produced by targeted disruptionof the FVIII gene at exons 16 and 17 encodingthe A3 domain. The E16 and E17 mice haveidentical phenotypes in studies to date. The micehave no measurable FVIII and have a bleedingdiathesis. However, they synthesize detectableamounts of FVIII antigen due to the fact that theFVIII gene is disrupted near to the middle of thesequence encoding protein. 55 In contrast, secretionof potentially tolerizing FVIII antigen may bethe exception, rather than the rule, in severehuman hemophilia A. Human hemophilia A isextremely heterogeneous at the genetic level,which probably influences the variability in theimmune response to FVIII. There could be similardifferences in the immune responses in hemophiliaA mice if they are produced by lesions in differentregions of the FVIII gene.When E16 or E17 is infused intravenously withhuman FVIII using a dosage schedule and dosageon a body weight basis that mimics treatment ofsevere hemophilia A in man, almost all develophigh titer inhibitors after two or three doses. 56These mice have a T-cell proliferative response toFVIII, demonstrating that the immune response isT-cell dependent. 56,57 A T-cell proliferativeresponse has also been demonstrated in humanhemophilia A. 58 The murine antibody response isprimarily IgG1-mediated, and to a lesser extentIgG2-mediated. 57 In contrast, when hemophilia Amice are crossed with B7.2 knockout mice, theydo not develop anti-FVIII antibodies or produce aT-cell proliferative response. 59 The fusion proteinCTLA-Ig, which blocks the B7-CD28 interaction,prevents the primary response to FVIII. 59 Theseresults indicate that the B7-CD28 co-stimulatorypathway is necessary for the immune responseto FVIII and suggest a possible therapeuticapproach. Anti-CD40L antibodies also block theprimary response to FVIII in this model. 60,61The murine hemophilia A model has beenadopted widely for development of somatic cellgene therapy of hemophilia A, in which immunogenicityof FVIII 55,62,63 and potential induction oftolerance 64 are major areas of interest.FundingSupported by grants from the National Institutesof Health, R01-HL40921.References1. Laver WG, Air GM, Webster RG, Smith-Gill SJ. Epitopeson protein antigens: misconceptions and realities. Cell1990;61:553-6.2. Jin L, Fendly BM, Wells JA. High resolution functionalanalysis of antibody-antigen interactions. J Mol Biol1992;226:851-65.3. Watts C, Lanzavecchia A. Suppressive effect of antibodyon processing of T cell epitopes. J Exp Med 1993; 178:1459-63.4. Benjamin DC, Berzofsky JA, East IJ, Gurd FR, HannumC, Leach SJ, et al. The antigenic structure of proteins: areappraisal. Ann Rev Immunol 1984;2:67-101.5. Caton AJ, Brownlee GG, Yewdell JW, Gerhard W. Theantigenic structure of the influenza virus A/PR/8/34hemagglutinin (H1 subtype). Cell 1982;31:417-27.6. Diaz M, Klinman NR. Relative roles of somatic and Darwinianevolution in shaping the antibody response.Immunol Res 2000;21:89-102.7. Mamula MJ, Jemmerson R, Hardin JA. The specificity ofhuman anti-cytochrome c autoantibodies that arise inautoimmune disease. J Immunol 1990; 144:1835-40.8. Staudt LM, Gerhard W. Generation of antibody diversityin the immune response of BALB/c mice to influenzavirus hemagglutinin. I. Significant variation in repertoireexpression between individual mice. J Exp Med 1983;157:687-704.9. Davies DR, Padlan EA. Antibody-antigen complexes.Ann Rev Biochem 1990;59:439-73.10. Cunningham BC, Jhurani P, Ng P, Wells JA. Receptorand antibody epitopes in human growth hormone identifiedby homolog-scanning mutagenesis. Science 1989;243:1330-6.11. Slayter H, Loscalzo J, Bockenstedt P, Handin RI. Nativeconformation of human von Willebrand protein. Analysisby electron microscopy and quasi-elastic light scattering.J Biol Chem 1985;260:8559-63.12. Kaufman RJ, Pipe SW, Tagliavacca L, Swaroop M, MoussalliM. Biosynthesis, assembly and secretion of coagulationfactor VIII. Blood Coagul Fibrinol 1997;8 Suppl2:S3-14.13. Lollar P, Parker CG. Subunit structure of thrombin-activatedporcine factor VIII. Biochemistry 1989;28:666-74.14. Vehar GA, Davie EW. Preparation and properties ofbovine factor VIII (antihemophilic factor). Biochemistry1980;19:401-10.15. Parker ET, Pohl J, Blackburn MN, Lollar P. Subunit struc-haematologica vol. 88(supplement n. 12):september 2003

16P. Lollarture and function of porcine factor Xa-activated factorVIII. Biochemistry 1997;36:9365-73.16. Kasper CK, Aledort LM, Counts RB, Edson JR, FrantatoniJ, Gree D, et al. A more uniform measurement of factorVIII inhibitors. Thromb Diathes Haemorrh 1975; 34:869-72.17. Verbruggen B, Novakova I, Wessels H, Boezeman J, vanden Berg M, Mauser-Bunschoten E. The Nijmegen modificationof the Bethesda assay for factor VIII:Cinhibitors: improved specificity and reliability. ThrombHaemost 1995;73:247-51.18. Lusher JM, Arkin S, Abildgaard CF, Schwartz RS. Recombinantfactor VIII for the treatment of previouslyuntreated patients with hemophilia A. Safety, efficacy,and development of inhibitors. Kogenate PreviouslyUntreated Patient Study Group. N Engl J Med 1993; 328:453-9.19. Kessler CM, Ludlam CA. The treatment of acquired factorVIII inhibitors: worldwide experience with porcinefactor VIII concentrate. International Acquired HemophiliaStudy Group. Semin Hematol 1993;30:22-7.20. Biggs R, Austen DE, Denson KW, Borrett R, Rizza CR.The mode of action of antibodies which destroy factorVIII. II. Antibodies which give complex concentrationgraphs. Br J Haematol 1972;23:137-55.21. Biggs R, Austen DE, Denson KW, Rizza CE, Borrett R. Themode of action of antibodies which destroy factor VIII.I. Antibodies which have second-order concentrationgraphs. Br J Haematol 1972;23:125-35.22. Gawryl MS, Hoyer LW. Inactivation of factor VIII coagulantactivity by two different types of human antibodies.Blood 1982;60:1103-9.23. Fulcher CA, Mahoney SD, Roberts JR, Kasper CK, ZimmermanTS. Localization of human factor FVIII inhibitorepitopes to two polypeptide fragments. Proc Natl AcadSci USA 1985;82:7728-32.24. Scandella D, Mahoney SD, Mattingly M, Roeder D, TimmonsL, Fulcher CA. Epitope mapping of human factorVIII inhibitor antibodies by deletion analysis of factorVIII fragments expressed in Escherichia Coli. Proc NatlAcad Sci USA 1988;85:6152-6.25. Scandella D, Mattingly M, de Graaf S, Fulcher CA. Localizationof epitopes for human factor VIII inhibitor antibodiesby immunoblotting and antibody neutralization.Blood 1989;74:1618-26.26. Scandella D, Timmons L, Mattingly M, Trabold N, HoyerLW. A soluble recombinant factor VIII fragment containingthe A2 domain binds to some human anti-factorVIII antibodies that are not detected by immunoblotting.Thromb Haemost 1992;67:665-71.27. Scandella D, Mattingly M, Prescott R. A recombinantfactor VIII A2 domain polypeptide quantitatively neutralizeshuman inhibitor antibodies which bind to A2.Blood 1993;82:1767-75.28. Prescott R, Nakai H, Saenko EL, Scharrer I, Nilsson IM,Humphries JE. The inhibitor antibody response is morecomplex in hemophilia A patients than in most nonhemophiliacswith factor VIII autoantibodies. Recombinateand Kogenate Study Groups. Blood 1997;89:3663-71.29. Duffy EJ, Parker ET, Mutucumarana VP, Johnson AE,Lollar P. Binding of factor VIIIa and factor VIII to factorIXa on phospholipid vesicles. J Biol Chem 1992; 267:17006-11.30. Bloom JW. The interaction of rDNA factor VIII, factorVIIIses-797-1562, and factor VIIIses-797-1562-derivedpeptides with phospholipid. Thromb Res 1987;48:439-48.31. Lollar P, Parke R, Curtis JE, Helgerson SL, Hoyer LW,Scott ME, Scandella D. Inhibition of human factor VII-Ia by anti-A2 subunit antibodies. J Clin Invest 1994; 93:2497-504.32. Fay PJ, Scandella D. Human inhibitor antibodies specificfor the factor VIII A2 domain disrupt the interactionbetween the subunit and factor IXa. J Biol Chem 1999;274:29826-30.33. Arai M, Scandella D, Hoyer LW. Molecular basis of factorVIII inhibition by human antibodies. Antibodies thatbind to the factor VIII light chain prevent the interactionof factor VIII with phospholipid. J Clin Invest 1989;83:1978-84.34. Saenko EL, Scandella D. The acidic region of the factorVIII light chain and the C2 domain together form thehigh affinity binding site for von Willebrand factor. JBiol Chem 1997;272:18007-14.35. Saenko EL, Shima M, Rajalakshmi KJ, Scandella D. Arole for the C2 domain of factor binding to von Willebrandfactor. J Biol Chem 1994;269:11601-5.36. Shima M, Nakai H, Scandella D, Tanaka I, Sawamoto Y,Kamisue S, et al. Common inhibitory effects of humananti-C2 domain inhibitor alloantibodies on factor VIIIbinding to von Willebrand factor. Br J Haematol 1995;91:714-21.37. Healey JF, Barrow RT, Tamim HM, Lubin IM, Shima M,Scandella D, et al. Residues Glu2181-Val2243 contain amajor determinant of the inhibitory epitope in the C2domain of human factor VIII. Blood 1998;92:3701-9.38. Healey JF, Lubin IM, Nakai H, Saenko EL, Hoyer LW,Scandella D, et al. Residues 484-508 contain a majordeterminant of the inhibitory epitope in the A2 domainof human factor VIII. J Biol Chem 1995;270:14505-9.39. Lubin IM, Healey JF, Scandella D, Lollar P. Analysis ofthe human factor VIII A2 inhibitor epitope by alaninescanning mutagenesis. J Biol Chem 1997;272:30191-5.40. Pratt KP, Shen BW, Takeshima K, Davie V, Fujikawa K,Stoddard BL. Structure of the C2 domain of human factorVIII at 1.5 A resolution. Nature 1999; 402:439-42.41. Jacquemin MG, Desqueper BG, Benhida A, Vander ElstL, Hoylaerts MF, Bakkus M, et al. Mechanism and kineticsof factor VIII inactivation: study with an IgG4 monoclonalantibody derived from a hemophilia A patientwith inhibitor. Blood 1998;92:496-506.42. Shima M, Scandella D, Yoshioka A, Nakai H, Tanaka I,Kamisue S, et al. A factor VIII neutralizing monoclonalantibody and a human inhibitor alloantibody recognizingepitopes in the C2 domain inhibit factor VIII bindingto von Willebrand factor and to phosphatidylserine.Thromb Haemost 1993;69:240-6.43. Spiegel PC Jr, Jacquemin M, Saint-Remy JM, StoddardBL, Pratt KP. Structure of the factor VIII C2 domainimmunoglobulinG4κ Fab complex: identification of aninhibitory antibody epitope on the surface of factor VIII.Blood 2001;98:13-9.44. Zhong D, Saenko EL, Shima M, Felch M, Scandella D.Some human inhibitor antibodies interfere with factorVIII binding to factor IX. Blood 1998;92:136-42.45. Lenting PJ, van de Loo JW, Donath MJ, van Mourik JA,Mertens K. The sequence Glu1811-Lys1818 of humanblood coagulation factor VIII comprises a binding sitefor activated factor IX. J Biol Chem 1996;271:1935-40.46. Barrow RT, Healey JF, Gailani D, Scandella D, Lollar P.Reduction of the antigenicity of factor VIII toward complexinhibitory plasmas using multiply-substitutedhybrid human/porcine factor VIII molecules. Blood2000; 95:557-61.47. Jacquemin M, Benhida A, Peerlinck K, Desqueper B,Vander EL, Lavend'homme R, et al. A human antibodydirected to the factor VIII C1 domain inhibits factor VIIIcofactor activity and binding to von Willebrand factor.Blood 2000;95:156-63.48. van den Brink EN, Timmermans SM, Turenhout EA,Bank CM, Fijnvandraat K, Voorberg J, et al. Longitudinalanalysis of factor VIII inhibitors in a previouslyuntreated mild haemophilia A patient with anArg593→Cys substitution. Thromb Haemost 1999; 81:723-6.49. Shibata M, Shima M, Morichika S, McVey J, TuddenhamEG, I. Tanaka H, et al. An alloantibody recognizingthe FVIII A1 domain in a patient with CRM reducedhaemophilia A due to deletion of a large portion of theA1 domain DNA sequence. Thromb Haemost 2000; 84:442-8.50. Nogami K, Shima M, Giddings JC, Hosokawa K, NagataM, Kamisue S, et al. Circulating factor VIII immunecomplexes in patients with type 2 acquired hemophiliaA and protection from activated protein C-mediated proteolysis.Blood 2001;97:669-77.51. McMillan CW, Shapiro SS, Whitehurst D, Hoyer LW, RaoAV, Lazerson J. The natural history of factor VIII:Cinhibitors in patients with hemophilia A: a national cooperativestudy. II. Observations on the initial developmentof factor VIII:C inhibitors. Blood 1988;71:344-8.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 1752. Sawamoto Y, Prescott R, Zhong D, Saenko EL, Mauser-Bunschoten E, Peerlinek K, et al. C2 domain restrictedepitope specificity of inhibitor antibodies elicited by aheat pasteurized product, factor VIII CPS-P, in previouslytreated hemophilia A patients without inhibitors.Thromb Haemost 1998;79:62-8.53. Laub R, Di Giambattista M, Fondu P, Brackmann HH,Lenk H, Saenko EL, et al. Inhibitors in German hemophiliaA patients treated with a double virus inactivatedfactor VIII concentrate bind to the C2 domain of FVIIIlight chain.\par. Thromb Haemost 1999;81:39-44.54. Bi L, Lawler AM, Antonarakis SE, High KA, Gearhart JD,Kazazian Jr HH. Targeted disruption of the mouse factorVIII gene produces a model of haemophilia A. Nat Genet1995;10:119-21.55. Sarkar R, Gao GP, Chirmule N, Tazelaar J, Kazazian HHJr. Partial correction of murine hemophilia A with neoantigenicmurine factor VIII. Hum Gene Ther 2000; 11:881-94.56. Qian J, M Borovak, Bi L, Kazazian Jr HH, Hoyer L.Inhibitor antibody development and T cell response tohuman factor VIII in murine hemophilia A. ThrombHaemost 1999;81:240-4.57. Wu H, Reding M, Qian J, Okita DK, Parker E, Lollar P,et al. Mechanism of the immune response to humanfactor VIII in murine hemophilia A. Thromb Haemost2001;85:125-33.58. Reding MT, Wu H, Krampf M, Okita DK, Diethelm-OkitaBM, Christie BA, et al. Sensitization of CD4 + T cells tocoagulation factor VIII: response in congenital andacquired hemophilia patients and in healthy subjects.Thromb Haemost 2000;84:643-52.59. Qian J, Collins M, Sharpe AM, Hoyer LW. Preventionand treatment of factor VIII inhibitors in murine hemophiliaA. Blood 2000;95:1324-9.60. Rossi G, Sarkar J, Scandella D. Long-term induction ofimmune tolerance after blockade of CD40-CD40L interactionin a mouse model of hemophilia A. Blood 2001;97:2750-7.61. Qian J, Burkly LC, Smith EP, Ferrant JL, Hoyer LW, ScottDW, Haudenschild CC. Role of CD154 in the secondaryimmune response: the reduction of pre-existing splenicgerminal centers and anti-factor VIII inhibitor titer. EurJ Immunol 2000;30:2548-54.62. Balague C, Zhou J, Dai Y, Alemany R, Josephs SF,Andreason G, et al. Sustained high-level expression offull-length human factor VIII and restoration of clottingactivity in hemophilic mice using a minimal adenovirusvector. Blood 2000;95:820-8.63. Connelly S, Andrews JL, Gallo AM, Kayda DM, Qian J,Hoyer L, et al. Sustained phenotypic correction ofmurine hemophilia A by in vivo gene therapy. Blood1998;91:3273-81.64. Evans GL, Morgan RA. Genetic induction of immunetolerance to human clotting factor VIII in a mouse modelfor hemophilia A. Proc Natl Acad Sci USA 1998; 95:5734-9.haematologica vol. 88(supplement n. 12):september 2003

[Immunological Aspects]review paperAnalysis of factor VIII inhibitorsusing phage displayhaematologica 2003; 88(suppl. n. 12):18-25http://www.haematologica.org/free/immunotolerance2001.pdfJAN VOORBERG, WENDY S. BRIL, EDWARD N. VAN DEN BRINK *Department of Plasma Proteins, Amsterdam, the Netherlands*Current address: Crucell BV, Leiden, the NetherlandsInhibitory antibodies that develop in patients withhemophilia A bind to restricted regions in the A2, A3and C2 domain of factor VIII. Functional studies haveshown that anti-A2 and anti-A3 antibodies interferewith assembly of the factor VIIIa-factor IXa complex.Binding of inhibitors to the C2 domain precludesbinding of factor VIII to phospholipids. We have usedphage display to isolate a large number of humanmonoclonal antibodies from the immunoglobulinrepertoire of hemophilia A patients with an inhibitor.Epitope mapping studies suggest that the majorityof human monoclonal antibodies bind to previouslyidentified epitopes on factor VIII. Inspection of theamino acid sequence of the variable heavy chain(VH) domains of human anti-factor VIII antibodiesreveals some striking features. Anti-A2 and anti-A3-C1 antibodies incorporate VH gene segments thatare frequently used for assembly of human IgG molecules.This may explain the presence of antibodieswith this specificity in a large number of inhibitorpatients. Anti-C2 antibodies are derived from twoclasses of VH gene segments, both belonging to theVH1 family, that bind to two distinct antigenic sitesin the C2 domain. Our findings suggest that B-cellsexpressing immunoglobulin molecules that compriseVH gene segments with the above-mentioned characteristicsare selectively amplified from the totalrepertoire following exposure to antigenic determinantsin the C2 domain of factor VIII.Correspondence: Jan Voorberg, Department of Plasma Proteins,Plesmanlaan 125, 1066 CX Amsterdam, The Netherlands.Phone: international +33.20.5123120. Fax: international+33.20.5123680. E-mail: j_voorberg@clb.nlFactor VIII participates in the intrinsic pathwayof blood coagulation. The physiologicimportance of factor VIII is illustrated by theX-linked bleeding disorder hemophilia A whichis due to functional absence of factor VIII. Thebleeding tendency in patients with hemophilia Acan be corrected by the administration of factorVIII concentrates. In approximately 20-40% ofpatients with severe hemophilia A inhibitory antibodiesdevelop usually after 5-12 exposure days. 1Inhibitor formation hampers further treatmentof patients with factor VIII. Activated prothrombincomplex concentrates and activated factor VIIare used for treatment of bleeding episodes ininhibitor patients. 1 Based on internal sequencehomology factor VIII is divided into a series ofhomologous domains which are interspersed byshort spacer regions that are rich in acidic aminoacids. 2 In plasma factor VIII circulates as a metal-ionlinked heterodimer. The heavy chain consistsof the domains A1-a1-A2-a2-B whereas thelight chain is composed of the domains a3-A3-C1-C2. Due to proteolysis at various sites in theB domain the size of factor VIII heavy chainranges between 90 and 220 kDa. In plasma factorVIII is bound to von Willebrand factor (VWF)which protects factor VIII from proteolytic degradation.2 Upon activation by thrombin or factorXa, factor VIII dissociates from VWF and assemblyof the factor VIIIa-factor IXa complex onphospholipid surfaces can occur. 2 Activation offactor X by factor IXa is significantly enhanced bythe non-enzymatic cofactor factor VIII. Functionalsites on factor VIII involved in binding tofactor IXa and phospholipids have been defined inconsiderable detail. 3 Concomitantly, knowledgeon the binding sites for factor VIII inhibitors onfactor VIII has grown rapidly during the last fiveyears. 4 Based on different experimental approachesthree major epitopes have been defined on factorVIII. In the A3 domain of factor VIII a bindingsite for factor VIII inhibitors has been localizedto residues Gln 1778 -Met 1823 . 5,6 Binding ofinhibitory antibodies to this site in the A3 domaininterferes with complex assembly of the factorIXa-factor VIIIa which is mediated by this part ofthe factor VIII light chain. 5-7 Residues Arg 484 -Ile 508comprise a major determinant of a binding sitefor factor VIII inhibitors in the A2 domain. 8 Singlealanine replacements in this part of the A2domain suggested that Tyr 487 is a residue criticalfor binding of inhibitory antibodies to this part ofhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 19the A2 domain of factor VIII. 9 Antibodies directedtoward residues Arg 484 -Ile 508 prevent the stimulatoryeffect of isolated A2 domain on the catalyticactivity of factor IXa. 10 The inhibitory propertiesof anti-A3 and anti-A2 antibodies are inagreement with our current model of the assemblyof the factor VIIIa-IXa complex. Initial highaffinity binding is mediated by interaction of thefactor VIII light chain with residues present in theEGF1-and EGF2 domains of factor IXa whereasstimulation of factor VIII cofactor activity resultsfrom interaction of the A2 domain with the proteasemoiety of factor IXa. 3 Apparently, factor VIIIinhibitory antibodies can block the interaction offactor VIIIa with factor IXa via two distinct mechanisms.Conflicting data have been reported on the epitopefor factor VIII inhibitors in the C2 domain.Recombinant factor VIII fragments expressed inEscherischia coli revealed the presence of a bindingsite within region Val 2248 -Ser 2312 . 11 Evaluationof the functional inhibition of a panel of anti-C2inhibitors against a series of human/porcinehybrids suggested that residues Glu 2181 -Val 2243 areinvolved in binding of factor VIII inhibitors. 12 Theapparent discrepancies between the two studiesmay be explained by the different approachesthat have been used to characterize factor VIIIinhibitors. Antibodies directed toward the C2domain interfere with binding of factor VIII tophospholipids. 11,13 Current data suggest that arestricted number of 3-4 major binding sites forinhibitory antibodies are present on factor VIII.At present it is not clear why only a limited numberof sites on factor VIII is targeted by theimmune system. This may be explained by theuse of only a selected proportion of the totalavailable immunoglobulin repertoire for the generationof antibodies that interact with immunodominantregions on factor VIII. To explore thisissue we analyzed the anti-factor VIII repertoireof a number of well-characterized inhibitorpatients by phage display.Assembly of human immunoglobulin repertoiresImmunoglobulins are composed of a light anda heavy chain that contain both variable andconstant regions. The variable or V-regions areencoded by two and three gene segments for thelight and heavy chain, respectively. The heavychain locus is present on chromosome 14 whereasloci encoding κ and λ light chains are presenton chromosome 2 and 22, respectively. The individualloci contain large numbers of differentgene segments that are assembled into functionalantibody molecules during maturation of Bcells. The heavy chain locus consists of over 120variable heavy chain gene segments, 27 diversity(D) segments and 6 joining (J) segments 14,15which are followed by gene segments that encodeconstant parts of immunoglobulin heavy chains(µ, δ, γ3, γ1, α1, γ2, γ4, ε, α2). The κ light chainlocus on chromosome 2 comprises 91 variablelight (VL) chain gene segments and 5 J segments.The λ light chain on chromosome 22 contains>45 variable light chain and four J segments. Notall variable heavy and light chain segments areused for assembly of human antibodies. Approximately50-75% of VH and VL segments are nonfunctional.This may be due to the absence of anopen reading frame or lack of recombinationsites flanking the variable segments. 16 Assemblyof immunoglobulin molecules proceeds in ahighly regulated manner. The first event thatoccurs is the joining of a D segment with a J segmentin the variable region of the immunoglobulinlocus. Subsequently, a VH gene segmentand a µ constant region is added giving riseto a functional heavy chain which is co-expressedtogether with a surrogate light chain on a B-cellprecursor in the bone marrow. 17 Next, a VL segmentis fused with a JL segment resulting in afunctional light chain segment that replaces thesurrogate light chain. Following negative selectionof self-reactive B-cells in the bone marrow,immature B-cells are transported to the peripherywhere they express both IgM and IgD. 17 Internalizationof antigen via surface immunoglobulinson B-cells is followed by proteolytic processingof antigens into peptides that are amenablefor presentation by MHC class II molecules onthe surface of B-cells. T helper cells specificallyrecognizing the presented peptide then help B-cells to proliferate. Proliferating B-cells migrate tothe lymph nodes where germinal centers arisethat provide an environment for adequate T-cellhelp for the generation of high affinity antibodies.Affinity maturation proceeds via amino acidreplacements in the variable parts of the antibody,a process termed somatic hypermutation.Additionally, class switching from IgM to IgA,IgG or IgE occurs at this stage. Following affinitymaturation, B-cells develop into plasma cellsthat produce antibodies that are present in plasma.Alternatively, B-cells develop into memoryB-cells, which are present for extended periods inthe periphery thereby facilitating secondaryresponses to incoming antigens.Analysis of factor VIII inhibitors by phagedisplay: general outlineUsing peripheral B lymphocytes as a source ofRNA we isolated immunoglobulin repertoires ofpatients with an inhibitor using a series of polymerasechain recations (PCR) that targeted thevariable domains of the immunoglobulin heavychain. 18 Several restrictions were applied to preferentiallyamplify the anti-factor VIII repertoireof these patients. The majority of factor VIIIinhibitors are of subclass IgG4: therefore, anIgG4 specific amplification step was introducedin our experimental protocol. The concentrationof IgG4 is approximately 5% of the total quantityof IgG in human serum. Although this numberdoes not necessarily correspond to the percentageof IgG4 positive B cells in the peripheralblood, it is anticipated that the resultinghaematologica vol. 88(supplement n. 12):september 2003

20J. Voorberg et al.immunoglobulin repertoires are enriched foranti-factor VIII antibodies. A further simplificationinvolves the use of a immunoglobulin lightchain repertoire that has been amplified fromperipheral blood lymphocytes of healthy individuals.19,20 Different immunoglobulin light chainscan pair to a single heavy chain and therefore itis anticipated that VH domains derived frompatients with hemophilia A will pair with a suitablelight chain from the available non-immunerepertoire. Because of our focus on the variablepart of the heavy chain we could only determinethe characteristics of this part of immunoglobulinmolecules that bind to factor VIII. The variableheavy chain locus encodes 51 functionalvariable heavy chain segments that can be classifiedinto seven families (see Figure 1). Somefamilies, such as VH2, 5, 6 and 7 contain onlyone or a few members whereas others (VH1, VH3and VH4) harbor more then 10 members. Usageof the different families of VH germline gene segmentsin the peripheral repertoire roughly correspondsto the number of VH gene segments presentwithin a single family (see Figure 1; based ondata in ref. 21). It has been shown that the variableheavy chain gene segments contribute to theoverall fold of the VH domain of immunoglobulins.22 Fusion of a VH gene with a D and a J segment,a process that involves addition and deletionof nucleotides at the sites of junction, createsantigen-independent diversity on this initialfold. Suitably assembled IgG molecules make upthe naïve B cell repertoire. The presence of antigenthen stimulates selective outgrowth and maturationof B cells that express surface immunoglobulinsthat can interact with antigen. We haveassessed the characteristics of the variable heavychain domains of human antibodies directedagainst factor VIII. At present we have isolatedhuman antibodies directed against majorinhibitor epitopes in the A2, A3-C1 and C2domains of factor VIII. 23-27 In the following paragraphsan initial analysis of the characteristics ofthe variable heavy chain (VH) domains of theseantibodies will be presented.Characteristics of anti-A2 and anti-A3-C1antibodies obtained by phage displaySince the characteristics of anti-A2 and anti-A3antibodies display some common features we willfirst discuss this subset of human anti-factor VIIIantibodies. Four different human antibodiesdirected against the A2 domain have been isolatedof which three bind to the residues Arg 484 -Ile 508 , an immunodominant region on factorVIII. 24,27 One anti-A2 antibody, designated VK41,bound to the acidic region that follows the A2domain. Two out of four antibodies were derivedfrom VH segments DP-47 (3-23) whereas theother ones were encoded by DP-10 (1-69) andDP-58 (gene segment not yet mapped). Amongthe six human antibodies directed against theA3-C1 domains of factor VIII that have been isolated5, were directed against residues Gln 1778 -Met 1823 , a major binding site for factor VIIIinhibitors in the A3 domain of factor VIII. 5,6,25Figure 1. (A) Schematic representation of a phylogenetic tree of human variable heavy (VH) gene segments. Based on sequencehomology human VH gene segments can be divided into seven families. The number of individual gene segments present withineach family is given in brackets (adapted from ref. #16). (B) Variable heavy chain segments found in peripheral IgG-positiveB-cells. VH gene segments are classified into seven families (x-axis). On the y-axis the percentage of IgG-positive B cells expressingan IgG with a VH gene segments from one of these families is depicted (data drawn from ref. #21).haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 21Figure 2. Variable heavy chain gene segments of the VH3 family present in peripheral IgG-positive B cells. On the x-axis the 22gene segments belonging to the VH3 family are listed. On the y-axis the percentage of IgG-positive B cells expressing an IgGcontaining one of these gene segments is depicted (data drawn from ref. #21). Anti-A2 and anti-A3-C1 antibodies are derivedfrom gene segments 3-21, 3-23 and 3.30 which are commonly expressed in peripheral IgG-positive B-cells.All isolated antibodies were derived fromgermline gene segments of the VH1 and VH3family. Two of the anti-A3-C1 antibodies werederived from germline gene segment DP-49 (3-30), two of DP-77 (3-21), one of DP-14 (1-18)and one of DP-15 (1-8). At first sight there is littleto be learned from the origin of variable heavychain of the anti-A2 and anti-A3 antibodies.Germline gene segments used are derived fromboth the VH1 and VH3 family which does notcome as a surprise since the VH segments fromthese families are used by 70% of peripheral IgG +B-cells (see Figure 1). Closer inspection of theoccurrence of the different VH gene segments inthe normal repertoire yields some remarkablefeatures (Figure 2). About 25% of the humanIgG + repertoire is derived from VH gene segmentsDP-47 (3-23), DP-49 (3-30) and DP-77 (3-21).These findings show that anti-A2 and anti-A3antibodies use VH gene segments that are preferentiallyexpressed in IgG molecules in the normalrepertoire. The observed preference suggeststhat epitopes present in the A2 and A3-C1domains of factor VIII are accessible forimmunoglobulins that have incorporated VHsegments DP-47 (3-23), DP-49 (3-30) and DP-77 (3-21). The high concentration of IgG + B cellsin the periphery that are derived from these VHgene segments may be due to some inherent flexibilityof IgG molecules with these VH gene segmentsto bind to antigenic sites on a variety ofproteins. Alternatively, IgG molecules containingthese VH gene segments may more efficientlycope with selective processes that occur duringmaturation of B-cells. A relatively large numberof immature B-cells containing surface IgGderived from these VH gene segments is thenavailable for incoming antigen. Independently ofthe underlying mechanism, the presence of DP-47 (3-23), DP-49 (3-30) and DP-77 (3-21) genesegments in anti-A2 and anti-A3-C1 antibodiessuggests that antibodies with this specificity arefrequently observed in plasma of inhibitorpatients. Indeed a large study concluded thatanti-A2 or anti-A3 antibodies are present in virtuallyall patients with factor VIII inhibitors. 28The above analysis provides an attractively simpleexplanation for the frequent occurrence ofanti-A2 and anti-A3 antibodies in inhibitorpatients. However, some caution is warranted.So far only a limited number of patients has beenincluded in our analysis. Also, our studies suggestthat the epitopes in the A2 and A3-C1 domainsare more complex then previously anticipated.24,25 More detailed studies on epitope specificityand VH gene usage are required in order todetermine unequivocally whether a particular VHgene segment is preferentially incorporated intoan IgG molecule binding to an antigenic site inthe A2 or A3-C1 domain of factor VIII.Characteristics of anti-C2 antibodiesobtained by phage displayThe plasma of most inhibitor patients containsantibodies that react with the C2 domain of factorVIII. 28 At present we have isolated 16 differentscFv reactive with this domain from therepertoire of patients with mild and acquiredhaematologica vol. 88(supplement n. 12):september 2003

22J. Voorberg et al.hemophilia A. 23,26 Independently, Arai and coworkersisolated 3 human antibodies from aphage library of an unrelated patient. 29 Epstein-Barr immortalization of peripheral B-lymphocyteswas used to isolate a human monoclonalantibody directed against the C2 domain of factorVIII. 30 Sequence analysis of the VH domain ofthese clones reveals some interesting features.Germline gene segment DP-5 (1-24) of the VH1family is used by about half of the anti-C2 antibodiesthat have been isolated so far. 23,26,29,30 TheVH domain of the remainder of human antibodiesdirected against the C2 domain is derivedfrom germline gene segments DP-10 (1-69), DP-14 (1-18) and DP-88 (1-e). 23,26 Phylogeneticanalysis of gene segments belonging to the VH1family shows that DP-10 (1-69), DP-14 (1-18)and DP-88 (1-e) are present within one branchof this family. 26 Consistent with the more extensivedifferences in nucleotide sequence, DP-5 (1-24) is present within a distinct branch of theVH1 family (see Figure 3). The available data suggestthat based on the characteristics of their VHdomains, anti-C2 antibodies can be subdividedin two different classes. In a previous section itwas proposed that anti-A2 and anti-A3 antibodiesare derived from VH gene segments that arecommonly used in the immunoglobulin repertoire.Both DP-10 (1-69) and DP-14 (1-18) areused in approximately 3-5% of peripheral IgG + Bcells whereas DP-5 (1-24) and DP-88 (1-e)occur in less then 1% the total IgG + repertoire(data extracted from ref. #21). These percentagesindicate that these germline gene segmentsare not dominantly expressed in IgG + peripheralB-cells. The restricted usage of VH gene segmentsDP-5 (1-24), DP-10 (1-69), DP-14 (1-18) andDP-88 (1-e) for assembly of anti-C2 antibodiessuggests that B cells expressing human antibodieswith these VH gene segments are positivelyselected following exposure of naïve B-cells tofactor VIII.Structural elements present within the VH genefamilies may facilitate interaction of antibodiescontaining these gene segments with the C2domain. Based on the three-dimensional structureof a large collection of antibodies an overalltopology of antigen binding sites has been proposedby Chothia and co-workers. 22 Six antigenbinding loops derived from both heavy (H1, H2,H3) and light chains (L1, L2, L3) have beendefined that partially overlap with the hypervariablecomplementary determining regions(CDRs) 15,31 (Figure 4A). Both the H1 and H2region are contained within the VH gene segmentof an antibody. For region H1 three socalledcanonical structures have been defined thatdiffer in the number of amino acids containedwithin the H1 region. Similarly, the existence offour different canonical structures for the H2region has been proposed. 22 Interestingly, bothgermline gene segments DP-10 (1-69), DP-14(1-18) and DP-88 (1-e) possess the H1-H2canonical structure 1-2. This combination ofcanonical structures is observed in only 7 out of51 VH gene segments. Four of them [DP-3 (1-f),DP-10 (1-69), DP-14 (1-18) and DP-88 (1-e)]belong to the VH1 family whereas DP-73 (5-51),5-a and DP-21 (7-4.1) belong to the VH5 andVH7 family. 15 Although regions H1 and H2determine only part of the architecture of anantigen binding site, the presence of canonicalstructure 1-2 may facilitate interactions of VHdomains with antigenic determinants present inthe C2 domain. A canonical structure for the H2region of germline gene segment DP-5 (1-24),could not be defined due to the presence of anglutamic acid instead of the more usual threonine,alanine or leucine at position 71. 22,32 Interestingly,in three of out of four human antibodiesthat are encoded by the germline gene segmentDP-5 (1-24) amino acid substitutions havetaken place at this position 26,30 (Figure 4A).Replacement of the glutamic acid for an alanineFigure 3. Phylogenetic tree of the VH1 family. Anti-C2 antibodies are derived from two classes of VHgene segments. The first class comprises VH genesegments DP-10 (1-69), DP-14 (1-18) and DP-88 (1-e). The second class of anti-C2 antibodies incorporatesVH gene segment DP-5 (1-24).haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 23at position 71 may affect the canonical structureof the VH domains of the DP-5 (1-24) encodedantibodies which now becomes identical to thatof the VH domains of the other anti-C2 antibodies(i.e. 1-2). The above analysis suggests thatrestricted usage of VH segments may endow anti-C2 antibodies with unique structural featuresthat may play a role in creating an appropriateparatope.In the previous section evidence was presentedfor the existence of two classes of human antibodiesthat bind to the C2 domain of factor VIII.The first class contains VH domains that arederived from the closely related VH gene segmentsDP-10 (1-69), DP-14 (1-18) and DP-88(1-e) whereas the second class of human antibodiesis characterized by the presence ofgermline gene segment DP-5 (1-24). 26 Epitopemapping studies using a series of human/porcineand factor V/VIII hybrids reveals that these twoclasses of antibodies bind to distinct sites withinthe C2 domain of factor VIII. Residues contributingto the epitope of human antibodiesencoded by VH segments DP-10 (1-69), DP-14(1-18) and DP-88 (1-e) resides in the carboxylterminal part of the C2 domain (residues 2243-2332). In contrast, human antibodies containingVH segment DP-5 (1-24) interact with residuesat the amino- and carboxyl terminal part of theC2 domain. Recently, the three-dimensionalstructure of the C2 domain of factor VIII wasdetermined. 33 Based on the structure it was proposedthat three loops containing hydrophobicresidues intersect the phospholipid bilayer whereasa layer of positively charged residues can interactwith negatively charged phosphate moietiesof phosphatidylserine. 33 In a follow-up study thethree-dimensional structure of a complex of thehuman monoclonal antibody BO2C11 and theC2 domain was determined at 2.0Å 34 (Figure 4B).The VH domain of BO2C11 is encoded by VHgene segment DP-5. 30 Figure 4A shows an alignmentof the amino acid sequence of BO2C11with the DP-5 (1-24) encoded antibodies wehave recently isolated. 26 All four DP-5 (1-24)encoded antibodies have a relatively small CDR3which comprises 8-11 amino acids. The patternof somatic hypermutation varies considerablybetween BO2C11 and the clones isolated byphage display, which is not surprising, since theseantibodies were derived from different patients.A common replacement between 3 of the 4 antibodiesis a substitution of glutamic acid at position71 for an alanine. This suggests that clonesharboring this amino acid replacement are preferentiallyselected from the total pool of C2domain reactive B-cells. A possible structuralbasis for this selective enrichment has been discussedin a previous paragraph. Based on thethree-dimensional structure of BO2C11 and theC2 domain, multiple atomic contacts have beendetected between the heavy chain of BO2C11and the C2 domain (Figure 4A/4B). 34 Comparisonof the amino acid sequence of the WR cloneswith that of BO2C11 reveals that the majority ofcontact residues are conserved in the sequence ofthe WR clones (see Figure 4A). This suggests thatthe interactive surface of the complex betweenBO2C11 and the C2 domain is representative forother DP-5 (1-24) encoded anti-C2 antibodies.Small differences between binding of individualDP-5 (1-24) encoded antibodies to the C2domain can, however, occur. The DP-5 (1-24)encoded residue Asp 52 of BO2C11 that interactswith Arg 2215 in the C2 domain has been substitutedfor an asparagine or alanine in the VHdomains of the WR clones (see Figure 4A). Alsoresidue Asp 99 present in the CDR3 of BO2C11(indicated by an arrow; Figure 4A) is not conservedamong the WR clones. Since several additionalnegatively charged amino acids are presentin both CDR2 and CDR3 of WR1, WR16 andWR17, the effect of these substitutions on theoverall architecture of the interactive surface ismost likely limited. Single amino acid substitutionsin the C2 domain have recently been evaluatedfor their reactivity with factor VIIIinhibitors. 35 A reduction in functional inhibitionof human antibodies was observed when residuesMet 2199 /Phe 2200 (Loop I; Figure 4B) and Leu 2252(Loop II; Figure 4B) were substituted for an Ile,Leu and Phe, respectively. 35 Combination of thesesubstitutions resulted in a further decrease in theantigenicity of the C2 domain. The significanceof electrostatic interactions mediated by Arg 2210and Arg 2215 for binding to factor VIII inhibitorshas not yet been assessed. Residues Arg 2210 andArg 2215 are not critical for binding of factor VIIIto phospholipids. 36 Factor VIII variants harboringsubstitutions at these sites may, therefore, displaya reduced antigenicity while maintainingtheir phospholipid binding properties.Conclusions and implicationsOverall, our data show that phage display is auseful technology for isolating large numbers ofhuman monoclonal antibodies from the repertoireof patients with inhibitors. Inspection ofthe primary amino acid sequence of these antibodiessuggest that only a restricted number ofvariable heavy chain segments is used for theassembly of human antibodies that react withthe C2 domain. In contrast, less restriction isobserved for anti-A2 and anti-A3 antibodies, atleast at the level of VH gene segments. The availabilityof a large panel of human antibodiesderived from different patients is likely toincrease our knowledge on the number and complexityof B-cell epitopes on factor VIII. Subsequentmodification of antigenic sites within theA2, A3 and C2 domains may provide a basis forthe reduction of the antigenicity and perhaps alsoimmunogenicity of factor VIII. 4 Alternatively, itmay be possible to develop antibody-basedreagents that interfere with the binding ofinhibitory antibodies to factor VIII. In this respectit is worth mentioning that the majority of variabledomain antibody fragments do not interferehaematologica vol. 88(supplement n. 12):september 2003

24J. Voorberg et al.ABFigure 4. (A) Deduced amino acid sequences of variable heavy chain domains of DP-5 (1-24) encoded human anti-C2 antibodies.FR, framework region; CDR, complementary-determining region; H1, H2 and H3, antigen binding loops defined based onstructural homology of immunoglobulins. 22 Dashes indicate sequence identity to germline gene segments. Lower case lettersindicate amino acid substitutions originating from the PCR primer. The crystal structure of a complex of BO2C11 in contactwith the C2 domain has been resolved. 34 Residues in BO2C11 in contact with the C2 domain are underlined. Aspartic acid 52and 99 are indicated by an arrow. (B) Three-dimensional model of a complex between BO2C11 and the C2 domain of factorVIII. At the left, part of the VH domain (residues 1-113) of BO2C11 is depicted, the C2 domain is positioned at the right. Thelight chain of BO2C11 is not included in this figure. Multiple contacts exist between BO2C11 and the C2 domain. 34 Loop I representcontacts between residues Met 2199 /Phe 2200 of the C2 domain with residues Ser 50 /Ile 59 of BO2C11. Loop II representscontacts between Leu 2251/2252 of C2 with residues Val 2 /Tyr 98 of BO2C11. Residues Arg 2215 and Arg 2220 (red) of C2 that interactwith negatively charged residues Asp 52 and Asp 99 (blue) in BO2C11 are highlighted.with factor VIII co-factor activity. We have previouslyshown that the non-inhibitory singlechain variable domain antibody fragment EL-14can partially neutralize factor VIII inhibitors inan in vitro assay. 18 Further study is required toestablish whether the concept of masking antigenicsites on factor VIII by antibody-derivedreagents can be translated into a new therapeuticoption for treatment of patients withinhibitors.AcknowledgementsParts of our studies are supported by a grant fromthe Netherlands Organization of Science (NWOgrant 902-26-204). We thank Dr. Sander Meijerand Gunny van Stempvoort for help with the preparationof Figure 4B.References1. Tuddenham EGD, Mannucci PM. The hemophilias—from royal genes to gene therapy. N Engl J Med 2001;344: 1773-9.2. Lenting PJ, van Mourik JA, Mertens K. The life cycle ofcoagulation factor VIII in view of its structure and function.Blood 1998; 92:3983-96.3. Mertens K, Celie PH, Kolkman JA, Lenting PJ. FactorVIII-factor IX interactions: molecular sites involved inenzyme-cofactor complex assembly. Thromb Haemost1999; 82:209-17.4. Lollar P. Characterization of factor VIII B-cell inhibitoryepitopes. Thromb Haemost 1999; 82:505-8.5. Fijnvandraat K, Celie PH, Turenhout EA, ten Cate JW,van Mourik JA, Mertens K, et al. A human allo-antibodyinterferes with binding of factor IXa to the factorVIII light chain. Blood 1998; 91:3701-9.6. Zhong D, Saenko EL, Shima M, Felch M, Scandella D.Some human inhibitor antibodies interfere with factorVIII binding to factor IX. Blood 1998; 92:136-42.7. enting PJ, van de Loo JW, Donath MJ, van Mourik JA,haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 25Mertens K. The sequence Glu1811-Lys1818 of humanblood coagulation factor VIII comprises a binding sitefor activated factor IX. J Biol Chem 1996; 271:1935-40.8. Healey JF, Lubin IM, Nakai H, Saenko EL, Hoyer LW,Scott ME, et al. Residues 484-508 contain a major determinantof the inhibitory epitope in the A2 domain ofhuman factor VIII. J Biol Chem 1995; 270:14505-9.9. Lubin IM, Healey JF, Barrow RT, Scandella D, Lollar P.Analysis of the human factor VIII A2 inhibitor epitopeby alanine scanning mutagenesis. J Biol Chem 1997;272:30191-5.10 Fay PJ, Scandella D. Human inhibitor antibodies specificfor the A2 domain disrupt the interaction between thesubunit and factor IXa. J Biol Chem 1999; 274:29826-30.11 Scandella D, Gilbert GE, Shima M, Nakai H, EaglesonC, Felch M, et al. Some factor VIII inhibitor antibodiesrecognize a common epitope corresponding to C2domain amino acids 2248 through 2312, which overlapa phospholipid- binding site. Blood 1995; 86:1811-9.12 Healey JF, Barrow RT, Tamim HM, Lubin IM, Shima M,Scandella D, et al. Residues Glu2181-Val2243 containa major determinant of the inhibitory epitope in the C2domain of human factor VIII. Blood 1998; 92:3701-9.13. Arai M, Scandella D, Hoyer LW. Molecular basis of factorVIII inhibition by human antibodies. Antibodies thatbind to the factor VIII light chain prevent the interactionof factor VIII with phospholipid. J Clin Invest 1989;83: 1978-84.14 Matsuda K, Ishii K, Bourvagnet P, Kuma KI, HayashidaH, Miyata T, et al. The complete nucleotide sequence ofthe human immunoglobulin heavy chain variableregion locus. J Exp Med 1998; 188:2151-62.15 Tomlinson IM, Williams SC, Ignatovitch O, Corbett SJ,Winter G. V-BASE sequence directory. MRC Centre forProtein Engineering, Cambridge, UK.16 Cook GP, Tomlinson IM. The human immunoglobulinrepertoire. Immunol Today 1995; 16:237-42.17 Rajewski K. Clonal selection and learning in the antibodysystem. Nature 1996; 381:751-8.18 Voorberg J, van den Brink EN. Phage display, a tool toexplore the diversity of inhibitors to blood coagulationfactor VIII. Semin Thromb Haemost 2000; 26:143-50.19 Schier R, Bye J, Apell G, McCall A, Adams GP,Malmqvist M, et al. Isolation of high-affinity monomerichuman anti-c-erB-2 single chain Fv using affinitydrivenselection. J Mol Biol 1996; 255:28-43.20 Griffin HM, Ouwehand WH. A human monoclonalantibody specific for the leucine-33 (P1A1, HPA-1a)form of platelet glycoprotein IIIa from a V gene phagedisplay library. Blood 1995; 86:4430-6.21 de Wildt RM, Hoet RM, van Venrooij WJ, TomlinsonIM, Winter G. Analysis of heavy and light chain pairingsindicates that receptor editing shapes the human antibodyrepertoire. J Mol Biol 1999; 285:895-901.22 Chothia C, Lesk AM, Gherardi E, Tomlinson IM, WalterG, Marks JD, et al. Structural repertoire of thehuman VH segments. J Mol Biol 1992; 227:799-817.23 Van den Brink EN, Turenhout EA, Davies J, BovenschenN, Fijnvandraat K, Ouwehand WH, et al. Human antibodieswith specificity for the C2 domain are derivedfrom VH1 germline genes. Blood 2000; 95:558-63.24 Van den Brink EN, Turenhout EA, Bank C, FijnvandraatK, Peters M, Voorberg J. Molecular analysis of humananti-factor VIII antibodies by V gene phage dsiplay identifiesa new epitope in the acidic region following the A2domain. Blood 2000b; 96:540-5.25 van den Brink EN, Turenhout EA, Bovenschen N, HeijnenBG, Mertens K, Peters M, et al. Multiple VH genesare used to assemble human antibodies directed towardthe A3-C1 domains of factor VIII. Blood 2001; 97:966-72.26 van den Brink EN, Bril WS, Turenhout EA, Zuurveld M,Bovenschen N, Peters M, et al. Two classes of germlinegenes both derived from the V(H)1 family direct theformation of human antibodies that recognize distinctantigenic sites in the C2 domain of factor VIII. Blood2002; 99:2828-34.27. Bril WS, Turenhout EA, Kaijen PH, van den Brink EN,Koopman MM, Peters M, et al. Analysis of factor VIIIinhibitors in a haemophilia A patient with an Arg593-→Cys mutation using phage display. Thromb Haemost2001; 86:247a[abstract].28 Prescott R, Nakai H, Saenko EL, Scharrer I, Nilsson IM,Humphries JE, et al. Recombinate and Kogenate studygroups. The inhibitor antibody responnse is more complexin hemophilia A patients than in most nonhemophiliacswith factor VIII autoantibodies. Blood1997; 89:3663-71.29. Arai M, Imai T, Yuguchi M, Nakashima, Kukutake K.Cloning and characterization of single chain Fv of antifactorVIII antibody derived from a hemophilia A patientwith factor VIII inhibitor. Thromb Haemost 1999; 82:238a[abstract].30 Jacquemin MG, Desqueper BG, Benhida A, van der ElstL, Hoylaerts MF, Bakkus M, et al. Mechanism and kineticsof factor VIII inactivation: study with an IgG4 monoclonalantibody derived from a hemophilia A patientwith an inhibitor. Blood 1998; 92:496-506.31 Kabat EA, Wu TT, Perry HM, Gottesman KS, Foeller C.Sequences of immunological interest. 1991; US Departmentof Health and Human Services: 5 th edition. Bethesda,MD, USA.32 Tramontano A, Chothia C, Lesk AM. Framework residue71 is a major determinant of the position abd conformationof the second hypervariable region in the VHdomains of immunoglobulins. J Mol Biol 1990; 215:175-82.33 Pratt KP, Shen BW, Takeshima K, Davie EW, FujikawaK , Stoddard BL. Structure of the C2 domain of humanfactor VIII at 1.5 Å resolution. Nature 1999; 402:439-42.34 Spiegel PC, Jacquemin M, Saint Remy JM, Stoddard BL,Pratt KP. Structure of a factor VIII C2 domain-immunoglobulinG4k Fab complex: identification of aninhibitor antibody epitope on the surface of factor VIII.Blood 2001; 98:13-9.35. Barrow RT, Healey JF, Jacquemin MG, Saint-Remy JM,Lollar P. Antigenicity of putative phospholipid membranebinding residues in factor VIII. Blood 2001;97:169-74.36 Gilbert GE, Kaufman RJ, Arena AA, Miao H, Pipe SW.Four hydrophobic amino acids of the factor VIII C2domain are constituents of both the membrane-bindingand von Willebrand factor-binding motifs. J Biol Chem2002; 277:6374-81.haematologica vol. 88(supplement n. 12):september 2003

[General Treatment of Bleeding Episodes]review paperThe Feiba NovoSevenComparative Study (FENOC)haematologica 2003; 88(suppl. n. 12):26-28http://www.haematologica.org/free/immunotolerance2001.pdfERIK BERNTORP, JAN ASTERMARKDepartment of Coagulation Disorders, Malmö UniversityHospital, Malmö, SwedenThe development of an inhibitor is still the mainthreat to the health of hemophiliacs. By-passingagents, such as activated prothrombin complex concentrateand recombinant factor VIIa, are widelyused to treat acute hemorrhages in hemophilia Ainhibitor patients. The efficacy of these productsvaries among patients and the treatment is expensive.In the Feiba NovoSeven Comparative Study(FENOC) a comparison is made between the activatedprothrombin complex concentrate Feiba andrecombinant factor VIIa NovoSeven. The study is aprospective, randomized, multi-center study, inwhich two bleeds in each of 60 patients will be randomizedand treated with the products. The primaryend-point is evaluation of the hemostatic effect at6 hours. Dose and dose interval will be as recommendedby the manufacturers. Using an in vitrothrombin generation test, an attempt to predict efficacyof the product will be made. The FENOC studyshould be considered as a basic study in order togain more knowledge about by-passing agents andhas the potential to create a platform for future studieson the treatment of acute bleeds in inhibitorpatients.Correspondence: Erik Berntorp, Department of CoagulationDisorders, Malmö University Hospital, SE-205 02 Malmö,Sweden. Phone: international +46.40.332392.Fax: international +46.40.336255.E-mail: erik.berntorp@medforsk.mas.lu.seThe incidence of inhibitors in patients withsevere hemophilia A is around 30% inrecent prospective studies, 1-3 although arange from zero up to 52% has been reported. 4,5There are two main options in the treatment ofan inhibitor patient, the main goal being to renderthe patient tolerant to replacement therapywith factor VIII by induction of immune tolerance.The other option, which can also be usedduring immune tolerance induction before thetolerated state has been achieved, is to treat acutehemorrhages using by-passing agents, which tosome extent improve hemostasis. Among theseagents, the activated prothrombin complex concentrateFeiba has been widely used for manyyears. 6,7 More recently, recombinant factor VIIa(NovoSeven) has been added to the therapeuticarmamentarium. 8,9 Both these products havebeen found to be effective in up to 80-90% of allhemorrhages in non-randomized studies. Theexact by-passing hemostatic mechanism(s) foreach product is not fully understood and thedosing also needs to be further explored. Thenumber of injections given for a bleed hasranged greatly in different reports and it isunknown so far whether one of the productsmight have a better effect in certain patients.Treatment with by-passing agents is costly.Therefore, it is important to perform controlledstudies with these products. As yet, a comparativein vivo study of effect as well as cost-efficacyis lacking. The FENOC study is designed toassess the hemostatic effects of Feiba and Novo-Seven in a randomized, controlled study. Thestudy is multi-center, multi-national and is organizedfrom Malmö with the authors as principalinvestigators. Additional principal investigatorshave been appointed for Italy (AlessandroGringeri, Milan) and North America (Donna DiMichele, New York City, USA).Study materialSixty patients with congenital hemophilia Awith an inhibitor and the need for by-passingagents in the case of joint bleeds will be recruited.The expected bleeding frequency is at leastthree joint bleeds per year. The lower limit of agefor eligibility is two years. Exclusion criteria areother congenital and acquired bleeding disorders,symptomatic liver disease with an INR >1.5haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 27and/or a platelet count

28E. Berntorp et al.and the scheduled 60 patients is still a substantialsized cohort for this kind of disease. Anotherapproach to gain more study material wouldbe to study more than two bleeds in each patient.This approach, however, would introduce anotherbias as the patient then would have previousexperience of both drugs provided that each bleedis randomized and, depending on the response tothe treatment, the patients may prefer one of theproducts when the third and forth bleeds occur.The FENOC study should be considered as abasic study in order to gain more knowledgeabout bypassing agents, their effects and the possibilityof predicting efficacy based on in vitroassays. The FENOC study, which is in progress ina number of sites in Europe and the US, has thepotential to provide a platform for future studieson the treatment of acute bleeds in patients withinhibitors.AcknowledgmentsThis study was supported by grants from BaxterBioScience, Malmö University Hospital andthe county of Scania.References1. Gruppo R, Bray GL, Schroth P, Perry M, Gompers ED.The Recombinant factor VIII (Recombinate) in previouslyuntreated patients (PUPs): a 6.5 year update.Thromb Haemost 1997; 162 Suppl:a[abstract PD-663].2. Rothschild C, Laurian Y, Satre EP, Borel Derlon A, ChambostH, Moreau P, et al. French previously untreatedpatients with severe hemophilia A after exposure torecombinant factor VIII : incidence of inhibitor and evaluationof immune tolerance. Thromb Haemost1998;80:779-83.3. Lusher JM, Shapiro A, Gruppo R, Bedrosian CL, NguyenK. Safety and efficacy in previously untreated patients(PUPs) treated exclusively with B-domain deleted factorVIII (BDD rFVIII). The ReFacto PUP Study Group.Thromb Haemost 2001; Suppl:a[abstract 2558].4. Ehrenforth S, Kreuz W, Scharrer I, Linde R, Funk M,Gungor T, et al. Incidence of development of factor VIIIand factor IX inhibitors in haemophilias. 1992;339:594-8.5. Yee TT, Williams MD, Hill FGH, Lee CA, Pasi JK.Absence of inhibitors in previously untreated patientswith severe haemophilia A after exposure to a singleintermediated purity factor VIII product. ThrombHaemost 1997;78:1027-9.6. Hilgartner M, Aledort L, Gill AJ, and the members of theFEIBA Study Group. Efficacy and safety of vapor-heatedanti-inhibitor coagulant complex in hemophiliapatients. Transfusion 1990;30:626-30.7. Negrier C, Goudemand J, Sultan Y, Bertrand M, RothschildC, Lauroua P. Multicenter retrospective study onthe utilization of the FEIBA in France in patients withfactor VIII and factor IX inhibitors. The members of theFrench Feiba study group. Thromb Haemost 1997;77:1113-9.8. Ingerslev J, Thykjaer H, Kudsk Jensen OK, Fredberg U.Home treatment with recombinant factor VIIa: resultsfrom one centre. Blood Coagul Fibrinolysis 1998; Suppl1:S107-10.9. Lusher J, Ingerslev J, Roberts H, Hedner U. Clinical experiencewith recombinant factor VIIa. Blood Coagul Fibrinolysis1998;9:119-28.10. Hemker HC, Beguin S. Thrombin generation in plasma:its assessment via the endogenous thrombin potential.Thromb Haemost 1995;74:1388.11. Hedner U. Recombinant factor VIIa (NovoSeven) as ahemostatic agent. Dis Mon 2003;49:39-48.12. Abshire TC, Brackmann HH, Scharrer I, Hoots K,Gazengel C, Powell JS, et al. Sucrose formulated recombinanthuman antihemophilic Factor VIII is safe andefficacious for treatment of hemophilia A in home therapy(results of a multinational, international, clinicalinvestigation). Thromb Haemost 2000; 83:811-6.13. Bray GL, Gomperts ED, Courter S, Gruppo R, GordonEM, Manco-Johnson M, et al. A multicenter study ofrecombinant factor VIII (recombinate): safety, efficacy,and inhibitor risk in previously untreated patients withhemophilia A. The Recombinate Study Group. Blood1994;83:2428-35.14. Berntorp E. Second generation, B-domain deletedrecombinant factor VIII. Thromb Haemost 1997; 78:256-60.haematologica vol. 88(supplement n. 12):september 2003

[Round Table on Immune Tolerance Treatment]review paperAntibody reactivity to factorVIIa may impede the effect ofby-passing agents in patientswith hemophilia Bhaematologica 2003; 88(suppl. n. 12):29-33http://www.haematologica.org/free/immunotolerance2001.pdfJAN ASTERMARK, ERIK BERNTORPDepartment for Coagulation Disorders, University of Lund,Malmö, SwedenWe studied patients with hemophilia B and highrespondingfactor IX inhibitors to detect possiblecross-reactivity with other vitamin K-dependent procoagulantfactors due to the structural similarity ofthe proteins. The aim was to explain why the hemostaticeffect of by-passing agents such as activatedprothrombin complex concentrates (aPCC) andrecombinant factor VIIa is inadequate in somepatients. Immunoglobulins (Igs) from three patientswere separately purified on protein A sepharose andthen subjected to immunoaffinity chromatographyon factor IX- and factor VIIa sepharose. All three Igfractions, but not commercially available Ig, containedantibodies that bound to both gels. The Igfractions and the immunoaffinity purified anti-factorIX and VIIa antibodies inhibited thrombin formationin a plasma system using Feiba‚ (12 mU/mL) asactive enzyme, but had little effect in the presenceof NovoSeven‚ (18 U/mL). Minor inhibition of factorXa formation also occurred in a system containingpurified proteins. The divergent effect of the twoproducts may be due to the availability of varyingamounts of factor VIIa for neutralization by the antibodies.Our findings are of unknown clinical significancebut may suggest characterization of the antibodyprofile of individual inhibitor patients to optimizethe type and dose of the agent used to treatsevere bleeds.©2003, Ferrata Storti FoundationKey words: hemophilia B, inhibitors, antibodies,by-passing agents, factor VIIa.Correspondence: Jan Astermark, M.D. Ph.D., Department forCoagulation Disorders, University Hospital, SE-205 02 Malmö,Sweden. Phone: international +46.40.332392. Fax: international+46.40.336255. E-mail:jan.astermark@medforsk.mas.lu.seDevelopment of inhibitory antibodies inpatients with hemophilia still representsa major challenge in regard to both eradicationof the inhibitor and treatment of bleedingepisodes. 1,2 In the case of major bleeds, bypassingagents such as activated prothrombincomplex concentrates (e.g. Feiba ® ‚ and Autoplex® , Baxter) or recombinant factor VIIa(NovoSeven‚ ® Novo-Nordisk) are needed andboth agents seem to be effective in 80-90% of allhemorrhages in inhibitor patients. However,some patients still do not achieve an adequatehemostatic response. 3-7 The reason for this is notyet known.Factor IX is structurally very similar to the othervitamin K-dependent procoagulant factors i.e.factor VII, factor X and prothrombin. 8 In addition,inhibitory antibodies that may develop inpatients with hemophilia in response to giventreatment are known to be synthesized by differentclones. Thus, from a structural point ofview, it would not be surprising if antibodiesagainst factor IX in patients with hemophilia Bwere to cross-react with one or more of the othervitamin K-dependent coagulation factors andthereby to some extent possess the capacity toreduce the hemostatic effect of factor VIIaand/or activated prothrombin complex concentrates(aPCCs). We conducted an investigationto ascertain whether Ig fractions from hemophiliaB patients with high-responding factor IXinhibitors exhibit antibody reactivity to other vitaminK-dependent proteins including factorVIIa. We also evaluated the effect of these antibodieson the formation of factor Xa and thrombinin vitro before and after immunoaffinitychromatography.Materials and MethodsMaterialsRecombinant factor VIIa (NovoSeven ® ) wasfrom Novo Nordisk A/S. Recombinant albuminfreefactor VIII was kindly provided by Dr. V. Foster,Baxter, and BeneFIX ® and Feiba ® were purchasedfrom the same company. Purified recombinantfactor VIIa was a generous gift from Dr.Persson, Novo Nordisk A/S. Intravenous immu-haematologica vol. 88(supplement n. 12):september 2003

30J. Astermark et al.noglobulins and CNBr-activated sepharose 4 Bwere obtained from Pharmacia & Upjohn.Sepharose 6B coupled with factor IX was kindlyprovided by Dr. Freiburghaus, Excorim. Recombinantfactor X was purchased from EnzymeResearch Laboratories and prothrombin fromICN Biochemicals Inc. Recombinant lipidatedtissue factor was obtained from American DiagnosticaInc. The fibrinogen polymerizationinhibitor H-glycine-proline-arginine-proline-OH(GPRP) was from Calbiochem. The chromogenicsubstrates S-2238 and S-2222 were purchasedfrom Haemochrom Diagnostica AB, factor IXdepletedplasma from Biopool AB and humanserum albumin from Sigma.SamplesImmunoglobulin fractions were isolated fromplasma from three patients with severe hemophiliaB (designated P1, P2 and P3) and a highrespondinginhibitor at the age of 5-6 years usingprotein A sepharose as previously described. 9 Theeluted fractions were dialyzed against 0.008 mMHepes containing 0.139 M NaCl, pH 7.3. TheMalmö inhibitor titer was determined asdescribed elsewhere; 10 one Malmö inhibitor unit(MU) is approximately equal to three Bethesdaunits. Western blotting was performed usingstandard techniques. 11Immunoaffinity purification of factor IX antibodieswas performed using sepharose 6 MB towhich factor IX had been coupled as previouslydescribed. 12 Factor VIIa antibodies were isolatedusing CNBr-activated sepharose 4B to which 3mg of human recombinant factor VIIa was coupledaccording to the instructions of the manufacturer.The factor VIIa gel was pre-eluted with0.1 M Tris-HCl containing 0.5 M NaCl, pH 2.2,and then equilibrated in the same Tris-buffer, pH7.0, before mixing with the antibody fractionovernight at 4°C. The gel was washed with Trisbufferand then eluted using a low pH. The eluatewas immediately neutralized to pH 7.0 andthen dialyzed against the Hepes buffer describedabove.In vitro assaysFormation of thrombin was studied in a plasmasystem at 37°C essentially as described byGallistl et al. 13 In short, we used 2 mM GPRP, 50mL of factor IX-deficient or patient’s plasma andrecombinant relipidated tissue factor at a finalconcentration of 0.5 ng/mL in 0.008 M Hepescontaining 0.139 M NaCl, pH 7.3. The final concentrationof NovoSeven ® ‚ was 18 U/mL correspondingto approximately 50% of the expectedpeak value after bolus injection of 100 µg/kg. 14,15The final concentration of Feiba ® ‚ was titrated togive a similar amount of thrombin formed in theabsence of antibodies as that for NovoSeven ® .The various antibody preparations and/or Hepesbuffer were added to a final volume of 218 µLand the reaction was started by adding 0.1 MCaCl2 to a final concentration of 5 mM. Sampleswere collected at various time points for up to 30min and added to the chromogenic substrate S-2238 (final concentration of 0.22 mM). Afterfive minutes the reaction was terminated byadding 160 µL of 50% HAc and the absorbancewas measured at 405 nm. The amount of thrombinformed at each time point was expressed asa percent of the maximal amount of thrombinformed in the factor IX-deficient control plasma.Each experiment was repeated three times on differentoccasions.Formation of factor Xa was studied in a systemcontaining purified proteins (total volume 1 mL)in 0.008 M Hepes containing 0.139 M NaCl, pH7.3, using factor VIII and factor X at a final concentrationof 1 U/mL and 5 µg/mL, respectively.The concentration of recombinant relipidatedtissue factor was 0.5 ng/mL, NovoSeven ® ‚ 0.25U/mL and Feiba ® ‚ 0.5 U/mL. The various antibodypreparations and/or Hepes buffer wereadded and the reaction mixture incubated for20 min at 37°C. The reaction was started byadding 0.1 M CaCl2 to a final concentration of 5mM and then terminated by retransferring 150µL of the mixture to tubes containing EDTA(final concentration 10 mM) at 5,10, 20, 30, 40and 60 min. S-2222 was used at a concentrationof 0.6 nM and the amount of factor X formedmeasured by the absorbance at 405 nm. Theamount of factor Xa formed at each time pointwas then expressed as a percent of the maximalamount of factor Xa formed in the control mixturewithout antibodies. Each experiment wasrepeated three times on different occasions.ResultsIg fractions from three patients with highrespondingfactor IX inhibitors, i.e. a historicalpeak titer of >10 BU/mL, had been collectedbefore the start of immune tolerance induction(ITI) according to the Malmö model using proteinA sepharose. After dialysis, the inhibitoryeffect of each Ig fraction at a final inhibitor titerof 1 MU was measured in a thrombin assay basedon factor IX-deficient plasma. All three Ig fractionswere found to exert an inhibitory effect onthrombin formation in the presence of Feiba®,whereas no major effect was seen when Novo-Seven®‚ was used as active enzyme (not shown).Each Ig fraction was then subjected to immunoaffinitychromatography on a factor IX- and afactor VIIa sepharose. Antibodies were elutedfrom both gels, and Western blotting confirmedthe reactivity of the eluted antibodies to each factor(Figure 1). None of the purified preparationsexhibited any reactivity to factor VIII, X or IIa.In the pure system with Feiba®, using a factorIX inhibitor titer of 1 MU, factor Xa formationwas inhibited 30% by the anti-factor IX antibodiesfrom patient P2 but was only slightly curtailedby the antibodies from the other two patients(not shown). In the presence of NovoSeven®,no major inhibition was seen. All three isolateshaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia31Figure 1. Western blots of the antibody reactivity to Novo-Seven ® (rFVIIa) and BeneFIX ® (rFIX) exhibited by the proteinA sepharose isolate (A) and immunoaffinity-purified anti-factorIX (B) and anti-factor VIIa (C) antibodies from patientP2. Molecular weight (MW) markers are shown to the left.Figure 2. Thrombin formation in the absence () and presenceof factor IX antibodies isolated from Ig fractions frompatients P1 (), P2 () and P3 (), respectively. The finalconcentration of Feiba ® was 12 mU/mL (A), NovoSeven ®18 U/mL (B) and tissue factor 0.5 ng/mL. The reactionswere initiated by adding 5 mM CaCl2. The amount of thrombinformed was measured using a chromogenic substrateand expressed as a percent of the maximal amount of eachenzyme formed in control plasma without added inhibitor.The antibodies were added to a final anti-factor IX inhibitortiter of 1 MU/mL in Hepes buffer.inhibited thrombin formation in the presence ofFeiba ® , but only patient P3 had a minor effect inthe presence of NovoSeven ® ‚ (Figures 2A and B).The immunoaffinity-purified antibodiesagainst factor VIIa were used at a final concentrationof approximately 0.1 mg/mL. Factor Xaformation in the pure system with Feiba®‚ wasslightly inhibited by the antibodies from patientsP1 and P2 but not by those from the third patient(Figures 3A and B). None of the anti-factor VIIapreparations exerted any inhibitory effect in thepure system with NovoSeven ® . In the plasmabasedthrombin assay, all three preparations hadan inhibitory effect on Feiba ® , but no impact inthe presence of NovoSeven ® (Figures 3A and B).No antibodies to either factor VIIa or factor IXwere isolated from 200 mg of commercially availableintravenous immunoglobulins.DiscussionIn clinical practice, the two by-passing agents,Feiba ® and NovoSeven ® , do not always seem tohave an immediate hemostatic effect, despite aneffective response in a majority of the treatedpatients. In at least 10-20% of the cases, thepatients continue to bleed and occasionallyrequire multiple doses. So far there have been noreports of the appearance of antibodies thatinteract with the two therapeutic agents. Nonetheless,considering that the vitamin K-dependentfactors VII, IX and X, and to some extentprothrombin, are highly similar in structure, theoretically,it is possible that polyclonal antibodiesto factor IX cross-react with and therebyimpede the hemostatic effect of the others.Therefore, we wanted to ascertain whether plasmafrom hemophilia B patients with highrespondingfactor IX inhibitors show antibodyreactivity to the other vitamin K-dependent procoagulantfactors.As expected, reactivity to factor IX was found inall three Ig fractions, but not in commerciallyavailable intravenous Ig. However, antibodies tofactor VIIa were also isolated by immunoaffinitychromatography. The anti-factor VIIa antibodiescross-reacted with factor IX (Figure 1), which isnot surprising considering the structural simi-haematologica vol. 88(supplement n. 12):september 2003

32J. Astermark et al.Figure 3. Factor Xa (A and B) and thrombin (C and D) formationin the absence () and presence of anti-factor VIIaantibodies isolated from Ig fractions from patients P1 (),P2 () and P3 (), respectively. In the factor X assays, thefinal concentration of Feiba‚ was 0.5 U/mL (A), NovoSeven ®0.25 U/mL (B) and tissue factor 0.5 ng/mL, whereas theconcentrations used in the thrombin assay were as indicatedin Figure 2. Each reaction was initiated by adding 5mM CaCl2. The amount of factor Xa and thrombin formed ateach time point was measured and expressed as described.The final concentration of the antibodies in Hepes bufferwas approximately 0.1 mg/mL.larity between the two factors. We do not knowwhether the antibodies may have developed secondaryto treatment with factor VII(a) or factorIX concentrates, since all three subjects had beenextensively treated with blood products beforestarting ITI.The antibodies exerted only minor inhibitoryeffects in the system that used purified proteinsto measure the amount of factor Xa formed. Onthe other hand, pronounced inhibition was seenwith both the immunoaffinity-purified anti-factorIX and the anti-factor VIIa antibodies in theplasma-based thrombin assay when using Feiba ®as the active enzyme. Minor or no effects wereseen in the presence of factor VIIa. Consideringthat the reaction was dependent on the additionof tissue factor and that Feiba ® contains muchless factor VIIa than NovoSeven ® it is possiblethat the amount of antibodies in the assays withNovoSeven ® were saturated with an excess of factorVIIa, whereas the smaller amount of factorVIIa in Feiba ® was not enough to saturate and toovercome the neutralizing effect of the antibodies.This would also explain the lower degree ofinhibition seen in the factor Xa assay, since theamount of Feiba ® used was approximately 40-fold higher than in the thrombin assay. Moreover,the 70-fold lower concentration of Novo-Seven ® used in the factor Xa assay may have beenresponsible for the inhibitory effects that weobserved in this system but not in the correspondingthrombin assay.Little is known about the hemostatic mechanism(s)of Feiba ® and the amount of thrombingenerated in our in vitro assay may be dependentnot only on the amount of factor VIIa present,but also on the activity of factor IX in the concentrate.Therefore, the inhibition of thrombinformation that occurred in the presence of thefactor IX antibodies may have been due to directinhibition of factor IX.Taken together, our data support the conceptthat antibody reactivity to factor VIIa might occurin the plasma of hemophilia B patients withhigh-responding inhibitors. These antibodiescould be developed in response to treatment withfactor VII(a) containing products and/or identifieddue to cross-reactivity. Theoretically, theextent to which such antibody reactivity to factorVIIa might influence the hemostatic effect ofboth Feiba ® and NovoSeven ® could depend onthe epitope profile of the patient being treated. Itis difficult to estimate the exact concentration ofthe activated factor(s) at the site of bleeding.Therefore, it is possible that the amount of antibodiesactually present in vivo at a site of injuryis sufficient to exert an inhibitory effect that renderssome patients less responsive to therapy orthat the pharmacokinetic profile of the infusedfactor VIIa could be affected. Our findings suggestthat screening for antibody reactivity to vitaminK-dependent coagulation factors otherthan factor IX might be informative in hemophiliaB patients with high-responding inhibitorsin order to optimize and individualize the typeand dosage of the therapeutic agent used.References1. White GC, Roberts HR. The treatment of factor VIIIinhibitors: a general overview. Vox Sang 1996; 70 Suppl1:19-23.2. Ingerslev J. Hemophilia. Strategies for the treatment ofinhibtor patients. Haematologica 2000; 85 Suppl 10:15-20.3. Hilgartner,M, Aledort L, Gill A. Efficacy and safety ofvapor-heated anti-inhibitor coagulant complex inhemophilia patients. The FEIBA Study Group. Transfusion1990; 30:626-30.4. Negrier C, Goudemand J, Sultan Y, Bertrand M, RothschildC, Lauroua P. Multicenter retrospective study onthe utilization of FEIBA in France in patients with factorVIII and factor IX inhibitors. The French Feiba studygroup. Thromb Haemost 1997; 77:1113-9.5. Ingerslev J, Thykjaer H, Kudsk JOK, Fredberg U. Hometreatment with recombinant factor VIIa: results fromone centre. Blood Coagul Fibrinolysis 1998; Suppl 1:S107-10.6. Lusher J, Ingerslev J, Roberts H, Hedner U. Clinical experiencewith recombinant factor VIIa. Blood Coagul Fibrinolysis1998; 9:119-28.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia337. Key NS, Aledort LM, Beardsley D, Cooper GD, EwensteinBM, Gilchrist GS, et al. Home treatment of mildand moderate bleeding episodes using recombinant factorVIIa (Novoseven) in haemophiliacs with inhibitors.Thromb Haemost 1998; 80:912-8.8. Furie B, Furie BC. The molecular basis of blood coagulation.Cell 1988; 53:505-18.9. Freiburghaus C, Berntorp E, Ekman M, Gunnarsson M,Kjellberg BM, Nilsson IM. Immunoadsorption forremoval of inhibitors: update on treatments in Malmö-Lund between 1980 and 1995. Haemophilia 1998; 4:16-20.10. Nilsson IM, Hedner U. Immunosuppressive treatmentin haemophiliacs with inhibitors to factor VIII and factorIX, Scand J Haematol 1976; 16:369-82.11. Astermark J, Sottile J, Mosher DF, Stenflo J. Baculovirusmediated expression of the EGF-like modules of humanfactor IX fused to the factor XIIIa transamidation site infibronectin. Evidence for a direct interaction betweenthe N-terminal EGF-like module of factor IXab and factorX. J Biol Chem 1994; 269:3690-7.12. Nilsson IM, Freiburghaus C, Sundqvist SB, Sandberg H.Removal of specific antibodies from whole blood in acontinuous extracorporeal system. Plasma Ther TransplTechnol 1984; 5:127-34.13. Gallistl S, Cvirn G, Muntean W. Recombinant factorVIIa does not induce hypercoagulability in vitro. ThrombHaemost 1999; 81:245-9.14. Lindley CM, Sawyer WT, Macik BG, Lusher J, HarrisonJF, Baird-Cox K, et al. Pharmacokinetics and pharmacodynamicsof recombinant factor VIIa. Clin PharmacolTher 1994; 55:638-48.15. Erhardtsen E. Pharmacokinetics of recombinant activatedfactor VII (rFVIIa). Semin Thromb Hemost 2000;26:385-91.haematologica vol. 88(supplement n. 12):september 2003

[Round Table on Immune Tolerance Treatment]Twenty years of experience withlow dose immune tolerancereview paperhaematologica 2003; 88(suppl. n. 12):34-39http://www.haematologica.org/free/immunotolerance2001.pdfEVELIEN P. MAUSER-BUNSCHOTEN,H. MARIJIKE VAN DEN BERG, GORIS ROOSENDAALVan Creveldkliniek, Univerisity Medical Center Utrecht,The NetherlandsCorrespondence: Evelien P. Mauser-Bunschoten, Van Creveldkliniek,University Medical Center Utrecht, Postbox 85000,C01-425, 3508 CX Utrecht, The Netherlands.Phone:international +31.30.2508449. Fax:international+31.30.2503854. E-mail: E.Mauserbunschoten@digd.azu.nlSome patients with inhibitors develop a hightiter antibody, having a brisk anamnesticresponse following infusion with any factorVIII material. In some of these patients theinhibitor disappears spontaneously after stoppingfactor VIII, but will relapse after the nextfactor VIII infusion. 1,2 For this reason factor VIIItherapy was stopped as soon as a hemophiliapatient developed an inhibitor until 1980.In 1974 Brackmann made a first serious protocolfor the irradiation of inhibitors in hemophiliaA by designing the so called Bonn protocolfor induction of immune tolerance in thesepatients. 3 The protocol originally daily infusionswith high dose factor VIII were given in combinationwith activated prothrombin complexconcentrate (APPC, FEIBA). This regimen wassuccessful in patients with low and extreme hightiter inhibitors. Since the beginning of the 1980svarious regimens for the introduction ofimmune tolerance were introduced. 4,8 Sometimesplasmapheresis, cyclophosphamide, gammaglobulinor corticosteroids were added tothese regimens. 9,10 All these strategies had incommon that they were not based on any quantitativeresearch and that nobody understood themechanism of successful treatment.Low dose immune tolerance therapyLow dose immune tolerance therapy wasdeveloped because this was thought to be lessdemanding for patients and staff becausepatients have to be infused only 2-3,5 timesweekly. Also the amount of factor VIII infused islow, making it for economical reasons moreattractive.In the Netherlands this low dose immune toleranceregimen was first introduced in 1981 inthree young patients with life threatening bleeds.Dosage regimenIn patients in whom factor VIII was startedwith the only aim of obtaining immune tolerance,the factor VIII dosage was 25-50 Units factorVIII per kilogram bodyweight (U FVIII/kgbw) every other day or three times a week, inde-pendent of their inhibitor titer. Initially in veryyoung children in whom venous access was difficult,factor VIII was injected twice weekly. Since1990 in these patients an intravenous catheter(Port-a-Cath system, PAC) is implanted in orderto obtain adequate venous access.)When factor VIII treatment was startedbecause of an operation or life-threateningbleeding and the inhibitor was less than 10BU/mL, an initial high dose of factor VIII wasgiven to neutralize the antibodies. The neutralizingdosage was calculated as follows: 112 × BW 80 x (100-Ht) × I100where BW is body weight in kilograms, Ht ishematocrit, and I is inhibitor in BethesdaUnits/mL.The initial high dose was followed by infusion25 U FVIII/kg bw twice daily for one or twoweeks, depending on the clinical status of thepatients and the anamnestic response to factorVIII. After this period factor VII was continued 3times a week or every other day in a dosage of25-50 U FVIII/kg bw.Dose adjustmentWhen factor VIII antibodies decreased and factorVIII recovery was restored, or when ananamnestic response was lacking, factor VIII wastapered down each time the absolute factor VIIIrecovery was higher than 30% until a standardprophylactic dosage of 10-15 units FVIII/kg bwwas reached. Since 1997 standard prophylacticdose in children may be, depending of its clinicaleffect, a similar doses of 25 U FVIII/kg bw 3times a week. 12In patients in whom the inhibitor titer showedno tendency to decrease over a period of 6months or longer and in patients with a highbleeding frequency, the factor VIII dosage wasincreased to 50-100 U factor VIII 3 times a weekor every other day. In those patients in whomthe inhibitor remained high despite 2 years ofhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia35low dose ITT, high dose ITT was started accordingto the Malmö protocol. 10Choice of productThe choice of factor VIII product varied withtime. Different factor VIII products were used:cryoprecipitate and intermediate purified product,with and without vonWillebrand factor, aswell as monoclonal-purified and recombinantfactor VIII.PatientsPatients are considered to have a type Ainhibitor when recovery of 50% or less is measured,with or without the clinical evidence of aninhibitor. Patients are considered to have transient(type B) inhibitors when the second sampletested negative for antibodies and a normalrecovery was found. 13 Type B inhibitors wereexcluded from the study. Twenty-seven patientswith persistent antibodies were included into thestudy.Informed consentFrom all patients informed consent wasobtained.Laboratory AssaysPlasma sampling. Plasma samples for factor VIIIand inhibitor assays were collected according tostandard techniques; 4.5 mL of venous blood wasdrawn with a disposable needle into a siliconetreated Vacutainer in which 0.5 mL of 3.8%(0.13M) sodium citrate was added. Immediateafter collection, samples were carefully mixedand centrifuged at 3000 × g for 15 minutes at4°C. The platelet poor plasma was carefullypipetted off and stored in a plastic tube at minus20°C. All samples were analyzed at the coagulationlaboratory of the University Medical CenterUtrecht (head Prof. dr. JWN Akkerman).Inhibitor assayInhibitor measurements were performed usingthe Bethesda method as described by Kasper etal. 14 Inhibitor titers of 1 Bethesda Unit per milliliter(BU/mL) or more were considered positive. 5After 1997 inhibitor measurement were performedusing the modified Nijmegen assay. 15Using this method, inhibitors higher than 0.3BU/mL are considered positive.Stored blood samples from all patients tolerizedbefore 1997 were tested for inhibitors usingthe Njmegen method. The results were comparablewith those using the original Bethesdainhibitor assay.In patients with positive inhibitor tests, bloodsamples for inhibitor measurement were subsequentlytaken every 4-8 weeks.Factor VIII assayFactor VIII assays were performed using the onestage method based on the kaolin-activated partialprothrombin time and expressed as a percentageof factor VIII present in pooled humanplasma. 16In vivo recoveryBlood samples for factor VIII assays were takenbefore and 15 minutes after infusion with factorVIII. Recovery was defined as the percentageof factor VIII measured and the expected levelcalculated by the method according to Lee et al. 17Definition of successOriginally immune tolerance was consideredto be clinically successful when the inhibitordecreased to < 2 BU/mL, with a factor VIII recoveryof at least 50% of normal and a half-life of 6hrs or more and the absence of an anamnesticresponse after infusion with factor VIII. Since1997, after the introduction of the inhibitormeasurement according to the modifiedNijmegen method, an inhibitor of less than 0.4BU/mL was taken as cut-off. Clinical success waschosen as an endpoint because these patients canbe treated with prophylaxis to prevent bleeds, andbleeds can be treated adequately with factor VIII.Complete success was defined as absence ofinhibitor, normal recovery and half-life time.Follow-upAfter start of ITT patients were seen at leastevery month. During follow-up visits samples forantibody test were taken. When the inhibitor wasless than 2 BU/mL recoveries were also performedafter infusion with the dosage factor VIIIa patient was currently using for ITT. When theinhibitor was less than 1 BU/mL (Kasper) or lessthan 0.4 BU/mL in the Nijmegen assay, and arecovery of more than 50% was measured, halflifewas performed after infusion with 50 U/kgbw. Once a patient was tolerized, blood samplesfor inhibitor measurement and recovery weretaken at least twice a year. The date of the lastinhibitor assay was taken as the end-point forevaluation.Treatment of bleeds during ITTBleeds in patients with active inhibitors weretreated with 50 U/kg bw APCC (FEIBA) or 90 µgfactor VIIa (NOVOSEVEN). When a patient hada factor VIII recovery, bleeds were treated with(increased) doses of factor VIII. Infusion withclotting factor was repeated depending on theclinical situation of the patient.Statistical analysisProbabilities of disappearance of the inhibitorover time were estimated with the product limitof Kaplan and Meier and were compared usingthe log-rank statistic. The time lapse until disappearanceof the inhibitor was also examined byunivariate stepwise Cox regression analyses. Allvariables found to have p values of less than 0.10in univariate were considered candidate variablesfor multivariate analyses.haematologica vol. 88(supplement n. 12):september 2003

36E.P. Mauser-Bunschoten et al.ResultsThe group consisted of 27 patients with severehemophilia A. Patient data are summarized inTables 1 and 2. The median age at inhibitordevelopment was 3 years. Inhibitors developedafter a median of 34 exposures. The median ageat start immune tolerance therapy was 13 years.The total period of follow-up since start ITT was336 years with a median of 13 years per patient.Two patients were lost for follow-up, and 3patients died from Aids.In patient 1-11 factor VIII was continued afteran inhibitor developed, whereas in patient 12 –27 factor VIII was discontinued for at least oneyear.During immune tolerance induction patientswere seen at least every month. We checked thediaries kept by the patients against the amountsof factor VIII supplied to them. Based on thesedata compliance was almost 100%.In 21 patients (78%) success was obtainedwith low dose therapy, in 4 patients dose had tobe adjusted to 75-100 U FVIII/kg bw 3 times aweek before success was seen. They were consideredto be failures. In 2 other patients therapyfailed completely.Success was obtained after 2–28 months. TheKaplan Meier plot of the presence of inhibitor isalmost linear in the first two years of ITT, indicatinga constant chance of disappearance ofinhibitor. Even after 3 years of therapy there is achance the inhibitor will disappear. So farimmune tolerance therapy totally failed in twopatients after 48 and 50 months respectively. Inthese two patients therapy failed even when therapywith high dose factor VIII, intravenous gammaglobulinand cyclophosphamide as describedby Nilsson et al. 10 was given.Logistic regression analyses revealed (Table 3)that there existed a relation between the highestinhibitor level and successful ITT and the timeneeded before success was obtained. Patientswith low inhibitor titers did better. All patientswith an inhibitor level of less than 40 BU/mLwere treated successfully. Patients in whom therapycompletely failed or in whom the dose wasadjusted had inhibitors between 44 BU/mL and753 BU/mL. In patients with maximum inhibitortiter of less than 40 BU/mL success was obtainedwithin 10 months (median 6 months). Inpatients with maximum titers over 40 BU/mLthe median time before success was obtained was18 months , but even after 36 months ITT wassuccessful.Whether therapy was started directly or manyyears after the inhibitor development did notseem to make much difference in this group ofpatients. Treatment with a neutralizing dose atstart of immune tolerance induction did notimprove the results. Furthermore the type ofproduct used to obtain immune tolerance did notaffect the results. Tolerance was also obtainedusing monoclonal purified and recombinantproduct (Table 2).Table 1. Demographic data of patients on immune tolerancetherapy.medianrangeAge at inhibitor development (years) 3 0,5 – 23Number of exposures before inhibitor development 34 8 – 53Age at start ITT (years) 13 1 – 43Period of follow-up (years) 13 4 – 20ComplicationsThe most frequent complication during ITT wasthe occurrence of bleeds. In the beginning bleedswere treated with APCC, and since 1998 withfactor VIIa. When the inhibitor was low and afactor VIII recovery was measured bleeds weretreated with factor VIII. Venous access was problematicin some small children. Therefore in fivepatients Port-a-Cath systems (PAC) wereimplanted at start of ITT when the inhibitor waslow or during ITT under coverage with increaseddose of factor VIII, porcine factor VIII, APCC orcontinuous infusion with factor VIIa. During ITTwe have seen in 3 patient 6 PAC infections. FourPACs were replaced for this reason.Until 1985 seven of 12 patients treated withITT were infected with HIV, 3 of them died fromAids related diseases.In one patient (patient 4) a relapse of theinhibitor was observed one year after he wastolerized. During this period a maximuminhibitor titer of 1 BU/mL was found, withabsence of factor VIII recovery. This patient wastreated with a second course of low dose ITT withgood result. Ten years later he has had no secondrelapse, is on normal dose prophylaxis and hasno spontaneous bleeds.Present statusAt their last visit (Table 4) all patients weretreated with prophylaxis, in most patients bleedsare prevented adequately. In one patient withsevere arthropathy high bleeding frequency isobserved, which probably is caused by the poorphysical condition of this patient. Also in youngpatients with risk behavior a higher bleeding frequencyis seen. However these patient do not sufferfrom spontaneous bleeds.In one patient prophylaxis was stopped duringa period of 10 years without recurrence of theinhibitor. After successful ITT 27 surgical interventionswere performed in 12 patients under factorVIII coverage. Bolus injection as well as continuousinfusion was used with good hemostaticeffect. In none of the patients post operativebleeding occurred.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia37Table 2. Course of inhibitor before, during and after immune tolerance therapy (ITT).Pat. 1 st inhibitor Highest inhibitor level Inhibitor level Highest anamnestic Time needed Product used Follow-upin BU/mL before ITT at onset ITT response at ITT for success for ITT since start ITTin BU/mL in BU/mL in BU/mL (months) (years)1 * 1.3 8.7 8.7 None 6 Intermediate purified 152 * 1.1 1.1 1.1 None 2 Cryoprecipita-te 183 # ,** 8.1 8.1 4.0 11 6 Intermediate purified conc 164 ** 2.5 2.2 2.2 68 24 Intermediate purified conc. 165 ** 3.9 3.9 3.9 None 3 Cryoprecipita-te 196 * 4.2 4.5 4.5 4.6 8 Intermediate purified conc. 157 * 1.1 1.3 1.3 11.0 6 Interm purified conc with vW 88 ** 9.7 9.7 9.7 24 6 Intermediate purified conc 109 ** 8.0 8.0 8.0 220 36 (dose adjusted) Intermediate and mono-clonal conc 1010 * 2.0 25 5.0 753 24 (dose adjusted) Recombinant 511* 5.0 5.0 5.0 None 6 Recombinant 412* Pos 23 0.6 2.3 6 Monoclonal 413 #,* Pos 177 0.4 17 12 (dose adjusted) Intermediate purified conc. 1814 #,* Pos 160 4.6 66 27 (dose adjusted) Intermedied purified conc 1115 #,* Pos 3.5 0.7 3.3 10 Intermediate purified conc. 1616 #,* Pos 94 2.8 34 12 Intermediate purified conc. 617 #,* Pos 15.5 2.5 None 6 Intermediate purified conc 2018 * Pos 15.6 1.0 2.7 9 Intermediate purified 2019 #,* Pos 164 1.2 1.6 14 Intermediate purified 1620 #,* 52 76 1.7 8.5 18 Intermediate purified 1821 #,* Pos 83 5.7 83 18 Intermediate purified conc. 1522 * 8.0 10.4 0.3 None 3 Cryoprecipitate 1923 * 5.0 64 7.3 31 28 Intermediate purified 1324 #,* 3.0 3.0 1.1 None 2 Intermediate purified conc. 425 #,* 4.3 51 1.5 19.3 8 Monoclonal purified 826 #,* Pos 33 2.8 450 Malmö protocol failure Intermediate purified and monoclonal 527 * Pos 25 0.6 44 Malmö protocol failure Intermediate purified and monoclonal 7#= initial high doses of factor VIII to neutralise the antibodies, followed by twice daily 25 U/kg bw factor VIII for 1 –2 week; *= Factor VIII 25-50 U/kg bw 3 times a week orevery other day; **= Factor VIII 25-50 U/kg bw 2 times a week.DiscussionLow dose immune tolerance therapy is successful.The duration of treatment is determined bythe maximum inhibitor level. Within 10 monthsafter start of ITT 100% success was seen inpatients with a maximum inhibitor titer of lessthan 40 BU/mL. 5, 18 Also in patients with a hightiter inhibitor success was obtained. However inthese patients time needed for success waslonger. But even after more than 24 months oftreatment success could be obtained.Low dose ITT is less demanding for a patientand his parents. A disadvantage of low dose ITTmay be the longer period of time needed beforetolerance is obtained as compared to the ITT regimensthat use daily high dose FVIII. 19 This maybe a problem for patients with a high bleedingtendency or for patients in whom surgery isrequired.Since the introduction of ITT in the beginningof the 1980s success has been obtained with differentdosage regimens. Mostly 100 U FVIII perday are given, with or without the addition ofcorticosteroids, gammaglobulin or plasmapheresis.4-10 Success rates varied between 76 and 89%and success was obtained after 1 – 18 months. Insome studies doses were compared. Haya et alfound that doses < 100 U FVIII/kg bw per daydid better than higher doses. 20 But in general,success rates in various studies are the same, onlythe median time to success is different. 19 However,definitive conclusions about the scheduleof preference in relation to the inhibitor titer canonly be drawn after prospective, randomized controlledstudies. Hopefully the study as proposedby Hay et al will give an answer to this. 21haematologica vol. 88(supplement n. 12):september 2003

38E.P. Mauser-Bunschoten et al.Table 4. Present prophylactic regimen, bleeding frequency, last recovery and half-life time in immune tolerized patients.Pat. Year of birth Prophylactic dose Number of Number of Last inhibitor titer Recovery (percentage) Half-life time (hours)(U FVIII/kg bw) bleeds per year surgical (BU/mL)interventions1 1975 † 1996 0 Neg 80% 12 hrs2 1982 10 U/kg bw 2× week 0-5 0 Neg (0.0) 100% 14 hrs3 1983 15 U/kg bw 3× week 0-5 0 Neg (0.0) 110% 15 hrs4 1984 10-15 U/kg bw 3× week 0-5 1 Neg (0.0) 110% 18 hrs5 1979 10 U/kg bw 3×week 0-5 0 Neg (0.1) 100% 12 hrs6 1984 10-15 U/kg bw 3×week 5-10 0 Neg (0.0) 100% 12 hrs7 1993 25 U/kg bw 3×week 5-10 1 Neg (0.1) 100% 10 hrs8 1989 20 U/kg bw 3×week 5-10 3 Neg (0.1) 110% 6 hrs9 1985 15 U/kg bw 2×week 0-5 0 Neg (0.0) 80% 8 hrs10 1995 25 U/kg bw 3×week 0-5 2 Neg (0.0) 100% 8 hrs11 1996 25 U/kg bw 3×week 0-5 1 Neg (0.1) 100% 6 hrs12 1967 † 1993 1 Neg 100% 12 hrs13 1955 20 U/kg bw 3×week 10-25 5 Neg (0.0) 120% 12 hrs14 1967 † 1993 3 Neg 100% 12 hrs15 1964 10-15 U/kg bw 2×week 0-5 0 Neg (0.0) 120% 12 hrs16 1967 1990 lost for follow-up 0 Neg 60% 7 hrs17 1967 15 U/kg bw 3×week 0-5 1 Neg (0.0) 130% 14 hrs18 1967 10 U/kg bw every other day 0-5 0 Neg (0.0) 90% 12 hrs19 1971 10 U/kg bw 3×week 0-5 0 Neg (0.0) 100% 19 hrs20 1971 10 U/kg bw 3×week 0-5 5 Neg (0.0) 80% 10 hrs21 1975 10-15 U/kg bw 3×week 0-5 0 Neg (0.0) 110% 10 hrs22 1976 10-15 U/kg bw 3×week 0-5 1 Neg (0.0) 100% 8 hrs23 1967 15-20 u/kg bw 3×week 0-5 0 Neg (0.2) 140% 6 hrs24 1974 1985 lost for follow-up 0 Neg 100% 12 hrs25 1987 25 U/kg bw 3×week 0-5 3 Neg (0.0) 100% 9 hrsTable 3. Success of low dose ITT in relation to inhibitor titer.Max titer < 40 BU/mLMax titer > 40 BU/mLNumber of patients 15 12Success 15 6Success after dose adjustment 2Failure 0 2Time needed before success(months)median 6 18range 2-10 12-28ConclusionLow dose ITT (25-50 u FVIII/kg bw 3 × week)is appropriate for patients with low titerinhibitors, for young children in whom aninhibitor just has developed and for patients withhigh titer inhibitors who have a low bleeding frequency.Also for patients in whom ITT has notbeen attempted due to the high costs of otherschedules low dose ITT may be indicated. 18References1. Bloom AL. The treatment of factor VIII inhibitors.Thromb Haemost 1987;58:447-71.2. Lusher JM. Management of patients with factor VIIIinhibitors. Transf Med Rev 1987;1:123-30.3. Brackman HH, Gormesen J. Massive factor VIII infusionin a haemophiliac patient with factor VIII inhibitor,high responder. Lancet 1977;2:933.4. Wensley RT, Burn AM, Redding OM. Induction of toleranceto factor VIII in hemophilia with inhibitors usinglow doses of factor VIII. Thromb Haemost 1985;54:227-30.5. Mauser-Bunschoten EP, Nieuwenhuis HK, RoosendaalR, van den Berg HM. Low-dose immune toleranceinduction in hemophilia A patients with inhibitors.Blood 1995;86:983-8.6. Ewing NP, Sanders NL, Dietrich SL, Kasper CK. Inductionof immune tolerance to factor VIII in hemophiliacswith inhibitors. JAMA 1988;259:65-8.7. Oldenburg J, Schwaab R, Brackmann HH. Induction ofimmune tolerance in haemophilia A inhibitor patientsby the “Bonn protocol”: Predictive parameter for therapyduration and outcome. Vox Sang 1999;77 Suppl1:49-54.8. Kreuz W, Mentzer D, Auerswald G. Successful immunetolerance therapy of FVIII-inhibitor in children afterchanging from high to intermediate purified FVIII concentrate.Haemophilia 1996;2 Suppl 1:19a[abstract].haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia399. Aznar JA, Jorquera JI, Peiro A, Garcia I. The importanceof corticoids added to continued treatment with factorVIII concentrates in the suppression of inhibitors inhaemopihilia A. Thromb Haemost 1984;51:217-21.10. Nilsson IM, Berntorp E, Zetterval O. Induction ofimmune tolerance to factor VIII in hemophiliacs withinhibitors, N Eng J Med 1988;318:947-50.11. Van Leeuwen EF, Mauser-Bunschoten EP, van DijkenPJ, Kok AJ, Sjamsoedin-Visser EJ, Sixma JJ. Low doseimmune tolerance induction in hemophilia A patientswith inhibitors. Br J Haematol 1986;64:291-7.12. Van den Berg HM, Fischer K, Mauser-Bunschoten EP,Beek FJA, Roosendaal G, van der Bom JG, et al. Longtermoutcome of individulized prophylactic treatment ofchildren with severe haemophilia. Br J Haematol 2001;112:561-5.13. Rosendaal FR, Nieuwenhuis HK, van den Berg HM, HeijboerH, Mauser-Bunschoten EP, van der Meer J, et al.A sudden increase in factor VIII development in multitransfusedhemophilia A patients in the Netherlands.Blood 1993;81:2180-6.14. Kasper CK, Aledort LM, Counts RB, Edson JG, FratnatoniJ, Green D, et al. A more uniform measurement offactor VIII inhibitors. Thromb Diath Haemorrh 1975;43: 469.15. Verbruggen B, Novakova I, Wessels H, Boezeman J, vanden Berg M, Mauser-Bunschoten E. The Nijmegen modificationof the Bethesda assay for factor VIII:cinhibitors: Improved specificity and reliability. ThrombHaemost 1995;73:247-51.16. Bouma BN, Starkenborg AE. Dilution of haemophilicplasma used as reagent in the determination of antihaemophilicfactor A (factor VIII). Haemostasis 1974;3:94-7.17. Lee ML, Gomperts ED, Kingdom HS. A note on the calculationof recovery for factor VIII infusions. ThrombHaemost 1993; 69:87.18. El Alfy MS, Tnatawy AAG, Ahmed MH, Abdin IA. Frequencyof inhibitor development in severe haemophiliaA children treated with cryoprecipitate and low-doseimmune tolerance induction. Haemophilia 2000; 6:635-8.19. Dimichele DM. Immune tolerance: a synopsis of internationalexperience. Haemophilia 1998; 4:568-73.20. Haya S, Lopez MF, Aznar JA, Battle J. Immune tolerancetreatment in haemophilia patients with inhibitors: theSpanish registry. The Spanish immune tolerance group.Haemophilia 2001; 7:154-9.21. Hay CR. Immune tolerance induction: prospective trials.Haematologica 2000; 85 Suppl:52-6.haematologica vol. 88(supplement n. 12):september 2003

[Round Table on Immune Tolerance Treatment]review paperPossible advantages ofprotocols with a waiting phaseto treat inhibitors againstfactor VIIIhaematologica 2003; 88(suppl. n. 12):40-41http://www.haematologica.org/free/immunotolerance2001.pdfR. KOBELTHämophiliezentrum, Wabern and Childrens Hospital of theUniversity of Berne, SwitzerlandAfter the detection of an inhibitor in apatient with hemophilia A , the immediatestart of an immune tolerance induction(ITI) is usually recommended. 1 This can be difficultto carry out for smaller centers, lacking therequired experience with this type of treatment.The registry of the Immune Tolerance Study Group(ITSG) shows that about 40% of these patientshave, in fact, only been placed on ITI after alonger delay. No details are known about thetherapeutic measures used in the time of delay. 2The question will be discussed of how an adequateprotocol could not only keep the patientsin good health during this waiting phase but perhapsmight even be of some advantage for them.Retrospective evaluations from ITI registrieshave shown a number of variables positivelyaffecting the duration and the outcome of ITI.Some of them are:— low inhibitor titer at the onset of ITI; 2-4— no repeated boosting of the inhibitor or lowhistorical inhibitor titer; 2,3,4— no interruptions of ITI; 5— High-dose ITI-regimen. 2These variables can be influenced by an appropriatetreatment protocol: early detection of aninhibitor by frequent testing during the first daysof treatment. This prevents further boostingfrom being given, which could, at least in somepatients, lead to increasingly high titers of theinhibitor. Therefore, in case of a positive result,the administration of standard factor concentrateshas to be immediately stopped and treatmenton demand with recombinant factor VIIa(rFVlla) has to be started. This product usuallyshows a very good efficacy in the type of bleedsseen in young patients mostly affected byCorrespondence: R. Kobelt, Hämophiliezentrum, Seftigenstrasse240, CH-3084 Wabern and Childrens Hospital of the Universityof Berne, Switzerland.inhibitors. 6-8 This waiting phase (delay of ITI)allow transient inhibitors to be excluded and theoptimal moment to start the ITI to be reached:the often difficult venous access of the childrencan be improved, and if necessary a port-systemcan be installed before starting ITI. All personsinvolved, including the family of the patient, canbe thorougly prepared for the treatment. 9,10Finally, the inhibitor can be allowed to fall,hopefully to a titer of 10 BU or less, before treatmentwith a high-dose ITI-regimen according tothe Bonn-protocol is initiated.This protocol is intended to optimize ITI bypreventing further boosting of the inhibitor titerand by delaying the start until an uninterruptedcourse can be expected. Starting at a low level ofthe inhibitor probably additionally improves theoutcome of the treatment and might optimizethe costs by reducing the duration and thus theconsumed quantity of FVIII product. 7,11 To datethere are only a few documented cases treatedaccording to such a protocol. 8,12 To prove thishypothesis more cases would be necessary.References1. Brackmann HH, Lenk H, Scharrer I, Auerswald G, KreuzW. German recommendations for immune tolerancetherapy in type A haemophiliacs with antibodies. Haemophilia1999; 5:203-6.2. Mariani G, Kroner B. Immune tolerance in hemophiliawith factor VIII inhibitor: predictors of success.Haematologica 2001; 86:1186-93.3. Lenk H anf the ITT Study Group. The German registryof immune tolerance treatment in haemophilia - 1999update. Haematologica 2000; 85(suppl to n. 10):45-7.4. DiMichele D, Kroner BL. Analysis of the North-AmericanImmune Tolerance Registry (NAITR) 1993-1997:current practice implications. Vox Sang 1999; 77(suppl1):31-2.5. Oldenburg J, Schwaab R, Brackmann HH. Induction ofimmune tolerance in haemophilia A inhibitor patientsby the 'Bonn Protocol': predictive parameters for therapyduration and outcome. Vox Sang 1999; (Suppl.1):49-54.6. Key NS, Aledort LM, Beardsley D, et al. Home treatmentof mild to moderate bleeding episodes usingrecombinant factor Vlla (NovoSeven) in haemophiliacswith inhibitor. Thromb Haemost 1998; 80:912-8.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia417. Ekert H, Brewin T, Boey W, Davey P, Zjlden D. Cost-utjlityanaJysis of recombinant factor Vlla (NovoSeven) insix children with long-standing jnhibjtors to factor VIIIor IX. Haemophilia 1999; 7:279-286.8. Brackmann HH, Effenberger E, Hess L, Schwaab R, OldenburgJ. NovoSeven in immune tolerance therapy.Blood Coagul Fibrinolysis 2000; 11 suppl 1:S39-44.9. Butler RB. Immune tolerance: a nursing perspective.Haematologica 2000; suppl to n. 10:78-80.10. Gargallo B. Immune tolerance: the parents perspective.Haematologica 2000; suppl to n. 10:81-2.11. Aledort LM. Factor VIII inhibitors. Immune toleranceinductjon: is it cost effective? We know too little. SeminThromb Hemost 2000; 26:189-93.12. Kobelt R. A Swiss Treatment Concept for the Use of rFVllain the Context of Immune Tolerance Therapy. Recombinantfactor Vlla, current and future indications, I.Scharrer, M. von Depka (eds.) 2000:50- 7.haematologica vol. 88(supplement n. 12):september 2003

[Round Table on Immune Tolerance Treatment]High infection rate in portsused for immune tolerance.The Spanish Registryreview paperhaematologica 2003; 88(suppl. n. 12):42-44http://www.haematologica.org/free/immunotolerance2001.pdfJ. TUSELL, A. VILLAR, MF. LOPEZ-FERNANDEZ, S. HAYA,D. BRITO, R. SOSA, F. RODRIGUEZ-MARTORELLSpanish Ports Study Group, SpainThere is no doubt that port-a-caths (ports)are an extremely useful tool in the currenttreatment of hemophilia. Many hemophiliacchildren with problems of venous access areundergoing a home prophylactic treatment programmeor have completed immune toleranceinduction (ITI) treatment thanks to thesedevices. Their use is widespread in a number ofcountries. However, they are not free of certaincomplications, of which the most frequent areinfections and thromboses.In Spain, port use in hemophilia has most frequentlybeen for ITI programmes in patients withinhibitors. The results presented here, from theSpanish Registry, are for these pediatric patients,in whom a high infection rate has been observed.AimsTo evaluate port use for facilitating ITI treatmentto factor VIII in a pediatric population ofhemophiliacs with inhibitors. To evaluate thepresence of infections and their potential influcencein achieving tolerance.MethodsCompletion of a questionnaire on port insertion,use, and associated complications. Thequestionnaire is distributed to all treatment centresin the country for patients with bleeding disorders.It is completed when a port is inserted,after which it is updated yearly.ResultsThirty ports have been used in 21 severehemophilia A patients with inhibitors, in orderto carry out ITI treatment. Nine patients requireda second port. Insertion was at an average age of4.2, ranging from 16 months to 16 years. Theinhibitor titer at the moment of insertion wasbetween 1 and 789 BU, with an average of 92BU. Factor VIIa was used for surgical coverage in16 cases, in a further 8 cases factor VIII was usedand in 6 other cases activated prothrombincomplex concentrates (aPCC) were used. On 22occasions the ports were inserted in the subclavianvein, 6 were inserted in the external jugular,and 2 more in the saphenous. There wereone or more infections in 20 out of the 20 ports(66%). The rate of first infection per 1000 daysof use and per port was 2.24. The most frequentgerms, on order of frequency, were Staphylococcusepidermidis, Staphylococcus aureus, Pseudomonas,Klebsiella, Proyteus and Escherichia Coli.In nine cases, removal and replacement of theport was required due to the persistence of theinfection, despite local and systemic antibiotictreatment. Two of the second ports also becameinfected. Of the 19 patients who completed theITI treatment, 14 successfully achieved tolerancewhile the other 5 failed. Out of the 19 cases, theport presented infections in 9 while 10 remainedinfection free. Of the 5 cases in which tolerancewas not achieved, 4 presented infections. Of the10 cases free from infection, tolerance wasachieved in 9 and there was one failure (Table 1).DiscussionThere is a wide range of results (expressed inpercentage of infected ports) in relation to portinfections in hemophilia, varying from 10 to60%, and in the infection rate per 1000 days ofport use, varying from 0.14 to 0.76 (Table 2). 1–15It must be born in mind that some studies areTable 1. Success and failure of ITI in relation to the presenceof infections in ports.Patient (%) Infection (%) No infection (%)Tolerance 14 (74) 5 (36) 9 (64)No tolerance 5 (26) 4 (80) 1 (20)Completed ITI 19 9 10haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia43Table 2. Percentage of port infections and infection rateper 1000 days of use, as published in literature.Table 3. Published percentage of infected ports used inhemophilic patients with inhibitor.% Rate%Rothschild 1 10 0.15Miller 2 13 0.14Ljung 3 17 0.19Manco-Johnson 4 22_Perkins 5 33 0.7Collins 6 34_Ragni 7 38_Warrier 8 40_Shapiro 9 45 0.26Geraghty 10 45_Blanchette 11 47 0.7Liesner 12 48 0.67Bollard 13 48_Santagostino 14 50 0.73Tusell 15 60 0.76Ljung 17 50Ragni 7 52Collins 6 58Tusell* 66Ljung 18 71Liesner 12 83Average 63Average 36 0.46still at a preliminary stage, so the number ofpatients and ports are low and the follow-upperiod short, causing values to vary over time andwith the addition of more patients. The resultsshould therefore be interpreted with caution. Onsome occasions, as in the Spanis register, the highincidence is due to port being used for ITI inpatients with inhibitors. It is well known that inthese cases there is a much higher infection rate,probably because treatment involves greater useof the port and there is no normalization inhemostasis. This leads to frequent hematomaswhere the skin is punctured, which, in turn,facilitates the presence of infections at this siteand the entrance of germs into the system. 16There is also a wide range of results (50-83%) inthe much higher percentages of infections inthese cases (Table 3).Although the serious consequences of theseinfections have not been described, they mustnot be underestimated, as they are bacterialprocesses, some of which are potentially very seriousgram-negative germs, 19 resistant to local andsystemic treatment, and, in many cases, requiringthe removal and replacement of the port.The role of infection in inhibitor developmentand their influence in the results of ITI are a subjectof some debate. 20 In the Spanish register, ofthe 5 patients who did not achieve tolerance, 4had presented infections in the port-a-cath, andof the 10 who did not suffer infections in theports, 9 achieved tolerance. Although the numberof cases is low, a negative effect of infectionson the results of immune tolerance can still bedetected. The inhibitor titer before and afterinfections may also be worth observing. Thrombosisrepresents a further complication withports and one that is of growing concern.Although no cases have yet been referred to theSpanish register, a prevalence of up to 18% 21 hasbeen published and there is a general increase incases, particularly when the period of use is prolongedfor more than 4 years.ConclusionsThere is no doubt that port-a-caths are a genuinelyefficacious tool in current hemophiliatreatment and especially in ITI. However, thehigh rate of complications - particularly infections- associated with port use in patients withinhibitors, as observed in the Spanish and otherstudies, needs to be evaluated. Infections maynegatively influence the results of ITI and someof them are potentially serious. It may be wise toexhaust the possibilities of administration viaperipheral veins before taking a decision onwhether to implant a port for ITI. Should implantationbe required then the most thorough antisepticmeasures must be taken.References1. Rotschild C, Mercks J, Torchet MF, Pertuiset N. Experiencewih implantable venous access devices inhaemophilia children treated intensively: a low rate ofcomplications. Haemophilia 2000; 6:276.2. Miller K, Buchanan GR, Zappa S, Cochran C, LaufenbergJ, Medeiros D, et al. Implantable venous accessdevices in children with hemophilia: a report of lowinfection rates. J Pediatr 1998; 132/934-8.3. Ljung R, van den Berg M, Petrini P, Tengborn L, ScheibelE, Kekomaki R, et al. Port-A-Cath usage in children withhaemophilia: experience of 53 cases. Acta Paediatr1998; 87:1051-4.4. Manco-Johnson M, Nuss R, Abshire T, Richardson L,Shapiro A, Valentinos L. Port infections in children withhaemophilia. The Hemophilia Joint Preservation StudyGroup. Haemophilia. 2000; 6:12a[abstract].5. Perkins JL, Johnson VA, Osip JM, Christie BA, Nelsonhaematologica vol. 88(supplement n. 12):september 2003

44J. Tusell et al.SC, Moertel CL, et al. The use of implantable venousaccess devices (IVADs) in children with hemophilia. JPediatr Hematol Oncol 1997; 19:339-44.6. Collins PW, Khair KS, Liesner R, Hann IM. Complicationsexperienced with central venous catheters in childrenwith congenital bleeding disorders. Br J Haematol1997; 99:206-8.7. Ragni MV, Hord JD, Blatt J. Central venous catheterinfection in haemophiliacs undergoing prophylaxis orimmune tolerance with clotting factor concentrates.Haemophilia 1997; 3:90-5.8. Warrior I, Baird-Cox K, Lusher J Use of central venouscatheters in children with haemophilia: one haemophiliatreatment centre experience. Haemophilia 1997;3:194-8.9. Shapiro AD, Donfeld SM. Infectious complications ofcentral venous access devices (CVAD) in children withhaemophilia. Haemophilia 2000; 6:278.10. Geraghty S, Kleinert D. Additional data on the morbidityof central venous access devices in patients withhaemophilia. Haemophilia 1998; 4:66.11. Blanchette VS, Al-Musa A, Stain AM, Ingram J, Fille RM.Central venous access devices in children with hemophilia:an update. Blood Coagul Fibrinolysis 1997; Suppl1:S11-4.12. Liesner RJ, Vora AJ, Hann IM, Lilleymann JS. Use of centralvenous catheters in children with severe congenitalcoagulopathy. Br J Haematol 1995; 91:203-7.13. Bollard CM, Teague LR, Berry EW, Ockelford PA. Theuse of central venous catheters (portacaths) in childrenwith haemophilia. Haemophilia 2000; 6:66-70.14. Santagostino E, Muca-Perga M, Coluccia P, BernardinelliL, Mannucci PM. Study of long-term safety of aport-a-cath use in children with haemophilia: whichalternative? Haemophilia 2000; 6:272.15. Tusell J, Villar A, Lopez MF, Aznar JA, Brito D, Sosa R, etal. High index of port-a-cath (PAC) infections in childrenwith congenital coagulopathies. Spanish Registry.Pediatr Res 2001; 6:883.16. Van Den Berg HM, Fischer K, Roosendaal G, Mauser-Bunschoten EP. The use of the Port-A-Cath in childrenwith haemophilia: a review. Haemophilia 1998; 4:418-20.17. Ljung R, Petrini P, Lingreen AK, Tengborn L. The longtermfeasibility of venous access in children with severhaemophilia. Semin Hematol 1994; Suppl 2:16-8.18. Ljung R, Van Den Berg HM, Petrini P, Tengborn L,Scheibel E, Effenberger W. Experience with the port-acathin children with severe hemophilia A: a multicentrestudy. Haemophilia 1996; Suppl 1:17.19. Hothi DK, Kelsall W, Baglin T, Williams DM. Bacterialendocarditis in a child with haemophilia B: risks of centralvenous catheters. Haemophilia 2001; 7:507-10.20. Yamamoto K, Niiya K, Shigematu T, Kiguchi T, TakenakaK, Shinagawa K, et al. Transient factor VIII inhibitorin a hemophilia patient after staphylococcal septic shocksyndrome. Int J Hematol 2000; 72:517-9.21. Journeycake JM, Quinn CT, Miller KL, Zajac JL,Buchanan GR. Catheter-related deep venous thrombosisin children with hemophilia. Blood 2001; 98:1727-31.haematologica vol. 88(supplement n. 12):september 2003

[Round Table on Immune Tolerance Treatment]review paperThe Malmö International BrotherStudy (MIBS) – An updatehaematologica 2003; 88(suppl. n. 12):45-47http://www.haematologica.org/free/immunotolerance2001.pdfJAN ASTERMARK, ERIK BERNTORP AND THE MIBS STUDY GROUP*Department for Coagulation Disorders, University of Lund,Malmö, SwedenThe Malmö International Brother Study (MIBS) wasinitiated in 1996 to study the issue of predisposingfactors to inhibitor development in hemophiliapatients. As of August 2001, 528 families haveaccrued, of whom 451 suffer from hemophilia A and77 hemophilia B. Twenty-five of the brother pairs aretwins. The inhibitor incidence in all families withsevere hemophilia A was 29.8%. In 36 of the 110inhibitor families (32.7%), at least two brothers hada history of inhibitors. In 24 of these families, theinhibitor was also of the same type, i.e. either highorlow-responding. The overall concordance within249 severe hemophilia A families was found to be78.3% compared to an expected figure of 68.0%and 58.0% using an inhibitor incidence of 20 and30%, respectively (p5 BU and a low-respondinginhibitor was one with a peak titer of ≤ 5 BU.Severe hemophilia A corresponds to a factor VIIIclotting activity of 5%, respectively.Statisticsχ 2 analysis was used for statistical evaluationof expected versus observed frequencies ofinhibitors. An expected number of inhibitor subjectsof at least five was required in each group.A p value below 0.05 was considered to indicatestatistical significance.haematologica vol. 88(supplement n. 12):september 2003

46The MIBS registryTable 1. Characteristics ofthe 528 brother pairs.Table 2. Incidence of inhibitor families stratified accordingto type of hemophilia and ethnicity. The number of familiesin each subgroup is shown in Table 1.ResultsAs of August 2001, data concerning 1,108 subjectsin 528 families had been accrued (451hemophilia A and 77 hemophilia B) from 27hemophilia centers in Europe and North America.The characteristics of these families areshown in Table 1. Twenty-five are twins of whommonozygosity was confirmed in 15 pairs. Twenty-oneof the twins suffer from hemophilia A (17severe and 4 mild) and 4 from hemophilia B (2severe and 2 mild).A history of inhibitors was described in 13.3%of all subjects and in 110 families. Eighty-five ofthese families were of Caucasian origin as shownin Table 2. The inhibitor incidence in all severehemophilia A individuals was 20.6% correspondingto 29.8% of the families.The distribution of low- and high-respondinginhibitors in each family is shown in Table 3. Ahistory of inhibitor was reported in more thanone sibling in 36 of the 110 inhibitor families. In24 of these 36 families the same type of inhibitorresponse was reported. Six of the 25 twins reporteda history of inhibitor. Three of these were confirmedmonozygotic twins, of whom both brothershad a history of a high-responding inhibitorin two cases. In the remaining discordant pair,one brother was a high-responder whereas theother had no history of an inhibitor.The observed inhibitor history within 249 ofthe families with two siblings suffering fromsevere hemophilia was significantly concordantusing expected inhibitor incidences of 20 and30%, (Table 4), corresponding to an overall concordanceof 78.3%. 13 The corresponding figurefor the small group of 17 twins with severehemophilia A was 88.2%. The risk of a siblingwith hemophilia developing antibodies if an olderbrother had a history of inhibitor was found tobe 48% (95% CI 35-62%), whereas the risk inthe case of no previous known inhibitor in thefamily was only 15% (95% CI 11-21%) correspondingto a relative risk of 3.2 (95% CI 2.1-4.9).DiscussionCharacterization of genetic markers affectingthe risk of developing inhibitory antibodies is adifficult but highly warranted task. So far, themost significant factor has been the underlyinggenetic defect, in that large rearrangements havebeen associated with an increased incidence ofinhibitors. 5,6 However, even on this matter contradictorydata exist. 7 In our MIBS registry, wefound significantly more siblings, both twins andnon-twins, with an inhibitor than could beexplained by chance alone. This finding is inagreement with previous reports. 2–4 We alsofound that the risk of a sibling developing aninhibitor was three-times higher in families witha previous known inhibitor (CI 95% 2.1-4.9). Asin all retrospective studies of inhibitor incidence,some transient low-responding inhibitors mighthave been missed and the true number of patientswith inhibitors among the MIBS siblings is probablyunderestimated. The issue of why somepatients develop high- and others low-respondinginhibitors is also unresolved, but is of majorimportance since the clinical and economicimpact, as well as the outcome, is so different.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia47FundingThe study was supported by grants from theResearch Fund at Malmö University Hospital.ReferencesTable 3. The distribution of high- and low-respondinginhibitors in the 110 families with a history of inhibitors.Table 4. Observed vs. expected incidence of severe hemophiliaA families with inhibitors using a calculated inhibitorincidence of 20 and 30%, respectively. 13Comparing the genetic profile between siblingswith the same underlying factor VIII/IX genedefect but a different anamnestic response, mayhelp to address this issue.A study protocol to evaluate polymorphism(s)of genes known to be of importance for theimmune response has been developed and thecollection of samples within Europe will soon beinitiated. In addition, a large genetic screeningof families with hemophilia is being plannedtogether with the NCI and after the protocol hasbeen finalized at the end of this year, data andsample collection will start in the middle of nextyear. It is hoped that these studies will shed somelight on the intriguing issue of inhibitor development.1. Vermylen J. How do some haemophiliacs developinhibitors? Haemophilia 1998; 4:538-42.2. Frommel D, Allain JP. Genetic predisposition to developfactor VIII antibody in classic hemophilia. ClinImmunol Immunopathol 1977; 8:34-8.3. Shapiro SS. Genetic predisposition to inhibitor formation.Prog Clin Biol Res 1984; 150:45-55.4. Gill JC. The role of genetics in inhibitor formation.Thromb Haemost 1999; 82:500-4.5. Schwaab R, Brackmann HH, Meyer C, Seehafer J,Kirchgesser M, Haak A, et al. Haemophilia A: mutationtype determines risk of inhibitor formation. ThrombHaemost 1995; 74:1402-6.6. Tuddenhamn EG, Mcvey JH. The genetic basis ofinhibitor development in haemophilia A. Haemophilia1998; 4:543-5.7. Antonarakis SE, Rossiter JP, Young M, Horst J, de MoerlooseP, Sommer SS, et al. Factor VIII gene inversions insevere hemophilia A: Results of an international consortiumstudy. Blood 1995; 86:2206-12.8. Hay CR, Ollier W, Pepper L, Cumming A, Keeney S,Goodeve AC, et al. HLA class II profile: a weak determinantof factor VIII inhibitor development in severehaemophilia A. Thromb Haemost 1997; 77:234-7.9. Oldenburg J, Picard JK, Schwaab R, Brackmann HH,Tuddenham EG, Simpson E. HLA genotype of patientswith severe haemophilia A due to intron 22 inversionwith and without inhibitors of factor VIII. ThrombHaemost 1997; 77:238-42.10. Mayr WR, Lechner K, Niessner H, Pabinger-Fasching I.HLA-DR and Factor VIII antibodies in hemophilia A.Thromb Haemost 1984; 51:293.11. Aly AM, Aledort LM, Lee TD, Hoyer LW. Histocompatibilityantigen patterns in haemophilic patients with factorVIII antibodies. Br J Haematol 1990; 76:238-41.12. Lippert LE, Fisher LM, Schook LB. Relationship of majorhistocompatibility complex class II genes to inhibitorantibody formation in hemophilia A. Thromb Haemost1990; 64:564-8.13. Astermark J, Berntorp E, White GC, Kroner BL. TheMalmö International Brother Study (MIBS): furthersupport for genetic predisposition to inhibitor developmentin hemophilia patients. The MIBS Study Group.Haemophilia 2001; 7:267-72.14. Mariani G, Brackmann HH. Immune tolerance and thetreatment of haemophiliacs with an inhibitor. Vox Sang1996; 70 Suppl 1:1-80.haematologica vol. 88(supplement n. 12):september 2003

[Immunobiology of Tolerance Induction]review paperCatalytic antibodies to factorVIIIhaematologica 2003; 88(suppl. n. 12):48-51http://www.haematologica.org/free/immunotolerance2001.pdfSÉBASTIEN LACROIX-DESMAZES, MICHEL D. KAZATCHKINE,SRINI KAVERIINSERM U430, Hôpital Broussais, Paris, FranceThe occurrence of factor VIII (FVIII) inhibitors is themajor complication associated with the administrationof exogenous FVIII to patients with hemophilia A.FVIII inhibitors have been shown to neutralize the procoagulantactivity of FVIII in a passive manner, i.e. bysteric hindrance they prevent the interaction betweenFVIII and other molecules of the coagulation cascade.We have described in multitransfused patients withhemophilia A, the presence of anti-FVIII IgG antibodiesthat hydrolyze FVIII. The estimated kineticparameters derived for FVIII cleavage by anti-FVIIIantibodies are in line with the previously describedcatalytic antibodies. The identified cleavage sitesare evenly spread throughout the FVIII molecule andare located after an arginine or a lysine in most cases.We have recently shown that the catalytic antibodiesare highly prevalent among hemophilia Apatients with FVIII inhibitors. Catalytic antibodies toFVIII are the first example in which the hydrolysis ofthe target molecule by hydrolytic antibodies may bedirectly relevant to the etiology of the disease. Thecharacterization of FVIII inhibitors as site-specificproteases may provide novel strategies in the designof therapy against FVIII inhibitors in patients withhemophilia A.©2003, Ferrata Storti FoundationKey words: hemophilia A, factor VIII,catalytic antibodies, factor VIII inhibitors.Correspondence: Sébastien Lacroix-Desmazes, INSERM U430,Hôpital Broussais, 96, rue Didot, F-75014 Paris, France. Fax:international +33.1.45459059. E-mail:sebastien.lacroix@brs.ap-hop-paris.frAnti-FVIII antibodies arise in 20 to 50% ofpatients with hemophilia A following therapeuticadministration of exogenous FVIIIto treat bleeding episodes. Some anti-FVIII antibodies,referred to as FVIII inhibitors, neutralizethe pro-coagulant activity of FVIII and precludethe further administration of FVIII to thepatients. The occurrence of FVIII inhibitorsremains a major therapeutic challenge andmuch effort has been dedicated in the last 15years to understanding the nature of FVIIIinhibitors and the mechanisms by which theyinhibit FVIII activity.The antibody response to FVIII in patients withhemophilia A is highly heterogeneous. 1-4 FactorVIII inhibitors are exclusively of the IgG isotype.Although early work had suggested that anti-FVIIIalloantibodies are predominantly of the IgG4subclass, more recent observations using polyclonalpreparations of affinity-purified anti-FVIIIantibodies indicate that the isotypic distributionof FVIII inhibitors follows the physiologic profileof IgG subclasses. 3 The inhibitory titer of FVIIIinhibitors is represented as Bethesda units (BU),1 BU being defined as the inverse of the IgG concentrationinhibiting 50% of total FVIII activity.There is evidence to suggest that the occurrenceof inhibitors correlates with the genetic abnormalityunderlying the disease. 5 Thus, 35% ofpatients with large gene deletions in the FVIIIgene, gene inversions and stop mutations developinhibitors whereas inhibitors occur in only 5to 7% of patients with missense mutations andsmall deletions in the FVIII gene.FVIII epitopes that are targeted by inhibitorshave been mapped by immunoblotting, 2,4 by immunoprecipitationof conformational epitopesexpressed by recombinant fragments of FVIII 6and by competition experiments between polyclonalpreparations of affinity-purified humananti-FVIII antibodies and monoclonal murineantibodies for binding to FVIII. 3,7 Three majorclusters of B-cell epitopes have been delineatedand span the A2 domain on the heavy chain(amino acids 373-740), 8,9 epitopes located in theA3 domain, 3 and in the C2 domain. 10-12 A monoclonalhuman FVIII inhibitor has also beenhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia49shown to recognize an epitope in the C1 domainof FVIII. 13Most FVIII inhibitors described to date neutralizethe pro-coagulant activity of FVIII byblocking functional epitopes by steric hindrance.FVIII inhibitors that bind to the heavy-chain ofFVIII prevent the cleavage of FVIII by thrombin orby activated factor X and the subsequent activationof FVIII. 4,8,14,15 Light-chain specific inhibitorsprevent the interaction of FVIII with activatedfactor IX, 16 or phospholipids and/or von Willebrandfactor (vWF). 10,17-19 A second category ofFVIII inhibitors bind to neo-epitopes formed bythe association of FVIII and vWF, thus reducingthe dissociation rate of FVIII from vWF. 20 In additionto FVIII inhibitors, the plasma of inhibitorpositivepatients contains anti-FVIII antibodiesdirected to non-functional determinants of themolecule. 2,3 Such antibodies, together with antibodiesspecific for functional epitopes, may beinvolved in the formation of immune complexesin multitransfused patients with haemophiliaA 21 and result in an accelerated in vivo clearanceof FVIII. 3 Altogether the latter mechanisms arepassive mechanisms of inhibition of FVIII thatare mediated by the passive binding of antibodiesto the molecule. We wondered whether someanti-FVIII antibodies may neutralize FVIII in anactive manner, may be endowed with enzymaticactivity and behave as catalytic antibodies.The concept of catalytic antibodies was firstintroduced in the early 1940's by Linus Pauling. 22At the time, Pauling hypothesized that if the antigenbinding sites of immunoglobulins are randomlygenerated, which was later to be confirmed,then the possibility exists that some ofthem may structurally mimic the active site ofenzymes and thus be endowed with catalyticactivity. It was not until the advent of hybridomatechnology in the 1970's that it became possibleto confirm Pauling's hypothesis formally. In1986, Tramontano et al. demonstrated that it ispossible to raise monoclonal murine antibodiesendowed with esterase activity upon activeimmunization of mice using an analog of thetransition state structure of the ester bond cleavagechemical reaction. In humans the presenceof catalytic antibodies with hydrolytic activity hassince been demonstrated in several instances. In1989, S. Paul et al. purified and characterizedantibodies capable of cleaving the vasoactiveintestinal peptide from the plasma of patientswith asthma. 23 Later, antibodies able to hydrolyzeDNA and thyroglobulin were purified from theplasma of patients with lupus erythematosus andof patients with Hashimoto's thyroiditis. 24,25Antibodies able to hydrolyze and activate prothrombininto thrombin were found amongmonoclonal light chains of patients with multiplemyeloma. 26 It is interesting to note that innone of these pathologic manifestations is thehydrolysis of the target molecules directly relevantto the clinical manifestations. We wonderedwhether, if some of the anti-FVIII antibodies ofpatients with hemophilia A are able to hydrolyzethe molecule, a direct correlation between theoccurrence of catalytic antibodies and the clinicalmanifestations could be demonstrated for thefirst time.IgG were isolated from the plasma of patientswith severe hemophilia A who had developedhigh titers of FVIII inhibitors following perfusionof therapeutic FVIII. IgG were purified from plasmaby ammonium sulfate precipitation, followedby chromatography on protein G. In our firstseries of experiments, anti-FVIII antibodies werefurther purified by affinity-chromatography on ahuman FVIII-coupled sepharose matrix. RadiolabeledFVIII presented a characteristic migrationprofile, with protein bands ranging from 300 to70 kD. Incubation of FVIII in the presence ofanti-FVIII IgG of 2 of 3 patients resulted in thehydrolysis of high molecular weight bands andappearance of bands of molecular weight lowerthan 70 kD, whereas incubation of FVIII withIgG of a third patient did not result in FVIIIhydrolysis. 27 Similarly, incubation of FVIII withnormal polyclonal human IgG (IVIg, Sandoglobulin®)and with a control human monoclonalIgG directed to cytomegalovirus did not result inFVIII proteolysis.The following lines of evidence confirmed thatFVIII hydrolysis was not due to contaminatingproteases: 27 (i) anti-FVIII IgG from patients BO1and WA1 exhibited different kinectics of FVIIIhydrolysis and different digestion patterns,whereas that of patient CH1 did not cleave FVIII,suggesting that FVIII hydrolysis is mediated bythe variable regions of antibodies; (ii) the catalyticactivity to FVIII was co-eluted with anti-FVIII IgG, whereas IgG not retained on the FVIIImatrix did not cleave FVIII; (iii) removal of IgGfrom the preparations of affinity-purified anti-FVIII antibodies by chromatography on protein Gresulted in the complete loss of the hydrolyzingcapacity; (iv) co-incubation of FVIII and anti-FVIII IgG in the presence of several proteaseinhibitors (i.e., E-64, pepstatin, leupeptine) didnot prevent FVIII hydrolysis; (v) size-exclusionchromatography of urea-treated affinity-purifiedanti-FVIII IgG yielded a major peak that wasdevoid of contaminants and retained the catalyticactivity to FVIII; (vi) F(ab')2 fragments preparedby pepsin digestion of affinity-purifiedanti-FVIII IgG were able to cleave FVIII, furthersuggesting that FVIII hydrolysis is mediated bythe variable regions of the antibodies.More recently, we have purified IgG from theplasma of 24 patients with severe hemophilia Aand FVIII inhibitors. 28 The patients' IgG were allsubjected to size-exclusion chromatography inthe presence of 8 M urea followed by renaturationby extensive dialysis. Significant hydrolyticactivity was detected among IgG of 13 patientsout of 24. In contrast, IgG from 4 patients withsevere hemophilia A without detectable inhibitordid not cleave FVIII, suggesting a correlationbetween the presence of FVIII-hydrolyzing anti-haematologica vol. 88(supplement n. 12):september 2003

50S. Lacroix-Desmazes et al.bodies and that of FVIII inhibitory IgG. Nevertheless,patients were heterogeneous when ratesof FVIII hydrolysis were compared with inhibitoryactivity: thus, 11 patients presented with FVIIIinhibitory activity in plasma but with no FVIIIhydrolyzingactivity among purified IgG; otherspresented with high rates of FVIII hydrolysis andlow inhibitory activity in plasma; a few patientshad both high rates of FVIII hydrolysis and highinhibitory activity in plasma. A moderate correlationbetween the rate of FVIII hydrolysis andthe FVIII-neutralizing activity in plasma mayreflect the complex and multiple mechanismsthat participate simultaneously in inactivatingFVIII in vivo.The cleavage fragments generated among IgGof different patients ranged between 70 kD andless than 30 kD. Smaller digestion fragmentswere not detected under the experimental conditionsused (i.e., 10% SDS-PAGE). N-terminalsequencing of the generated digestion fragmentsof FVIII allowed the identification of the cleavagesites in the case of four patients with FVIIIhydrolyzingantibodies. Overall, the cleavage sitesare evenly spread throughout the A1, A2, B, A3and C1 domains of the FVIII molecule and arelocated after an arginine or a lysine in 85% of thecases (unpublished data). Identification of thecleavage sites on the available 3 dimensionalmodels of the A1, A2, A3 and C1 domains indicatesthat most cleavage sites are located on theouter core of the molecule and are readily accessibleby the antibodies. Using both radiolabeledFVIII and synthetic peptides coupled to a fluorescentdye as substrates, we have calculated thekinetic parameters of the anti-FVIII hydrolyticantibodies. Vmax and Km were between 6 and600 fmol/min and 10 and 1500 µM, respectively.The estimated catalytic efficiency was in therange of 0.6 or 60 M -1 .sec -1 (unpublished data). Incalculating the kinetic parameters, we overestimatedthe quantity of hydrolytic IgG, assumingthat all IgG molecules present in the assay arecatalytic. Paul et al. estimated the amount ofantibodies hydrolytic to the vasoactive intestinalpeptide within pools of purified IgG to be equalto 73.4 fmol/mg of IgG (Paul et al., 1989). Thecatalytic efficiency calculated in the case of FVIIIhydrolyzingantibodies may thus be underestimatedby several orders of magnitude.Taken together, our data demonstrate thathydrolytic antibodies are present among anti-FVIII IgG of inhibitor-positive patients withhemophilia A. Hemophilia thus appears as thefirst human disease in which hydrolysis of thetarget molecule by catalytic antibodies mayaccount for the clinical manifestations. We hopethat the identification of molecules that specificallyblock the catalytic activity of anti-FVIII antibodieswill provide novel therapeutic approachesfor patients with inhibitors in the near future.References1. Algiman M, Dietrich G, Nydegger UE, Boieldieu D, SultanY, Kazatchkine MD. Natural antibodies to factorVIII (anti-hemophilic factor) in healthy individuals.Proc Natl Acad Sci USA 1992; 89:3795-9.2. Fulcher CA, de Graaf Mahoney S, Roberts JR, KasperCK, Zimmerman TS. Localization of human factor FVIIIinhibitor epitopes to two polypeptide fragments. ProcNatl Acad Sci USA 1985; 82:7728-32.3. Gilles JG, Arnout J, Vermylen J, Saint-Remy JM. AntifactorVIII antibodies of hemophiliac patients are frequentlydirected towards nonfunctional determinantsand do not exhibit isotypic restriction. Blood 1993;82:2452-61.4. Scandella D, DeGraaf Mahoney S, Mattingly M, RoederD, Timmons L, Fulcher CA. Epitope mapping ofhuman factor VIII inhibitor antibodies by deletionanalysis of factor VIII fragments expressed inEscherichia coli. Proc Natl Acad Sci USA 1988; 85:6152-6.5. Schwaab R, Brackmann HH, Meyer C, Seehafer J,Kirchgesser M, Haack A, et al. Haemophilia A: mutationtype determines risk of inhibitor formation. ThrombHaemost 1995; 74:1402-6.6. Scandella D, Timmons L, Mattingly M, Trabold N, HoyerLW. A soluble recombinant factor VIII fragment containingthe A2 domain binds to some human anti-factorVIII antibodies that are not detected by immunoblotting.Thromb Haemost 1992; 67:665-71.7. Gilles JG, Desqueper B, Lenk H, Vermylen J, Saint-RemyJM. Neutralizing antiidiotypic antibodies to factor VIIIinhibitors after desensitization in patients with hemophiliaA. J Clin Invest 1996; 97:1382-8.8. Lollar P, Parker ET, Curtis JE, Helgerson SL, Hoyer LW,Scott ME, et al. Inhibition of human factor VIIIa by anti-A2 subunit antibodies. J Clin Invest 1994; 93:2497-504.9. Scandella D, Mattingly M, Prescott R. A recombinantfactor VIII A2 domain polypeptide quantitatively neutralizeshuman inhibitor antibodies that bind to A2.Blood 1993; 82:1767-75.10. Saenko EL, Shima M, Rajalakshmi KJ, Scandella D. Arole for the C2 domain of factor VIII in binding to vonWillebrand factor. J Biol Chem 1994; 269:11601-5.11. Scandella D, Kessler C, Esmon P, Hurst D, Courter S,Gomperts E, et al. Epitope specificity and functionalcharacterization of factor VIII inhibitors. In: Aledort LM,editor. Inhibitors to coagulation factors. New York:Plenum Press; 1995. p. 47-63.12. Shima M, Fulcher CA, de Graaf Mahoney S, HoughtenRA, Zimmerman TS. Localization of the binding site fora factor VIII activity neutralizing antibody to amino acidresidues Asp1663-Ser1669. J Biol Chem 1988; 263:10198-203.13. Peerlinck K, Jacquemin MG, Arnout J, Hoylaerts MF,Gilles JG, Lavend'homme R, et al. Antifactor VIII antibodyinhibiting allogeneic but not autologous factor VIIIin patients with mild hemophilia A. Blood 1999; 93:2267-73.14. Foster PA, Fulcher CA, Houghten RA, de Graaf MahoneyS, Zimmerman TS. Localization of the binding regionsof a murine monoclonal anti-factor VIII antibody anda human anti-factor VIII alloantibody, both of whichinhibit factor VIII procoagulant activity, to amino acidresidues threonine351-serine365 of the factor VIIIheavy chain. J Clin Invest 1988; 82:123-8.15. Lubahn BC, Ware J, Stafford DW, Reisner HM. Identificationof a F.VIII epitope recognized by a humanhemophilic inhibitor. Blood 1989; 73:497-9.16. Zhong D, Saenko EL, Shima M, Felch M, Scandella D.Some human inhibitor antibodies interfere with factorVIII binding to factor IX. Blood 1998; 92:136-42.17. Arai M, Scandella D, Hoyer LW. Molecular basis of factorVIII inhibition by human antibodies. Antibodies thatbind to the factor VIII light chain prevent the interactionof factor VIII with phospholipid. J Clin Invest 1989;83: 1978-84.18. Jacquemin MG, Desqueper BG, Benhida A, Vander ElstL, Hoylaerts MF, Bakkus M, et al. Mechanism and kinet-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia51ics of factor VIII inactivation: study with an IgG4 monoclonalantibody derived from a hemophilia A patientwith inhibitor. Blood 1998; 92:496-506.19. Shima M, Scandella D, Yoshioka A, Nakai H, Tanaka I,Kamisue S, et al. A factor VIII neutralizing monoclonalantibody and a human inhibitor alloantibody recognizingepitopes in the C2 domain inhibit factor VIII bindingto von Willebrand factor and to phosphatidylserine.Thromb Haemost 1993; 69:240-6.20. Saenko EL, Shima M, Gilbert GE, Scandella D. Slowedrelease of thrombin-cleaved factor VIII from von Willebrandfactor by a monoclonal and a human antibody isa novel mechanism for factor VIII inhibition. J BiolChem 1996; 271:27424-31.21. Kazatchkine MD, Sultan Y, Burton-Kee EJ, Mowbray JF.Circulating immune complexes containing anti-VIIIantibodies in multi-transfused patients with haemophiliaA. Clin Exp Immunol 1980; 39:315-20.22. Pauling L. Nature of forces between large molecules ofbiological interest. Nature 1948; 161:707-9.23. Paul S, Volle DJ, Beach CM, Johnson DR, Powell MJ,Massey RJ. Catalytic hydrolysis of vasoactive intestinalpeptide by human autoantibody. Science 1989; 244:1158-62.24. Gololobov GV, Chernova EA, Schourov DV, SmirnovIV, Kudelina IA, Gabibov AG. Cleavage of supercoiledplasmid DNA by autoantibody Fab fragment: applicationof the flow linear dichroism technique. Proc NatlAcad Sci USA 1995; 92:254-7.25. Li L, Paul S, Tyutyulkova S, Kazatchkine MD, Kaveri S.Catalytic activity of anti-thyroglobulin antibodies. JImmunol 1995; 154:3328-32.26. Thiagarajan P, Dannenbring R, Matsuura K, TramontanoA, Gololobov G, Paul S. Monoclonal antibody lightchain with prothrombinase activity. Biochemistry 2000;39: 6459-65.27. Lacroix-Desmazes S, Moreau A, Sooryanarayana, BonnemainC, Stieltjes N, Pashov A, et al. Catalytic activityof antibodies against factor VIII in patients with hemophiliaA. Nat Med 1999; 5:1044-7.28. Lacroix-Desmazes S, Bayry J, Misra N, Horn MP, VillardS, Pashov A, et al. The prevalence of proteolytic antibodiesagainst factor VIII in hemophilia A. N Engl J Med2002; 346:662-7.haematologica vol. 88(supplement n. 12):september 2003

[Immunobiology of Tolerance Induction]review paperIdiotypic control of inhibitorsJEAN GUY GILLESCenter for Molecular and Vascular Biology,University of Leuven, Belgiumhaematologica 2003; 88(suppl. n. 12):52-54http://www.haematologica.org/free/immunotolerance2001.pdfThe major complication following factor VIII(FVIII) infusion in hemophilia A patients isthe development of a specific immuneresponse. It is usually admitted that 20% ofpatients under such therapy develop an antibodyresponse towards FVIII. 1 The actual incidence is,however, difficult to provide, with percentagesvarying from 0% to 45% in individual studies, 2depending of the patient population considered,the method used to detect inhibitors, the detectionthreshold level, and the frequency and timepoints (transient inhibitors) at which inhibitordetection is carried out.An important additional factor that furthercomplicates the evaluation of inhibitor incidenceis the fact that anti-idiotypic antibodiescan neutralize the activity of inhibitors in plasma.Idiotype refers to the ensemble of determinantsthat are located within the variable part ofantibodies (Abs1) or antigen-specific receptorsof T-cells. Therefore, anti-idiotypic antibodies(anti-Ids) are second-generation antibodies(Abs2) directed towards the variable part ofpathogenic antibodies and have the theoreticalcapacity to neutralize Ab1 activity and production.It is well established that tolerance to self-proteinis first induced at an early stage by clonaldeletion of self-reactive B- and T-cells in thebone marrow and the thymus, respectively.However, not all self-reactive lymphocytes areeliminated by central deletion. Auto-reactive B-cells are a common feature of peripheral blood,as well as low- or intermediate-affinity self-reactiveT-cells.A number of mechanisms by which such autoreactivecells are rendered non-functioning orare deleted in the periphery have been described.Anti-idiotypic antibodies could represent athird level of tolerance maintenance in this generalscheme, with their capacity to fine tune thefunction of antibodies and maintain a subtleequilibrium between complementary idiotypesCorrespondence: Dr. Jean Guy Gilles, Center for Molecular andVascular Biology, Katholiek Universiteit Leuven, Herestraat 49,B-3000 Leuven, Belgium. Phone: international+32.16.346018. Fax: international +32.16.345990.E-mail: jeanguy.gilles@med.kuleuven.ac.beexpressed on B- and T-cells. 3A good indication of how anti-idiotypîc antibodiescan indeed exert a regulatory mechanismin the periphery is provided by the demonstrationthat healthy individuals with normal levelsof FVIII produced significant titers of inhibitoryantibodies to FVIII, 4,5 the activity of which isundetectable in plasma because of the presenceof complementary anti-idiotypic antibodies.However, such a FVIII inhibitory activity can bereadily detected whenever anti-FVIII antibodiesare purified by a combination of chromatographyand specific immunoadsorption over aninsolubilized-FVIII column. 6 The FVIII inhibitorycapacity of these Abs was demonstrated to beequal to that of anti-FVIII Abs purified fromhemophilia A patient’s plasma with high levelof inhibitors, as measured by the Bethesdaassay. 6Such neutralizing anti-Id activity has also beendetected in a group of patients successfullydesensitized by administration of high doses ofFVIII. 7 The study demonstrated that the concentrationof anti-FVIII antibodies, purified by thesame procedure as for healthy donors, did notchange during desensitization and that antibodiesmaintained their capacity to inhibit the procoagulantfunction of FVIII, even though Bethesdaunit titration in plasma was reduced to undetectablelevels. This showed the potentiallyimportant function of anti-Id regulation in toleranceto the FVIII molecule. Therefore, any newtherapy inducing an increased production ofanti-Id Abs could be a method of choice in thetreatment of inhibitors. A first approach alongthese lines has been reported: patients weretreated by injections of immune complexes madeof FVIII and autologous specific antibodiestowards FVIII, which resulted in a significantreduction in the level of circulating FVIIIinhibitors that were neutralized by correspondinganti-Id Abs. 8 Such an approach could openup the way towards new therapeutic strategiesfor FVIII inhibitors, at potentially low cost comparedto classical desensitization.To get further insight into the mechanisms bywhich anti-FVIII antibodies are produced andthe role of the second generation of Abs (anti-Ids) in FVIII tolerance maintenance and/orhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia53induction, animal models have been developed.A first model, the FVIII-knockout mouse, 9 inwhich there is no circulating FVIII molecule,allowed us to show an immune response uponinjection of human FVIII without the possibleinterference due to the cross-reactivity betweenmouse and human FVIII. A second model, calledthe severe combined immunodeficiency (SCID)mouse 9,10 lacks the enzyme required to rearrangeDNA for immunoglobulin synthesis and antigenreceptor T-cells, can be used to reconstitute animmune response de novo. For example, it wasshown that by reconstituting mice with cellsfrom hemophilia A patients, followed by FVIIIinjection, it is possible to reproduce a humananti-FVIII immune response in this strain ofmice.Based on the results of idiotypic interactionsobtained in healthy donors, hemophilia Apatients and animal models, passive and activetherapies can be envisaged. Passive therapy couldbe helpful in the case of emergency and involvethe infusion of anti-Id Abs in patients withinhibitors. Such anti-Id Abs are present in intravenousimmunoglobulin (IVIg) solution 11 orcould be obtained from enriched or purifiedpreparations from a healthy donor plasma pool,since it has been demonstrated that plasma fromhealthy donors contains significant amounts ofanti-Id Abs. In this context, one can perhaps considerthat the efficacy of IVIg infusion in patientswith auto-immune inhibitors 12 is related to itscontent of anti-Ids Abs, even though anti-Id levelsare usually found to be very low. Anti-Id Abenrichedpreparations might represent a suitablealternative for passive immunotherapy. Activeimmunization with idiotype has already beensuccessfully attempted in man in the case oftumor-specific non-Hodgkin’s lymphomas. 13 Wealso observed that classical desensitizationtowards FVIII inhibitors through infusion of highdoses FVIII induced an increased production ofanti-Id Abs, a result also achievable at the lowestcost by injection of autologous FVIII/anti-FVIIIAbs complexes. 8The key question is to define what should beinjected to enhance a specific anti-Id immuneresponse. This question is particularly importantwith FVIII, with regard to the size of the FVIIImolecule and the number of recognized epitopes.In an attempt to address this major question, theanti-FVIII Abs repertoire is being explored at theclonal level. In our laboratory, we started producinghigh-affinity human monoclonal Abstowards FVIII using a method involving cellimmortalization with the Epstein-Barr virus. 14,15Peripheral blood lymphocytes are used to this endto ensure that the Abs obtained are part of thepatient’s antibody repertoire. Such monoclonalantibodies have then been sequenced and relevantidiotypes identified.Thus, Bo2C11, a human monoclonal antibodyto FVIII was injected into Balb/c mice to produceanti-Id mAbs. Eight clones were selected on theRestoration of FVIII activityAnti-Id mouse mAb conc. (µg/mL)Figure 1. The figure illustrates the capacity of anti-Id mousemonoclonal Abs 14C12 and 10E9 to neutralize the inhibitingactivity of the human monoclonal anti-FVIII in a chromogenicfunctional assay. At an anti-Id Abs/anti-FVIII Absmolar ratio from 10/1 to 100/1, for 14C12 and 10E9,respectively, 100% of the FVIII coagulation activity isrestored.basis of strong recognition of Bo2C11 insolubilizedon a ELISA plate; 2 of such clones (14C12and 10E9) also inhibited Bo2C11 binding toinsolubilized FVIII in ELISA. More interestingly,in a chromogenic assay 14C12 and 10E9 neutralizedup to 100% of Bo2C11 anti-FVIII activityat molar ratios varying from 10/1 to 100/1, asshown in Figure 1. Bo2C11 is representative ofantibodies inhibiting the binding of FVIII tophospholipids. We, therefore, tested the capacityof the two anti-Id mAbs to neutralize the anti-FVIII activity of purified polyclonal antibodiesisolated from the plasma of the Bo2C11 patient.14C12 and 10E9 neutralized 70% and 20% ofthe anti-FVIII activity of the polyclonal antibodies,respectively.Based on these data indicating that anti-Id Abscould be of high specificity and affinity we used14C12 in a FVIII knockout mouse model, in anattempt to restore hemostasis after successiveinjections of FVIII and Bo2C11. Once again,administration of anti-Id Abs was able to neutralizethe anti-FVIII inhibiting activity and tofully restore hemostasis.These data demonstrate that anti-Ids havefunctional properties that would be useful in caseof emergency situations for neutralizing anti-FVIII antibodies. It also indicates that increasingthe production of such anti-Ids by active immunizationcould be of interest in the regulation ofthe anti-FVIII immune response. To this latterend, further experiments in animal modelshaematologica vol. 88(supplement n. 12):september 2003

54J.G. Gilles(FVIII-KO mice and SCID mice) are now ongoingto evaluate how anti-FVIII production can becontrolled by an anti-Id immune response, andmore precisely, to design and evaluate new formsof specific idiotype-base therapies.References1. Briet E, Rosendaal FR, Kreuz W, Rasi V, Peerlinck K,Vermylen J, et al. High titer inhibitors in severehaemophilia A. A meta-analysis based on eight longtermfollow-up studies concerning inhibitors associatedwith crude or intermediate purity factor VIII products.Thromb Haemost 1994; 72:162-4.2. Ehrenforth S, Kreuz W, Scharrer I, Linde R, Funk M,Gungor T, et al. Incidence of development of factor VIIIand factor IX inhibitors in haemophiliacs. Lancet 1992;339:594-8.3. Saint-Remy JM, Jacquemin MG, Gilles JG. Anti-idiotypicantibodies: from regulation to therapy of factorVIII inhibitors. Vox Sang 1999; 77 Suppl 1:21-4.4. Gilles JG, Vanzieleghem B, Saint-Remy JM. Factor VIIIInhibitors. Natural autoantibodies and anti-idiotypes.Semin Thromb Hemost 2000; 26:151-5.5. Algiman M, Dietrich G, Nydegger UE, Boieldieu D, SultanY, Kazatchkine MD. Natural antibodies to factorVIII (anti-hemophilic factor) in healthy individuals.Proc Natl Acad Sci USA 1992; 89:3795-9.6. Gilles JG, Saint-Remy JM. Healthy subjects produce bothanti-factor VIII and specific anti-idiotypic antibodies. JClin Invest 1994; 94:1496-505.7. Gilles JG, Desqueper B, Lenk H, Vermylen J, Saint-RemyJM. Neutralizing antiidiotypic antibodies to factor VIIIinhibitors after desensitization in patients with hemophiliaA. J Clin Invest 1996; 97:1382-8.8. Gilles JG, Arnout J, Peerlinck K, Vermylen J, Saint-RemyJM. Antigen-antibody complexes made of FVIII andautologous specific antibodies down-regulate the productionof anti-FVIIIantibodies. XXI International Congressof the World Federation of Haemophilia 1994April; [abstract].9. Gilles JG, Vanzieleghem B, Saint-Remy JM. Animalmodels to explore mechanisms of tolerance induction toFVIII: SCID mice and SCID-FVIII-KO mice. Haematologica2000; 85 Suppl 10:103-7.10. Vanzieleghem B, Gilles JG, Desqueper B, Vermylen J,Saint-Remy JM. Humanized severe combined immunodeficientmice as a potential model for the study of toleranceto factor VIII. Thromb Haemost 2000; 83:833-9.11. Rossi F, Kazatchkine MD. Antiidiotypes against autoantibodiesin pooled normal human polyspecific Ig. JImmunol 1989; 143:4104-9.12. Sultan Y, Kazatchkine MD, Maisonneuve P, NydeggerUE. Anti-idiotypic suppression of autoantibodies to factorVIII (antihaemophilic factor) by high-dose intravenousgammaglobulin. Lancet 1984; 2:765-8.13. Nelson EL, Li X, Hsu FJ, Kwak LW, Levy R, Clayberger C,et al. Tumor-specific, cytotoxic T-lymphocyte responseafter idiotype vaccination for B-cell, non-Hodgkin'slymphoma. Blood 1996; 88:580-9.14. Jacquemin MG, Desqueper BG, Benhida A, Vander ElstL, Hoylaerts MF, Bakkus M, et al. Mechanism and kineticsof factor VIII inactivation: study with an IgG4 monoclonalantibody derived from a hemophilia A patientwith inhibitor. Blood 1998; 92:496-506.15. Jacquemin M, Benhida A, Peerlinck K, Desqueper B,Vander Elst L, Lavend'homme R, et al. A human antibodydirected to the factor VIII C1 domain inhibits factorVIII cofactor activity and binding to von Willebrandfactor. Blood 2000; 95:156-63.haematologica vol. 88(supplement n. 12):september 2003

[Immunobiology of Tolerance Induction]review paperMurine models for the study offactor VIII inhibitorshaematologica 2003; 88(suppl. n. 12):55-63http://www.haematologica.org/free/immunotolerance2001.pdfBIRGIT M. REIPERT,*° MARIA SASGARY,* CHRISTINA HAUSL,°ELISABETH MAIER,* RAFI U. AHMAD,* PETER L. TURECEK,*HANS P. SCHWARZ*°*Baxter BioScience, Vienna, Austria; °Center forBioMolecular Therapeutics, Vienna, AustriaCorrespondence: Hans Peter Schwarz, MD, Baxter,BioScience Industriestraße 67 A-1220 Vienna, Austria.Phone: international +43.1.20100-2067. Fax: international+43.1.20100-525. E-mail: schwarh@baxter.comRegulation of the immune response to a complexprotein antigen like factor VIII (FVIII)requires interaction between different typesof immune cells and migration of these cellsbetween several compartments of the immunesystem. Access to such cells and the compartmentsthey traverse is limited in humans for obviousethical reasons. Hence, in vivo animal modelsare needed to understand the mechanisms ofantibody formation and to develop new therapeuticapproaches for inducing immune toleranceto FVIII in humans. As imperfect as any animalmodel is, important advances in human therapyhave resulted from appropriate use of suchmodels. Our present knowledge of how neutralizingantibodies develop against FVIII is based predominantlyon clinical findings in hemophilicpatients. The presence of FVIII inhibitors makespatients refractory to replacement therapy and is,therefore, a serious complication of modern treatment.The risk of patients with severe hemophiliaA developing anti-FVIII antibodies is clearlyassociated with the type of FVIII gene mutationthey have. Patients affected with nonsense mutations,large deletions and intrachromosomalrecombinations (inversions) in the FVIII geneshow the highest inhibitor incidence 1-3 (HAM-STeRS: The hemophilia A mutation, structure,test, resource site: http://europium.csc.mrc.ac.uk).These mutations are predicted to cause a completedeficit of any endogenous FVIII production.Therefore, human FVIII is probably a protein foreignto the immune system in these patients.Other mutations such as missense mutations andsmall deletions are associated with a much lowerincidence of inhibitor formation. 1-3 In these cases,non-functional FVIII antigen may circulate inthe blood 4 and render the immune system tolerantto the mutant FVIII protein. Native FVIII, presentin FVIII products could, therefore, be recognizedas an altered self protein by the immunesystem in these patients. Apart from the type ofFVIII gene mutation, other genetic factors such asthe HLA haplotype seem to be important for thedevelopment of FVIII inhibitors. To investigatehow important the HLA haplotype is in an animalmodel, animals with different genetic backgrounds,e.g. different inbred strains of mice, areneeded.The application route and dose of an antigendetermines the subsequent immune response. Inpatients, FVIII is given intravenously. Therefore,the spleen is probably the major location for thedevelopment of anti-FVIII immune responses.An animal model should reflect this and developdetectable anti-FVIII immune responses afterintravenous injection of FVIII doses equivalent tothe doses used for patients.Factor VIII is a protein antigen with some extrafeatures that distinguish it from other proteinantigens. In its activated form, activated FVIII(FVIIIa), is an essential cofactor in blood coagulation.In the presence of negatively chargedphospholipids, FVIIIa directly interacts with theserine protease factor IXa and forms a complexthat converts factor X into activated factor X(factor Xa). Factor Xa, in the presence of activatedfactor V (factor Va), phospholipids andcalcium ions, is then able to convert prothrombinto its active enzymatic form, thrombin. Apartfrom their function in blood coagulation, bothfactor Xa and thrombin have been shown in vitroto activate one or more of the protease-activatedreceptors (PAR receptors) expressed onendothelial cells. 5-7 This activation inducesproinflammatory stimuli. Therefore, each timeFVIII is given it is possible that it induces proinflammatorystimuli that directly influence theregulation of the immune response to FVIII. Ananimal model in which a severe bleeding diathesisis associated with a lack of FVIII would beoptimal for immunological studies.Given the diversity of the disease and the heterogeneityof the genetic background of thepatients with hemophilia A, who develop anti-FVIII antibodies, no single animal model couldcover all aspects of the immune response againstFVIII. Therefore, different models are needed toresearch the various facets of the response.Murine models in which the mechanism andhaematologica vol. 88(supplement n. 12):september 2003

56B. M. Reipert et al.genetic of the immune response is extremely wellcharacterized have potential advantages and meetmany of the above requirements.Factor VIII knockout miceIn 1995, Bi et al. described two murine modelsof hemophilia A in which a targeted gene disruptionin exon 16 (E-16) or exon 17 (E-16) ofthe FVIII gene resulted in a complete deficiencyof FVIII. 8 These mice express a typical phenotypeof hemophilia A 9,10 which can be corrected byhuman FVIII. 11 Qian et al. showed that E-16 andE-17 knockout mice develop anti-FVIII antibodiesafter intravenous injection of human FVIII intherapeutic doses and that this immune responseis dependent on the induction of FVIII-specific T-lymphocytes. 12 Human FVIII is certainly moreforeign to hemophilic mice than to humans withhemophilia and can be expected to induce astronger immune response in mice than murineFVIII would do. The lack of a convenient sourceof murine FVIII limits its use but human FVIIIcan be used instead because it interacts with themurine proteins of the coagulation system 13 dueto its sequence homology with the murineFVIII. 14 As mentioned above, this interaction inan environment of a severe bleeding syndromemight be important for regulating the immuneresponse. Therefore, human FVIII can be considereda suitable model antigen in the search fornew strategies to induce immune tolerance.We used the E-17 model in a series of differentstudies. In our hands, all E-17 mice developeddetectable anti-FVIII antibodies after two doses ofhuman FVIII (200 ng recombinant FVIII, freefrom albumin) that increased in titer after subsequentdoses. 15, 16 Titers of total anti-FVIII antibodiesanalyzed by ELISA correlated with titers ofneutralizing anti-FVIII antibodies measured byBethesda assays, 16 (Figure 1). Anti-FVIII antibodysecreting cells (ASC) first appeared in the spleenwhere they were detectable after two doses ofFVIII, 17 (Figure 2a). Their appearance correlatedwith that of anti-FVIII antibodies in blood plasma(Figure 2b). Anti-FVIII ASC in bone marrowwere detectable after three doses of FVIII (Figure2a). These cells had probably formed initially inthe spleen and then migrated to the bone marrow.We did not see any formation of anti-FVIIIASC in lymph nodes confirming that the spleenis the major development location for immuneresponses against blood-borne antigens. The IgGsubclassdistribution of anti-FVIII ASC was similarin spleen and bone marrow and matched thesubclasses of anti-FVIII antibodies in blood plasma(Figure 3, Table 1), indicating that bothorgans contribute to circulating antibodies in theblood. The IgG1 and IgG2a subclasses dominatedthe anti-FVIII antibody response. After FVIIItreatment had terminated, anti-FVIII antibodiespersisted for at least 22 weeks (Figure 2b). Thepersistence of antibodies correlated with thelong-term persistence of anti-FVIII ASC (Figure2a). These ASC could be either long-living ASC asFigure 1. Relation of total anti-FVIII antibody titers (ELISAtiter) to titers of FVIII-neutralizing antibodies (Bethesdatiter) in plasma obtained from hemophilic mice after onedose (), two doses () or four doses () of FVIII. Eachpoint represents values for an individual mouse. Blood sampleswere obtained 1 week after each dose. ELISA titers andBethesda titers were analyzed as described (16). From Sasgaryet al. 16 with permission.described by Slifka et al. 18 and Manz et al. 19 orcells continuously formed by antigen-driven differentiationof memory B cells as described byOchsenbein et al. 20 Future studies using celltransfer experiments should be able to showwhich model is best for explaining the maintenanceof high titers of anti-FVIII antibodies inhemophilic mice, and possibly also in patients.The outcome of such studies could have considerableimplications for creating new strategiesaimed at inducing immune tolerance to FVIII.The development of anti-FVIII antibodies in E-17mice correlated with the appearance of FVIII-specificCD4 + T cells. Of these, the most prominenttype that could be detected were CD4 + T cells producingIFN-γ, followed by T cells producingIL10 16 (Figure 4).The CD40/CD40L interaction is a key event inthe initiation of humoral immune responsesagainst T-cell-dependent antigens. 21 Previousstudies have shown that a blockade of CD40/CD40L interactions can achieve prolonged survivalof allografts in rodents and monkeys 22,23and prevent graft-versus-host disease 24,25 andautoimmunity in rodent models. 26,27 These effectsare probably due to tolerance induction in theCD4 + T-cell population 28 and tempt the speculationthat anti-CD40L antibodies can induce lastingT-cell tolerance to FVIII in hemophilia A. Asanti-FVIII antibody formation is T-cell dependent,inducing FVIII-specific T-cell toleranceshould prevent their formation. Using the E-17mouse model we could show that the blockade ofCD40-CD40 ligand interactions prevents theinduction of an anti-FVIII immune response, 29haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia57AFigure 2. Long-term persistence of anti-FVIII antibody-secreting cells in spleen and bone marrow and of anti-FVIII antibodies inblood plasma of hemophilic E-17 mice after treatment with human FVIII. From Hausl et al. (17) with permission. Figure 2a: Kineticof anti-FVIII antibody-secreting cells in spleen and bone marrow of hemophilic mice after one, two, three or four doses of FVIIIas well as 14 and 22 weeks after the fourth dose of FVIII. Cells were analyzed by Elispot as shown in Figure 3. Presented arethe medium values and standard deviations, calculated from the results obtained with three different cell dilutions.(Figure 5). This initial prevention did not, however,induce a permanent tolerance to FVIII, 29(Figure 5). These results confirm earlier data publishedby Qian et al. 30 Rossi et al. 31 suggested thattreating hemophilic E-16 mice with FVIII accompaniedby anti-CD40L antibodies induces lastingimmune tolerance to FVIII under certain circumstances,at least in some animals. Their datado not, however, indicate if the immune tolerancewould really be long lasting with frequentFVIII treatment. In another study, Qian et al. 32explored the importance of the B7-CD28 interactionfor the anti-FVIII immune response. Theywere able to show that blocking this interactioncompletely prevented anti-FVIII antibody developmentin hemophilic E-16 mice. Initial blockingof the co-stimulatory interaction between B7and CD28 did not, however, induce lastingimmune tolerance to FVIII. We believe, therefore,that the blockade of a single co-stimulatory interactionmight not be sufficient to induce a stabletolerance to FVIII.Hemophilic FVIII knockout mice certainlyresemble the situation in patients in someaspects, but not in others. Sarkar et al. 33 recentlyshowed that E-17 mice produced nonfunctionalFVIII heavy chain proteins. Such heavy chain proteinsmight induce immunological tolerance toepitopes of the heavy chain of murine FVIII.Therefore, the disruption of the FVIII gene inhemophilic E-17 mice might not be severeenough to stimulate the full repertoire ofimmune responses against murine FVIII. Consequentlyadditional knock-out models, whichlack any endogenous FVIII synthesis, have to bedeveloped. Likewise it is essential that a recombinantprotein is developed to provide a convenientsource of murine FVIII. Even with betterknockout models and the availability of murineFVIII, the genome of a mouse is not that of ahuman. Therefore, we have to be aware that therewill always be differences between hemophilicmice and patients with hemophilia A.Normal mice and ratsBefore the hemophilic mouse model existed,different groups tried to use normal mice 34 orrats 35,36 to study FVIII inhibitors. Dazzi et al. 34 didhaematologica vol. 88(supplement n. 12):september 2003

58B. M. Reipert et al.BFigure 2. Long-term persistence of anti-FVIII antibody-secreting cells in spleen and bone marrow and of anti-FVIII antibodies inblood plasma of hemophilic E-17 mice after treatment with human FVIII. From Hausl et al. (17) with permission. Figure 2b: Kineticof anti-FVIII antibody titers in blood of hemophilic mice after one, two, three or four doses of FVIII as well as 5, 10, 14 and22 weeks after the fourth dose of FVIII. Anti-FVIII antibody titers were analyzed by ELISA as described in Hausl et al. (17). Eachpoint represents the result obtained from an individual mouse. Three mice were analyzed per time point.not find any anti-FVIII antibody response afterintravenous application of human FVIII in normalmice, but detected anti-FVIII antibodies afterintraperitoneal application. This antibodyresponse was dependent on the induction ofFVIII-specific T cells. Jarvin et al. 35 and Levin etal. 36 used normal rats and immunized them withhuman FVIII in adjuvant. This treatment scheduleresulted in the induction of an anti-FVIIIimmune response that was characterized by neutralizingantibodies that recognized some of theepitopes that were also found in patients withFVIII inhibitors. The application route thatinduced the anti-FVIII immune response in thesestudies cannot, however, be compared to that forFVIII in patients. Furthermore, treating normalmice and rats with mouse or rat FVIII should notinduce any immune response at all because themurine proteins represent a self protein. Analyticalcomparison of the amino-acid sequence ofmurine and human FVIII has shown that thehomology in the functionally important A and Cdomains is 84-93%, but only 42-70% in thenonfunctional B domain and the acidic regions. 14Therefore, anti-FVIII antibodies developedagainst human FVIII can be expected to be directedpredominantly against the more divergentregions and, accordingly, not to reflect the wholerepertoire of possible antibody responses thatwould be seen in FVIII knockout mice.Another major disadvantage is the lack of ableeding syndrome in normal mice. As mentionedabove the cross-talk between the immunesystem and the coagulation system might beimportant for regulating the immune response toFVIII and should, therefore, be considered in ananimal model.Nevertheless, recent experimental study usingnormal mice should be mentioned. Chao andWalsh reported the loss of neutralizing anti-FVIIIantibodies in normal C57BL/6 mice after sustainedexpression of human FVIII. 37 Antibodieshaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia59Figure 3. Representative Elispots showing the frequency of anti-FVIII antibody-secreting cells (total IgG) as well as the IgGsubclassdistribution of anti-FVIII antibody-secreting cells (IgG1, IgG2a, IgG2b, IgG3) in bone marrow and spleen of hemophilicmice after four intravenous doses of 200 ng FVIII (80 U/kg). Serial dilutions of cell samples were incubated with FVIII immobilizedto the solid phase of PVDF-bottomed Multiscreen-IP filtration plates. Each spot represents an individual anti-FVIII antibody-secretingcell. Analyses of antibody-secreting cells were done as described. 17 From Hausl et al. 17 with permission.haematologica vol. 88(supplement n. 12):september 2003

60B. M. Reipert et al.Table 1. Median anti-factor VIII antibody titers of differentIgG subclasses in hemophilic E-17 mice (n=3).Anti-factor VIII antibody titersIgG1 IgG2a IgG2b IgG3One dose FVIII n.d. n.d. n.d. n.d.Two doses FVIII 1280 2560 2560 120Three doses FVIII 40960 40960 15360 640Four doses FVIII 122700 122700 61440 512022 weeks after the 20480 81920 5120 640fourth dosen.d. – not detectable; IgG subclasses of anti-FVIII antibodies were analyzed asdescribed in Hausl et al. 17 Briefly, PolySorp multiwell plates were coated with FVIII(free from albumin). Plates were incubated with serial dilutions of plasma samplesfrom mice treated with FVIII. IgG subclasses of anti-FVIII antibodies bound to theimmobilized FVIII were detected by incubation with isotype-specific, horseradish-peroxidase-labeledsecondary monoclonal antibodies (Southern Biotechnologies, Birmingham,AL, USA) and subsequent substrate development. Antibody titers wereexpressed as the highest dilution of plasma samples showing a positive result (opticaldensity >0.3) in the assay. From Hausl et al. 17 with permission.developed within 7 to 14 days of intraportalinjection of adeno-associated virus (AAV) carryinghuman FVIII. Bethesda titers of anti-FVIIIantibodies (>100 BU/mL) persisted relativelyunchanged for 9 to 10 months. At 10 monthsafter injection of the virus, FVIII inhibitors disappearedand FVIII protein became detectable inthe blood. These results suggest that immune toleranceto human FVIII can be induced by sustainedexpression of human FVIII in a mousemodel.SCID mice and SCID/Factor VIII knockoutmiceMice with severe combined immunodeficiency(SCID) lack functional B and T cells 38 due toan arrest of B and T lineage committed cells inearly development. This arrest is caused by a genemutation that impairs the recombination ofantigen receptor genes. The SCID-related geneencodes a DNA-dependent protein kinase catalyticsubunit and is located in chromosome16. 39 The SCID mice are unable to produce antibodiesor to reject allogeneic skin grafts 38 and,therefore, do not develop a functionally activeimmune system. Furthermore, spleen cells ofSCID mice fail to show a proliferative response toB-cell and T-cell mitogens. 38 Other hematopoieticcells such as macrophages and NK cellsdevelop and function normally. Because of theirsevere immune deficiency, SCID mice are able toaccept xenotransplants and can, therefore, bereconstituted with peripheral blood mononuclearcells isolated from human blood donors. 38 Themajor advantage of such a model is the possibilityof studying at least part of the humanimmune system in a mouse model. The majordisadvantages are the lack of functional lymphoidorgans and difficulty in eliciting primaryimmune responses.Laulan et al. 40 grafted SCID mice with peripheralblood mononuclear cells from healthydonors or donors with hemophilia A. They injectedreconstituted mice intraperitoneally with 100U (10 mg) of human FVIII. A specific response toFVIII only developed in mice that received bloodcells from patients with hemophilia who hadFVIII inhibitors and not in those that receivedcells from healthy donors or patients with hemophiliabut no FVIII inhibitors. These results suggestit is possible to induce secondary but not primaryimmune responses against human FVIII inreconstituted SCID mice. The mice received,however, only one FVIII injection and this was ata dose (100 U per mouse equals about 4000U/kg) which is not comparable to that used inpatients with hemophilia A. In similar experimentsby Vanzieleghem et al. 41 SCID mice werereconstituted with peripheral blood mononuclearcells from healthy blood donors and treated withone intraperitoneal injection of 50 U (5 mg)human FVIII followed by three of 25 U (2.5 mg).All reconstituted mice spontaneously producedanti-FVIII antibodies in the absence of any treatmentwith FVIII. These antibodies were onlydetectable after affinity purification. Treatingmice with up to four doses of FVIII did not inducean increase in antibody titers detectable afteraffinity purification. Furthermore, treatmentwith FVIII did not stimulate antibodies detectablewithout affinity purification. When SCID micewere reconstituted with peripheral blood mononuclearcells from patients with hemophilia Awho had FVIII inhibitors, treating the mice withFVIII induced a detectable secondary immuneresponse. 42 In an attempt to improve the SCIDmouse model, Gilles et al. created a SCID/FVIIIknockout mouse by cross-breeding SCID micewith FVIII knockout mice. 42 Future studies willreveal the advantages and limitations of this newmodel.To summarize the results obtained with SCIDmice, this model might be suitable for investigatingspecific questions of secondary anti-FVIIIimmune responses in an environment that containsparts of the human immune system.DiscussionAn animal model suitable for developing newapproaches for inducing immune tolerance toFVIII should encompass features that as much aspossible resemble those found in patients withhemophilia A while still accommodating sophisticatedanalysis and manipulation of the ongoingimmune response. Hemophilia A is a hereditaryX-linked bleeding disorder caused by theincomplete function of circulating clotting FVIIIor its total absence due to a mutation in the FVIIIgene. The consequent coagulation deficiency ishaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia61Figure 4. Kinetic of FVIII-specific T cells producing IL-2, IFN- γ, IL-4 or IL-10 in the spleen of hemophilic E-17 mice after treatmentwith one, two or four doses of human FVIII. Dot plots show intracellular cytokine staining in CD4 + T cells after in vitrorestimulation of splenic T cells with FVIII. Results shown in the upper right-hand corner of each dot plot represent the percentageof CD4 + T cells that stained positive for the cytokine indicated. The negative cell population (lower left-hand corner) wasdefined by using cells that were stained with appropriate isotype-matched negative control antibodies. T cells were isolated,restimulated and stained as described. 16 From Sasgary et al. 16 with permission.haematologica vol. 88(supplement n. 12):september 2003

62B. M. Reipert et al.Figure 5. Relation of neutralizing anti-FVIII antibodies (Bethesda titer) and total anti-FVIII antibodies (ELISA titer) in hemophilicE-17 mice after treatment with FVIII together with or without anti-CD40 ligand antibody. 200 mg anti-mouse CD40 ligand antibodyMR1 were given intravenously 24 hours before treatment with FVIII as described in Reipert et al. 29 Each point representsvalues for an individual mouse. Mice received the treatments as indicated at weekly intervals. 200 ng FVIII (80 U/kg) weregiven per dose. All analyses were done as described. 29 From Reipert et al. 29 with permission.treated with infusions of FVIII concentrates.Depending on the type of gene mutation, theFVIII in these concentrates might be recognized bythe patient’s immune system as a foreign proteinor an altered self-protein. In addition, othergenetic factors such as the HLA haplotype arebelieved to influence the anti-FVIII immuneresponse. Considering this heterogeneous situation,obviously no single animal model canaccommodate all the different aspects of the anti-FVIII immune response in patients.Therefore, we have to be aware that each animalmodel that we use, and will use in the future,to search for new approaches for inducing toleranceto FVIII has its advantages and its limitations.Furthermore, it must not be forgotten thatexperimental studies in animal models can onlyprovide new ideas, proof of concepts and strategiesfor therapeutic approaches and it is the clinicaltrial that will always be the ultimate tool forleading to therapeutic advances.AcknowledgmentsWe are grateful to Howard M. Reisner for his criticalreview and Elise Langdon-Neuner for editingthe manuscript.References1. Schwaab R, Brackmann HH, Meyer C, Seehafer J,Kirchgesser M, Haack A, et al. Haemophilia A: Mutationtype determines risk of inhibitor formation. ThrombHaemost 1995;74:1402-6.2. Tuddenham EGD, McVey JH. The genetic basis ofinhibitor development in hemophilia A. Haemophilia1998;4:543-5.3. Fakharzadeh SS, Kazazian HH. Correlation between factorVIII genotype and inhibitor development in hemophiliaA. Sem Thromb Hemost 2000;26:167-71.4. McGinniss MJ, Kazazian HH, Hoyer LW, Bi L, Inaba H,Antonarakis SE. Spectrum of mutations in CRM-positiveand CRM-reduced hemophilia A. Genomics 1993;15:392-8.5. Vergnolle N, Wallace JL, Bunnett NW, Hollenberg MD.Protease-activated receptors in inflammation, neuronalsignalling and pain. Trends Pharmacol Sci 2001; 22:146-52.6. Camerer E, Huang W, Coughlin SR. Tissue factor- andfactor X-dependent activation of PAR2 by factor VIIa.Proc Natl Acad Sci USA 2000;97:5255-60.7. Asokananthan N, Graham PT, Fink J, Knight DA,Bakker AJ, McWilliam AS et al. Activation of proteaseactivatedreceptor (PAR)-1, PAR-2, and PAR-4 stimulatesIL-6, IL-8, and prostaglandin E2 release fromhuman respiratory epithelial cells. J Immunol 2002;168:3577-85.8. Bi L, Lawler AM, Antonarakis SE, High KA, Gearhart JD,Kazazian Jr HH. Targeted disruption of the mouse factorVIII gene produces a model of haemophilia A. Nathaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia63Genet 1995;10:119-21.9. Bi L, Sarkar R, Naas T, Lawler AM, Pain J, Shumaker SL,et al. Further characterization of factor VIII-deficientmice created by gene targeting: RNA and protein studies.Blood 1996;88:3446-50.10. Muchitsch EM, Turecek PL, Zimmermann K, Pichler L,Auer W, Richter G, et al. Phenotypic expression ofmurine hemophilia. Thromb Haemost 1999;82:1371-3.11. Connelly S, Andrews JL, Gallo AM, Kayda DB, Qian J,Hoyer L, et al. Sustained phenotypic correction ofmurine hemophilia A by in vivo gene therapy. Blood1998;91:3273-81.12. Qian J, Borovok M, Bi L, Kazazian Jr HH, Hoyer LW.Inhibitor development and T cell response to humanfactor VIII in murine haemophilia A. Thromb Haemost1999;81:240-4.13. Balague C, Zhou J, Dai Y, Alemany R, Josephs SF,Andreason G, et al. Sustained high-level expression offull-length human factor VIII and restoration of clottingactivity in hemophilic mice using a minimal adenovirusvector. Blood 2000;95:820-8.14. Elder B, Lakich D, Gitschier J. Sequence of the murinefactor VIII cDNA. Genomics 1993;16:374-9.15. Reipert BM, Ahmad RU, Turecek PL, Schwarz HP. Characterizationof antibodies induced by human factor VIIIin a murine knockout model of hemophilia A. ThrombHaemost 2000;84:826-32.16. Sasgary M, Ahmad RU, Schwarz HP, Turecek PL, ReipertBM. Single cell analysis of factor VIII-specific T-cells inhemophilic mice after treatment with human factorVIII. Thromb Haemost 2002;87:266-72.17. Hausl C, Maier E, Schwarz HP, Ahmad RU, Turecek PL,Dorner F, Reipert BM. Long-term persistence of antifactorVIII antibody-secreting cells in hemophilic miceafter treatment with human factor VIII. ThrombHaemost 2002;87:840-5.18. Slifka MK, Antia R, Withmire JK, Ahmed R. Humoralimmunity due to long-lived plasma cells. Immunity1998; 8:363-72.19. Manz RA, Löhning M, Cassese G, Thiel A, Radbruch A.Survival of long-lived plasma cells is independent ofantigen. Intern Immunol 1998;11:1703-11.20. Ochsenbein AF, Pinschewer DD, Sierro S, Horvath E,Hengartner H, Zinkernagel RM. Protective long-termantibody memory by antigen-driven T help-dependentdifferentiation of long-lived memory B-cells to shortlivedplasma cells independent of secondary lymphoidorgans. Proc Am Acad Sci USA 2000;97:13263-8.21. Foy TM, Aruffo A, Bajorath J, Buhlmann JE, Noelle RJ.Immune regulation by CD40 and its ligand GP39. AnnRev Immunol 1996;14:591-617.22. Larsen CP, Alexander DZ, Hollenbaugh D, Elwood ET,Ritchie SC, Aruffo A, et al. CD40-gp39 interactions playa critical role during allograft rejection: suppression ofallograft rejection by blockade of the CD40-gp39 pathway.Transplantation 1996;61:4-9.23. Kirk AD, Burkly LC, Batty DC, Baumgartner RE, BerningJD, Buchanan K. Treatment with humanized monoclonalantibody against CD154 prevents acute renalallograft rejection in nonhuman primates. Nat Med1999;5:686-93.24. Durie FH, Aruffo A, Ledbetter J, Crassi KM, Green WR,Fast LD, et al. Antibody to the ligand of CD40, gp39,blocks the occurrence of the acute and chronic forms ofgraft-vs-host-disease. J Clin Invest 1994;94:1333-8.25. Saito K, Sakurai J, Ohata J, Kohsaka T, Hashimoto H,Okumura K, et al. Involvement of CD40 ligand-CD40and CTLA4-B7 pathways in murine acute graft-versushostdisease induced by allogeneic T cells lacking CD28.J Immunol 1998;160:4225-31.26. Gerritse K, Laman JD, Noeller RJ, Aruffo A, Ledbetter JA,Boersma WJA, et al. CD40-CD40 ligand interactions inexperimental allergic encephalomyelitis and multiplesclerosis. Proc Natl Acad Sci USA 1996;93:2499-504.27. Early GS, Zhao W, Burns CM. Anti-CD40 ligand antibodytreatment prevents the development of lupus-likenephritis in a subset of New Zealand black x NewZealand white mice. Response correlates with theabsence of an anti-antibody response. J Immunol 1996;157:3159-64.28. Taylor PA, Friedman TM, Korngold R, Noelle RJ, BlazarBR. Tolerance induction of alloreactive T cells via ex vivoblockade of the CD40:CD40L costimulatory pathwayresults in the generation of a potent immune regulatorycell. Blood 2002;99:4601-929. Reipert BM, Sasgary M, Ahmad RU, Auer W, TurecekPL, Schwarz HP. Blockade of CD40/CD40 ligand interactionsprevents induction of factor VIII inhibitors inhemophilic mice but does not induce lasting immunetolerance. Thromb Haemost 2001;86:1345-52.30. Qian J, Burkly LC, Smith EP, Ferrant JL, Hoyer LW, ScottDW, et al. Role of CD154 in the secondary immuneresponse: the reduction of pre-existing splenic germinalcenters and anti-factor VIII inhibitor titers. Eur JImmunol 2000;30:2548-54.31. Rossi G, Sarakar J, Scandella D. Long-term induction ofimmune tolerance after blockade of CD40-CD40Linteraction in a mouse model of hemophilia A. Blood2001;97:2750-6.32. Qian J, Collins M, Sharpe AH, Hoyer LW. Preventionand treatment of factor VIII inhibitors in murine hemophiliaA. Blood 2000;95:1324-9.33. Sarkar R, Gao GP, Chirmule N, Tazelaar J, Kazazian HH.Partial correction of the murine hemophilia A with neoantigenicmurine factor VIII. Hum Gene Ther 2000;11:881-94.34. Dazzi F, Rosato A, Tison T, Vianello F, Radossi P, GirolamiA. An animal model to explore the molecular basisof factor VIII (FVIII) inhibitor formation: evidence ofanti-FVIII T-cell response and importance of administrationroute (abstract). Thromb Haemost 1995; 73:1026.35. Jarvis MA, Levin LG, Harrison JA, De Pianto DJ, SuzukiCM, Ziaja CL, et al. Induction of human factor VIIIinhibitors in rats by immunization with human recombinantfactor VIII: a small animal model for humanswith high responder inhibitor phenotype. ThrombHaemost 1996;75:318-25.36. Levin LG, Jarvis M, Powell J, Harrison JA, Reisner HM.Induction of human factor VIII inhibitors in rats 2: finemapping of rat anti-human rFVIII antibodies. ThrombHaemost 1996;76:998-100337. Chao H, Walsh CE. Induction of tolerance to humanfactor VIII in mice. Blood 2001;97:3311-238. Bosma GC, Carroll AM. The SCID mouse mutant: definition,characterization and potential uses. Annu RevImmunol 1991;9:323-50.39. Araki R, Fujimori A, Hamatani K, Mita K, Saito T, MoriM, et al. Nonsense mutation at Tyr-4046 in the DNAdependentprotein kinase catalytic subunit of severecombined immune deficiency mice. Proc Natl Acad SciUSA 1997;94:2438-83.40. Laulan A, Sauger A, Germain C, Montembault AM, SansI, Potentini-Esnault C, et al. Frequency of anti-FVIIIantibodies in humanized SCID mice elicited by recombinantdeleted factor VIII and by plasma derived factorVIII. J Immunol Methods 1997;210:205-14.41. Vanzieleghem B, Gilles JG, Desqueper B, Vermylen J,Saint-Remy JM. Humanized Severe Combined ImmunodeficientMice as a Potential Model for the Study ofTolerance to Factor VIII. Thromb Haemost 2000; 83:833-9.42. Gilles JG, Vanzieleghem B, Saint-Remy JM. Animalmodels to explore mechanisms of tolerance induction toFVIII: SCID mice and SCID-FVIII-KO mice. Haematologica2000; 85 Suppl:103-7.haematologica vol. 88(supplement n. 12):september 2003

[Immunobiology of Tolerance Induction]review paperT-lymphocytes in the anti-factorVIII immune responsehaematologica 2003; 88(suppl. n. 12):64-65http://www.haematologica.org/free/immunotolerance2001.pdfMARC G. JACQUEMIN, JEANMARIE R. SAINT-REMYCenter for Molecular and Vascular Biology, Universityof Leuven, BelgiumThe immune response towards factor VIII(FVIII), as for any other soluble glycoprotein,requires the participation of B- and T-lymphocytes. In the absence of specific T-cells nohelp is provided to B-cells to allow them tomature and differentiate into antibody secretingcells. Somewhat surprisingly, the involvement ofT-cells has been more or less neglected by mostinvestigators, when at the same time most effortwas been devoted to the characterization of B-cell epitopes and antibodies.Why is it so important to characterize specificanti-FVIII T-cells? First of all, T-cell recognitionof an antigen is the first specific event in themounting of an immune response. When anantigen enters the body, it has to be taken up bycells specialized in the presentation of it. Thisinvolves the intracellular digestion of the proteinfollowed by presentation of short peptidesderived from the digestion in the context ofMHC-class II molecules. The T-cell makes thefirst specific reading of the antigen, i.e. peptidebound to MHC class II alleles, which allows it tobecome activated and to provide B-cells with thehelp necessary for activation and the productionof antibodies.Second, T-cells recognize the peptide derivedfrom the antigen only when it is presented in theappropriate MHC class II molecule. This restrictionmakes it plausible that some alleles of MHCclass II molecules are preferentially associatedwith FVIII peptides. Therefore, it can be assumedon a theoretical basis that the capacity to produceinhibitor antibodies is associated with theexpression of a particular set of class II molecules.If so, this could represent a means to identifypatients at risk of producing inhibitory antibodies.Third, T-cells could represent an ideal target forthe design of an immunotherapy for FVIIIinhibitors. This is inherent to the actual characteristicsof the T-cell repertoire. T-cells are producedby the bone marrow, but must proceedthrough the thymus to acquire full maturation.The T-cell pool is essentially acquired at birth andmost of the thymus function disappears duringinfancy. Besides, the mechanism of somatichypermutation, which is responsible for a largeCorrespondence: JeanMarie R. Saint-Remy, Center for Molecularand Vascular Biology, University of Leuven, Belgium.E-mail: jeanmarie.saint-remy@med.kuleuven.ac.bepart of the diversity of B-cells, does not play arole at the level of T-cells. Controlling the T-cellarm of the anti-FVIII immune response could,therefore, result in a long-standing state of unresponsiveness.One caveat is however that the sizeof the T-cell repertoire is such that it has thecapacity to respond to any imaginable T-cell epitopein the universe. FVIII being a large molecule,it should always be possible to find a T-cellreacting towards it.Evidence for T-cell involvement in theanti-FVIII immune responseThe evidence is two-fold. In animal models,and in particular in the hemophilia A mousemodel, specific T-cells are already observed 3 daysafter the first administration of FVIII, well beforethe first antibodies can be detected. 1 It is alsoknown in such models that an interruption inthe B – T-cell dialog by, for instance, antibodiesto a surface molecule such as CD40L expressedon T-cells, prevents the development of theresponse. The second series of data comes fromobservations made in patients. A few years ago itwas observed that in HIV infected patients thetiter of anti-FVIII inhibitors declined togetherwith the number of CD4 + T-cells. 2More direct evidence has been obtained in manthrough the demonstration of FVIII-reactive T-cells in the peripheral blood of hemophilia Apatients with inhibitors. More recently, a veryexhaustive investigation was carried out using alarge number of synthetic peptides covering theentire FVIII sequence. These peptides were incubatedwith peripheral blood T-cells of hemophiliaA patients with or without inhibitors. A largenumber of FVIII peptides activating CD4 + T-cellswere found. However, screening hemophilia Apatients without inhibitor and healthy individualsas controls, it was found that individuals withnormal levels of FVIII also had specific T-cells intheir circulation. 3The latter observation raised questions aboutthe meaning of these findings. In vivo, theimmune system is confronted by full proteins andnot by peptides derived from them. T cell epitopeson a protein are organized according to a hierarchy.Major epitopes are dominant in the sensethat they are recognized by the large number andas a first line of recognition; these are also theepitopes towards which tolerance is establishedby selective deletion in the thymus. Minor epitopesare recognized less predominantly and oftenhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia65by a more restricted number of subjects. Crypticepitopes are not recognized when presentationstarts from the whole protein. However, crypticepitopes can be activated by peptides. As the tuningof the T-cell repertoire in the thymus is carriedout by deletion-selection on proteins, T-cells recognizingcryptic epitopes can be found in theperiphery.The quest for pathologically relevant T-cellsFrom the above, it appears that the identificationof T-cells of pathologic relevance, that is tosay T-cells with the capacity to drive the productionof inhibitor antibodies, could require furtherscrutiny. Nature offers us three situations withregards to FVIII. Healthy individuals with normallevels of FVIII have — presumably — had a chanceto eliminate strongly reactive T-cells from theirrepertoire during thymus maturation. T-cells toimmunodominant epitopes should, therefore, beabsent in the periphery. On the other hand, insevere hemophilia A, the absence of FVIII precludesthe negative selection of T-cells in thethymus: therefore in the periphery all T-cellsreacting to FVIII should be present. There is, however,an intermediate situation, in which patientssuffering from a mild degree of hemophilia A produceinhibitor antibodies towards normal wildtype FVIII, while remaining tolerant to their ownFVIII. Under such circumstances, the immunesystem should be purged from strongly reactive T-cells as for healthy individuals, except for these T-cells reacting against the site where the aminoacid substitution is located. Starting from suchpatients could offer a good opportunity to restrictthe population of T-cells greatly.T-cells were therefore cloned for the first timefrom T-cells of a mild hemophilia A patient carryinga mutation in the carboxy-terminal end ofthe C1 domain. 4 This patient produced inhibitorantibodies upon administration of wild-type FVIII,given to help the patient undergo a surgical intervention.T-cell cloning and characterizationA preparation of dendritic cells was made fromperipheral blood monocytes of the same patientfrom which T-cells had to be cloned. The latterwere selected by surface marker to obtain a pureCD4 + population. After several cycles of stimulationT-cells were cloned and expanded. Once theclones had been obtained, the search for specificitywas initiated by showing absence of proliferationand/or cytokine production in the presenceof autologous (mutated) FVIII, while strongactivation could be obtained by presentation ofwild-type FVIII.Synthetic peptides of various lengths centeredon the mutation site in the C1 domain were constructedand tested in the activation assay system.Three T-cell clones reacted towards the sameregion with subtle differences in the actual epitoperecognized. Recombinant FVIII moleculescarrying mutations within the carboxy-terminalend of the C1 domain were then tested for reactivitywith T-cell clones, some of which reactedstrongly.FVIII-specific T-cells: perspectivesWith such material at hand, it is now possibleto evaluate the MHC class II restriction of theFVIII recognition by T-cells, and perhaps identifyspecific alleles associated with activation of suchcells. This, as described above, could offer a markerpredisposing to the development of inhibitoryantibodies.A number of strategies aiming at making specificT-cells ignorant or unresponsive, or even todeleting such T-cells have been described in animalmodels in alternative research fields. It ishoped that the information gained in animalmodels will also hold true for FVIII. This is currentlybeing tested in mouse models of hemophiliaA. It can be expected that specific immunotherapyfor FVIII inhibitors targeting specificT-cells will become available in the near future.References1. Qian J, Collins M, Sharpe AH, Hoyer LW. Preventionand treatment of factor VIII inhibitors inmurine hemophilia A. Blood 2000; 95:1324-9.2. Bray GL, Kroner BL, Arkin S, Aledort LW, HilgartnerMW, Eyster ME, et al. Loss of high-responderinhibitors in patients with severe hemophilia A andhuman immunodeficiency virus type 1 infection: areport from the Multi-Center Hemophilia CohortStudy. Am J Hematol 1993; 42:375-9.3. Reding MT, Wu H, Krampf M, Okita DK, Diethelm-Okita BM, Christie BA, et al. Sensitization of CD4 +T cells to coagulation factor VIII: response in congenitaland acquired hemophilia patients and inhealthy subjects. Thromb Haemost 2000; 84:643-52.4. Jacquemin M, Burny W, Vantomme V, Chaux P,Lavend’homme R, Gilles JG, et al. High levels ofIFNγ production by FVIII-specific T cells from amild/moderate hemophilia A patient with inhibitor.42 nd Annual Meeting of the American Society forHematology, San Francisco, CA, USA, December2000.haematologica vol. 88(supplement n. 12):september 2003

[Immunobiology of Tolerance Induction]Perspectives of immunotherapyfor inhibitor patientsreview paperhaematologica 2003; 88(suppl. n. 12):66-68http://www.haematologica.org/free/immunotolerance2001.pdfJEAN-MARIE R. SAINT-REMYCenter for Molecular and Vascular BiologyUniversity of Leuven, BelgiumThe ultimate goal for the problematic ofinhibitor for Factor VIII (FVIII) is to preventor suppress an immune response by anentirely antigen-specific approach. Obviously,such goal is also identifiable for other commonhuman pathologies, in particular for immunediseases. It is therefore no surprise that knowledgeis accumulating so rapidly on the mechanimsby which it will become feasible to achievesuch a prevention and/or suppression.For the sake of clarity, it remains useful todivide the current attempts towards the eliminationof inhibitors in three broad categories:immune ignorance, anergy or unresponsivenessinduction, and deletion. Basic knowledge inimmunology tells us that each of these three categoriescontains a number of potentialapproaches, which are at different levels of elaborationand therefore of clinical application.Ignorance, anergy and deletion should be consideredat the level of both B and T cells, not toforget dendritic cells and the fact that B and Tcells exchange signals, which, when interrupted,can also abort an immune response.The concept of ignorance means that FVIIIbecomes invisible for the immune system. Thus,in theory, one could eliminate the main bindingsites for antibodies on the FVIII molecule. Inpractice, however, as epitopes recognized aremostly conformational, any significant changein B cell epitope is bound to be accompanied bysignificant changes in the 3-D conformation ofthe FVIII molecule, with the risk of altering itsfunction and to create novel B cell epitopes. Fortunately,only a limited number of B cell epitopesof FVIII bear a direct relevance for its function:these are the epitopes which are directlyCorrespondence: JeanMarie R. Saint-Remy, Center for Molecularand Vascular Biology, University of Leuven, Belgium.E-mail: jeanmarie.saint-remy@med.kuleuven.ac.beinvolved in the interactions of FVIII with it physiologicalpartners, such as von Willebrand factor(VWF), phospholipids (PL), FIX and so. In otherwords, reducing antibody binding to FVIIIcould be obtained by modifying only the regionscontaining the clusters of B cell epitopes.We should however keep in mind that the Bcell repertoire is continuosly renewed in man,over his entire life span, which means that everyday, a number of B cells emerge from the bonemarrow, which have the potential to react withFVIII. Whatever the situation in practice, the elegantstudies carried out by the group of Pete Lollarhave convincingly demonstrated that it is possibleto alter and reduce the binding of preformedantibodies to FVIII, by introducing mutations atkey locations, or by replacing sequences of FVIIIby their homologue sequence taken from anothermammal, such as pig. We hope to see in thevery next future that such alterations of FVIIIwould also lead to reduced immunogenicity,namely reduced capacity to elicit an immunereponse upon injection to an animal and, in fine,to man.Ignorance can also be applied at the T cell level.In fact, this corresponds to a physiologicalmechanism. Thus, in healthy individuals,autoreactive T cells are present in the circulation,which are, however, not activated. Activationdepends on a number of parameters,including the affinity of the T cell receptor (TCR)for its cognate peptide, the density of the peptidea the surface of an antigen-presenting cell(APC), the type of APC that presents FVIII, etc…Ignorance is the end result of the lack of sufficientavidity in the recognition of specific peptidesby T cells. But ignorance is rapidly lost if anincrease in TCR avidity occurs, as it would happenin inflammation.Whether or not, it would be possible to alterthe FVIII molecule in such a manner as to reduceT cell recognition is not known. On the onehand the size of the T cell repertoire is such thatit has the capacity to recognize virtually any epitopein the universe. In addition, the degeneratehaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia67nature of the TCR provides a single TCR with thepossibility of recognizing a number of differentpeptides (up to 10 6 , according to some authors!).On the other hand, T cells recognize stretches ofaminoacids with no tertiary conformation. Thismeans that single mutation/deletion in the FVIIImolecule could significantly alter T cell recognitionwith no or only minimal alteration of the3-D conformation of FVIII. The experimentalapproach of reducing T cell immunogenicity isnevertheless hampered by the fact that relevantT cell epitopes are hard to localize. Using sets ofpeptides overlapping the entire FVIII moleculepermits the identification of T cell epitopes. However,the relevance of these epitopes in the formationof inhibitor antibodies is questionable.In search for relevant T cell epitopes, it is essentialto start from the entire FVIII molecule, i.e. tomimic as closely as possible what happens in reallife. However, the frequency of T cell precursorsin peripheral blood is very low, at the verge ofdetection.Anergy should be defined as unresponsivenessto an antigen administered in a normalimmunogenic manner to an otherwise normallyimmunocompetent host. It is therefore antigen-specificand distinct from the generalimmune asthenia observed in some diseases orresulting from therapy. However, even when seenin this restricted meaning, the concept of anergycovers many different situations, the commondenominator of which being that it results fromactive cell signaling.Signaling can occur through soluble factorssuch as cytokines or by direct cell-to-cell contact.B cells are physiologically induced to anergy attwo stages. The first is just before they leave thebone marrow, in process of eliminating functionallyB cells reacting with too strong an affinitywith autoantigens. The second stage at whichanergy is induced is during the secondaryresponses occuring in secondary follicles ofperipheral lymph nodes. During this response, Bcells undergo a process called affinity maturation,which depends on random mutationsintroduced in the antibody variable parts, followedby selection. As it is random, the processcarries the risk of generating B cells with highaffinity for autoantigens, luckily prevented byanergy induction. Cross-linking of antibody variableparts at the surface of B lymphocytes is themain mechanism leading to anergy.Another way by which anergy could beobtained at the B cell level is through recognitionof antibody variable parts (idiotype) by specificanti-idiotypic antibodies. The occurrence of antiidiotypicantibodies in the regulation of the anti-FVIII response has been amply demonstrated,both in haemophilia A patients with or withoutinhibitor and in healthy individuals. There is littledoubt that these antibodies exert a selectivepressure on specific antibodies and that they areassociated with the down-regulation of anti-FVIIIantibodies in some autoimmune diseases. However,their role in the regulation of antibodiestowards allogeneic FVIII is far from being clear.Recent data have however indicated that it is possibleto generate anti-idiotypic antibodies withsufficient affinity as to neutralize the inhibitoryactivity of inhibitors with the highest affinity.Anti-idiotypic antibodies are often seen asreagents for passive administration in a setting inwhich they would neutralize circulatinginhibitors. This type of clinical application wouldby no means be devoid of interest, for instancewhen inhibitor patients have to undergo a surgicalprocedure. However, we think that the possibleclinical applications of anti-idiotypic antibodiesare much broader. Indeed, anti-idiotypicantibodies can transduce a signal within cellscarrying the corresponding idiotype at their surface.Such signal involves tyrosine kinases andprotein phosphorylation. There is strong suggestionthat the transduced signal can lead to cellapoptosis, which would then occur in exquisitelyspecific manner. Genetic engineering of antibodiesoffers many alternative possibilities: antiidiotypicantibodies can be handled in such a wayas to reinforce their affinity, their capacity to activatethe complement system or to interact withFcγ receptors at the surface of natural killer cells.Under each of those three circumstances, theresult could be B cell anergy or death.Induction of anergy at the T cell level is againa mechanism of high physiological importance.Seen from the other side, T cells require a numberof criteria to be activated. These include ahigh enough affinity of the T cell receptor, athreshold density of MHC class II molecules atthe surface of antigen-presenting cell and thepresence of co-stimulatory molecules. AutoreactiveT cell, including T cells specific for FVIII, arepresent in the peripheral blood of healthy individuals,despite effective elimination in the thymus.On top of this, haemophilia A patients haveT cells reacting also towards major T cell epitopesof FVIII. Once a sufficient threshold for activationis reached, an immune response is initiated.This can occur whenever a pathological mechanismis triggered by for instance a viral infectionor, possibly, tissue bleeding. Whether inductionof anergy can be used as a strategy to eliminateFVIII inhibitors remains to be seen. Some possibilitieshave in fact already been tested, even inthe clinics. They include the use of anti-CD40Lantibodies, antibodies to B7 molecules or CTLA-4IgG constructs, alone or in association.Although of interest, the likelihood is weak thatsuch strategies would lead to long-lasting anergy,the rule being that normal responsiveness isrestored as soon as the inhibiting agent is nolonger administered together with FVIII.One promising approach is to use peptidesderived from FVIII and carrying T cell epitopes.Administration of such peptides could result inanergy induction, provided co-stimulation is preventedat APC level. Many questions first requirean answer before embarking on this, the mainhaematologica vol. 88(supplement n. 12):september 2003

68J.M.R. Saint-Remyone being to decide which epitope to use. As statedabove, the size of the T cell repertoire is suchthat a T cell can be found to any possible epitope.However, only major epitopes have to be hit, anddefining which epitope is a major one is not aneasy task. Peptides have a short half-life, so waysto increase their resistance to proteolytic enzymesand/or prolong their half-life will have to befound. Such peptides also offer the possibility toproduce variants with which it is possible tomodulate the affinity of MHC class II determinants.A special case should be made concerning regulatoryT cells, and specifically a recently identifiedsubpopulation of CD4 + T cells carrying theCD25 (IL-2) receptor, and which are activelyselected in the thymus. Such CD4 + CD25 + T cellsexert their regulatory properties by a combinationof mechanisms including secretion of suppressivecytokines such as IL-10 and TGF-β andcell-to-cell contact. It is too early to decidewhether such regulatory T cells will be useful ineliminating FVIII inhibitors, but we can alreadyanticipate a number of studies in this direction.Perhaps the ultimate goal in the therapy ofFVIII inhibitors would be to eliminate FVIII-specificimmunocompetent cells. Actual deletion ofB and T cells occurs as part of the homeostasis ofthe immune system. Deletion occurs as a resultof either neglect or overstimulation. An immuneresponse is triggered only whenever there is asubtle equilibrium between activation and suppressionsignals. If activation goes too far, thenspecific B as well as T cells are eliminated. Thebest example of this is the so-called activationinducedcell death, which drives T cells into committingsuicide via the overexpression of FAS andFas-ligand. It is however too early to decidewhether it would be possible to master such aphenomenon for therapeutic purposes. One mayperhaps argue that treating patients withinhibitor by infusing large doses of FVIII alreadyrepresents an example of specific T cell deletion.Experiments are ongoing to determine whetherthis is the case.This short overview of the prospects forimmune therapy of FVIII inhibitors should leavethe reader with the strong feeling that many possibilitiesexist, which, within the forthcoming tenyears, should be reducible to practice.haematologica vol. 88(supplement n. 12):september 2003

[New Aspects in Treatment of Hemophilia B Patients]review paperImmune tolerance induction inhemophilia B. Internationalimmune tolerance registryhaematologica 2003; 88(suppl. n. 12):69-70http://www.haematologica.org/free/immunotolerance2001.pdfI. WARRIERDivision of Hematology/Oncology, Children's Hospital ofMichigan, Wayne State University School of Medicine,Detroit, Michigan, USAFIX inhibitorsUnlike FVIII inhibitors in hemophilia A ourknowledge about FIX inhibitors in hemophilia Bis limited. The main reason for this is the smallernumber of severe FIX deficient patients.Inhibitor development is also less common withFIX deficiency. 1,2 In addition, anaphylaxis occurringsimultaneously with inhibitor developmentis a unique problem receltly described in hemophiliaB patients. 3 Since the last ITSG (immunetolerance study group) meeting in Palermo, Sicily,we have developed a FIX inhibitor registryfrom, which is available on the ISTH web site orcan be obtained by calling 1.888.877.0767. Thedata presented here is obtained from several publishedreports and questionnaires in addition tothe registry. 4,5When one looks at the data presented in Table1, it is apparent that FIX inhibitor patients withallergywere selectively reported to the registryand not all patients with hemophilia B andinhibitors. Thirty of the 46 patients withinhibitors and allergy had provided informationdetailing inhibitor development (Table 2). Someof the other features included: i)affection of allethnic and racial groups and ii) reported use ofvarious factor concentrates including PCC,APCC, monoclonal FIX, and rFIX at the time ofinhibitor development and allergy.Immune tolerance inductionExperience with ITI in hemophilia B inhibitorpatients is limited due to low prevalence, thrombogenicityof PCC and APCC, unavailability ofultrapure products until recently and finally lackof a standard method for ITI.From the limited information available, ITI hasbeen attempted in 21/30 patients (70%). ITIwas successful in only 2/21 patients (9.5%).Several different protocols were used for ITIincluding some with plasmapheresis and treatmentwith cyclophosphamide and IVIG.The only published data regarding ITI in hemo-Table 1. Inhibitors in hemophilia B.Total number of inhbitors 81In USA 52Outside USA 29Inhibitors with allergy 46In USA 24 (46%)Outside USA 22 (76%)Table 2. Characteristics of FIX inhibitor patients.Median age at INH 19.5 months (9-156 months)Median exposure days prior to INH 11 days (2-180 days)Peak inhibitor titer 30 BU (1-950 BU)Complete gene deletion 50%philia B patients without allergy comes from Dr.Berntrop’s Treatment Center and he will be discussingthat in his paper. A unique complicationof ITI seen in hemophilia B inhibitor patientswith allergy is the development of NephroticSyndrome during ITI. 6 This will be discussed indetail by Dr. Bruce Ewenstein.The genetics of inhibitor development inhemophilia B will be discussed by Dr. Lillicrap inhis paper. I will end my short report with a pleato hemophilia treaters for reporting all thehemophilia B inhibitor patients to the InternationalImmune Tolerance Registry. The toll freenumber for the Registry is 1.888.877.0767.References1. Briët E. Factor IX inhibitors in hemophilia B patients:their incidence and prospects for development withhigh purity FIX products. Blood Coagul Fibrinolysis1991; Suppl 2:47-50.haematologica vol. 88(supplement n. 12):september 2003

70I. Warrier2. Katz J. Prevalence of factor IX inhibitors among patientswith hemophilia B; results of a large scale North Americansurvey. Haemophilia 1996;2:28-31.3. Warrier I, Ewenstein BM, Koerper MA, Shapiro A, KeyN, DiMichele D, et al. Factor IX inhibitors and anaphylaxisin hemophilia B. J Pediatr Hematol Oncol1997;19:23-74. Berntorp E, Bjorkman S, Carlsson M, Lethagen S, NilssonIM. Biochemical and in vivo properties of high purityfactor IX concentrates. Thromb Haemost 1993;70:768-73.5. Warrier I. ITI in hemophilia B; possibilities and problems.IMH 2000;8:3-6.6. Ewenstein BM, Takemoto C, Warrier I, Lusher J, Saidi P,Eisele J, et al. Nephrotic syndrome as a complication ofimmune tolerance in hemophilia B. Blood 1997; 89:1115-6.haematologica vol. 88(supplement n. 12):september 2003

[New Aspects in Treatment of Hemophilia B Patients]review paperThe Malmö immune toleranceexperience in hemophilia Bhaematologica 2003; 88(suppl. n. 12):71-74http://www.haematologica.org/free/immunotolerance2001.pdfERIK BERNTORPDepartment of Coagulation Disorders, Malmö UniversityHospital, Malmö, SwedenHemophilia B is a rare disorder with a comparativelylow incidence of inhibitors. Therefore, the experienceof immune-tolerance induction in hemophiliaB is restricted. The Malmö protocol for immune-toleranceinduction entails prednisone, factor IX,cyclophosphamide and intravenous immunoglobulinand if the inhibitor titer is >10 Bethesda units at thestart of treatment, the titer is lowered by using extracorporealadsorption to protein A. Nine patients withhemophilia B, complicated by an inhibitor, havebeen treated using a total of 13 attempts. Tolerancewas achieved in six of the eight high-respondingpatients and one patient relapsed after six months.The mean dose of factor IX during the treatmentepisodes was 182,000 units and the mean durationof treatment (success) was 25 days. The Malmö protocolgives a high response rate in the treatment ofhigh-responding factor IX inhibitor patients and, giventhe short treatment time, could be an option inorder to reduce treatment complications such asdevelopment of a nephrotic syndrome.Key words: haemophilia B, inhibitor, factor IX, immunetolerance inductionCorrespondence: Erik Berntorp, Department of CoagulationDisorders, Malmö University Hospital, SE-205 02 Malmö,Sweden. Phone: international +46.40.332392. Fax : international+46.40.336255.E-mail: erik.berntorp@medforsk.mas.lu.se.Hemophilia B is a rare disorder comparedto hemophilia A and the incidence ofinhibitors is less: 0-3.8% in hemophiliaB 1-4 compared to about 30% in hemophilia A. 5-7Inhibitors in severe hemophilia B are often associatedwith anaphylaxis or severe allergic manifestationson exposure to factor IX-containingproducts. 8 These complications jeopardize thepossibility of applying immune tolerance inductionin hemophilia B. Immune tolerance induction(ITI) has not been studied in as much detailin hemophilia B as it has been in hemophilia A.The reason for this is that the number ofinhibitor patients with haemophilia B is smaller9 and their treatment has been hampered bythe risk of thromboembolic complications withprothrombin complex concentrates given inhigh doses. The advent of purified factor IX concentratesin recent years has probably abolishedthis risk.Recent experience from immune toleranceinduction in hemophilia B has shown that thereis a risk of nephrotic syndrome developing inpatients who have had allergic reactions toinfused factor IX in close association with thedevelopment of an inhibitor to factor IX. 10Success has been reported with the Malmöprotocol in six of seven patients with highresponding inhibitors to factor IX. 11-13A modified Malmö protocol using continuousinfusion of factor IX with a dose of about300 IU/kg body weight daily for three weeks hasalso been tried in two patients without success. 14Today there is less agreement on how to treathemophilia B inhibitor patients than on how totreat hemophilia A ones, but the use of theMalmö regimen is one option as the course oftreatment is short, which should minimize therisk of development of nephrotic syndrome. Inthis report, an update is given of the experiencein Malmö using the Malmö protocol forimmune tolerance induction in hemophilia B.haematologica vol. 88(supplement n. 12):september 2003

72E. BerntorpMaterials and MethodsStudy materialThe characteritics of the patients studied arelisted in Table 1. Nine patients with severe hemophiliaB (factor IX:C 300 394 4 >600 105 38 120 466 6 100 227 1 >300 58 1 429 1.69 27 3 29From the fourth day after the onset of treatment,gammaglobulin is given intravenously at dailydoses of 0.4 g/kg body weight for 5 days. Theinhibitor usually reappears after 5-6 days, andthe administration of factor concentrates mustbe intensified. For persistent inhibitors recombinantfactor VIIa (NovoSeven ® , Novo Nordisk) ispreferred for treatment of acute bleeds as acti-Figure 1. The Malmö model for Immune Tolerance Induction.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia73Table 2. Malmö Treatment Model. Treatment and results ofITI in patients with severe hemophilia B.Patient No. of ITI Protein A ads Time inhibitor Toleranceattempts included detection to first ITI (y) achieved1 1 Yes 2 Yes2 2 Last ITI 9 Yes, relapseafter 6 months3 1 Yes 24 Yes4 1 Yes 10 Yes5 1 No 46 Yes6 1 No 22 Yes7 3 First ITI 2 No8 2 Last ITI 2 No9 1 No 26 NoTable 3. Malmö Treatment Model. Follow-up of patients withsevere hemophilia B (31 Dec., 2002).PatientFollow-up time (years)1 112 Relapse after 6 months3 19.54 175 Not known6 12.57 Not tolerant8 Not tolerant9 Not tolerantvated prothrombin complex concentrates shouldgive a booster effect in factor IX-deficientpatients. The definition of a tolerant state is asfollows: no measurable inhibitor, normal recoveryand half-life, and efficient prophylaxis withstandard doses, i.e. 25-40 IU/kg body weighttwice a week in hemophilia B. If tolerance isachieved and considered successful, the patientcontinues on ordinary prophylaxis. If tolerance isnot successful, the treatment can be repeatedafter 6-12 months when the inhibitor productionhas leveled out and reached a steady-state.A new attempt at a shorter interval will considerablyjeopardize the possibility of reducing theinhibitor by extracorporeal protein A adsorption,increase the cost of the treatment and probablydecrease the possibility of success.Concentrates usedProthrombin complex concentrate was used infive patients (Preconativ ® , Kabi; Prothromplex ® ,Immuno). Purified factor IX was used in fourpatients (Nanotiv ® , Pharmacia; Mononine ® ,Aventis Behring; Immunine ® , Baxter and Alphanine® , Alpha).ResultsNine patients were treated according to theMalmö protocol using a total of 13 attempts. Insix of the patients only one attempt was made.Protein A adsorption was used in six attempts.Tolerance was achieved in six of the eight highrespondingpatients. One patient relapsed aftersix months. This patient was treated again witha full treatment schedule including protein Aadsorption. A prothrombin complex concentratewas used. Two weeks after the start of treatmentthe patient developed an acute myocardial infarctionand died from heart failure two years later.Other events associated with the ITI includeheadache and one patient also experienced arterialbleeding in connection with puncture of theradial artery and developed a compartment syndrome.The patient recovered. The patients’ follow-upis indicated in Table 3. Thus, four patientshave been in a tolerant state for 10-20 years now.The treatment during the six successful ITIepisodes reveals that the total dose of factor IXused was (mean and range) 182,000 IU (92,000-430,000). The duration of treatment was 25 days(15-39) and the daily dose per kg body weight170 IU (91-291).DiscussionIn Malmö, we have treated nine patients withsevere hemophilia B complicated by inhibitors ofwhom eight had a high-responding inhibitor.Tolerance was achieved in six patients after onetreatment attempt. In three patients (one ofwhom had a low-responding inhibitor), thetreatment failed using six treatment episodes. Inone of the successful patients, a relapse was seenafter six months. Thus, the Malmö ITI-modelgives a high response rate in the treatment ofhigh-responding factor IX inhibitor (6/8patients), although one relapse was seen. Thesefigures do not include two patients who wereunsuccessfully treated with a modified protocol. 14No serious complications were seen due to theproduct treatment during the treatment episodes,in those cases in which purified factor IX concentrateswere used and none of the patientsdeveloped the nephrotic syndrome. One patientreported previous allergic reactions. This patient,who did not respond to the protocol, was treatedlater on with a high-dose, long-term factor IXregimen and developed proteinuria. Althoughaware that our case series is small, we would liketo draw the following tentative conclusion. TheMalmö ITI-model gives a high-response rate inthe treatment of high-responding factor IXinhibitor patients. Only purified factor IX concentratesshould be used in order to avoid thromboemboliccomplications and the short treatmenttime (mean: 25 days) minimizes the risk ofdeveloping a nephrotic syndrome using theMalmö protocol. The Malmö model for ITI inhemophilia B may be considered as a startingpoint for a larger prospective study to evaluateand compare cost-effectiveness and the rate ofhaematologica vol. 88(supplement n. 12):september 2003

74E. Berntorpadverse events of this regimen and those of tohigh-dose, long-term protocols.AcknowledgementsThis study was supported by funds from MalmöUniversity Hospital and the region of Scania, Sweden.References1. Ehrenforth S, Kreuz W, Scharrer I, Linde R, Funk M,Gungor T, et al. Incidence of development of factor VIIIand factor IX inhibitors in haemophiliacs. Lancet1992;339:594-8.2. Sultan Y. Prevalence of inhibitors in a population of3435 hemophilia patients in France. French HemophiliaStudy Group. Thromb Haemost 1992;67:600-2.3. Katz J. Prevalence of factor IX inhibitors among patientswith haemophilia B: results of a large-scale North Americansurvey. Haemophilia 1996;2:28-31.4. Soucie JM, Evatt B, Jackson D. Occurrence of hemophiliain the United States. The Hemophilia SurveillanceSystem Project Investigators. Am J Hematol 1998; 59:288-94.5. Gruppo R, Bray GL, Schroth P, Perry M, Gomperts ED.Safety and immunogenicity of recombinant factor VIII(Recombinate) in previously untreated patients (PUPs):a 6.5 year update. The Recombinate PUP Study Group.Thromb Haemost 1997;162 Suppl 6:a[abstract PD-663].6. Rothschild C, Laurian Y, Satre EP, Borel Derlon A,Chambost H, Moreau P, et al. French previously untreatedpatients with severe hemophilia A after exposureto recombinant factor VIII : incidence of inhibitorand evaluation of immune tolerance. Thromb Haemost1998;80:779-83.7. Lusher JM, Shapiro A, Gruppo R, Bedrosian CL, NguyenK. Safety and efficacy in previously untreated patients(PUPs) treated exclusively with B-domain deleted factorVIII (BDD rFVIII). The ReFacto PUP Study Group.Thromb Haemost 2001;Suppl:2558a[abstract].8. Warrier I, Ewenstein BM, Koerper MA, Shapiro A, KeyN, DiMichele D, et al. Factor IX inhibitors and anaphylaxisin hemophilia B. J Pediatr Hematol Oncol1997; 19:23-7.9. Katz J. Prevalence of factor IX inhibitors among patientswith haemophilia B: results of a large-scale North Americansurvey. Haemophilia 1996;2:28-31.10. Warrier I. Factor IX inhibitors and anaphylaxis. In:Rodriguez-Merchan, Lee CA, editors. Inhibitors inPatients with Haemophilia. Oxford: Blackwell ScienceLtd. 2002. p. 87-91.11. Nilsson IM, Berntorp E, Zettervall O. Induction of splittolerance and clinical cure in high-responding hemophiliacswith factor IX antibodies. Proc Natl Acad SciUSA 1986;83:9169-73.12. Nilsson IM, Berntorp E, Rickard KA. Results in threeAustralian haemophilia B patients with high-respondinginhibitors treated with the Malmö model. Haemophilia1995;1:59-66.13. Freiburghaus C, Berntorp E, Ekman M, Gunnarsson J,Kjellberg BM, Nilsson IM. Immunoadsorption forremoval of inhibitors: update on treatments in Malmö-Lund between 1980 and 1995. Haemophilia 1998; 4:16-20.14. Tengborn L, Berntorp E. Continuous infusion of factorIX concentrate to induce immune tolerance in twopatients with haemophilia B. Haemophilia 1998;4:56-9.haematologica vol. 88(supplement n. 12):september 2003

[New Aspects in Treatment of Hemophilia B Patients]Factor IX mutations andinhibitor development inhemophilia Breview paperhaematologica 2003; 88(suppl. n. 12):75-77http://www.haematologica.org/free/immunotolerance2001.pdfDAVID LILLICRAPDepartments of Pathology and Medicine, Queen’s University,Kingston, Ontario, CanadaThe development of neutralizing antibodies(inhibitors) to the infused therapeutic coagulationprotein in hemophilia represents themost serious therapy-related complicationencountered in the clinical management ofhemophilia. In severe hemophilia A, inhibitorshave been reported in 15-50% of patients surveyedin different populations. 1,2 In contrast, theincidence of anti-factor IX allo-antibody generationin severe hemophilia B patients has beendocumented to be 30 nucleotides),21 patients with factor IX inhibitors are detailed.These 21 patients have eight different factor IXmutations: four different nonsense mutations,three frameshift mutations and a single missensemutation. In addition to the information derivedfrom the Mutation Database, several other studieshave also documented the association of partialor complete factor IX deletion mutations withthe development of inhibitors (Table 1 and Figure1). Thus, as with hemophilia A, the propensityfor inhibitor development is far greater in patientswith factor IX mutations that eliminate or severelydisrupt protein synthesis. In contrast, hemophiliaB patients with missense mutations veryrarely develop inhibitors.This pattern of genotype/phenotype associationis well illustrated by recent studies fromSweden where all of the hemophilia B patientswith inhibitor had either deletion or nonsensehaematologica vol. 88(supplement n. 12):september 2003

76D. LillicrapTable 1. Role of factor IX genotype in anaphylaxis (Thorland,Hemophilia 1999).mutations, whereas all patients with severe missensemutations were inhibitor-free. 14,15 Interestingly,and presently unexplained, is the fact thatthe prevalence of factor IX inhibitors in theSwedish population of patients with severehemophilia B is 23%, an approximately five-foldhigher prevalence than that reported in otherpopulations.Management of factor IX inhibitorsIn the light of the severe adverse reactions experiencedby hemophilia B patients who developallo-antibodies, a good case can be made for earlygenotyping of new cases of severe hemophiliaB. These studies should most effectively proceedwith polymerase chain reaction amplification ofthe eight factor IX exons followed by sequencingof the amplicons. Failure to amplify one or severalexons is highly likely to be indicative of afactor IX deletion mutation. With currentlyavailable information, the risk of factor IXinhibitor development with a factor IX deletionmutation is between 25-50%. With a missensemutation the inhibitor risk is negligible and witha nonsense mutation the risk is probably between5-25%.If a new factor IX deletion or nonsense mutationis detected, two revisions to routine treatmentmerit consideration. Firstly, that the initial15-25 infusions of factor IX be performed in amedically supervised environment with access todrugs for the treatment of an anaphylactic reaction.The second treatment revision that deservesevaluation is either the co-administration ofpulse immunosuppressive therapy during theearly phase of factor IX administration or the useof an alternative hemostatic agent, most probablyrecombinant factor VIIa, to treat bleedingduring the first two to three years of life. If a factorIX inhibitor develops and immune toleranceinduction is attempted, careful monitoring forthe development of proteinuria should be performed.Figure 1. The factor IX geneand adjacent mcf2 proto-oncogenelocus. The extent of thirteen,partial and total factor IXgene deletion mutations isshown with documentation ofaccompanying inhibitor generation(I) or no inhibitor (NI*).haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 77Unfortunately, the number of patients whodevelop anti-factor IX inhibitors is so small thatan objective evaluation of any of these therapeuticstrategies will require the collation of informationfrom a multi-center, international study.AcknowledgmentsThe author’s genetic studies of hemophilia B arefunded, in part, by Health Canada. DL is a CareerInvestigator of the Heart and Stroke Foundation ofOntario and a recipient of a Canada Research Chairin Molecular Hemostasis.References1. Ehrenforth S, Kreuz W, Scharrer I, Linde R, Funk M,Gungor T, et al. Incidence of development of factor VIIIand factor IX inhibitors in haemophiliacs. Lancet1992;339:594-8.2. Colvin BT, Hay CR, Hill FG, Preston FE. The incidenceof factor VIII inhibitors in the United Kingdom, 1990-93. Inhibitor Working Party. United KingdomHaemophilia Centre Directors Organization. BritishJournal of Haematology. 1995;89:908-10.3. Briet E. Factor IX inhibitors in haemophilia B patients:their incidence and prospects for development with highpurity factor IX products. Blood Coagulation and Fibrinolysis.1991;2 Suppl 1:47-50.4. High KA. Factor IX: molecular structure, epitopes, andmutations associated with inhibitor formation. Adv ExpMed Biol 1995;386:79-86.5. Gitschier J, Wood WI, Goralka TM, Wion KL, Chen EY,Eaton DH, et al. Characterization of the human factorVIII gene. Nature 1984;312: 326-30.6. Yoshitake S, Schach BG, Foster DC, Davie EW, KurachiW. Nucleotide sequence of the gene for human factorIX (antihaemophilic factor B). Biochemistry 1985;24:3736-50.7. Warrier I, Lusher JM. Development of anaphylacticshock in haemophilia B patients with inhibitors. BloodCoagul Fibrinolysis 1998;9 Suppl 1:S125-S128.8. Warrier I. Antibodies to factor IX. Haematologica 2000;85(suppl. to n. 10):31-34.9. Hay CR, Colvin BT, Ludlam CA, Hill FG, Preston FE.Recommendations for the treatment of factor VIIIinhibitors: from the UK Haemophilia Centre Directors'Organisation Inhibitor Working Party. Blood CoagulFibrinolysis 1996;7:134-8.10. Dharnidharka VR, Takemoto C, Ewenstein BM, RosenS, Harris HW. Membranous glomerulonephritis andnephrosis post factor IX infusions in hemophilia B.Pediatr Nephrol 1998;12:654-7.11. Gill JC. The role of genetics in inhibitor formation.Thromb Haemost 1999;82:500-4.12. Schwaab R, Brackmann HH, Meyer C, Seehafer J,Kirchgesser M, Haack A, et al. Haemophilia A: mutationtype determines risk of inhibitor formation. ThrombHaemost 1995;74:1402-6.13. Thorland EC, Drost JB, Lusher JM, Warrier I, Shapiro A,Koerper MA, et al. Anaphylactic response to factor IXreplacement therapy in haemophilia B patients: completegene deletions confer the highest risk. Haemophilia1999;5:101-5.14. Ljung RC. Gene mutations and inhibitor formation inpatients with hemophilia B. Acta Haematologica 1995;94 Suppl 1:49-52.15. Ljung R, Petrini P, Tengborn L, Sjorin E. Haemophilia Bmutations in Sweden: a population-based study ofmutational heterogeneity. Br J Haematol 2001;113:81-6.haematologica vol. 88(supplement n. 12):september 2003

[Acquired Inhibitors in Non-Hemophiliacs]review paperModified Bonn-Malmö Protocol(MBM-P)haematologica 2003; 88(suppl. n. 12):78-85http://www.haematologica.org/free/immunotolerance2001.pdfL. HESS, H. ZEITLER, CH. UNKRIG, W. NETTEKOVEN,T. ALBERT, R. SCHWAAB, W. EFFENBERGER, J. OLDENBURG,H. VETTER, P. HANFLAND, H.H. BRACKMANNInstitute for Experimental Hematology and Transfusion Medicine,Bonn University Hospital, Hemophilia Center, Bonn,GermanyInhibitor formation against clotting factor VIII is alife-threatening condition with a mortality rate of upto 22 %. We treated 24 patients between 30 and 89years of age, all with high-titer factor VIII inhibitorand life-threatening bleeding, using conventionaltherapies and MBM Protocol, respectively. Fivepatients were treated using conventional treatmentmethods, i.e. immunosuppression or immunomodulation(prednisolone, cyclophosphamide, vincristine,IgG). Two of these patients received additionalimmune tolerance therapy according to the BonnProtocol. Permanent inhibitor elimination wasachieved in only one of the five conventionally treatedpatients. One patient had inhibitor remission fora duration of 12 months, followed by another lifethreateningevent. Thus, a total of four patients wereunsuccessfully treated with conventional therapiesfor nine to 86 months. In three patients, conventionaltherapy was discontinued, and treatmentaccording to the Modified Bonn-Malmö Protocol wasinitiated (successfully in two patients; one had todiscontinue treatment despite good response dueto secondary diseases).In total, we treated 22 patients diagnosed withacquired inhibitor and acute life-threatening bleedingwith the Bonn-Malmö Protocol (MPM-P). MBM-P consists of four treatment elements (immunoadsorption,antigen stimulation (factor VIII), immunoglobulinsubstitution and immunosuppression). Seventeenpatients had completely overcome theinhibitor after a median of 16 apheresis treatmentsand showed normal factor VIII plasma activity. A furtherfour patients had to discontinue treatment dueto secondary diseases. Treatment was unsuccessfulin only 1 patient. Median factor VIII consumptionwas 384,000 units. At the end of July 2001, themedian follow-up period for the 17 patients successfullytreated according to the MBM Protocol wasCorrespondence: Dr. Lothar Hess, Institute for ExperimentalHematology and Transfusion Medicine, Bonn University Hospital,Hemophilia Center, Bonn, Germany.31 months, without any indication of new inhibitoractivities. The follow-up period of conventionallytreated patients was between 13 months and nineyears. Factor consumption in patients treated conventionallywas 56,000 to 4.3 million factor VIII unitsand 777,000 to 15 million FEIBA units. The MBMProtocol offers an economical alternative to conventionalinhibitor treatment. The advantages of theMBM Protocol are rapid control of bleeding withinone or two aphereses (24 to 48 hours), quickinhibitor elimination (median of 16 apheresis treatments),stability of successful outcome (median follow-upof 31 months) and a success rate of 94 %(17 successful vs. 1 failure). Early application of factorVIII using apheresis treatment had an earlyimmunomodulatory effect as well as an early hemostaticeffect. This results in the prevention of secondarybleeding complications and the causal treatmentof existing bleedings.©2003, Ferrata Storti FoundationKey words: hemophilia A, factor VIII,catalytic antibodies, factor VIII inhibitors.Modified Bonn-Malmö Protocol (MBMProtocol)Acquired inhibitors, which are mostly directedagainst Factor VIII of the coagulation system, arean extremely rare disease, with an incidence of0.2 to 1/million/year, 1,2 They are usually oligoorpolyclonal autoantibodies type IgG I and IV;however, IgA and IgM inhibitors against factorVIII have been described as well. 3-5Inhibitor occurrence leads to a decompensationof the coagulation system, resulting in mostly lifethreateningsoft-tissue bleeding. Lethality used tobe 22%. 6,7 The fact that according to recent studieslethality is reduced to 7.5% may be attributableto the availability of coagulative APCC and factorVIIa concentrates. 8 The treatment regimensdirected against autoantibodies are based onimmunosuppression by cyclophosphamide, prednisolone,azathiophrine, vincristine and others.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 79Another treatment option is referred to as immunomodulationusing immunoglobulin substitutionin order to trigger inhibitor elimination. 9 Thepotential to treat acquired hemophilia with FVIIIand cyclophosphamide has been apparent for 30years. 10 In a randomized trial various immunosuppressiveregimens with prednisone and cyclophosphamidereduced inhibitor titers to undetectablelevels in 68% of patients. 11 However theseresults have been partially attributable to concomitanttherapy with coagulation factor concentrates.Furthermore, a complete response toimmunosuppressive therapy in acquired inhibitorpatients may require weeks or months. 11-14Assessment of the current data available is difficultgiven the small number of cases as well asthe inconsistent description of the differenttreatment protocols and experiences with respectto end point definitions of a successful therapy,treatment duration, undesirable effects and especiallythe occurrence of secondary bleedingevents and the amount of activated prothrombincomplex concentrate (aPCC) and recombinantfactor VIIa (rFVIIa) concentrate required. Thesame applies to cost aspects as far as they havebeen considered in publications so far.Availability of immunoadsorption and thus thepossibility of automated antibody reduction inthe patient´s plasma paved the way for the developmentof new treatment cycles. 15,16Based on our experience in the treatment ofpatients both with congenital and with acquiredinhibitors we at the Hemophilia Center in Bonndeveloped a new so-called Modified Bonn-Malmö Protocol (MBM Protocol) from a combinationof the Bonn Protocol (also referred to asImmune Tolerance Therapy, ITT) for the permanentelimination of FVIII inhibitor in patientswith congenital hemophilia A or B, 17-19 and amodified Malmö Protocol (with additional immunoadsorption,oral cyclophosphamide and IgGapplication). 15,20,22 The main focus is on antibodyreduction by means of immunoglobulin columnswhich allow adsorption of all IgG classes, as wellas on immunomodulatory treatment by antigenexposition using factor VIII concentrates 22,23 andapplication of foreign immunoglobulins undersimultaneous immunosuppressive treatmentwith corticosteroids and cyclophosphamide.Medicinal immunosuppression seemed indispensablegiven the experience in the history ofthe treatment of acquired inhibitor diseases.In the following, we will for the first time presentour therapy results achieved in patients withacquired factor VIII inhibitor, treated using conventionalimmunosuppressive or immunomodulatoryregimens as well as the MBM Protocol.MethodsThe MBM ProtocolThe treatment cycle consists of four componentswhich are applied during the course of one week.1. From day 1 to 5 long-term immunoadsorptionwith daily processing of 2.5 times thepatient’s plasma volume.2. Application of factor VIII concentrates,depending on the extent of inhibitor titer andbleeding situation, at a dosage of 100 units/kgbody weight every six hours (in exceptionalcases up to 200 units/kg). Standard dosereduction upon exceeding 100% factor VIIIplasma activity within a six or twelve hourrecovery under consideration of the clinicalrequirements.3. On day 5, 6 and optionally on day 7 applicationof immunoglobulins at a dosage of 0.4mg/kg body weight/day.4. From day 1 to 7 application of cyclophosphamide(2 mg/kg/BW) in combination withprednisolone (1 mg/kg/BW).Upon achieving inhibitor elimination immunosuppressionwas gradually reduced over a periodof six weeks.The treatment cycle is repeated in accordancewith clinical requirements and laboratory results.Application of rFVIIa or aPCC was usually performedprior to transportation to our hospital orprior to the placement of central venous cathetersunder difficult anatomical conditions.Immunoadsorption was accomplished byapheresis with a dual-column system (Ig-Thera-Sorb ® , PlasmaSelect AG, Teterow, Germany) ofsheep-derived polyvalent anti-human immunoglobulinbound to Sepharose CL 4B (AmershamPharmacia Biotech AB, Uppsala, Sweden).Conventional treatment strategies 1986 to 1997Prior to the development of the MBM Protocolour treatment of patients with acquiredinhibitors was guided by our own experience withthe Bonn Protocol as well as the guidelines andtherapy recommendations of current literatureand the clinical requirements.We used cyclophosphamide, prednisolone, vincristine,immunoglobulins, factor VIII concentratesand FEIBA in mono and in combinationtherapies. We also tried induction of immunetolerance by high-dose factor VIII application intwo patients, following therapeutic experiencewith hemophilia patients according to the BonnProtocol. Treatment was modified after sixmonths of unsuccessful therapy at the earliest.DiagnosisAcquired inhibitors were diagnosed by singlefactoranalysis in combination with a plasmaexchange and the Bethesda assay (BU) using theNijmegen method modification.In all patients and for the entire treatmentcycle, factor VIII was invariably determined usingone-stage clotting assay (Immuno) and chromogenicassay (Baxter). Lupus was excluded bymeans of lupus APTT and dRVVT (diluted RussellViper Venom Test).All treatment decisions were based on theresults of the one-stage clotting assay as due tothe high degree of dilution of the plasma samplehaematologica vol. 88(supplement n. 12):september 2003

80L. Hess et al.the chromogenic assay yielded a factor VIII activitywhich did not reflect in particular the clinicalaspect, i.e. the degree of bleeding diathesis. Thetwo measuring methods will only converge uponreaching a normal range of FVIII plasma activityin the one-stage clotting assay.The only treatment objective accepted was successfulinhibitor elimination. Successful inhibitorelimination was defined, for the above reasons, asthe normalization of factor VIII plasma levelsboth in the one-stage clotting assay and in thechromogenic assay. Factor activity had to bedemonstrated permanently over 100% for bothmethods without requiring any supportive factorVIII substitution or other drug therapy (immunosuppression,IgG administration).A single negative inhibitor test alone was notincluded in the definition as negative inhibitorresults were usually measured even at pathologicalfactor VIII activity.As soon as the patient’s clinical conditionallowed, diagnosis was performed with regard toinhibitor-associated disease, the main focusbeing on causal tumor disease.Total patient populationBetween 1986 and July 2001, we treated 24patients suffering from acutely life-threateninghigh-titer inhibitor against factor VIII (age: median69; mean 64.9; maximum 89; minimum 30years). Inhibitor titer at the beginning of treatment:median 61; mean 273.3; max. 3556; min.5 Bethesda units; women: 13, men: 11).All patients had a life-threatening bleedingwhich in most cases consisted in extensive truncalor soft-tissue bleedings. These caused problemsdue to displacement symptoms, e.g. compartmentsyndrome, obstruction of the airways,nerve lesions and secondary infection of hematomas.All 24 patients were examined forinhibitor-associated diseases. We diagnosed fourcollagenoses, three postpartal courses, three carcinomas(two prostate and one pulmonary carcinoma)as well as one Hashimoto’s autoimmunethyreoiditis as inhibitor-associated diseases.In 13 patients there was no indication oftypical inhibitor-associated disease.Patients treated using conventional strategiesFive patients with high-titer inhibitor againstfactor VIII (age: median 79, mean 76.7, maximum87, minimum 62) (inhibitors at initiationof treatment: median 51.5, mean 104, maximum308, minimum 5) (women: 3, men: 2) initiallyreceived immunosuppressive or immunomodulatorytreatment only. Two of those patients wereadditionally treated according to the Bonn Protocol.At the beginning of treatment, all patientsexhibited life-threatening bleeding. Threepatients who were refractory or suffered a relapsewere started on a therapy according to the MBMProtocol (Figure 1).Patients treated according to the MBMProtocolSince 1996 we have treated 22 patients accordingto the MBM Protocol, all of whom had hightiterinhibitor against factor VIII and sufferedfrom life-threatening bleeding at the beginningof treatment (Figure 1). The patient populationconsisted of ten men and twelve women with amean age of 64.7 years (median 69, maximum89, minimum 30 years). The highest inhibitortiter measured during inhibitor treatment was8,400 BU (median 63; maximum 8,400; minimum15).Results of the MBM ProtocolOf the 22 patients who suffered from high-titerfactor VIII inhibitor, 17 were able to completethe entire treatment protocol and were treateduntil total inhibitor elimination. Four patientshad to be excluded due to secondary underlyingdiseases even though they were responding wellto the therapy, with negative inhibitor test orreduction of inhibitor activity to below 10 BU.Despite all efforts, one patient could not be treatedsuccessfully under this treatment protocol.The following is the exemplary course of treatmentof 53-year-old female patient M.E., whosuffered from high-titer inhibitor (110 BU). Thepatient exhibited extensive soft-tissue hematomasof the trunk, the extremities and the face.In addition, she developed a compartment syndromeat the right lower leg during her stay at thereferring hospital. Rapid diagnosis and immediateinitiation of treatment using the MBM Protocolrendered the use of APCC or recombinantfactor VIIa unnecessary. The curve in Figure 2illustrates factor VIII increase, which had initiallybeen below 1% plasma activity, during thecourse of the MBM Protocol. The trianglesdepicted symbolize the days on which immunoadsorptionwas performed. The squares depictedsymbolize the days on which immunoglobulinswere applied. During the entire treatment cyclethe patient received oral prednisolone and cyclophosphamide.During the MBM Protocol wewere able to measure inhibitor titer reductionfrom 110 BU in parallel to the increase of factorVIII plasma activity. Factor VIII inhibitor was nolonger detectable on day 7 of treatment. Thedosages of factor VIII are illustrated in Figure 3.Application was performed depending on thepatient’s clinical condition as well as on the factorVIII plasma activity detected. The markeddecrease of factor VIII plasma activity upon stoppingthe application of exogenous factor VIII wascompensated for by repeating immunoadsorptionon days 19 and 21. The outcome of treatment,which is illustrated here, is representativefor all 17 patients who were able to complete theentire treatment protocol.In patients with life-threatening bleeding eventswe have to differentiate four stages until inhibitorelimination within the framework of the MBMhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 81Figure 1. Patients with high titer FVIII inhibitor.Protocol. First, the bleeding event is brought undercontrol within the first one or two apheresis treatmentdays. In the further course the inhibitor testis negative after an average of five apheresis treatmentdays. Finally, factor VIII application becameunnecessary after a mean 15.2 days. In 11 casesfurther apheresis treatment was required afterstopping factor application since upon discontinuationof exogenous factor VIII the factor VIIIactivity in the patient´s plasma decreased, as wasshown in the exemplary case of patient M.E. Onaverage, 2.5 additional apheresis days wererequired (median 3, maximum 5, minimum 1apheresis day).In total, the mean duration of treatment untillong-term inhibitor elimination is 16.6 apheresisdays (median 16, maximum 36, minimum 6apheresis days) (see Table 1).Therapy discontinuation in four patientsduring MBM protocolIn four patients with high-titer inhibitor againstfactor VIII and at the same time life-threateningbleeding events (soft-tissue bleedings) treatmentaccording to MBM Protocol had to be discontinued.In one case, this decision was made becauseof a diagnosis of pulmonary carcinoma. Withregard to the poor prognosis the patient decided todiscontinue therapy. In two patients 89 and 76years of age treatment had to be discontinued dueto extremely low cardio-pulmonary tolerance. Afourth patient, 81 years of age and with pre-existingarteriosclerosis, developed a stroke and had todiscontinue treatment in the light of her considerablyreduced general condition.In all four patients the bleeding had been undercontrol within one or two apheresis treatments.Table 1.Figure 2.haematologica vol. 88(supplement n. 12):september 2003

82L. Hess et al.36 months (median 28, minimum 1 month).None of the patients exhibited another bleedingtendency or inhibitor activity. Of these 17patients three died due to events not related tobleedings; one after 1 month, one after 10months, and one after 20 months.Figure 3.Treatment failures during MBM-ProtocolIn one case, inhibitor elimination failed despiteintensive therapeutic efforts. Unlike the previouslymentioned patients this 47-year-old manwas not in an acute bleeding interval at thebeginning of treatment. However, his medicalhistory revealed several severe bleedings in thearea of the lower extremities. Treatment extendedover a period of 75 days of hospitalizationincluding 62 apheresis days. The inhibitor withan initial value of 3,556 BU was reduced to 6 BU.One reason for the patient´s failure to respondto the therapy may have been the presence ofsevere adiposity (160 cm, 144 kg), which duringthe course of treatment progressed to 155 kg. Therequired factor VIII dosages per kg body weight aswell as the dosages of immunosuppressants werenot achieved; in addition the adsorption timesand as a result the plasma volumes were reducedto 50% of the target value.Follow-up of MBM-ProtocolFor the 17 successfully treated patients a meanfollow-up period (end of July 2001) of 31months was reached. The longest follow-up wasApplication of recombinant factor VIIa duringMBM ProtocolFactor VIIa was used in six cases, for a medianduration not exceeding three apheresis days(mean 3, minimum 1, maximum 5 days). Medianconsumption was 12,000 KIU (mean 15,000;maximum 30,000; minimum 2,200). Applicationwas indicated in five patients due to acutetreatment started in the peripheral hospital andcontinued until hospitalization in our clinic. Inone case, it was only indicated as a first-line treatmentand bleeding prophylaxis for the placementof a central venous catheter in the presence ofsevere adiposity.Factor VIII application during MBM ProtocolAs regards the life-threatening condition, allpatients exhibit a similar clinical picture and hightiterinhibitor irrespective of the treatment regimen.Factor VIII application was a mean of 552,000units in those patients (n=17) who received theentire scope of the MBM Protocol. One caserequired infusion of 2 million units. The median,which is 384,000 units, on the other handindicates a lower factor consumption. Onepatient required only 134,000 factor VIII units.Results of conventional inhibitor therapyOf the five patients who were treated usingimmunosuppressive or immunomodulatorystrategies inhibitor elimination was achieved inone patient (SG) after 15 months (see Table 2).The patient exhibited an initially life-threateningbleeding and subsequently four additional softtissuebleedings, some of them with an effect onhemoglobin, under the cyclophosphamide mono-Table 2.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 83therapy (1 mg/kg/BW) applied. In total 1.8 millionFEIBA units were applied, for vital treatmentas well as for prophylactic treatment of furtherbleedings. Today, 9 years after normalization ofplasma activity, the patient does not exhibit anynew inhibitor activity.Another patient (TT) received monotherapywith prednisolone as application of cyclophosphamidehad led to severe leukopenia in severaltreatment attempts. After overcoming the initiallylife-threatening bleeding event, the patientwas treated with FEIBA prophylaxis on an outpatientbasis for almost 10 years.Inhibitor elimination was not achieved withinthat period. Factor requirement during that timewas 15 million FEIBA units. Despite intensivebleeding prophylaxis a total of 15 bleeding eventswere observed in that patient. The patient died ofa disease not associated with any bleeding event.Another three patients (RA, PK, MU) were initiallytreated using conventional combinationtherapies (cyclophosphamide, prednisolone, vincristineand IgG). Two patients additionallyreceived factor VIII within the framework of anITT according to the Bonn Protocol. Initially,these patients also suffered from a life-threateningbleeding event. Duration of treatment was13-67 months (mean 36.7) until these threepatients were switched to treatment according tothe MBM Protocol. During conventional treatmentthe patients received 0.056 to 4.3 millionfactor VIII units (mean 1.78). FEIBA consumptionwas 0.77 to 5.8 million units; however, 1-28(mean 12) additional bleedings were observed.Only one patient (PK) experienced inhibitorremission for the duration of 12 months untilanother life-threatening bleeding occurred,which was the reason for starting MBM-P (seeTable 2).The treatment results of the three patients (RA,PK, MU) who after unsuccessful conventionaltreatment changed to the MBM Protocol areillustrated once more in table 3.Under the MBM Protocol long-term inhibitorelimination was achieved in patients RA and PKDue to the fact that she developed pulmonarycarcinoma as a secondary disease, patient MUdecided to discontinue the therapy, as alreadyexplained above.Comparison of factor concentrate consumption(FVIII and FEIBA) during the conventionaltreatment phase of the two patients RA and PKyields that the MBM Protocol reduces clottingfactor consumption by 1.18 million units (64%).When looking at the procurement costs for FEI-BA and factor VIII the difference between the twotherapies in terms of treatment costs is furtherincreased as the procurement costs for one FEI-BA unit are well above those for factor VIII, andduring the MBM Protocol, application of FEIBAor recombinant FVIIa is almost entirely dispensable.In addition, neither patient (RA or PK)experienced another bleeding event after conclusionof the MBM Protocol. The follow-up periodis currently between 41 and 62 months, duringwhich time no new bleedings and therefore nobleeding-related follow-up costs have occurred.DiscussionThe Modified Bonn-Malmö Protocol presentedhere can be characterized by the following properties.With the MBM Protocol, life-threateningbleeding events can be controlled within 24-48hours. The hemostatic effect of factor VIII underthe MBM Protocol can be described as promptand direct. The high success rate of 94.4% (n=18)illustrates the high reliability of the therapy presentedhere. Considering the long-term follow-upresults (median: 28 months) inhibitor eradicationcan be considered as permanent. Theextremely short treatment period of 16.5 apheresisdays until inhibitor elimination ensures theprevention of undesirable effects (compartmentsyndrome, fistula formation, etc.) as may occurduring immunosuppression with cyclophosphamideor prednisolone over several months. Inaddition, secondary infections, which tend tooccur particularly during long-term immunosuppression,can be avoided. We can also assume lowercosts for short-term treatment with factor FVIII(median: 384,000 units) within the frameworkof the MBM Protocol as compared to conventionaltherapies (up to 4.3 million FVIII units).Argumentation in favor of factor VIII applicationusing apheresis treatment is based on its veryearly hemostatic effect, 24 so that use of othercost-intensive coagulation concentrates (APCCand rFVIIa) can be reduced or entirely eliminated.The costs of an exclusively immunosuppressiveand immunomodulatory treatment aredetermined by the use of APCC and rFVIIa, whichare indispensable for conventional treatment regimens,especially in the presence of bleedingevents. 25,26As far as economic efficiency is concerned, theprocurement costs for APCC and rFVIIa have tobe taken into consideration, which are more thantwice the expenses for factor VIII. In addition,our experience shows that during conventionaltreatment the drugs APCC and rFVIIa have to beapplied in hemostatically effective dosages whichaccount for final costs far exceeding the treatmentcosts associated with the MBM Protocol.Our results also show that the use of APCCwithin the framework of a conventional therapydoes not effectively protect the patients from newbleedings (1-28 bleedings). Under the MBM Protocolon the other hand there was only the primarybleeding event, no re-bleedings.Apart from the hemostaseological necessity offactor VIII therapy the immunological competenceof factor VIII during the three weeks oftreatment according to the MBM Protocol is amatter worthy of discussion.In contrast to inhibitor patients with hemophilia,patients with acquired inhibitors still havethe ability to synthesize their own (endogenous)factor VIII. Exogenous factor application at reg-haematologica vol. 88(supplement n. 12):september 2003

84L. Hess et al.Table 3.ular intervals as specified in the MBM Protocol incombination with immunoadsorption and thecontinued endogenous FVIII synthesis resultedin detectable factor VIII plasma levels within thefirst 1-2 apheresis treatments.In addition, application of exogenous factorVIII creates an antigen stimulus, which ininhibitor patients both of acquired as well as congenitalgenesis results in a booster effect due toincreased lymphocyte proliferation. 27,28 Thisshould also be discussed with regard to an elevatedsensitivity of inhibitor clones towardsimmunosuppressants.A minor booster effect is also observed duringthe first treatment cycle of the MBM Protocol.This is of no clinical relevance in the presence ofdetectable FVIII recovery values. This short-termquantitative increase in inhibitor activity is neutralizedby the patient’s exogenous and endogenousFVIII levels as FVIII plasma activity remainsdetectable and there is clinical proof of a hemostaticeffect. In the end, booster effect is of lowclinical relevance as the inhibitor titers are nolonger detectable after a median of five apheresisdays. Whether or not booster effect is of any elementarysignificance for inhibitor eliminationcannot be decided at present. There is empiricalproof at least of limited effectiveness of prednisoloneand cyclophosphamide in the treatmentof hemophilia patients with alloantibodies asopposed to the treatment of patients withacquired autoantibodies. 25 In the combination ofmeasures (FVIII, immunoadsorption, IgG application)the additional immunosuppressive effectleads to permanent inhibitor elimination. TheFVIII levels measured, as in the case reported, are6-8-12-hour recovery values. We can assumethat after application of FVIII plasma levels arereached which sufficiently explain the earlyhemostatic effect we have observed, even at lowrecovery values. 24 Therapeutic response cantherefore be monitored by determination of FVIIIactivity, apart from the clinical parameters. Ourexperience suggests that the factor VIII activitiesmeasured by one-stage clotting assay are closelyrelated to bleeding tendency and its decrease duringthe MBM Protocol.This type of factor VIII determination allowsmonitoring the treatment of bleedings as well asthe success of therapy, so that treatment can beadjusted as appropriate. Comparable laboratorymonitoring is currently not available for APCC orrFVIIa therapy, or only at an unsatisfactory level.8,29The MBM Protocol consists of two components:the modified Bonn Protocol and the modifiedMalmö Protocol. In the Bonn Protocol, which wasdeveloped in 1974/75, 100-150 factor VIIIunits/kg BW are applied daily until permanentinhibitor elimination. For the Malmö Protocol,the Bonn Protocol was modified, and infusion ofcomparable amounts of factor VIII is preceded byimmunoadsorption if the inhibitor titer exceeds10 BU. As opposed to the Malmö Protocol theMBM Protocol involves long-term immunoadsorptionover several weeks. The target volume tobe absorbed within one apheresis day (2.5 timesthe plasma volume) is reached after 4-5 hours.This permits quick inhibitor reduction and, withthe Therasorb columns used, the binding of allIgG subclasses.In addition, immunosuppressive therapy waschanged from short-term bolus therapy usingcyclophosphamide within the Malmö Protocol tocontinuous oral administration in combinationwith oral prednisolone for up to 6 weeks afterinhibitor eradication.IgG is applied in a changed mode on a regularand standardized basis after several days ofimmunoadsorption.Conclusive assessment is problematic due tothe difficult data situation in current literature.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 85This is due to the low incidence of inhibitor formation.Suggested models for explaining themodes of action of the MBM Protocol presentedhere are of a rather speculative nature. The focusremains on empirical experience and the statedadvantages of the MBM Protocol over conventionaltherapies in the treatment of inhibitorpatients.The working group of Prof. R. Zimmermannand Dr. A. Huthe-Kühne in Heidelberg (Germany)came up with similar results for the treatmentof patients with acquired inhibitors. Theonly difference to the MBM Protocol, whichinvolves bolus injections, is continuous i.v. applicationof factor during the treatment cycle,which proved to be efficient.The life-threatening nature of acquired factorVIII inhibitors necessitates the use of highly efficienttreatment options. We think that if at all,reviewing individual treatment components of theMBM Protocol should be done with great cautionso as not to jeopardize the chances of recovery andthe treatment outcome. In addition, in view ofthe low incidence and the resulting low case numberswe advise against further modification of theprotocol for the time being, in favor of a growingpatient population. Testing the MBM Protocolpresented here at other centers would be appreciated.References1. Lusher JM. Factor VIII inhibitors. Etiology, characterization,natural history, and management. Ann NY AcadSci 1987;509:89-102.2. Cohen AJ, Kessler CM. Acquired inhibitors. BaillièresClin Haematol 1996;9:331-54.3. Lavergne JM, Meyer D, Reisner H. Characterization ofhuman antifactor VIII antibodies purified by immunecomplex formation. Blood 1976;48:931-6.4. Glueck HI and Hong R. A circulating anticoagulant inIgA-multiple myeloma: its modification by penicillin. JClin Invest 1965;44:1866-81.5. Hoyer LW, Gawryl MS and de la Fuente B. Immunochemicalcharacterization of factor VIII inhibitors. ProgClin Biol Res 1984;150:73-85.6. Kessler CM. An introduction to factor VIII inhibitors:the detection and quantitation. Am J Med 1991;91:1S-5S.7. Green D, Lechner K. A survey of 215 non-hemophilicpatients with inhibitors to Factor VIII. Thromb Haemost1981;45:200-3.8. Hay CR, Negrier C and Ludlam CA. The treatment ofbleeding in acquired haemophilia with recombinant factorVIIa: a multicentre study. Thromb Haemost 1997;78:1463-7.9. Sultan Y, Kazatchkine MD, Nydegger U, Rossi F, DietrichG, Algiman M. Intravenous immunoglobulin inthe treatment of spontaneously acquired factor VIII:Cinhibitors. Am J Med 1991;91:35S-9S.10. Green D. Suppression of an antibody to factor VIII by acombination of factor VIII and cyclophosphamide.Blood 1971;37:381-7.11. Green D, Rademaker AW, Briet E. A prospective, randomizedtrial of prednisone and cyclophosphamide inthe treatment of patients with factor VIII autoantibodies.Thromb Haemost 1993;70:753-7.12. Spero JA, Lewis JH, Hasiba U. Corticosteroid therapy foracquired F VIII:C inhibitors. Br J Haematol 1981;48:635-42.13. Lian EC, Larcada AF, Chiu AY. Combination immunosuppressivetherapy after factor VIII infusion foracquired factor VIII inhibitor. Ann Intern Med 1989;110:774-8.14. Shaffer LG, Phillips MD. Successful treatment ofacquired hemophilia with oral immunosuppressivetherapy. Ann Intern Med 1997;127:206-9.15. Nilsson IM, Berntorp E, Zettervall O. Induction ofimmune tolerance in patients with hemophilia andantibodies to factor VIII by combined treatment withintravenous IgG, cyclophosphamide, and factor VIII. NEngl J Med 1988;318:947-50.16. Knöbl P, Derfler K, Korninger L, et al. Elimination ofacquired factor VIII antibodies by extracorporal antibody-basedimmunoadsorption (Ig-Therasorb). ThrombHaemost 1995;74:1035-8.17. Brackmann HH. Induced immunotolerance in factorVIII inhibitor patients. Prog Clin Biol Res 1984;150:181-95.18. Brackmann HH, Oldenburg J, Schwaab R. Immune tolerancefor the treatment of factor VIII inhibitors—twentyyears’ ‘Bonn Protocol’. Vox Sang 1996;70 Suppl 1:30-5.19. Oldenburg J, Schwaab R, Brackmann HH. Induction ofimmune tolerance in haemophilia A inhibitor patientsby the ‘Bonn Protocol’: predictive parameter for therapyduration and outcome. Vox Sang 1999;77:49-54.20. Nilsson IM, Freiburghaus C. Apheresis. Adv Exp MedBiol 1995;386:175-84.21. Freiburghaus C, Berntorp E, Ekman M, Gunnarsson M,Kjellberg BM, Nilsson IM. Immunoadsorption forremoval of inhibitors: update on treatments in Malmö-Lund between 1980 and 1995. Haemophilia 1998;4:16-20.22. Brackmann HH, Gormsen J. Massive factor-VIII infusionin haemophiliac with factor-VIII inhibitor, highresponder. Lancet 1977;2:933.23. Gilles JG, Desqueper B, Lenk H, Vermylen J, Saint-RemyJM. Neutralizing antiidiotypic antibodies to factor VIIIinhibitors after desensitization in patients with hemophiliaA. J Clin Invest 1996;97:1382-8.24. Francesconi M, Korninger C, Thaler E, Niessner H,Hocker P, Lechner K. Plasmapheresis: its value in themanagement of patients with antibodies to factor VIII.Haemostasis. 1982;11:79-86.25. Kessler CM. Acquired factor VIII autoantibodyinhibitors: current concepts and potential therapeuticstrategies for the future. Haematologica 2000;85 Suppl10: 57-63.26. Inhibitor Subcommittee of the association of hemophiliaclinic directors of Canada. Suggestions for themanagement of factor VIII inhibitors. Haemophilia2000;6 Suppl 1:52-9.27. Schroeder JO, Euler HH, Löffler H. Synchronization ofplasmapheresis and pulse cyclophosphamide in severesystemic lupus erythematosus. Ann Intern Med 1995;107:344-6.28. Jarreuse B, Blinchet P, Gayrand M et al. Synchronizationof plasma exchanges and cyclophosphamide in severesystemic diseases. Press Med 1993;22:293-8.29. Hedner U, Glazer S, Falch J. Recombinant activated factorVII in the treatment of bleeding episodes in patientswith inherited and acquired bleeding disorders. TransfusMed Rev 1993;7:78-83.haematologica vol. 88(supplement n. 12):september 2003

[Acquired Inhibitors in Non-Hemophiliacs]review paperManagement of severehemorrhage and inhibitor-eliminationin acquired hemophilia:the modified Heidelberg-Malmöprotocolhaematologica 2003; 88(suppl. n. 12):86-92http://www.haematologica.org/free/immunotolerance2001.pdfA. HUTH-KÜHNE, P. LAGES, H. HAMPEL, R. ZIMMERMANNKurpfalzkrankenhaus and Haemophilia Center, Heidelberg,GermanyThe overall goals in the treatment of patients withacquired hemophilia are the management of acutebleeds and permanent inhibitor elimination. A newtreatment option, represented by the modified Heidelberg-Malmöprotocol (MHMP), has been successfullyapplied to 8 patients with acquired hemophiliaand severe bleeding complications. It consistsof long-term immunoadsorption (IA) by Ig-apheresisin combination with high-dose factor VIII (FVIII),intravenous immunoglobulin (IVIG) between adsorptioncycles, and immunosuppressive therapy withcyclophosphamide and prednisolone to avoidautoantibody rebound. This treatment was continueduntil patients maintained normal FVIII:C levelswithout replacement therapy. Acute bleeding wascontrolled in all patients following the application ofrecombinant factor VIIa (rFVIIa). With this new protocol,which combines 4 immuno-modulatory strategies,inhibitors were eliminated in all cases and allpatients continue to be in remission.Correspondence: Dr A. Huth-Kühne, Kurpfalzkrankenhaus andHaemophilia Center, Bonhoefferstraße 5 D 69123 Heidelberg,Germany. Phone: international +49.06221.884088.Fax: international +49.06221.88401.E-mail: angela.huth-kuehne@kkh.srh.deAcquired hemophilia is a rare but sometimeslife-threatening bleeding condition causedby autoantibodies depleting circulatingFVIII. Autoantibodies to FVIII:C constitute themost common spontaneous inhibitor to anycoagulation factor and arise in patients with previouslynormal FVIII:C activity in association witha diverse variety of clinical settings. Inhibitor formationis rare with an incidence of 0.2-1 per onemillion persons per year, 1 although clinical experiencesuggests, that this condition was either previouslyunderdiagnosed or is increasing in incidence.These autoantibodies often occur as anepiphenomenon of autoimmune disorders andmay represent a component of immune dysfunction.Conditions associated with spontaneousinhibitor formation include rheumatoid arthritis,bronchial asthma, myasthenia gravis, systemiclupus erythematosus, as well as postpartum periodand malignancies. 2-5 However, in approximately50% of cases no underlying disease isfound. 1,6Concerning pharmacokinetics, there is a significantdifference between allo- and autoantibodies.Alloantibodies to FVIII:C (type I), whichcan develop in response to replacement therapy inpatients with congenital hemophilia, display atype I reaction kinetic with a linear and completeinactivation of FVIII. In contrast FVIII autoantibodies(type II) inactivate endogenous FVIII in anon linear and incomplete fashion and display acomplex type II inhibition kinetic. 7,8 As a resultthere may be no correlation between inhibitortiter, FVIII activity and the severity of bleeding.Inhibitors are quantified using the Bethesdaassay,which depends on the measurement ofresidual FVIII activity and a linear inhibitionkinetic. Therefore quantitation of type IIinhibitors is very arbitrary, because the inhibitorneither inactivates FVIII completely nor becomessaturated with FVIII and the in vivo inhibitorpotency may be underestimated. 7,8 These propertiescorrespond to the clinical situation, that lowlevels of FVIII may still be detectable despite theconcomittant presence of a high titer inhibitorand extensive bleeding. Recovery and half-life ofinfused FVIII may be significantly reduced evenhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 87in those patients with low levels of inhibitors.The clinical course of patients with acquiredhemophilia is characterized by progressive anddramatic hemorrhages with a different bleedingpattern compared to patients with congenitalhemophilia and inhibitors. Severe soft tissuebleeds, major life-threatening abdominal andretroperitoneal hemorrhages, as well as extensivemuscle bleeding predominate and may lead to afatal outcome in up to 22% of cases. 1,6The dual treatment objectives are the managementof acute bleeds and permanent inhibitorelimination to restore hemostasis. For the managementof severe and life-threatening hemorrhagedifferent treatment options are availableincluding hemostatic agents such as high-doseFVIII or porcine FVIII and inhibitor bypassingproducts, including PCCs with a reported efficacyof 50% and APCCs with an efficacy up to 70%. 9,10A new treatment option is recombinant factor VIIin its activated form (rFVIIa, NovoSeven®) witha reported efficacy rate of 80-90%. 11,12 ExtracorporealIA is another option, when an immediatereduction in the level of autoantibodies isrequired. 13-18For the management of permanent inhibitorelimination, different treatment modalities havebeen applied with varying degrees of success. Theseinclude a range of immunosuppressive drugs givenalone or in combination, with or without priorhigh-dose FVIII 19,20 and IVIG. 21,22 Immune toleranceinduction (ITI) with high-dose FVIII, asapplied in the Malmö protocol 14-16 and the Bonnprotocol, 23,24 is a form of immunemodulation,which has been successfully applied to patientswith congenital hemophilia and inhibitors. (TheMalmö protocol consists of high-dose FVIIIadministered several times a day as bolus injectionswith an initial dose of IVIG (2.5-5 g),cyclophosphamide for 8-10 days, first intravenouslyfor 2 days then orally, as well as IVIG0.4 g/kg for 5 days beginning on day 4 of treatment.If the initial inhibitor level exceeds 10 BU,treatment is preceded by IA to protein A columnsto remove the antibodies).To date, ITI protocols have rarely been applied topatients with acquired hemophilia. The need for amore effective treatment of patients with acquiredhemophilia led to the development of the modifiedBonn-Malmö protocol (MBMP), referred toinitially as modified Bonn protocol. It has beenintroduced in 1997 and promising data haverecently been reported. 25,26 It represents a newtherapeutic approach to inhibitor-elimination andmanagement of acute hemorrhage in patientswith acquired hemophilia and consists of longtermIA, high-dose FVIII, immunosuppressivetherapy and IVIG between adsorptions.We modified this protocol in our center in Heidelbergwith regard to FVIII dosage and application.In addition all patients received rFVIIa tocontrol major hemorrhages. Our protocol isreferred to as modified Heidelberg- Malmö protocol(MHMP). 27,28Design and MethodsWe treated 8 patients with acquired hemophiliaand severe life-threatening bleeds with thismodified Heidelberg- Malmö- treatment protocolas follows:1. IA (Therasorb‚) on 2-5 consecutive days eachweek. Length of treatment was based on a Mondayto Friday schedule and was therefore determinedby the day of patient admission.2. IVIG (0.3 g/kg) on 3 consecutive days (Fridayto Sunday) in between adsorption cycles, startingon the day of the last IA.3. High-dose FVIII starting after the first IA.4. Immunosuppressive therapy until remission,beginning at the same time as IA and includingcyclophosphamide 2 mg/kg per day and prednisolone1 mg/kg per day to avoid autoantibodyrebound.5. rFVIIa applied as continuous infusion (0.75-1 KIU/kg per hour) following 1 or 2 initial bolusinjections (4.5-6.0 KIU/kg), aiming at FVII:C levelsof 10-15U/mL. When FVIII:C levels increasedto >30% under this specific therapy, treatmentwith rFVIIa was discontinued.For dosing of FVIII, we administered a bolus of200 IU/kg FVIII followed by 200 IU/kg per day ascontinuous infusion starting after the first IA. Assoon as plasma FVIII:C levels increased to normal(>70%), the FVIII dose was gradually reduced insteps of 50 IU/kg. The treatment cycles were continueduntil patients reached normal FVIII:C levelswithout replacement therapy (complete remission,CR). Following CR, immunosuppressivetherapy with cyclophosphamide and prednisolonewas continued for 2-4 weeks and IVIG twice aweek for 4 weeks.Extracorporeal antibody-based IABlood was taken from either peripheral venousaccess or a central venous catheter at a flow rateof 40-60 mL/minute. Heparin was added as ani.v. bolus of 2000 U, which rose to 5000 U accordingto the increase in FVIII:C activity, followed by500-750 U/hour during IA. In addition, 0.15 Mcitrate (ACD-A anticoagulant citrate dextrose, formulaA, Baxter) was used as an anticoagulant. Thevolume ratio of citrate to blood was maintained atbetween 1:18 and 1:22. Plasma separation wasperformed with an Autopheresis-C TherapeuticPlasma System (Baxter). Two columns, each containing150 mL sepharose coupled with polyclonalsheep antibodies to human immunoglobulin(IgG, IgA and IgM) heavy and light chains (Ig-Therasorb ® ), were used for immunoglobulinremoval. Each column has an immunoglobulinbinding capacity of approximately 4 g. The plasmais directed to one column and the immunoglobulinsof the patient are bound to the immobilizedsheep antibodies. In one adsorption cycle (lasting15 minutes) 400 mL of plasma were loaded ontoone column, while the other column was regenerated.A total of 12-20 cycles were performed inone IA session. Thus, a total plasma volume of4.8-8 L passed through the columns. Columnshaematologica vol. 88(supplement n. 12):september 2003

88A. Huth-Kühne et al.were regenerated by protein elution with glycinebuffer at p.H. 2.8, followed by washing cycles withphosphate-buffered saline and 9 g/L isotonic sodiumchloride solution. An adsorption-desorptionautomate system (ADA system, Baxter) was used tocontrol the flow of plasma and regenerating solutionsto the IA columns. After passing through thecolumns, the plasma and separated blood cellswere re-infused into the patient.AssaysClotting assays were performed with citratedplasma on an Amax Coagulometer (Amelung,Lemgo-Lieme, Germany). FVIII:C activity wasdetermined with a one-step clotting assay using ahuman FVIII:C deficient (Haemophilia A) plasma(Immuno, Austria) and a Dapttin reagent(Immuno, Austria) (normal range 0.7-1.1U/mL). Activated partial thromboplastin time(APTT) was determined using a Dapttin reagentconsisting of kaolin and sulphatide as surface activatorsand a blend of highly purified phospholipids(Immuno, Austria) (normal range 32-42seconds). FVIII:C inhibitory activity was quantifiedby the Bethesda assay. 29 The inhibitor titre wasexpressed as the reciprocal of the patient plasmadilution, which yielded 50% of the residual FVIII:Cactivity in the test system.ResultsBetween December 1997 and November 2000eight patients with autoantibodies to FVIII:C werereferred to the MHMP. All patients were transferredto our hospital because of severe bleedingcomplications. The median age of patients was 73years (range 60-82) with a gender distribution of5 males and 3 females. Six of the 8 patientsshowed an associated underlying medical condition.3 patients were taking immunosuppressivedrugs due to their underlying disease (Table 1).Maximum inhibitor titres, inhibitor titres andFVIII:C levels prior to IA are shown in Table 2. Theeffect of the immunomodulatory treatment oninhibitor titres and FVIII:C activity in 2 patients isshown in Figures 1A and 1B. After a median of 3.5adsorptions, patients FVIII:C levels increased to>30%, while the inhibitor was no longerdetectable in plasma. CR in all patients requiredcontinuous FVIII replacement for a median of 19days and 16 adsorptions. (Table 3). In contrast toother therapeutic strategies, CR in all patients wasachieved in less than 3 weeks in median.Therapywas well tolerated by all patients without severeside effects. To date, after a median follow-up periodof 47 months (range 8-52 months), 7/8patients continue to be in remission.One patientdied due to a plasmocytoma, which had been inremission at the time of the diagnosis of acquiredhemophilia.DiscussionTreatment of patients with FVIII or FIX autoantibodiesstill represents a major challenge. Thehemorrhagic manifestations in these patients aremore dramatic and often life-threatening as comparedto bleeds associated with inhibitory alloantibodies.In most cases, immediate intervention tocontrol the acute bleeding is necessary and anychoice of therapeutic modality should be based onan algorithm which considers primarily the clinicalpresentation and severity of bleeding. Therapeuticdecisions should not be based only on theinhibitor level, because the Bethesda-assay is lesspredictive of FVIII recovery in patients withacquired hemophilia and often underestimates thein vivo potency of type II inhibitors. 7,8 For themanagement of acute bleeds treatment optionsconsist of high-dose FVIII (plasma-derived orrecombinant), porcine FVIII depending on crossactivity,or administration of plasma-derivedbypassing agents, including PCCs and APCCs.A new treatment option to bypass the inhibitorand achieve hemostasis, is the administration ofrFVIIa, 11,12 which has been reported to be clinicallyeffective and safe in patients with congenitalhemophilia and inhibitors and in patients withacquired hemophilia. In healthy individuals FXactivation on the surface of activated plateletsand the resulting necessary burst of thrombin generationis ensured by the Tenase complex consistingof FIX and FVIII. If one part of this complex ismissing or compromised by autoantibodies, thenecessary amount of thrombin is not generated.Administration of rFVIIa in supraphysiologicaldoses activates FX on the surface of activatedplatelets attached to the injured vessel wall providingthe necessary amount of thrombin. 30 Comparedto conventional bypassing products, whichare known to cause systemic activation of coagulationand thromboembolism, 31-33 the risk ofthrombotic complications seems to be very lowafter administration of rFVIIa. Because of its modeof action, rFVIIa has much less thrombogenicpotential and although administered in supraphysiologicaldoses, to date only few side effectshave been reported. 34 There is no anamnesticresponse and it has been shown to be very effectiveirrespective of the inhibitor titer. 12 Clear disadvantagesare the high costs when applied asbolus infusion with a current recommended doseof 90–120 mg/kg every 2 hours and the lack ofany specific laboratory control other than FVII levels.Adequate hemostatic levels still remain to beestablished. Administration by continuous infusionis a new approach and may reduce FVIIarequirement up to 50%. 35,36Concerning permanent inhibitor eliminationthe most successful treatment to date is theadministration of immunosuppressive and cytotoxicagents. 1,37 Due to their myelosuppressiveeffects, the currently available cytotoxic agentsmay lead to severe side effects. Neutropenia andsubsequent life-threatening infection have beenreported, 20 even after a single treatment, 38 andmay contribute to death. 39FVIII autoantibodies may also be responsive toimmune-modulating strategies, including IVIGwith varying response rates 22 and IA. A newhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 89Table 1. Patients’ characteristics. Table 2. Patients’ data prior to the start of treatment. IA =immunoadsorption.promising approach to immunemodulation maybe treatment with the monoclonal anti-CD20antibody rituximab.This treatment eliminates theclones of lymphocytes responsible for autoantibodyproduction. 40,41The increased risk of severe infections and septicemiaduring treatment with immunosuppressiveand cytotoxic regimens and the relatively lowand varying response rates to IVIG, suggests thatfurther investigation of therapeutic alternatives isnecessary.A new therapeutic approach is represented bytwo protocols, the modified Bonn-Malmö protocol,the first ITI protocol for patients with acquiredhemophilia and a second protocol, the modifiedHeidelberg-Malmö protocol. 25-28 Both protocolsinclude high-dose FVIII and IA for permanentinhibitor-elimination. To date 20 patients (personalcommunication) have been successfullytreated according to the MBMP, we have included8 patients to our protocol. In contrast to theoriginal Malmö protocol, IA is considered as animmunemodulating strategy and applied as longtermtreatment in both modified protocols.Another modification is the introduction of anew extracorporeal adsorption technique (Ig-Therasorb ® ) based on the selective IA of specificproteins to polyclonal sheep antibodies immobilizedon sepharose columns. This antibody-antigenreaction currently represents the most selectiveapproach to extract plasma components andeliminates all 4 subclasses of IgG, as well as IgM,IgA and circulating immunecomplexes. ThisFigures 1A,B [left column]. The following text is for figures1a and 1b dosing schedules for FVIII, rFVIIa, prednisolone,cyclophosphamide, IVIG, and application of immunoadsorption(IA), for the first 14 days of therapy in patients 5 (1a)and 8 (1b). FVIII:C activity is expressed as percentage ofactivity in normal plasma. FVIII:C inhibitor titre is expressedin Bethesda units (BU).haematologica vol. 88(supplement n. 12):september 2003

90A. Huth-Kühne et al.Table 3. Details of treatment andoutcome. IA = immunoadsorption.adsorption technique, introduced by Knöbl et al.in 1995, 42 has been successfully applied to theelimination of apolipoprotein, 43-45 HLA antibodiesin renal transplant recipients 46 and auto- andalloantibodies to FVIII:C. 25-28We included 8 patients to the MHMP. Inhibitorelimination was achieved in all patients and, aftera period of 43 months in median, 7 patients continueto be in remission. One patient died due tohis underlying medical condition, 8 months afterthe end of treatment. Therapy was well toleratedwithout any serious side effects. For the managementof acute bleeding, patients received rFVIIaas first-line therapy applied by continuous infusion.This mode of application reduced rFVIIarequirements and bleeding was successfully controlledin all patients. In contrast to the treatmentfollowing the MBMP (bolus injections of rFVIII100–200 IU/kg q.i.d.), we applied a modified factorVIII regimen with a single bolus of 200 IU/kgfollowed by 200 IU/kg per 24 hours as continuousinfusion. This mode of application was as effectiveas the original, despite a 50% reduction in dose,and yielded additional cost benefits.Immunosuppressive therapy and immune-modulation,including IVIG, high-dose FVIII andextracorporeal IA comprise the MHMP protocol.Because each of these immune-modulatory strategies,described below, target a different aspect ofthe immune system and work, in part, synergistically,we believe this treatment protocol to be verypromising.Immunosuppressive and cytotoxic agentsCorticosteroids suppress immune responsesmediated by B- and T-lymphocytes. By inhibitinginterleukin-2 they prevent B-cells from respondingto T-helper lymphocytes, thus reducing immunoglobulinproduction. 47Cytotoxic agents kill actively proliferating cells,acting preferentially on cancer cells and cells ofthe immune system. The survey by Green andLechner, 1 which included 215 patients withacquired inhibitors to FVIII:C, reported efficacyrates of 30% for corticosteroid medication alone,57% for corticosteroids and cyclophosphamide,and 68% for corticosteroids and azathioprine,respectively.ImmunoglobulinsIVIG exerts its effect by complex formation withthe circulating antibodies.This is mediated viaanti-idiotype antibodies present in the normalantibody population and in the pooled plasmathat comprise IVIG. 48-51 These anti-idiotypic antibodiesalso bind and downregulate the B-cellreceptor for antigen thus decreasing autoantibodyproduction. There are several reports and oneprospective clinical trial, which demonstrate variabledegrees of responsiveness to IVIG treatmentin patients with acquired hemophilia. 52,53 The factthat the binding of Fab2 fragments from theimmunoglobulin preparation to the autoantibodywas responsible for the suppression of the autoantibodyprovided the first convincing evidence forthe manipulation of the immune system by antiidiotypicantibodies. 21 Another mechanism ofaction might be an acceleration in the rate of IgGcatabolism. Such a change may be induced by highdoses of exogenous IgG and results in the eliminationof individual IgG molecules in direct proportionto their relative concentration in plasma.54,55Re-induction of immune tolerance with highdoseFVIIIMechanisms by which high doses of FVIIIreduce the anti-FVIII:C antibodies are not clearlyunderstood. They can induce anti-idiotypic antibodies,which bind to the variable region of thehaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 91anti-FVIII:C antibodies and neutralize them. 56Alternatively, high-dose FVIII may stimulate thepathologic cell clone responsible for antibody synthesisvia antigen presentation and render it moresusceptible to immunosuppressive and cytotoxicagents. 19Extracorporeal IA (Ig-Therasorb ® )This new adsorption technique was introducedfor the treatment of acquired hemophilia by Knöblet al. 42 It leads to an extensive reduction of immunoglobulinsfrom intravascular space and changesthe distribution of antibodies and immune complexesbetween intra- and extravascular compartmentsleading to mobilization of antibodies frominterstitial sites, especially when applied as longtermtreatment.ConclusionsAcquired hemophilia caused by inhibitory autoantibodiesto FVIII or FIX is still associated withhigh morbidity and mortality. Since rFVIIa is available,therapy of acute and life-threatening hemorrhagehas been considerably enhanced. Thechoice of treatment for acute bleeds should bebased on the clinical presentation rather than onthe level of the inhibitor titer. rFVIIa should beconsidered as first-line therapy because of its highefficacy and better side-effect profile as comparedto PCCs, APCCs.High dose continuous FVIII infusion as appliedin the MHMP appears to be an effective alternativeto bolus infusion as a means of presentingsufficient antigen and to reduce FVIII requirement.Considering the side-effects of immunosuppresivetherapy, which should be part of any regimen forinhibitor elimination, a major advantage of ourprotocol is the relatively short time-course of immunosuppressivetherapy. We achieved inhibitorelimination in less than 3 weeks (median) and allpatients continue to be in remission. The combinationof immunosuppression and immunomodulatorystrategies proved very successful and representsan alternative therapeutic option inacquired hemophilia. Further evaluation in a largernumber of patients is required.References1. Green D, Lechner K. A survey of 215 non-hemophilicpatients with inhibitors to Factor VIII. Thromb Haemost1981;45:200-3.2. Shapiro SS, Hultin M. Acquired inhibitors to the bloodcoagulation factors. Semin Thromb Hemost 1975;1:336-85.3. Palascak JE. Autoantibodies Against Clotting Factors. In:Acquired Bleeding Disorders in Children: abnormalitiesof Hemostasis. Lusher JM, Barnhart MI, eds. New York:Masson; 1981. p. 99-113.4. Lusher JM, Hillman CRL. The Effect of Inhibitors on FactorAssays. In: Advances in Coagulation Testing: interpretationand application. Triplett DA, editor. Skokie, Illinois;College of American Pathologists. 1986. p. 73.5. Green D. The management of factor VIII inhibitors innon-hemophilic patients. Prog Clin Biol Res 1984; 150:337-52.6. Kessler CM, Ludlam CA. The treatment of acquired factorVIII inhibitors: worldwide experience with porcinefactor VIII concentrate. International Acquired HemophiliaStudy Group. Semin Hematol 1993;30(Suppl1):22-7.7. Kasper CK. Complications of hemophilia A treatment:factor VIII inhibitors. Ann N Y Acad Sci 1991;614:97-105.8. Gawryl MS, Hoyer LW. Inactivation of factor VIII coagulantactivity by two different types of human antibodies.Blood 1982;60:1103-9.9. Morrison AE, Ludlam CA. Acquired haemophilia and itsmanagement. Br J Haematol 1995;89:231-6.10. Hay CR. Innovative use of porcine factor VIII:C forimmune tolerance induction. Am J Med 1991;91 Suppl5A:27S-9S.11. Hedner U, Glazer S, Falch J. Recombinant activated factorVII in the treatment of bleeding episodes in patientswith inherited and acquired bleeding disorders. TransfusMed Rev 1993;7:78-83.12. Hay CR, Negrier C, Ludlam CA. The treatment of bleedingin acquired haemophilia with recombinant factorVIIa: a multicentre study. Thromb Haemost 1997;78:1463-7.13. Nilsson IM, Jonsson S, Sundqvist SB, Ahlberg A, BergentzSE. A procedure for removing high titer antibodies byextracorporeal protein-A-sepharose adsorption in hemophilia:substitution therapy and surgery in a patient withhemophilia B and antibodies. Blood 1981;58:38-44.14. Nilsson IM, Sundquist SB, Ljung R, Holmberg L, FreiburghausC, Björling G. Suppression of secondary antibodyresponse by intravenous immunoglobulin in apatient with haemophilia B and antibodies. Scand JHaematol 1983;30:458-64.15. Nilsson IM, Berntorp E, Zettervall O. Induction ofimmune tolerance in patients with hemophilia and antibodyto factor VIII by combined treatment with intravenousIgG, cyclophosphamide, and factor VIII. N Engl JMed 1988;318:947-50.16. Freiburghaus C, Berntorp E, Ekman M, Gunnarsson M,Kjellberg B-M, Nilsson IM. Tolerance induction using theMalmö treatment model 1982-1995. Haemophilia 1999;5:32-9.17. Negrier C, Dechavanne M, Alfonsi F, Tremisi PJ. Successfultreatment of acquired factor VIII antibody byextracorporeal immunoabsorption. Acta Haematol 1991;85:107-10.18. Watt RM, Bunitsky K, Faulkner EB, Hatt CM, Horan J,Ramstack JM, Viola JL, Yordy JR, and the HaemophiliaStudy Group. Treatment of congenital and acquiredhemophilia patients by extracorporeal removal of antibodiesto coagulation factors: a review of US clinical studies1987-1990. Transfus Sci 1992;13:233-53.19. Green D. Suppression of an antibody to factor VIII bycombination of factor VIII and cyclophosphamides.Blood 1971;37:381-7.20. Lian EC, Lacarda AF, Chiu AY. Combination immunosuppressivetherapy after factor VIII infusion for acquiredfactor VIII inhibitor. Ann Intern Med 1989;110:774-8.21. Sultan Y, Kazatchkine MD, Maisonneuve P, Nydegger VE.Antiidiotypic suppression of autoantibodies to factor VIII(antihaemophilic factor) by high-dose intravenous gammaglobulin.Lancet 1984;2:765-8.22. Schwartz RS, Gabriel D, Aledort LM, Green D, KesslerCM. A prospective study of treatment of acquired(autoimmune) factor VIII inhibitors with high-doseintravenous gammaglobulin. Blood 1995;86:797-80423. Brackmann HH, Gormsen J. Massive factor VIII infusionin haemophiliac with factor VIII inhibitor, high responder.Lancet 1977;2:993.24. Brackmann HH. Induced immunotolerance in factor VIIIinhibitor patients. Prog Clin Biol Res 1986;150:181-95.25. Zeitler H, Unkrig C, Brackmann H, Effenberger C, HanflandD, Stier S, Ko Y, Vetter H. An immunomodulatorytreatment of acquired hemophilia A with long-term IgGIA, immunosuppression, and antigen substitution – amodified Bonn protocol inducing immunotolerance.Blood 1997; 90 Suppl 1:36a[abstract].26. Hess L, Zeitler H, Unkrieg C, Nettekoven W, EffenbergerW, Hanfland P, et al. Successful inhibitor eradication inhaematologica vol. 88(supplement n. 12):september 2003

92A. Huth-Kühne et al.patients with acquired haemophilia: modified Bonn-Malmö protocol. Haemophilia 2000;6:3027. Huth-Kühne A, Gebhardt A, Zimmermann R. New therapeuticoptions in acqiured hemophilia- the modifiedMalmö protocol. Haemophilia 2000;6:31228. Huth-Kühne A, Zimmermann R, Uhle C. Treatment withlong-term Ig-immunoadsorption and high-dose continuousfactor VIII in patients with acquired hemophilia andhemophiliacs with alloantibodies Thromb Haemost 1999;91 Suppl 1:277a[abstract].29. Kasper CK, Aledort LM, Counts RB. Proceedings. A moreuniform measurement of factor VIII inhibitors. ThrombDiath Haemorrh 1975;34:612.30. Monroe DM,Hoffman M, Oliver JA, Roberts HR. Plateletactivity of high-dose factor VIIa is independent of tissuefactor. Br J Haematol 1997;99:542-731. Chavin SI, Siegel DM; Rocco TA, Olson JP. Acute myocardialinfarction during treatment with an activated prothrombincomplex concentrate in a patient with factorVIII deficiency and a factor VIII inhibitor. Am J Med 1988;85:245-932. Lusher JM. Use of prothrombin complex concentrates inthe management of bleeding in hemophiliacs withinhibitors - benefits and limitations. Semin Hematol1994;31:49-52.33. Hough RE, Hampton KK, Preston FE, Channer KS,West J, Makris M. Recombinant VIIa concentrate inthe management of bleeding following prothrombincomplex concentrate-related myocardial in patientswith haemophilia and inhibitors. Br J Haematol 2000;111:974-9.34. Gallistl S, Cvirn G, Muntean W. Recombinant FactorVIIa Does not Induce Hypercoagulability In Vitro.ThrombHaemost 1999;81:245-9.35. Tagariello G, De Biasi E, Risato R, Radossi P, Davoli G,Traldi A. Recombinant FVIIa (NovoSeven) continuousinfusion and total hip replacement in patients withhaemophilia and high titre inhibitors to FVIII: experienceof two cases. Haemophilia 2000;6:581-3.36. Schulman S. Safety, efficacy and lessons from continuousinfusion with rFVIIa. Haemophilia 1998;4:564-737. Green D, Rademaker AW, Briet E. A prospective,randomizedtrial of prednisone and cyclophosphamide inthe treatment of patients with factor VIII autoantibodies.Thromb Haemost 1993; 70:753-7.38. Lottenberg R, Kentro TB, Kitchesn CS. Acquired haemophilia:a natural history study of 16 patients with factorVIII inhibitors receiving little or no therapy. Arch Int Med1987;147:1077-81.39. Morrison AE, Ludlam CA, Kessler C. Use of porcine factorVIII in the treatment of patients with acquired hemophilia.Blood 1993;81:1513-20.40. Lee EJ, Kueck B. Rituxan in the treatment of cold agglutinindisease. Blood 1998;92:3490-1.41. Karwal MW, Schlueter AJ, Zenk DW, Davis RT.Treatmentof Acquired Factor VIII Deficiency with Rituximab. Blood2001; 98 Suppl:2232a[abstract].42. Knöbl P, Derfler K, Korninger L, Kapiotis S, Jäger U,Maier-Dobersberger T, et al. Elimination of acquired factorVIII antibodies by extracorporal antibody-based IA (Ig-Therasorb®). Thromb Haemost 1995;74:1035-8.43. Du Moulin A. LDL Immunoapheresis Technique. In:Hypercholesterolemia in the Prevention of CoronaryHeart Disease. Gotto AM, Mancini M, Richter WO,Schwandt P, eds. Proceedings of the 2 nd InternationalSymposium in Munich, 1989. Basel, Switzerland: Karger,1990;170a[abstract].44. Derfler K, Swoboda K, Hirschl MM, Gottsauner-Wolf M,Steger G, Sunder-Plassmann G, et al. Comparison ofplasma separation and immunospecific LDL eliminationin severe hypercholesterolemia. Int J Artif Organs 1992;15:383-4.45. Richter WO, Jacob BG, Ritter MM, Sühler K, ViermeiselK, Schwandt P. Three year treatment of familial heterozygoushypercholesterolemia by extracorporal lowdensity lipoprotein IA with polyclonal apolipoprotein Bantibodies. Metabolism 1993;42:888-94.46. Derfler K, Druml W, Mayr W, Mühlbacher F, Jansen M,Sauter T, et al. Combination of IgG columns and IgG i.v.in the Downmodulation of Cytotoxic Anti-HLA Antibodiesand Autoantibodies. In: Congress of ImmunoglobulinsIntravenous – IgG i.v.: current status of clinicalapplication and research yopics. Lisbon, Portugal: BiotechGroup – Hyland Division; 1993. 15a[abstract].47. Dwyer JM. Manipulating the immune system withimmune globulin. N Engl J Med 1992;326:107-16.48. Gilles JG, Saint-Remy-JM. Healthy subjects produce bothanti-factor VIII and specific anti-idiotypic antibodies. JClin Invest 1994;94:1496-505.49. Gautier P, Sultan Y, Parquet-Gernez A, Meriane F,Guerois C, Derlon A. Detection and IgG subclass analysisof antibodies to FVIII in multitransfused hemophiliacsand healthy individuals. Haemophilia 1996; 2: 88-94.50. Dietrich G, Algiman M, Sultan Y. Origin of antiidiotypicactivity against anti-FVIII antibodies in pools of normalhuman immunoglobulin G, IVIG. Blood 1992;79:2946-51.51. Rossi F, Sultan Y, Kazatchkine MD. Antiidiotypes againstauto-antibodies and allo-antibodies to VIII:c (antihaemophilicfactor) are present in therapeutic polyspecificnormal immunoglobulins. Clin Exp Immunol 1988; 74:311-6.52. Green D, Kwaan HC. An acquired factor VIII inhibitorresponsive to high-dose γ globulin. Thromb Haemost1987;58:1005-7.53. Zimmermann R, Kommerell B, Harenberg J, Eich W,Rother K, Schimpf Kl. Intravenous IgG for patients withspontaneous inhibitor to Factor VIII. Lancet 1985;273-4.54. Zhiya Y, Lennon VA. Mechanism of intravenousimmune globulin therapy in antibody-mediated autoimmunediseases. N Engl J Med 1999;340:227-8.55. Masson PL. Elimination of infectious antigens andincrease of IgG catabolism as possible modes of IVIG. JAutoimmun 1993;6:638-9.56. Gilles JG, Desqueper B, Lenk H, Vermylen J, Saint-RemyJM. Neutralizing antiidiotypic antibodies to FVIIIinhibitors after desensitization in patients with haemophiliaA. J Clin Invest 1996;97:1382-8.haematologica vol. 88(supplement n. 12):september 2003

[Acquired Inhibitors in Non-Hemophiliacs]review paperAcquired factor VIII and factorIX inhibitors: a survey of ItalianHemophilia CentersFRANCESCO BAUDO, GIANNI MOSTARDA,haematologica 2003; 88(suppl. n. 12):93-99http://www.haematologica.org/free/immunotolerance2001.pdfFRANCESCO DE CATALDO, WRITING COMMITTEEContributing hemophilia centers: Bari, Schiavon M; Bergamo,Finazzi G; Bologna, De Rosa V; Bolzano, Billio A; CastefrancoVeneto, Tagariello G; Catania, Musso R; Catanzaro, Santoro R;Firenze, Linari S; Milano Niguarda, Mostarda G; Milano Policlinico,Santagostino E; Palermo, Caracciolo C; Parma, TagliaferriAR; Pavia, Gamba G; Perugia, Berrettini M; Pescara, Dragani A;Reggio Emilia, Ghirarduzzi A; Roma, Mazzucconi G; Torino,Schinco P; Verona, Franchini M; Vicenza, Castaman G.Objective. To evaluate the clinical problems relatedto the acquired hemophilia syndrome.Study design. Data collected from the ItalianAcquired Hemophilia Register.Results. Ninety-six cases were registered, 95 withanti-factor VIII and 1 with anti-factor IX. Forty-fivecases (46.8%) were idiopathic, 51 (53.2%) associatedwith different clinical conditions. The overallfollow-up is 7 years (range 1-20). In 31 patients(32.2%) bleeding occurred during or after an intervention.A prolonged activated partial thromboplastintime (aPTT) was present in all the patients inwhom the test was carried out but the inhibitor wasidentified on the occasion of a bleed. Twenty-threepatients did not require treatment for bleeding; 66patients received different therapies and in 58patients bleeding was controlled. An anamnesticresponse occurred in 8 cases unrelated to pregnancy.Sixty-five of the ninety patients evaluable forresponse to the initial anti-hemorrhagic therapy arealso evaluable for the effect of the immunosuppressivetherapy used to suppress the inhibitor. The finalresults of induction therapy are 52 (80% ) completeremissions, 6 (9.3%) partial remissions and 7(10.7%) treatment failures. The majority of thepatients who achieved complete remission receivedsteroids. A 2nd remission was obtained in 10/11patients who relapsed with combined therapy. Theoverall mortality was 16.6%.Correspondence: Francesco Baudo, MD, Thrombosis and HemostasisUnit, Niguarda Hospital, p.za Ospedale Maggiore 3,20162 Milan, Italy. Phone: international +39.02.64442970,Fax: international +39.02.66103898. E-mail:md9821@mclink.itConclusions. Correct evaluation of coagulationscreening tests, in particular the prothrombin time,is mandatory. At present combined immunosuppressivetherapy seems to be a reasonable choice forthe treatment of cases unrelated to pregnancy.©2003, Ferrata Storti FoundationKey words: acquired FVIII and FIX inhibitors; survey ofItalian Hemophilia Centers.Acquired hemophilia is a clinical syndromecharacterized by the sudden onset of bleeding,which may be either spontaneous oroccur after surgery or trauma and is usuallysevere (87% of the cases), in patients with a negativefamily or personal history of hemophilia.The depletion of factor VIII (FVIII), and muchless frequently of factor IX (FIX), is immunemediatedby specific autoantibodies. The incidencereported in the literature varies between 1to 4 cases per million/population per year. 1,2 Theincidence increases with age (median age 60-67years) with an equal sex distribution except inthe younger group because of the relation topregnancy. Acquired hemophilia is commonlyassociated with autoimmune (systemic lupuserythematosus, rheumatoid arthritis, asthma)and neoplastic diseases, hypersensitivity to drugsand pregnancy but in 50% of the cases the conditionis idiopathic. 3-5 Common sites of bleedingare the skin (large ecchymoses), the mucosae(epistaxis, gengivorrhagia, metrorrhagia), themuscles and the retro-peritoneum (hematomata);hemarthroses are very rare. The bleeding canarise spontaneously or after a procedure (positioningof an intravenous catheter, surgery,intramuscular injections). If the bleeding occursin critical sites compression problems mayensue. The severity of the bleeding is not proportionallyrelated to the inhibitor titer. 6 Thereported mortality related to bleeding is 8-22%. 2-5The majority of deaths occur within the firstfew weeks after presentation.haematologica vol. 88(supplement n. 12):september 2003

94F. Baudo et al.Prolongation of the activated partial thromboplastintime (aPTT) with a normal prothrombintime (PT) is the hallmark of the laboratory diagnosis.The objectives of therapy are to control thebleeding and suppress the autoantibodies. In 1999the Italian Hemophilia Centers carried out a surveyof the cases of acquired factor VIII inhibitorsregistered in the previous 15 years. The data collectedby this survey are presented in this report.Design and MethodsA questionnaire was mailed to 42 ItalianHemophilia Centers. The following informationconcerning each patient was requested: the primarycondition (if present); the initiating causeof bleeding (e.g. trauma, surgery, other events) orits spontaneous occurrence; the site of bleeding;its severity evaluated by hemoglobin level and bytransfusion requirement; the type (anti FVIII oranti FIX) and the titer of the inhibitor; theanamnestic response if present; the therapyadministered to control the bleeding; agentsused, modality of administration and results; theimmunosuppressive therapy; drugs and modalityof their administration, side effects and results;for the pregnancy-related cases, temporal correlationwith bleeding. Complete remission (CR)was defined as a normal level of FVIII or FIX andno detectable inhibitor; partial remission (PR)was defined as an inhibitor titer < 10 Bethesdaunit (BU) or a decrease of 50% if the baselineinhibitor titer was 10BU or a decrease less than 50% of the baselinevalue.Mild bleeding was defined as self-limited ecchymosesand cutaneous hematomata. Severe bleedingwas defined as retroperitoneal, intracranial, orocular; a fatal hemorrhages; requirement of ≥ 2red blood cell packs; or a decrease of ≥ 2 g ofhemoglobin.Statistical analysisStatistical analyses were performed using theStudent’s t test (Program MEDCALC, release1999). p values < 0.05 were considered statisticallysignificant.Table 1. Characteristics of the patients and clinical conditionsassociated with acquired inhibitor development; wholepopulation.Patients 96Inhibitor FVIII/IX 95/1Sex: male/female 38/58Age (median-range) 65 (2-88)male 69 (5-85)female 40 (2-88)Clinical conditions* n (%)Idiopathic 45 (46.8)Concomitant conditions 51 (53.2)autoimmune 16(16.6)malignant tumor 9 (9.4)post-partum 20 (21)HCV hepatitis 3 (3.1)HBV hepatitis 1 (1)pemphigo 1 (1)sarcoidosis 1 (1)*Children: idiopathic 2 years old, (FIX inhibitor), nephrotic syndrome 5 years old(FVIII inhibitor).Table 2. Site of bleeding. The majority of the patients bledfrom more than one site.Sites of bleeding First episode Relapse96 patients % 31 patients %Muscle 59 61.4 16 51.6Central nervous system 0 -- 2 6.5Gastro-intestinal tract 5 5.2 4 13.0Urogenital system 13 13.5 2 6.5Mucosa 13 13.5 3 9.7Skin 36 37.5 5 16.1Abdomen 8 8.3 2 6.5Joint* 7 7.3 5 16.1*Incidence higher than previously reported.ResultsTwenty centers contributed information on atotal of 96 patients. The requested informationwas not available for all the patients, thereforethe data presented do not always refer to allpatients. Anti FVIII antibodies were identified in95 patients and anti FIX antibody in one; 38 ofthe patients were male and 58 were female. Themedian age was 65 years (range 2-88), 69 formales (range 5-85) and 40 for females (range 2-88). No underlying disease was shown in 45 cases(46.8%). Twenty cases (20.8%) were related topregnancy and 9 (9.4%) to malignant tumors: (2cases of lung tumor, one each of pancreas, seminoma,astrocytoma, tonsil, urinary bladder, idiopathicmyelofibrosis, multiple myeloma [Table1]). Muscles, skin, mucosae and the urogenitaltract were the most frequent sites of bleeding(Table 2), and the majority of patients had bleedingfrom more than one site. The incidence ofhemarthroses was 7%, higher than that previouslyreported. The overall median follow-up is 7years (range 1-20); in the post-partum cases 8years (range 0.5-15), in the idiopathic and theautoimmune cases 4 (range 1-13) and 6 (range1-11) years, respectively. In a few patients a previousmild bleeding was overlooked. In 65patients (67.7%) the bleeding occurred spontaneously,in 31 (32.3%) cases, 15 of which idiopathicand 16 secondary, the index bleedoccurred during or after an intervention (Table3). In 4 cases a hysterectomy was carried outbecause of uncontrollable metrorrhagia (3 caseshaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 95Table 3. In 31/96 (15 idiopathic, 16 with concomitant disease)the inhibitor was identified because of bleeding aftera procedure.Clinical conditionNumber of patientsElective surgery 12Dental extraction 3Post-partum hysterectomy 4Trauma 8Intramuscular injection (primary disease not reported) 3Bone marrow biopsy 1Total number 31Elective surgery: hiatus hernia; pace maker insertion; cholecystectomy;herniotomy; nephrectomy; appendectomy; hysterectomy for fibroma (2 cases);1 case of psoas hematoma and 1 case of corpus luteum-related bleeding weremisdiagnosed before surgery; the diagnosis was unknown in two cases.Table 4. Characteristics of the patients in relation to treatment.Table 5. Type of therapy and control of bleeding at the firstepisode.Single modalityMultiple modalities*Number of treated patients 26 32Type of therapyhigh dose immunoglobulins 2−−−DDAVP 4 1human FVIII 6 9porcine FVIII 5 6APCC 4 6rFVIIa 5 8immuno-adsorption−−−2Mean inhibitor titer BU (+ SD) 30.7 (8.1) 85.4 (363)^Mean Hb g/dl (+ SD) 7.7 (2.1) 7.4 (1.7)^^p = n.s.; APCC = activated prothrombin complex concentrate;DDAVP = vasopressin; *Last treatment that controlled the bleeding;patients treated previously with a variety of agents.Clinical condition Treated Not treatedIdiopathic 35 5Autoimmune 7 8Post-partum 13 7HBV/HCV hepatitis 3 1Malignant tumors 6 3Miscellaneous 2 0total number 66 24 §not evaluable for response 5* N.A.evaluable 61 N.A.Bleedingspontaneous 35 19induced 26 5mild 15 11severe 34 4not defined 12 9Mean Hb g/dL (SD) 8.2 (2.4) 10.9 (3.1)°Mean inhibitor titer BU (SD) 61.7 (262.4) 25.1 (38.4)^Exitus for bleeding 3 1N.A. = not applicable; °p

96F. Baudo et al.Table 6. Anamnestic response in 8 patients treated withporcine (7) or human (1) FVIII. Median value and range ofinhibitor titer.Type Type LR (1 patient) Number HR (7 patients)of of Inhibitor titer (BU) of Inhibitor titer (BU)treatment inhibitor Before After patients Before AfterH FVIII anti H 2.7 8.3 1 15 240anti P 0 0−−0 41P FVIII anti H−− −−4 14 (9-16) 31 (22-38)anti P−− −− −−2 (0-4) 16 (5-62)H FVIII+ anti H−− −−2 2, 4 22, 26P FVIII anti P−− −− −−1, 27 21, 73H = human; P = porcine; Low responder (LR) = increase of the inhibitor titer < 10BU/mL after treatment; High responder = HR.(hemopericardium, retroperitoneal hematoma)(Table 7). Modalities of treatment are reported inthe Table 8.ImmunosuppressionSixty-five of the 90 patients evaluable for theresponse to the initial anti-hemorrhagic therapyare also evaluable for the immunosuppressivetherapy. Three patients died before starting treatment,1 because of bleeding, 2 from the underlyingdisease. Eight patients with a low inhibitortiter (

IV International Workshop on Immune Tolerance in Hemophilia 97Table 9. Immunosuppressive therapy.Therapy Induction: final results Relapsen CR PR F n CR PR FSteroid 28 27 2 2 4 2 1 1Steroid + immunoglobulins 4 3 0 1 1 1 0 0Steroid + cyclophosphamide 13 10 1 2 2 1 1 0Cyclophosphamide 6 4 1 1 1 1 0 0Steroid + azathioprine 8 6 1 1 2 1 1 0Azathioprine 3 2 1 0 1 1 0 0Steroid + melphalan 1 1 0 0 0 0 0 0Cyclosporine 2 2 0 0 0 0 0 0Total 65 52 6 7 11 7 3 1CR = complete remission ; PR = partial remission ; F = failure ; R = relapse.Table 11. Immunosuppressive therapy in patients with postpartuminhibitor.Therapy n 1 st CR PR F R 2 nd CRNone 2 2Steroid 10 7 3 3* 3Steroid + cyclophosphamide 1 1Cyclophosphamide 1 1 1° 1Steroid + azathioprine 2 2 1 § 1Steroid + high-dose immunoglobulins 4 3 1 # 1^ 1Total 20 14 3 6 6CR = complete remission; PR = partial remission; F = failure; R = relapse.*CR obtained in PR patients, 1 with azathioprine, 1 withsteroid + cyclophosphamide, 1 with no therapy; # steroid + Ig for 6 months;°azathioprine; § steroid + azathioprine; ^steroid + cyclophosphamide.Table 10. Characteristics of the women with post-partuminhibitor formation in relation to the treatment for bleedingcontrol (median, range and mean±SD).Patients requiring Therapy N. therapyNumber 11 9Median (range) time to bleeding(relation to delivery) (days) 8.5 (1-150) 90 (1-150)Type of bleeding Methrorragia Mildhematomata ecchymosesMean (SD) inhibitor titer (BU/mL) 7.5 (7.0) 13.5 (11.8)^Mean (SD) Hb (g/dL) 7.5 (2.1) 10.3 (2.1)^Number of patients requiring transfusions 8 2^p = ns.delayed in the patients not requiring treatment. In2 patients the inhibitor disappeared spontaneouslyafter 2 and 4 months. Eighteen patientsreceived immunosuppressive therapy as detailedin Table 11: prednisone alone in 10 patients, combinedwith other agents in 7. CR and PR wereobtained in 14 (78%) and 3 (16%) patients,respectively (overall response rate 94%). In onepatient the inhibitor was unchanged (2.5 BU)after a follow up of 48+ months. The median timeto obtain the first complete remission was 2.7months (range 0.5-36) and the median time ofthe treatment was 4.0 (range 1-18) months. In 1patient the inhibitor persisted at a low titer duringthe treatment and disappeared thereafter. Sixwomen relapsed after a median time of 16months (range 6-24 months) and were rescuedwith additional treatment. The majority of thepatients who achieved 1 st or 2 nd CR receivedsteroids alone or in combination. The time to thefirst CR was independent of the baseline inhibitortiter: median 3 months (range 0.5-18) in 8patients with inhibitor titer 10 BU. The inhibitor titerand the time to response were not different in thepatients who relapsed (median inhibitor titer 6.5BU - range 3.2-140 and median time to response2.7 months - range 2-36) and in those who didnot relapse (median inhibitor titer 8.0 - range 1.6-32 and median time to response 2.0 months –range 0.5-15). The survey did not yield informationon the neonates.DiscussionThe total number of cases in our survey is lowerthan expected, considering the 15 year-period,probably because of underreferral to the centers.The results of this retrospective survey could,therefore, be biased. Age, percentage of idiopathiccases (47%) and primary diseases are similarto those in previous reports. 2,4,6 In 31 patients(33%) (15 with idiopathic inhibitors and 16with concomitant disease), bleeding ensued afteran intervention or after a trauma. The prolongedaPTT was known prior to surgery and also in thepost partum patients but was overlooked. Thecorrect evaluation could have avoided the surgicalbleeding and, possibly, the hysterectomies.Twenty-three patients, including 9 post partumwomen, did not require therapy for the bleeding.The bleeding was mild and self-limited as indicatedby clinical signs and hemoglobin level(Tables 4 and 6). In our survey the inhibitor titerwas not correlated with either the intensity ofbleeding or the response to therapy but such correlationshave been reported by other investigators.7,8 An anamnestic response was recorded in8 cases unrelated to pregnancy but is at variancewith other reports on the post-partum period. 8,9The variety of agents used to control the bleedingprecludes any evaluation of the appropriate initialtreatment. A similar observation refers to theimmunosuppressive therapy. Prednisone was thepreferred agent both in induction and in relapse.A response (CR+PR) was obtained in the majorityof cases. The data are in agreement with thosealready published. 2,9 Eight patients died becausehaematologica vol. 88(supplement n. 12):september 2003

98F. Baudo et al.of bleeding, 4 from the initial episode, 2 atrelapse and 2 during the immunosuppressivetherapy. All deaths occurred in the subgroupunrelated to pregnancy. The mortality related tobleeding is similar to that described in otherreports. 1,2,4,5 The high mortality of the cases unrelatedto pregnancy and the high percentage ofthe patients rescued suggest that other agentsalone or combined with prednisone might be areasonable choice. 2,6,9-13Pregnancy is a frequent concomitant conditionin the FVIII inhibitor syndrome. In our survey21% of the cases were associated with a pregnancy;other studies reported a lower incidence (7-11%). 2-4,6 The majority of the post-partum casesin our series were idiopathic (19/20). Theinhibitor occurred in the first pregnancy in 16out of 20 patients and did not recur in the fourwomen who had a subsequent pregnancy. Thesame observations were reported by others 7,14-16but in the survey by Solymoss the inhibitorrecurred in 3 women during pregnancies. 8 Theinhibitor was identified because of the occurrenceof bleeding in 10 women in the peripartum periodand in 10 women more than 30 days afterdelivery. In the series reported by Solymoss thebleeding occurred pre-partum in 2 women, within3 days from delivery in 3 and within 3-12months in 9. 8 In the review by Michiels bleedingoccurred during pregnancy in 1 woman, after anabortion in 1, immediately after delivery in 4,within 4 months in 8, and after a period of 4months to over 12 months in 27 women. 15 Thetime of the occurrence of bleeding is quite variable.The inhibitor is in general identified onoccasion of overt bleeding; the time of the developmentof the inhibitor in the absence of bleedingsigns cannot be determined retrospectively.The onset of bleeding in the 9 women who didnot require treatment was later (range 60-150days) than in those who required therapy (Table10). In the survey by Solymoss 3 out of 14 patientsdid not require therapy to control their bleeding. 8Mortality in post-partum patients is low. No fatalitieswere reported in our series or in that by Solymoss;4 out of 40 women (10%) died in Michiels’survey. 15Inhibitors may cross the placenta and may persistfor up to three months in the fetus, usuallywithout causing bleeding complications. 17-19 However,Ries reported a case of an intracranial hemorrhagein a neonate. 20 These data suggest thatdelivery should be managed as in hemophilia. 21In our survey information on the neonates wasnot requested.Immunosuppressive treatment with steroidsalone or in combination with other agents wasalso the preferred treatment (17/18) during thesepregnancy-related cases. The response rate washigh (94%), as too was the relapse rate (42%),but all the patients were rescued. Immunosuppressivetherapy may shorten the time to responsewithout influencing the response rate. Hauser etal. reviewed the post partum-data in the literaturecomparing immunosuppressive therapy versus notreatment. Time to response was shorter in thetreated patients but the overall response rate wasnot different. 7 Similar results were reported byMichiels. 15 These studies must be considered withcaution for a number of reasons. They are retrospective,refer to a small number of patients withheterogeneous characteristics and have no predefinedcriteria for treatment. Nevertheless steroidsmay be considered the treatment of choice. 21 Asgeneral consideration the final outcome of thepost-partum inhibitor syndrome is favorable,independently of the type of treatment. Thereforethe recognition of the syndrome is of utmostimportance.Acquired hemophilia usually occurs in generalhospitals. A prolonged aPTT should not be overlooked.In the presence of an unexplained oftensevere hemorrhage with an abnormal coagulationscreening test it is important to seek immediatespecialist advice. Because of the rarity of thedisorder, the treatment modality and the potentialrisk of severe bleeding, these patients shouldbe managed in hemophilia centers or under thesupervision of such centers. 22,23References1. Lottenberg R, Kentro TB, Kitchens CS. Acquired hemophilia.A natural history study of 16 patients with factorVIII inhibitors receiving little or no therapy. ArchIntern Med 1987;147:1077-81.2. Green D, Lechner K. A survey of 215 non-hemophilicpatients with inhibitors to Factor VIII. Thromb Haemost1981;45:200-3.3. Kessler CM, Ludlam CA. The treatment of acquired factorVIII inhibitors: worldwide experience with porcinefactor VIII concentrate. International Acquired HemophiliaStudy Group. Semin Hematol 1993;30 Suppl 1:22-7.4. Ludlam CA, Morrison AE, Kessler C. Treatment ofacquired hemophilia. Semin Hematol 1994;31 Suppl 4:16-9.5. Morrison AE, Ludlam CA. Acquired haemophilia andits treatment. Br J Haematol 1995;89:231-6.6. Bossi P, Cabane J, Ninet J, Dhote R, Hanslik T, ChosidowO, et al. Acquired hemophilia due to factor VIIIinhibitors in 34 patients. Am J Med 1998;105:400-8.7. Hauser I, Schneider B, Lechner K. Post-partum factorVIII inhibitors. A review of the literature with specialreference to the value of steroid and immunosuppressivetreatment. Thromb Haemost 1995;73:1-5.8. Solymoss S. Post partum acquired factor VIII inhibitors:results of a survey. Am J Hematol 1998;59:1-4.9. Spero JA, Lewis JH, Hasiba U. Corticosteroid therapy foracquired FVIII:C inhibitors. Br J Haematol 1991; 48:636-42.10. Green D, Rademaker AW, Briet E. A prospective, randomizedtrial of prednisone and cyclophosphamide inthe treatment of patients with factor VIII autoantibodies.Thromb Haemost 1993;70:753-7.11. Herbst KD, Rapaport SI, Kenoyer DG, Stanton W, FeinsteinDI. Syndrome of an acquired inhibitor of factorVIII responsive to cyclophosphamide and prednisone.Ann Intern Med 1981;95:575-8.12. Lian EC, Larcada AF, Chiu AY. Combination immunosuppressivetherapy after factor VIII infusion foracquired factor VIII inhibitor. Ann Intern Med 1989;110:774-8.13. Shaffer LG, Phillips MD. Successful treatment ofacquired hemophilia with oral immunosuppressive therapy.Ann Intern Med 1997;127:206-9.14. Neidhardt B, Bartels O, Hahn B. Postpartum hemo-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 99philia A with factor VIII inhibitor. Dtsch Med Wochenschr1985;110:799-802.15. Michiels JJ, Hamulyak K, Nieuwenhuis HK, NovakovaI, van Vliet HH. Acquired haemophilia A in womenpostpartum: management of bleeding episodes and naturalhistory of the factor VIII inhibitor. Eur J Haematol1997;59:105-9.16. Coller BS, Hultin MB, Hoyer LW, Miller F, Dobbs JV,Dosik MH, et al. Normal pregnancy in a patient with aprior postpartum factor VIII inhibitor: with observationson pathogenesis and prognosis. Blood 1981;58:619-24.17. Frick PG. Hemophilia-like disease following pregnancywith transplacental transfer of an acquired circulatinganticoagulant. Blood 1953;8:598-608.18. Broxson EH, Hathaway WE. Transplacental transfer ofacquired factor VIII: C inhibitor. Thromb Haemost1987;57:126.19. Vicente V, Alberca I, Gonzalez R, Alegre A. Normal pregnancyin a patient with a postpartum factor VIIIinhibitor. Am J Hematol 1987;24:107-9.20. Ries M, Wolfel D, Maier-Brandt B. Severe intracranialhemorrhage in a newborn infant with transplacentaltransfer of an acquired factor VII:C inhibitor. J Pediatr1995;127:649-50.21. Kulkarni R, Lusher J. Perinatal management of newbornswith haemophilia. Br J Haematol 2001;112:264-74.22. Hay CR, Baglin TP, Collins PW, Hill FG, Keeling DM.The diagnosis and management of factor VIII and IXinhibitors: a guideline from the UK Haemophilia CentreDoctors' Organization (UKHCDO). Br J Haematol2000;111:78-90.23. Kadir RA, Koh MB, Lee CA, Pasi KJ. Acquired haemophilia,an unusual cause of severe postpartum haemorrhage.Br J Obstet Gynaecol 1997;104:854-6.AddendumProspective Italian register on Web site:www.emonet.org.haematologica vol. 88(supplement n. 12):september 2003

[Acquired Inhibitors in Non-Hemophiliacs]review paperImmunological responsivenessof maternal T cells to selfantigens during pregnancy:pregnancy as a model to studyperipheral T cell tolerance andthe role of co-stimulatorymolecules in toleranceinductionhaematologica 2003; 88(suppl. n. 12):100-105http://www.haematologica.org/free/immunotolerance2001.pdfMELANIE S. VACCHIO, RICHARD J. HODES #Experimental Immunology Branch, National Cancer Institute;#National Institute of Aging, National Institutes of Health,Bethesda, MD, USAHow do maternal T cells respond to encounters withfetal antigens? Many earlier studies have attemptedto characterize changes induced by pregnancy inthe maternal T-cell repertoire of both humans andmice. Interpretation of these results has been difficultsince there has been no way to decipher whetherthe alterations were the result of encountering thefetal antigens or were non-specific changes relatedto pregnancy itself. However, the availability of T-cellreceptor (TCR) transgenic mice allows direct visualizationof the fate of maternal T cells that are reactiveto the fetus and provides a means to probe themechanisms by which maternal T cells respond tothe fetus. We propose that the fetus more closelyresembles self than a true allograft (non-self) in itseffect on the T-cell repertoire, resulting in the inductionof tolerance rather than an immune response.To test this hypothesis, we utilized H-Y-specific TCRtransgenic mice and demonstrate that maternal Tcells specific for fetal antigens decrease in an antigen-specificmanner during pregnancy and remaindepleted postpartum. In addition, the survivingclonotypic T cells have decreased responsiveness tosubsequent antigenic stimulation. Our study demonstratesthat specific recognition of fetal allogeneicantigens by maternal T cells results in toleranceinduction of reactive T cells via mechanisms similarto those seen when mature peripheral T cells aretolerized to self antigens. Interestingly, while ligationof CD28 provides a critical co-stimulatory signalfor activation of T cells, we found that CD28 isalso necessary for tolerance induction of peripheralCD8 T cells. This finding provides a potential focusfor development of tolerance induction therapies.Correspondence: Melanie S. Vacchio, Experimental ImmunologyBranch, National Cancer Institute; National Institutes ofHealth, Bethesda, MD, USA.T-cell recognition of antigenT cells recognize foreign antigen in the form ofpeptide complexed to self major histocompatibility(MHC)-encoded molecules. The specificityof the T-cell receptor (TCR) is generated byrearrangement of discontinuous V, D, and J genesegments resulting in synthesis of TCR α and βchains with unique antigen specificities. Therequirement for TCR binding to self-MHC,together with the random manner in which TCRspecificity is generated, dictates the need forthymic selection to eliminate non-functional orpotentially harmful T cells from the repertoire.Expression of the TCR α and β chains by the thymocyteduring differentiation exposes the thymocyteto a selection process that screens forexpression of TCR with affinity for self MHC. 1-3Thymocytes with appropriate affinity for selfMHC undergo a process called positive selectionthat rescues them from death by neglect, the fateof those thymocytes that either lack functionalTCR expression or fail to express a receptor withsufficient affinity for self-MHC. A proportion ofthymocytes will have a TCR of high affinity for aself MHC/peptide complex, making thempotentially harmful. T cells expressing TCRs withsuch an affinity for self antigens undergo negativeselection within the thymus, a process thatcan be mediated through clonal deletion, 4,5 functionalsilencing or anergy, 6 or by clonal arrest ofthe developing thymocytes. 7 While potent signalingthrough the TCR can initiate toleranceinduction, the mechanism of tolerance canpotentially be influenced by the type of antigenpresentingcell (APC) or the differentiation stageat which antigen is encountered.Ideally, elimination of all autoreactive T cellswould occur in the thymus resulting in a pool ofmature T cells reactive against only foreign proteinantigens. As the thymic cortex is relativelyinaccessible to systemically circulating proteins, 8thymocytes may not encounter, or be tolerizedto, many self proteins not expressed directly inthe thymus. An additional complication is thefact that most T-cell differentiation occurs dur-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 101Figure 1. T-cell recognition of antigens and co-stimulatorypathways.ing the fetal/neonatal period when certain developmentally-expressedself antigens are not yetpresent. Since thymic tolerance is ultimatelydependent on the array of proteins present in thethymus during the brief period of time that anyone thymocyte may reside there, mature T cellsmay not encounter developmentally expressedproteins until significantly later, in the peripheralimmune system when the proteins areexpressed during puberty, pregnancy, lactationor aging. Situations such as these result in thedistinct possibility that a mature T cell can leavethe thymus and at some point, encounter aperipheral self antigen with which its TCR willbind with sufficient avidity to cause activationand the potential for an autoreactive response.For this reason, extensive research has focusedon understanding if and how peripheral self-toleranceoccurs, not only to understand how thesemechanisms may go awry and result in autoimmunity,but also to identify potential manipulationsthat could be used to induce long-term toleranceto tissue transplants. The mechanisms ofperipheral T-cell tolerance have proven to be varied,and include clonal deletion, immune deviation(alteration of cytokine production andtherefore function), generation of regulatorycells, and functional unresponsiveness (includingTCR or co-receptor downregulation, orinduction of anergy). 2 In turn, many factors havebeen proposed to account for the reason a particularmechanism of tolerance occurs, includingaffinity of the MHC/TCR interaction, type ofAPC, environment in which the antigen isencountered (e.g. presence of cytokines orsteroids), and the concentration and duration ofexposure to an antigen. The impact of these factorson determining the mechanism of self toleranceremains to be elucidated.While the antigen specificity of T cells is determinedby the TCR αβ heterodimer, control of aT cell’s ability to become activated is regulated bya requirement for additional signals such as thoseprovided by cell-surface co-stimulatory molecules.One of the most important of these in theregulatory control of CD4 T cells is the highlystudied CD28/CTLA-4/B7 co-stimulatory pathway.T cells express cell surface CD28 that interactswith its ligands B7.1 or B7.2, which areexpressed primarily on antigen-presenting cells(APCs) such as macrophages, B cells and dendriticcells (Figure 1). 9 In vitro, TCR activation(signal one) in the absence of CD28 (signal two)has been shown to induce anergy, a state of unresponsivenessto subsequent encounters withantigen, 10,11 whereas TCR signaling in conjunctionwith costimulatory CD28 augments T-cellactivation by upregulating mechanisms thatenhance survival (bcl-2) and proliferation (IL-2production). 12-14The fetus: “self” or “non-self”?Pregnancy presents a situation in which maternalT cells have the potential to encounter antigensof fetal origin for which they may be specificbut which they have not encountered duringthymic differentiation and therefore to whichthey would not be tolerized. Mice and humansshare a similar means of placentation known ashemochorial, in which the maternal blood circulatesthrough the placenta in direct contactwith fetally-derived trophoblasts that wouldexpress allogeneic antigens. The giant cells on theouter region of the placenta directly abut thematernal decidua, the specialized region of theuterus surrounding the fetus, whereas thelabyrinthine trophoblasts act as the physicalinterface between maternal and fetal circulations.Forty years ago, Medawar attempted toexplain the mother’s lack of rejection of the fetusby arguing that the fetus may be antigenicallyimmature. 15 While this is clearly not the case, onemight still view the placenta as maintaining thefetus as antigenically separate. Indeed, separationof the maternal and fetal circulation via the placentaprevents access of maternal lymphocytesto the wealth of allogeneic antigens expressed inthe fetus itself. In addition, unlike the fetus, thereis no expression of MHC class II and tightly regulatedexpression of class I in the mouse placenta.In the mouse, polymorphic class I molecules,H-2 K, D and L, are expressed on trophoblaststhat can be found in contact with the maternaldecidua but are noticeably absent on thelabyrinthine trophoblasts. 16 The presence of limitedamounts of MHC class I on the placenta andthe notable absence of class II minimizes the possibilityof antigenic priming of CD4 T cells to thefetus and the likelihood of maternal attack.How might maternal T cells encounter fetalantigens? Virgin T cells do not circulate butrather are found predominantly in the spleenand lymph nodes where initial encounters withantigen occur. However, trophoblasts have beenshown to escape into the maternal blood at a ratethat has been estimated to be as high as 10 5cells/day in humans, 17,18 although detection offetally-derived male cells into maternal lymphoidorgans during gestation in mice has proven moredifficult. 19 Furthermore, human trophoblast cellshaematologica vol. 88(supplement n. 12):september 2003

102M.S. Vacchio et al.in circulation have been shown to be weakly positivefor expression of classical class I moleculessuggesting that upregulation may occur outsidethe placental environment. 17 Thus, maternal Tcells could conceivably encounter fetal antigensvia exported MHC-expressing trophoblasts presentin the maternal tissues (Figure 2a). Thiswould implicate the trophoblasts as potentialAPCs and provide a means by which peripheral Tcells could encounter fetally-derived antigens.Alternatively, cellular proteins from the fetallyderivedtissues may be presented to maternal Tcells after uptake by maternal professional APCs(dendritic cells, macrophages and B cells) andmigration to lymphoid organs, as has been proposedin the model of cross-presentation(Figure2b). 20There has been extensive discussion of Medawar’sproposal that the fetus represents nature’sallograft. 15 Viewing the fetus in the context of theimmune system's ability to distinguish self fromnonself, the fetus would fall under the category ofnon-self since many of the gene productsexpressed by the fetus may not even be encodedby the maternal genome and therefore, have neverbefore been encountered. Yet, the immune systemdoes not mount an immune responseagainst the fetus as it would a transplanted tissue.While there may be disparate MHC expressed,the dynamics of the interaction between themother and fetus are far different from that of aclassical allograft with the recipient immune system.Mature T cells in a non-inflammatory setting,for the most part, become tolerized uponencounter with antigen, whether self or foreign.Activation of T cells in response to antigen hasbeen proposed to depend on the presence of asecond danger signal, such as those generatedduring tissue destruction, 21,22 resulting in theactivation of APCs that would lead to inductionof T-cell effector function. However, during pregnancythere is not an immediate threat to themother posed by the fetus. The encroachmentand establishment of the fetus into the maternaluterus is invasive but with minimal tissue damageas compared to that of transplantation of atrue allograft, such as a transplanted kidney. Furthermore,the placental separation of maternal/fetalcirculatory systems and relatively lowlevels of allogeneic MHC class I and absence ofclass II antigens on the placenta would minimizethe amount of immediate antigenic challenge.This is in stark contrast to tissue transplants inwhich the recipient’s immune system is confrontedwith a potent antigenic challenge incombination with tissue damage and potentialadventitious pathogens. In view of the inflammatorysituation of tissue transplantation, theencounter of maternal T-cells with fetal antigenswould be more innocuous and seem to be littlemore than an encounter with a developmentally-regulatedor tissue-specific self-antigen.Encounter with the fetal tissues should thereforebe little different than encounter with any otherFigure 2. How do fetal antigens encounter maternal T cells?Maternal T cells can potentially encounter fetal antigens inseveral ways: either presentation via MHC-expressing trophoblaststhat escape into the maternal circulation (a) or“cross-presentation” of fetal antigens by maternal APCs (b).organ present in the mother. In addition, unlikea transplanted kidney, the fetus employs multiplemechanisms designed to ensure acceptanceby the maternal immune system. Given the noninflammatorynature of pregnancy, we proposethat maternal T-cell responses upon encounterwith fetal antigens during pregnancy may be representativeof how mature T cells respond uponencounter with self-antigens in the periphery.Therefore, we propose that maternal T cellswould become tolerized to fetal antigens ratherthan immunized against them during a normalpregnancy.T-cell tolerance to fetal antigensGiven our proposal that the fetal antigensmore closely resemble tissue-specific self-antigensthan alloantigens, we postulated that maternal Tcells should become tolerized to fetal antigens.Many people have attempted to characterize anylymphocytic changes that may occur in thematernal immune system in the hope of understandinghow these changes reflect maternalresponsiveness or tolerance to the fetus. Unfortunately,these studies have often yielded resultsthat are conflicting or difficult to interpret, withno firm consensus concerning the antigen-specificresponse to the fetus, primarily due to thedifficulty in interpreting changes in heterogeneousT-cell populations. 23-31 The generation ofTCR transgenic mice has allowed us to visualizethe fate of maternal T cells during pregnancy. Inthese mice, the majority, if not all of the T cellsexpress a transgenic TCR of known antigenicspecificity and MHC restriction. Therefore, thefate of T cells specific for a given antigen afterhaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 103antigen exposure can be detected using anticlonotypicantibodies and flow cytometry. Withthis simple approach, the fate of fetal-specific Tcells can be clearly evaluated. We utilized pregnantmice in which the T cells expressed a transgenicTCR specific for the male antigen H-Y presentedby H-2 D b32 and in which the recombinaseactivating gene-2 had been eliminated byhomologous recombination (RAG-2-/-), renderingthese mice incapable of expressingendogenous (non-transgenic) TCR. In this model,all the T cells expressed the transgenic TCRspecific for H-Y that would be expressed by themale fetuses. Monitoring of maternal T cellsexpressing the clonotypic H-Y-specific TCR inthese mice revealed that the number of cellsexpressing the clonotypic TCR in spleendecreased. 33 By 18 days of gestation, there was a40% decrease in overall cell number that correspondedwith a loss in splenic, clonotypic T cellsand by 5 day post-partum, the number of clonotype-expressingT cells was only 50% that of nonpregnantcontrols. There was no detectabledownregulation in co-receptor, or for that matterin TCR expression, nor was there a rapidrecovery to pre-pregnancy cell numbers after parturition.33 As no decrease was observed in pregnancieswith all females in the litter, 34 these datademonstrate that the elimination of T cellsobserved in pregnant H-Y-specific TCR trangenicmice was due to encounter with male antigenand not the result of non-specific effects of pregnancy.It is also important to note that pregnanciesproceeded normally with normal numbersof offspring, and no skewing of sex that wouldsuggest a bias against male offspring. 33While 50% of H-Y specific maternal T cells disappearedfrom spleen, it was unclear whetherthose cells remaining were naïve and simply hadnot encountered antigen, or whether encounterwith antigen had induced tolerance by anothermechanism. Analysis of the proliferativeresponse of clonotype-expressing T cells fromnon-pregnant or gestational day 18 TCR transgenicRAG-2-/- mice revealed deficient responsesto male antigen-presenting cells (APCs) or tolow concentrations of the H-Y peptide in thosecells isolated from pregnant mice, compared toresponses observed in non-pregnant controls. 33However, potent stimulation via either crosslinkingof the TCR with anti-TCRβ antibody or highconcentrations of the H-Y peptide induced equivalentproliferative responses from T cells isolatedfrom both non-pregnant or pregnant mice.Decreased responsiveness was not only observedin proliferative responses, but also in cytotoxic Tlymphocyte (CTL) responses, although interestingly,cytokine production was unaltered in thepregnant females. 34 Furthermore, decreasedresponsiveness could not be attributed todecreased avidity for the antigen/MHC complexas downregulation of the TCR or CD8 was notobserved. In addition, unresponsiveness persistedbeyond the time of parturition to at least 5days post-partum. This antigen-specific nonresponsivenessis not attributable to generalizednon-responsiveness to antigen during pregnancybecause deficient T-cell proliferative responsesto antigen were not observed in T cells isolatedfrom pregnant H-Y-specific TCR transgenicmice that had female-only litters (no H-Y antigenpresent). 34A requirement for CD28 co-stimulation in theinduction of peripheral tolerance in CD8 T cellshas not previously been assessed. To address thisquestion, we bred female CD28-deficient H-YspecificTCR transgenic mice and compared themto their CD28 WT counterparts for the inductionof tolerance to the H-Y antigen expressed bymale fetuses during pregnancy. In contrast towhat has been observed in CD4 T cells, there wasno strong requirement for CD28 in the activationof these CD8 cells in vitro. However, surprisingly,while CD28 co-stimulation is dispensiblefor in vitro activation, it is required in vivofor both the induction of clonal deletion and forantigen-specific unresponsiveness to the fetalantigen H-Y during pregnancy. 34 In marked contrastto their CD28 WT counterparts, H-Y-specificT cells from pregnant CD28 knockout micedid not exhibit either decreased proliferation ordecreased CTL activity. However, T cells from theday 18 CD28 knockout pregnant mice had significantlylower levels of interferon-γ-producingcells than did T cells from the CD28 knockoutnon-pregnant females, indicating that these cellsare not antigen ignorant but have seen antigen invivo and responded, albeit in an altered manner.The decrease in interferon-γ-producing cells is anot a non-specific effect of pregnancy since Tcells from day 18 CD28 wild type pregnant micedid not exhibit decreased IFN-γ-producing cells.This type of split anergy, altered cytokine productionbut normal CTL function, in the absence ofco-stimulation has been observed by others. 35However, the significance of this altered ability toproduce IFN-γ is not yet understood.The dependence on CD28 co-stimulation forthe induction of both clonal deletion and anergyinduction is surprising. At least for CD4 Tcells, it has been proposed that CD28 co-stimulationis required for clonal deletion of T cells toself-antigen because activation is required foractivation induced cell death. 36 The absence ofclonal deletion in CD28 knockout mice would,therefore, be consistent with this model. However,it is also important to consider that whileCD4 T cells have a requisite dependence onCD28 for activation, it appears that the H-Y-specificCD8 T cells utilized in our study do not.Therefore, it is surprising that clonal deletion ofthese CD8 cells is CD28 dependent while activationis not.The role of co-stimulation in the induction ofanergy has previously been addressed in studiesof CD4 T cells. It has been proposed that CTLA-4/B7 interactions may be particularly importantin the induction of anergy. 37,38 It has also beenhaematologica vol. 88(supplement n. 12):september 2003

104M.S. Vacchio et al.reported that two kinds of anergy can be inducedin vitro, one that is induced by TCR stimulationin the absence of CD28 and is reversible by interleukin-2(IL-2) and the other that is induced byCTLA-4 signals and is not reversible by IL-2. 39The antigen-specific unresponsiveness that wehave observed in H-Y specific transgenic T cellsresembles the second of these anergic states, inthat it is not reversible by IL-2. 33 However in ourmodel of peripheral tolerance, anergy inductiondoes not occur in CD28 knockout mice, despitethe fact that CTLA-4/B7 interactions remainintact in these animals.The dependence on CD28 co-stimulation forthe induction of peripheral tolerance in this systemis in striking contrast to what has beenobserved in CD4 T cells. The observation that H-Y specific T cells did not require CD28 co-stimulationfor activation, as well as the fact thatCD28 is required for peripheral tolerance, haveimportant implications for the development oftolerance-inducing therapies. Because the majorityof work studying the role of CD28 has beenperformed with CD4 T cells, many current therapiesare being designed with the approach thatblockade of CD28 could not only prevent T-cellactivation but induce T-cell anergy. Our workwith CD8 T cells suggests that activation canoccur in the absence of CD28. Moreover, CD28signals can actually be required for toleranceinduction. While our data are representative of asingle antigen-specific T cell, these finding suggestthat activation and tolerance-inductionrequirements can differ between CD4 and CD8T cells; a point to be considered in the design oftolerance-inducing therapies.ConclusionsWhile others have focused on the fetus as anallograft, we have tested the hypothesis that thefetus is responded to as a tissue expressing developmentally-regulatedself antigens, and that T-cell responses to the fetus may be representativeof peripheral mechanisms of T-cell tolerance.Accumulating experimental evidence using antigen-specificsystems to address the fate of maternalT cells upon encounter with fetally-derivedantigens has demonstrated that these cells canundergo several fates. The identified responses ofT cells to the fetal antigens include antigen-specificnon-responsiveness and clonal elimination.These mechanisms are observed in response toself-antigens in other peripheral self-tolerancemodels. Therefore, under normal circumstances,the maternal immune system is not poised toreject the allogeneic fetus, but activates mechanismsthat lead to acceptance of the fetus as self.Using pregnancy as a model to study peripheralT-cell tolerance, we have now attempted to dissectthe role of other co-stimulatory molecules intolerance induction. We find that peripheral toleranceinduction in CD8 T cells in vivo requiresCD28 co-stimulation. A model such as ours canprove useful in the elucidation of peripheral tolerancemechanisms that can ultimately bemanipulated for therapeutic purposes.References1. Kruisbeek AM, Amsen D. Mechanisms underlying T-cell tolerance. Curr Op Immunol 1996;8:233-44.2. Mondino A, Khoruts A, Jenkins MK. The anatomy of T-cell activation and tolerance. Proc Natl Acad Sci USA1996;93:2245-52.3. Kroemer, G. Clonal deletion of self-reactive T cells. InMechanisms of immunological self-tolerance. G. Kroemer,editor. R.G. Landis Co., Austin, Texas, USA. 1994.p. 16-50.4. Kisielow P, Teh HS, Bluthmann H, von Boehmer H. Positiveselection of antigen-specific T cells in thymus byrestricting MHC molecules. Nature 1988;335:730-3.5. Kappler JW, Roehm H, Marrack P. T cell tolerance byclonal elimination in the thymus. Cell 1987;49:273-80.6. Roberts JL, Sharrow SO, Singer A. Clonal deletion andclonal anergy in the thymus induced by cellular elementswith different radiation sensitivities. J Exp Med1990;171:935-40.7. Takahama Y, Shore EW, Singer A. Negative selection ofprecursor thymocytes before their differentiation intoCD4 + CD8 + cells. Science 1992;258:653-6.8. Nieuwenhuis P, Stet RJ, Wagenaar JP, Wubbena AS,Kampinga J, Karrenbeld A. The transcapsular route: anew way for (self-) antigens to by-pass the blood thymusbarrier? Immunol Today 1988;9:372-5.9. Lenschow DJ, Walunas TL, Bluestone JA. CD28/B7 systemof T cell costimulation. Ann Rev Immunol 1996;14:233-8.10. Mueller DL, Jenkins MK, Schwartz RH. Clonal expansionversus functional clonal inactivation: a costimulatorysignalling pathway determines the outcome of Tcell antigen receptor occupancy. Ann Rev Immunol1989;7:445-80.11. Schwartz RH. A cell culture model of T lymphocyte clonalanergy. Science 1990;248:1349-56.12. Boise LH, Minn AJ, Noel PJ, June CH, Accavitti M, LindstenT, et al. CD28 costimulation can promote T cellsurvival by enhancing the expression of Bcl-XL. Immunity1995;3:87-98.13. Linsley PS, Ledbetter JA. The role of CD28 receptor duringT cell responses to antigen. Ann Rev Immunol 1993;11:191-212.14. Thompson CB, Lindsten T, Ledbetter JA, Kunkel SL,Young HA, Emerson SG, et al. CD28 activation pathwayregulates the production of multiple T cell-derived lymphokines/cytokines.Proc Natl Acad Sci USA 1989;86:1333-7.15. Medawar PB. Some immunological and endocrinologicalproblems raised by the evolution of viviparity in vertebrates.Symp Soc Exp Biol 1954;7:320-38.16. Redline RW, Lu CY. Localization of fetal major histocompatibilitycomplex antigens and maternal leukocytesin murine placenta. Implications for maternal-fetalimmunological relationship. Lab Invest 1989;61:27-36.17. Chaouat G, Menu E. Maternal T cell reactivity in pregnancy?Curr Top Microbiol Immunol 1997;222:103-26.18. Beer AE, Kwak JYH, Ruiz JE. The biological basis of passageof fetal cellular material into the maternal circulation.In Ann NY Acad Sci 1993;8:21-35.19. Bonney EA, Matzinger P. The maternal immune system'sinteraction with circulating fetal cells. J Immunol1997;158:40-7.20. Carbone FR, Kurts C, Bennett SRM, Miller JFAP, HeathWR. Cross-presentation: a general mechanism for CTLimmunity and tolerance. Immunol Today 1998;368:368-73.21. Matzinger P. Tolerance, danger and the extended family.Ann Rev Immunol 1994;12:991-1045.22. Janeway CA, Goodnow CC, Medzhitov R. Immunologicaltolerance: danger - pathogen on the premises! CurrBiol 1996;5:519-22.23. Carter J, Newport A, Keeler KD, Dresser DW. FACSanalysis of changes in T and B lymphocyte populationsin the blood, spleen and lymph nodes of pregnant mice.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 105Immunology 1983;48:791-7.24. Couderc B, Mohammad M, Bouc AM, Mayer G,Chateaureynaud P. Maternal alloimmunization in ratpregnancy: in vivo and in vitro studies of T-cell-dependentimmunity to mating and third party alloantigens.Dev Comp Immunol 1992;16:485-92.25. Gambel PI, Cleland AW, Ferguson FG. Alterations inthymus and spleen cell populations and immune reactivityduring syngeneic pregnancy and lactation. J ClinLab Immunol 1980;3:115-9.26. Gottesman SR, Stutman O. Cellular immunity duringpregnancy. I. Proliferative and cytotoxic reactivity ofparaaortic lymph nodes. Am J Reprod Immunol 1980;1:10-7.27. Gottesman SR, Stutman O. Cellular immunity duringpregnancy. II. Response to T and B cell mitogens. Am JReprod Immunol 1981;1:78-82.28. Hoger TA, Tokuyama M, Yonamine K, Hayashi K,Masuko-Hongo K, Kato T, et al. Time course analysis ofα + β + T cell clones during normal pregnancy. Eur JImmunol 1996;26:834-8.29. Matthiesen L, Berg G, Ernerudh J, Hakansson L. Lymphocytesubsets and mitogen stimulation of blood lymphocytesin normal pregnancy. Am J Reprod Immunol1996;35:70-9.30. Tallon DF, Corcoran DJD, O'Dwyer EM, Greally JF. Circulatinglymphocyte subpopulations in pregnancy: alongitudinal study. J Immunol 1984;132:1784-7.31. Watanabe M, Iwatani Y, Kaneda T, Hidaka Y, MitsudaN, Morimoto Y, et al. Changes in T, B, and NK lymphocytesubsets during and after normal pregnancy. AmJ Reprod Immunol 1997;37:368-77.32. Teh HS, Kisielow P, Scott B, Kishi H, Uematsu Y, BluthmannH, et al. Thymic major histocompatibility complexantigens and the a b T-cell receptor determine theCD4/CD8 phenotype of T cells. Nature 1988;335:229-33.33. Jiang SP, Vacchio MS. Multiple mechanisms of peripheralT cell tolerance to the fetal "allograft". J Immunol1998;160:3086-90.34. Vacchio MS, Hodes RJ. CD28 costimulation is requiredfor in vivo induction of peripheral tolerance in CD8 Tcells. J Exp Med 2003;197:19-26.35. Otten GR, Germain RN. Split anergy in a CD8 + T cell:receptor-dependent cytolysis in the absence of interleukin-2production. Science 1991;251:1228-31.36. van Parijs L, Perez VL, Abbas AK. Mechanisms of peripheralT cell tolerance. Novartis Found Symp 1998;215:5-14.37. Perez VL, Van Parijs L, Biuckians A, Zheng XX, Strom TB,Abbas AK. Induction of peripheral T cell tolerance invivo requires CTLA-4 engagement. Immunity 1997;6:411-7.38. Greenwald RJ, Boussiotis VA, Lorsbach RB, Abbas AK,Sharpe AH. CTLA-4 regulates induction of anergy invivo. Immunity 2001;14:145-55.39. Wells AD, Walsh MC, Bluestone JA, Turka LA. Signalingthrough CD28 and CTLA-4 controls two distinctforms of T cell anergy. J Clin Invest 2001;108:895-903.haematologica vol. 88(supplement n. 12):september 2003

[Acquired Inhibitors in Non-Hemophiliacs]review paperTen years experience withimmune tolerance inductiontherapy in acquired hemophiliahaematologica 2003; 88(suppl. n. 12):106-110http://www.haematologica.org/free/immunotolerance2001.pdfLASZLO NEMES,* ERVIN PITLIK°*National Hemophilia Center, National Medical Center°Semmelweis University, 2 nd Department of InternalMedicine, Budapest, HungaryObjective. The primary aim of long-term managementin acquired hemophilia is to eradicate the FVIIIautoantibody so that further bleeding can be prevented.ITI regimens with human FVIII concentratesuntil recently were rarely implemented in adultpatients with autoantibody-inhibitors even thoughthey have been used with increasing frequency foralloantibody suppression primarily in young childrenwith congenital hemophilia. In the ITI treatment ofacquired hemophilia the FVIII administration servesto enhance the stimulation of the autoantibody-producinglymphocyte clones and is a useful adjuvant toimmunosuppressive therapy. For successful ITI inautoantibody patients, small, repeated FVIII dosesseem to provide the adequate stimulation for thesubsequent successful immunosuppression andthere is no obvious need for the exhaustive high-doseFVIII administration, as in the original Bonn andMalmö protocols. We evaluated the results of 20consecutive non-hemophiliac patients with factor VIIIautoantibody treated in a single center with our ITIprotocol between 1992 and 2001, comparing themto 6 historical control patients treated with the traditionalimmunosuppressive therapy (steroid and/orcyclophosphamide) between 1988 and 1992 in thesame setting. The sex ratio, mean age at the diagnosis,the initial and peak inhibitor titers and residualFVIII:C values were similar in the two groups.Study design. Our ITI protocol consists of 3 weeks oftreatment with 1) human FVIII concentrates (30U/kg/day for the1st week, 20 U/kg/day for the 2ndweek, and 15 U/kg/day for the 3rd week, plus 2) iv.cyclophosphamide (200 mg/day to a total dose of2-3 grams), plus 3) methylprednisolone (100mg/day iv. for one week and then tapering down thedose gradually over the next two weeks). After thedisappearance of the inhibitor no further maintenanceimmunosuppression was given. The laborato-Correspondence: Dr. Laszlo Nemes, National Hemophilia CenterNational Medical Center Vagany u. 2. 1135, Budapest,Hungary. Phone: international +36.1.32034973504760/ext.1882, 1884. Fax: international +36.1.3297097. E-mail:lnemes@hiete.hury follow up consisted of aPTT and mixing tests beforeand after two hours of incubation, Bethesda inhibitorassay, porcine FVIII cross-reactivity, FVIII:C before andafter FVIII administration (recovery), three times aweek. The definition of success was the disappearanceof the inhibitor in the Bethesda assay systemand the persistent normalization of the FVIII:C value(i.e. >70% activity of the normal).Results. Eradication of the inhibitor occurred in18(19)/20 patients in the ITI group versus 4/6patients in the control group. (One patient achievedcomplete remission only after a second course ofITI.)Conclusions. The main difference between the twogroups was in the time needed for the complete disappearanceof the inhibitor (4,7 weeks for ITI vs. 28.3weeks for controls). No bleeding-related mortalityoccurred in the ITI group in contrast to that of 33%in the controls.©2003, Ferrata Storti FoundationKey words: acquired hemophilia, factor VIIIautoantibodies, immunosuppression,immune tolerance induction therapy.Acquired hemophilia is a rare, usually severebleeding disorder characterized by the formationof IgG1-IgG4 autoantibodiesagainst the factor VIII (FVIII) procoagulant activity.The FVIII autoantibodies occur either spontaneously(idiopathic cases) or in connection withpregnancy (postpartum cases), various autoimmune,dermatological and malignant diseasesand drug therapy. 1,2 The management of acquiredhemophilia serves for dual goals: the promotionof hemostasis by the treatment of the acute bleedingepisodes and the elimination of the FVIIIautoantibody by long-term eradication therapyfor consequent cure of the condition.Whereas the principles of the management ofacute bleedings are comparable in allo- andautoantibody patients, the methods used in theeradication therapy differ more fundamentally.In contrast to congenital hemophilia withalloantibody inhibitors, acquired inhibitors tohaematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 107FVIII represent a rather heterogeneous group ofpatients with different basic and accompanyingdiseases and prognostic factors. Therefore, thedetailed assessment of the patient’s individualcharacteristics (e.g. location and severity of thebleeding at presentation, underlying disorder, ageand clinical condition of the patient, accompanyingdiseases, expected response to immunosuppression,and initial and peak human and porcineinhibitor titers) is essential before the decision onthe particular therapeutic interventions is made. 3The primary aim of long-term management inacquired hemophilia is to eradicate the FVIIIautoantibody so that further bleeding can be prevented.This can be achieved through immunomodulationby immunosuppressive drugs, intravenousgammaglobulin (IVIG), and by immunetolerance induction (ITI) regimens. AlthoughFVIII autoantibodies may remit spontaneously, 2clinical studies indicate that early initiation oferadication treatment is advantageous. 1,4 Dramaticand life-threatening bleeding complicationsare experienced in 80-90% of patients at sometime in the course of their disease. The 10-22%mortality rate for this disorder is directly or indirectlyattributable to the inhibitor. 1,4 Most publishedtherapeutic guidelines and algorithms recommendthat eradication therapy should be institutedas soon as the diagnosis has been established.5,6 It is possible that different strategies forlong-term management may be suitable for thevarious subgroups of patients. 7 A conservativewatch and wait approach for children, peripartum,and drug-induced cases in whom spontaneousremission can be reasonably expected may bemore appropriate than the combined immunomodulatorytherapies for idiopathic, autoimmune-and malignancy-associated cases. 3 In anycase, individuals presenting with acquired hemophiliaand severe hemorrhage need rapid andeffective treatment.Immunosuppressive therapy with steroid or thecombination of steroid plus cyclophosphamid hasbeen the mainstream of traditional eradicationtreatment for FVIII autoantibody patients. 8,9 ITIregimens with human FVIII concentrates untilrecently were rarely implemented in adult patientswith autoantibody-inhibitors even though theyhave been used with increasing frequency foralloantibody suppression primarily in young childrenwith congenital hemophilia. 10 Although ITIcould be utilized in the management of both conditions,the actual mechanisms of the effect anddosage schedules should be fundamentally different.ITI for alloantibodies is a typical desensitizingtherapy in the immunological sense; so large dailydoses of FVIII are given for a prolonged periodaiming at exhausting the alloantibody-producingclones. The duration of ITI is generally betweensome months to one year, and the addition of vigorousimmunosuppressive therapy is of doubtfulimportance. On the other hand, in the ITI treatmentof acquired hemophilia the FVIII administrationserves to enhance the stimulation of theautoantibody-producing lymphocyte clones and isa useful adjuvant to immunosuppression. Theduration of therapy is limited to some weeks. Forsuccessful ITI in autoantibody patients, small,repeated FVIII doses seem to provide the adequatestimulation for the subsequent successful immunosuppressionand there is no obvious need forthe exhaustive high-dose FVIII administration, as inthe original Bonn 11 and Malmö 12 protocols.The theoretical basis for the development of ITIin acquired hemophilia was derived from the successesof plasma exchange therapy for progressiveautoimmune disorders unresponsive to conventionalimmunosupressive treatment. 13,14,15 It washypothesized that plasmapheresis induced proliferationof the pathogenic clones and subsequentpartial clonal depletion could be produced by givinglarge doses of cytotoxic drugs during theassumed period of increased B-cell vulnerability.The stimulation induced by plasma exchange wassynchronized with pulse immunosuppressive therapy.14,15 Similarly, in ITI of autoantibodyinhibitors, exogenous FVIII administration mayresult in additional stimulation with a correspondingincrease in the susceptibility of theimmunocytes to the effect of the cytotoxic drugs. 16In other words, the repeated administration of theantigen (FVIII) causes extra stimulation of theinhibitor-producing B-cell clones, making themmore susceptible to immunosuppression. In theoriginal case report of Green, 16 massive doses ofFVIII were given simultaneously with 1.5 g intravenouscyclophosphamide to an acquiredinhibitor patient previously unresponsive to combinedimmunosuppressive medication. In thisreport, the distinction between the acute treatmentof bleeding and eradication of inhibitor asthe aim of the administration of high-dose FVIIIwas not yet completely clear. In a later trial in1989, Lian et al. 17 used a combination of singlehigh-dose FVIII bolus followed by a modifiedcyclophosphamide, vincristine, and prednisonecytostatic protocol for the treatment of seriousacute bleeding in acquired hemophilia. 17 In 1996,two other successful applications of the same protocolwere published. 18,19 On the basis of these earlierexperiences, we have developed a new aggressiveprotocol for the management of patients withacquired FVIII inhibitor.Design and MethodsWe evaluated the results of 20 consecutive nonhemophiliacpatients with factor VIII autoantibodytreated in a single center with our ITI protocolbetween 1992 and 2001, comparing them to6 historical control patients treated with the traditionalimmunosuppressive therapy (steroidand/or cyclophosphamide) between 1988 and1992 in the same setting.In the two study groups similar treatmentmodalities were used for the management of acutebleeding episodes (traditional and activated prothrombincomplex concentrates, recombinantactivated FVII, porcine and human FVIII concen-haematologica vol. 88(supplement n. 12):september 2003

108L. Nemes et al.trates, plasmapheresis and transfusions).Our ITI protocol consists of 3 weeks of treatmentwithI. human FVIII:1 st week: 30 U/kg/day;2 nd week: 20 U/kg/day;3 rd week: 15 U/kg/day;II. cyclophosphamide:200 mg/day to a total dose of 2-3 grams;III. methylprednisolone:1 st week: 100 mg/day intravenously;2-3 rd week: tapering of the dose gradually.I and II were immediately discontinued ifFVIII:C has been normalized within the 3 weeks oftreatment. After the disappearance of the inhibitorno further maintenance immunosuppression wasgiven. Different high purity (Beriate-P, HaemoctinSDH, Koate-HP, Koate DVI) and ultra-high purity(Octonativ-M, Hemofil-M) FVIII concentrateswere used for the ITI. The laboratory follow upconsisted of aPTT and mixing tests before andafter two hours of incubation, Bethesda inhibitorassay, porcine FVIII cross-reactivity, FVIII:C beforeand after FVIII administration (recovery), threetimes a week.The definition of success was the disappearanceof the inhibitor in the Bethesda assay system andthe persistent normalization of the FVIII:C value(i.e. >70% activity of the normal). The sex ratio,mean age at the diagnosis, the initial and peakinhibitor titers and residual FVIII:C values weresimilar in the two groups. The summarized meanvalues for the entire cohort of the 26 inhibitorpatients are shown in Table 1. The characteristicsof the control and the ITI groups are demonstratedin Tables 2 and 3.Table 1. Summary of the patients’ mean values.Female/male 13/13Age at diagnosis 62 years (27-85 years)Residual FVIII:C 3.94% (1-16%)Initial human inhibitor titer 277,3 BU (2,2-3200 BU)Peak human inhibitor titer 288,5 BU (8-3200 BU)(maximal value after FVIII challenge)Initial porcine cross-reactivity 12,4 IU/mL (0-104 IU/mL)(measured in 11 patients)Peak porcine cross-reactivity 12,7 IU/mL (0-104 IU/mL)(measured in 11 patients)ResultsEradication of the inhibitor occurred in18(19)/20 patients in the ITI group versus 4/6patients in the control group. (One patientachieved complete remission only after a secondcourse of ITI.) The comparison of the results of thecontrol and ITI groups is shown in the Table 4.The main difference between the two groupswas in the time needed for the complete disappearanceof the inhibitor (4,7 weeks for ITI vs.28,3 weeks for controls). In the ITI group we haveobserved only two relapses during the relative longmean follow up period (26,2 months), in whichcases the same re-induction protocol was successfulagain. No bleeding-related mortalityoccurred in the ITI group in contrast to that of33% in the controls. Apart from the well-knownadverse effects of glucocorticoid therapy, we haveobserved only one patient with transient cytopenia,which resolved spontaneously without anyfurther consequence. We have not seen anyadverse event, which could be attributed to theuse of FVIII concentrates.DiscussionOn the basis of some earlier experiences withFVIII administration in autoantibody patients, anew aggressive protocol has been developed in1992 for the ITI treatment of acquired hemophiliapatients presenting with serious bleeds. 20 ThisBudapest protocol consists of three weeks of treatmentwith 1) human FVIII concentrates (30IU/kg/day for the 1 st week, 20 IU/kg/day for the2 nd and 15 IU/kg/day for the 3 rd week), plus 2)intravenous cyclophosphamide (200 mg/day to aTable 2. Control group.initials sex F/M age at Dx underlying condition initial bleeding res.FVIII:C(%) initial BIA rem.time* rem.duration° eradication1. J.H. F 36 PSS sc.,musc.hematomas 1 3200 − 3 CPH+steroid2. M.O. F 66 idiopathic sc.,musc.hematomas 2 610 8 36 steroid3. B.K.V F 27 postpartum uterine 2 5.5 3 1 steroid4. M.M. F 85 idiopathic hematuria 1 15 107 12 steroid5. S.B. M 69 idiopathic pharyngeal 5 20 2 3 steroid6. A.Cs. M 62 idiopathic femoral 1 437 − − CPH+steroidMean 57.5 2 714 28.25 13*remission time: time needed to achieve remission in weeks; °remission duration: duration of remission (follow-up) in months; BIA: Bethesda inhibitor assay; CPH: cyclophosphamide;PSS: progressive systemic sclerosis.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 109Table 3. ITI group.initials sex F/M age at Dx underlying condition initial bleeding res.FVIII:C(%) initial BIA rem.time* rem.duration°1. J.M. M 70 idiopathic sc.,musc.hematomas 3,5 21 3 932. I.T. F 53 idiopathic sc.,musc.hematomas 4 20 4 233. P.D. F 62 idiopathic sc.,musc.hematomas 3,5 8 3 534. F.R. M 74 gastric cc. sc.,musc.hematomas 1,4 19 2 205. Gy. Sz. M 55 renal cc.+IFNα tx retroperitoneal 5 320 10 246. P.F. M 79 gastric cc. retroperitoneal 4 22 10 367. L.L. M 65 idiopathic sc.,musc.hematomas 2 30 3 358. E.B. M 58 psoriasis pharyngeal 1 60 3 439. I.J. F 65 psoriasis retroperitoneal 1 1128 − − remission after 2 nd course10. J.B. F 49 idiopathic+MGUS retroperitoneal 7 64 12 1511. K.L. F 75 idiopathic femoral hematoma 14 28 3 2512. M.K. F 60 PSS brachial hematoma 16 10,3 2 2613. J.A. F 57 idiopathic CNS, intraabdominal 9 6 4 714. A.P. M 74 idiopathic femoral hematoma 2 58 5 1515. J.P. F 30 postpartum hematuria,fem.hemat. 1 33 2 1916. M.B M 50 idiopathic sc.musc.hematoma 1 1050 − −17. E.N. F 64 idiopathic thoracal 1 9 3 14 relapsus 1×18. J.T. M 80 idiopathic femoral hematoma, GI 6 2,2 9 3 relapsus 1×19. K.B. M 22 TTP, pheresis sc.musc.hematoma 6 4,7 2 1920. J. B. M 70 idiopathic, MGUS femoral hematoma 2 28 4 2Mean 60,6 4,52 146,06 4,67 26,22*remission time: time needed to achieve remission in weeks; °remission duration: duration of remission (follow-up) in months; BIA: Bethesda inhibitor assay; cc.: carcinoma;IFN: interferon; MGUS: monoclonal gammopathy of unknown significance; TTP: thrombotic thrombocytopenic purpura.total dose of 2-3 grams), plus 3) methylprednisolone(100 mg/day intravenously for the 1stweek and then tapering the dose gradually overthe next two weeks). After the completion of thethree weeks of ITI treatment, no further maintenanceimmunosuppression is given. In contrastto the earlier reports mentioned above, in thisregime FVIII is administered daily in lower doses,simultaneously with cyclophosphamide andsteroids, aiming for the rapid disappearance of theinhibitor. ITI resulted in eradication of the autoantibodyin 95% (19/20) of the cases. The maindifference between patients treated by ITI versustraditional immunosuppression was in the timeneeded for complete disappearance of theinhibitor (4.7 weeks versus 28.3 weeks). Nobleeding-related mortality occurred.We concluded that the ITI protocol describedabove is highly effective for the treatment ofacquired hemophilia, induces exceptionally rapidtherapeutic responses and advantageously influencesthe underlying autoimmune disorder. HoweverITI should be reserved for the eradication ofidiopathic, autoimmune- and malignancy-associatedFVIII autoantibodies in patients presentingwith severe bleeding or as a second line therapy ofother autoantibody inhibitors resistant to moreconservative approaches. It will be necessary toconfirm these initial promising results in furthermulti-center prospective studies.In the late 1990s three leading German hemophiliacenters (Bonn, Frankfurt and Heidelberg)also adopted the concept of ITI in the managementof acquired hemophilia. Brackmann et al. 11introduced a modified Malmö protocol consistingof long-term immunoadsorption by Ig-apheresis,plus high-dose (100-200 IU/kg/day) humanFVIII, plus high-dose IVIG, plus cyclophosphamide,plus steroids, and rFVIIa for acute hemorrhages.In the Heidelberg modification of thisregime FVIII is given as a 200 IU/kg bolus followedby high-dose continuous infusion aiming toachieve a FVIII:C level greater than 60% of normal.Recently, an oral ITI regimen also has beenreported in acquired hemophilia 21 . These preliminarydata indicate that ITI regimens may beapplied successfully in acquired hemophilia, butfurther studies are warranted to establish feasibilityand cost-benefit relationships. 22Table 4. Comparison of the control and ITI groups.ControlNumber of patients 4/6 18/20who achieved remission (19/20)Time needed 28,3 (2-107) 4,67 (2-12)to achieve remission (weeks)Duration of remission 13 (1-36) 26,22 (2-93)follow-up (months)Mortality rate 3/6 4/20Bleeding-related mortality 2/6 0/20ITIhaematologica vol. 88(supplement n. 12):september 2003

110L. Nemes et al.References1. Green D, Lechner K. A survey of 215 non-hemophilicpatients with inhibitors to factor VIII. Thromb Haemost1981; 45:200-3.2. Lottenberg R, Kentro TB, Kitchens CS. Acquired hemophilia:A natural history study of 16 patients with factorVII inhibitor receiving little or no therapy. Arch InternMed 1987; 147:1077-81.3. Kessler CM, Nemes L. Acquired inhibitors to factor VIII.In: Rodriguez-Merchan EC, Lee CA eds. Inhibitors inpatients with haemophilia. Blackwell, 2002.4. Morrison AE, Ludlam CA, Kessler CM. Use of porcinefactor VIII in the treatment of patients with acquiredhemophilia. Blood 1993; 81:1513-20.5. Hay CR, Colvin BT, Ludlam CA, Hill FGH, Preston FE.Recommendations for the treatment of factor VIIIinhibitors: from the UK Haemophilia Centre Directors’Organisation Inhibitor Working Party. Blood Coagul Fibrinolysis1996; 7:134-8.6. Rubinger M, Rivard GM, Teitel J, Walker IL. Suggestionsfor the management of factor VIII inhibitors. Haemophilia2000; Suppl 1 6:52S-9S.7. Morrison AE, Ludlam CA. Acquired haemophilia and itsmanagement. Br J Haematol 1995; 89:231-6.8. Green D, Rademaker AW, Briet E. A prospective, randomizedtrial of prednisone and cyclophosohamid in thetreatment of patients with factor VIII autoantibodies.Thromb Haemost 1993; 70:753-7.9. Shaffer LG, Phillips MD. Successful treatment of acquiredhemophilia with oral immunosuppressive therapy. AnnIntern Med 1997; 127:206-9.10. Cohen AJ, Kessler CM. Acquired inhibitors. In: Lee CA,editor. Haemophilia. Bailliéres Clin Haematol 1996; 9:331-54.11. Brackmann HH, Gormse J. Massive factor-VIII infusionin haemophiliac with factor-VIII inhibitor, high responder.Lancet 1977; 2:933.12. Nilsson IM, Berntorp E, Zettervall O. Induction ofimmune tolerance in patients with hemophilia and antibodiesto factor VIII by combined treatment with intravenousIgG, cyclophosphamide, and factor VIII. N Engl JMed 1988; 318:947-50.13. SchroederJO, Euler HH, Loffler H. Synchronization ofplasmapheresis and pulse cyclophosphamid in severe systemiclupus erythematosus. Ann Intern Med 1987; 107:344-6.14. Lupus Plasmapheresis Study Group. Plasmapheresis andsubsequent pulse cyclophosphamid versus pulse cyclophosphamidalone in severe lupus: design of the LPSG trial.J Clin Apheresis 1991; 6:40-7.15. Jarreusse B, Blinchet P, Gayrand M. Synchronization ofplasma exchanges and cyclophosphamide in severe systemicdiseases. Presse Med 1993; 22:293-8.16. Green D. Suppression of an antibody to factor VIII by acombination of factor VIII and cyclophosphamide. Blood1971; 37:318-7.17. Lian ELY, Larcada AF, Chiu AYZ. Combination immunosuppressivetherapy after factor VIII infusion for acquiredfactor VIII inhibitor. Ann Intern Med 1989; 110:774-8.18. Bucalossi A, Marotta G, De Regis F, Galieni P, DispensaE. A case report of acquired idiopathic hemophilia successfullytreated with immunosuppressive drugs and factorVIII concentrates. Clin Appl Thromb/Hemost 1996;2: 222-3.19. Durig J, de Wit M, Fiedler W, Marx G, Hossfeld DK. ImmunsuppressiveBehandlung einer spontanen HemmkörperhämophilieA mit Cyclophoshamid, Vincristin undPrednison nach vorangegangener Faktor-VIII-Stimulation.Schweiz Med Wochenschr 1996; 126: 20026-31.20. Nemes L, Pitlik E. New protocol for immune toleranceinduction in acquired hemophilia. Haematologica 2000;Suppl 10: 85:64S-8S.21. Lindren A, Wadenvik H, Tarkowski A, Tengborn L. Doesperoral administration of factor VIII induce oral tolerancein patients with acquired haemophilia A? ThrombHaemost 2000; 83:632-3.22. Kessler CM. Acquired factor VIII autoantibody inhibitors:current concepts and potential therapeutic strategies forthe future. Haematologica 2000; Suppl 10: 85:57-63.haematologica vol. 88(supplement n. 12):september 2003

[Gene Therapy]review paperThe host immune response andrisk of inhibitor developmentfollowing adenoviral genetherapy for hemophiliahaematologica 2003; 88(suppl. n. 12):111-114http://www.haematologica.org/free/immunotolerance2001.pdfDAVID LILLICRAPDepartment of Pathology, Richardson LaboratoryQueen’s University, Kingston, Ontario, CanadaBackground and current status ofhemophilia gene therapy trialsHemophilia remains a leading candidate diseasefor the successful application of somatic cellgene therapy. Indicative of this potential, thepast year has witnessed the completion of threephase I/II human clinical trials. Two of thesestudies have involved the delivery of a factor VIIItransgene by either the systemic administrationof a replication defective onco-retrovirus 1 or byex vivo delivery via electroporation to autologousfibroblasts. 2 The third study has involved deliveryof a factor IX transgene through intramuscularinjection of a recombinant serotype 2 Adeno-Associatedviral (AAV) vector. 3 None of thesestudies was associated with any adverse eventsand all three studies documented evidence ofintermittent, transient, low levels of expressionof the transgene product. No evidence ofinhibitor development has been seen in any ofthe twenty seven patients treated in these studies.Following the completion of these initialstudies, two new phase I/II trials have now beenstarted, the first involving hepatic delivery of afactor IX transgene by a serotype 2 AAV vectorand the second a gutless (minimal) adenovirusdelivering factor VIII. These two trials are still inthe very early stages of patient recruitment butboth trials have experienced temporary stoppagesdue to detection of vector in the semen of thefirst two patients in the AAV trial (no vectorfound in sperm) and transient thrombocytopeniain the first patient treated in the adenovirustrial.The objective of this report is to summarizeinformation pertaining to the nature of the hostimmune response associated with the use ofrecombinant adenoviral vectors for hemophiliaCorrespondence: David Lillicrap, Department of PathologyRichardson Laboratory, Queen’s University Kingston, OntarioCanada K7L 3N6. Phone: international +1.613.5481304.Fax: international +1.613.5481356E-mail: lillicrap@cliff.path.queensu.cagene therapy. This discussion also highlights thedifferences in the immune presentation of clottingproteins in the context of gene therapy comparedto routine protein infusion therapy (Figure1).Introduction to adenoviral gene therapyReplication incompetent adenoviruses havenow been used as gene therapy vectors for severalyears. They have a number of important advantagesfor this purpose, the most critical of whichare their ability to very efficiently transduce bothdividing and non-dividing host cells, their capacityfor packaging large transgene cassettes andfinally, the relative ease with which high titers ofthe vector can be produced. 4 All of these featurescontinue to make adenovirus an attractive candidatefor hemophilia gene therapy.However, a significant, ongoing concern associatedwith the delivery of adenoviral vectors isthe nature and magnitude of the host immuneresponse. 5-7 There is ample evidence to indicatethat this response involves early, innate reactivityfollowing within hours of vector delivery anda later, acquired, response involving both cellmediatedand humoral elements of the hostimmune system (Figure 2). The further discussionof these responses requires the separate considerationof so-called early generation vectors inwhich a limited number of early viral genes aredeleted from the vector (ie. E1+E2/E3/E4) andthe most recently generated helper-dependent or“gutless” vectors (also know as minimal or highcapacity vectors). 8,9The host innate immune response toadenoviral gene deliveryThe early, innate immune response to adenoviralgene delivery appears to be against componentsof the viral capsid. While adenovirus isan extremely efficient transducing agent formany different cell types, it is clear that viralentry is accompanied by significant cellular perturbation.Indeed, in preliminary studies of thepatterns of gene expression following adenoviralhaematologica vol. 88(supplement n. 12):september 2003

112D. LillicrapFigure 1. A comparison of the host immune response withcoagulation protein delivery by either exogenous infusionor by gene therapy.Figure 2. Elements of the early, innate immune response followingadenoviral gene delivery and the later adaptive immunityinvolving cell-mediated and humoral responses to severalcomponents of the gene therapy protocol.transduction, using DNA microarray technology,it has been shown that the expression of severalhundreds of genes are altered, some of whichcorrespond to early response genes that are activatedunder acute-phase circumstances (unpublisheddata, Stilwell JL and Samulski RJ). In animalsand patients receiving adenoviral vectors,this early immune response is most often manifestthrough, transient hepatotoxicity (elevatedALT values), transient thrombocytopenia andelevations of the acute phase cytokines IL-6 andTNFα. These innate responses have been extensivelydocumented with early generation adenoviralvectors, and there is at least some evidenceto indicate that a similar, albeit reduced, innateresponse accompanies delivery of helper-dependentvectors. Whether this response to helperdependentvectors relates to factors such as themagnitude of helper virus contamination of thevector preparation remains to be resolved.Another feature of this early response which hasstill to be resolved is the potential for significantinter-individual variability in the magnitude ofthe response and the influence of pre-existinganti-adenoviral antibodies. 10,11 It is possible thatthese variables might significantly complicate theprediction of vector doses that are both effectiveand safe.Acquired immunity following adenoviralgene therapyThe second component of the immuneresponse following adenoviral gene deliveryinvolves an acquired cell mediated and humoralresponse. There is now good evidence to indicatethat these responses will differ for the earlyand later generations of adenoviral vectors. Thus,while residual viral transcription and presentationof virally-derived peptides will occur withearly generation vectors, in which a number ofadenoviral genes are preserved, this will not complicatethe delivery of helper-dependent adenoviralvectors. This has the consequence of minimizingthe risk of a subsequent cell-mediatedresponse following the delivery of helper-dependentvectors. In theory, this should serve to protecttransduced cells from immune attack andshould therefore enhance the likelihood of a positiveoutcome from adenoviral transgene delivery.In contrast, humoral responses to the viral capsidproteins will likely occur with equal frequencyfor the early generation and latest forms ofAdenovector.Immune responses to the transgeneproduct in adenoviral-mediatedhemophilia gene therapyThe development of neutralizing anti-clottingfactor antibodies (inhibitors) following hemophiliagene therapy remains a potential treatmentcomplication that requires further investigation.Clearly, the generation of inhibitors in a substantialproportion of individuals treated withgene therapy would represent a profound concernfor the continuation of this therapeuticapproach.As stated above, no inhibitors have been detectedin any of the 27 patients enrolled in the firstthree, phase I/II clinical trials. However,inhibitors have been documented, in manyinstances, in pre-clinical animal protocols forhemophilia gene therapy. Among the factors thatcontribute to this propensity are the use of heterologoustransgenes, the employment of pro-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 113Figure 3. Components of the gene therapy protocol thatinfluence the likelihood of inhibitor development.Figure 4. Features of the transgene recipient and vectordelivery protocol that are likely to reduce the risk of inhibitordevelopment.moters that mediate ubiquitous transgeneexpression, high vector doses, the study of animalswith an inherited predisposition forinhibitor development and the use of deliverysystems that incite a significant host immuneresponse (Figure 3). As described above, the useof early generation adenovectors is clearly associatedwith the generation of both innate andadaptive host immune responses and these haveincluded the development of inhibitor antibodiesfollowing the delivery of a factor VIII transgene.However, this phenomenon is not consistentlyobserved, and the delivery of factor VIIItransgenes with early generation adenovectors tothe immunotolerant C57BL/6 strain of hemophilicmice has resulted in long-term factor VIIIexpression without the generation ofinhibitors. 12,13 It is too early to predict whetherthe tendency for inhibitor development will bereduced with helper-dependent adenoviral vectorsbut at least one report has already documentedtheir occurrence in some hemophilicmice treated with one of these vectors. 14Strategies to minimize inhibitordevelopment following adenoviralgene therapy for hemophiliaCurrent pre-clinical evidence suggests thateven with the use of homologous transgenes andthe latest generation of helper-dependent adenoviralvectors, inhibitory antibodies can stilldevelop. To minimize the likelihood of this complication,several approaches can be employed inthe selection of transgene recipients and the conductof the gene delivery protocol (Figure 4).Potential transgene recipients with a knownfamilial (strain) propensity for antibody developmentshould be avoided and, where known,recipients with hemophilic genotypes that areassociated with higher risks for inhibitor developmentshould also be excluded from initial trialsof hemophilia gene therapy. In terms of thegene delivery protocol, use of a transgene promoterto prevent transgene expression in antigenpresenting cells and the use of vector dosesthat optimize the balance between transgeneexpression and host immune activation will bothreduce the likelihood of inhibitor development.With specific reference to adenovector delivery,strategies to eliminate or at least minimize theearly innate immune response to the vector willalso likely reduce the subsequent development ofinhibitors.The future of adenoviral gene therapy forhemophiliaWith the development of the latest generationof helper-dependent adenovectors, the potentialfor using these vectors to deliver clotting factortransgenes has been revisited. While residualquestions remain concerning the innate immunitygenerated by these vectors and the durationof transgene expression, their relative ease of production,high transduction efficiency and opportunitiesfor readministration following serotypeswitching with different helper viruses suggeststhat further evaluation of this vector system iswarranted. Nevertheless, until the issues of theinnate immune response and the inter-individualvariability of immune responsiveness are betterunderstood, the evaluation of these vectors, inthe context of hemophilia gene therapy, shouldprobably be restricted to the pre-clinical setting. 15haematologica vol. 88(supplement n. 12):september 2003

114D. LillicrapFundingThe author's hemophilia gene therapy studies arefunded by the Canadian Institutes of HealthResearch (grant MT-10912), the Canadian HemophiliaSociety and the Bayer/Canadian Blood ServicesPartnership Fund. DL is a Career Investigatorof the Heart and Stroke Foundation of Ontario andholds a Canada Research Chair in Molecular Hemostasis.References1. Powell JS, Ragni MV, White GC, Lusher J, Hillman-Wiseman C, Cole V, et al. Results from one year followup of a phase I trial of FVIII gene transfer for severehemophilia A using a retroviral construct administeredby peripheral intravenous infusion. Blood 2001; Suppl:a2896[abstract].2. Roth DA, Tawa NE, Jr, O'Brien JM, Treco DA, Selden RF.Nonviral transfer of the gene encoding coagulation factorVIII in patients with severe hemophilia A. N Engl JMed 2001; 344:1735-42.3. Kay MA, Manno CS, Ragni MV, Larson PJ, Couto LB,McClelland A, et al. Evidence for gene transfer andexpression of factor IX in haemophilia B patients treatedwith an AAV vector. Nat Genet 2000; 24:257-61.4. Graham FL. Adenovirus vectors for high-efficiency genetransfer into mammalian cells. Immunol Today 2000;21:426-8.5. Assessment of adenoviral vector safety and toxicity:report of the National Institutes of Health RecombinantDNA Advisory Committee. Hum Gene Ther 2002; 13:3-13.6. Lozier JN, Csako G, Mondoro TH, Krizek DM, MetzgerME, Costello R, et al. Toxicity of a first-generation adenoviralvector in rhesus macaques. Hum Gene Ther2002; 13:113-24.7. Morral N, O'Neal WK, Rice K, Leland MM, Piedra PA,Aguilar-Cordova E, et al. Lethal toxicity, severe endothelialinjury, and a threshold effect with high doses of anadenoviral vector in baboons. Hum Gene Ther 2002;13:143-54.8. Parks RJ, Chen L, Anton M, Sankar U, Rudnicki MA,Graham FL. A helper-dependent adenovirus vector system:removal of helper virus by Cre-mediated excisionof the viral packaging signal. Proc Natl Acad Sci USA1996; 93:13565-70.9. Schiedner G, Morral N, Parks RJ, Wu Y, Koopmans SC,Langston C, et al. Genomic DNA transfer with a highcapacityadenovirus vector results in improved in vivogene expression and decreased toxicity. Nat Genet 1998;18:180-3.10. Harvey BG, Hackett NR, El Sawy T, Rosengart TK,Hirschowitz EA, Lieberman MD, et al. Variability ofhuman systemic humoral immune responses to adenovirusgene transfer vectors administered to differentorgans. J Virol 1999; 73:6729-42.11. Crystal RG, Harvey BG, Wisnivesky JP, O'DonoghueKA, Chu KW, Maroni J, et al. Analysis of risk factors forlocal delivery of low- and intermediate- dose adenovirusgene transfer vectors to individuals with a spectrum ofcomorbid conditions. Hum Gene Ther 2002; 13:65-100.12. Connelly S, Andrews JL, Gallo AM, Kayda DB, Qian J,Hoyer L, et al. Sustained phenotypic correction ofmurine hemophilia A by in vivo gene therapy. Blood1998; 91:3273-81.13. Gallo-Penn AM, Shirley PS, Andrews JL, Kayda DB,Pinkstaff AM, Kaloss M, et al. In vivo evaluation of anadenoviral vector encoding canine factor VIII: high-level,sustained expression in hemophiliac mice. HumGene Ther 1999; 10:1791-802.14. Balague C, Zhou J, Dai Y, Alemany R, Josephs SF,Andreason G et al. Sustained high-level expression offull-length human factor VIII and restoration of clottingactivity in hemophilic mice using a minimal adenovirusvector. Blood 2000; 95:820-8.15. Nasto B. Questions about Systemic Adenovirus Delivery.Mol Ther 2002; 5:652-3.haematologica vol. 88(supplement n. 12):september 2003

[Gene Therapy]review paperGene therapy for hemophilia A:immune consequences of viralvectormediated factor VIIIgene transferhaematologica 2003; 88(suppl. n. 12):115-121http://www.haematologica.org/free/immunotolerance2001.pdfTHIERRY VANDENDRIESSCHE, DESIRE COLLEN,MARINEE K.L. CHUAHCenter for Transgene Technology and Gene Therapy,Flanders Interuniversity Institute for Biotechnology-Universityof Leuven, BelgiumObjective. Hemophilia A is potentially amenable fortreatment by gene therapy. We have been exploringthe use of different viral vectors including onco-retroviral,lentiviral and high-capacity adenoviral (HC-Ad),each with their own advantages and limitations, forhemophilia A gene therapy.Study design. Injection of onco-retroviral vectorsencoding the B-domain deleted human FVIII cDNAinto neonatal hemophilia A mice resulted in longtermexpression of therapeutic and even supra-physiologicFVIII levels that stably corrected the bleedingdiathesis in 50% of the recipient mice. The lack ofneutralizing antibodies specific for human FVIII inthese animals, may have been due to the inductionof neonatal tolerance. However, in the remainingrecipient mice, a humoral and possibly also a cellularimmune response developed which thwarted phenotypiccorrection. Since onco-retroviral transductionis restricted to rapidly dividing target cells, efficienthepatic gene transfer could only be achievedin neonates but not in adult mice. To overcome thislimitation, lentiviral vectors were employed instead.Results. Non-dividing hepatocytes in adult recipientmice can be efficiently transduced (5-10%) usingimproved lentiviral vector designs, leading to longtermtransgene expression with only limited and transienthepatotoxicity. This underscores the potentialof lentiviral vectors for hemophilia gene therapy.However, efficient gene transfer was also apparentin antigen-presenting cells (APCs), particularly inKupffer cells, splenic macrophages and B-lymphocytes.Since it has previously been shown that inadvertenttransgene expression in APC triggers thedevelopment of neutralizing antibodies, the use ofCorrespondence: Thierry VandenDriessche, MD, Center forTransgene Technology and Gene Therapy, Flanders InteruniversityInstitute for Biotechnology, University of Leuven, 49 HerestraatB-3000 Leuven, Belgium. Phone: international+32.16.346144. Fax: international +32.16.34599. E-mail:thierry.vandendriessche@med.kuleuven.ac.be ormarineekhim.chuah@med.kuleuven.ac.behepatocyte-specific promoters may be warranted tocircumvent this potential risk. Finally, we have shownthat high-capacity adenoviral vectors can be used toachieve unprecedented high levels of human FVIIIexpression in hemophilic mice but that humoral andpossibly also cellular immune responses precludedlong-term gene expression.Conclusions. In conclusion, the use of different replication-deficientviral vectors that do not encode viralantigens can be used to express therapeutic levelsof human FVIII in hemophilic mice with varying efficiencies.However, the development of neutralizingantibodies following gene therapy in conjunctionwith the induction of cellular immune responsesagainst the transduced target cells, remains a concernthat will need to be addressed further in largeanimal models, such as hemophilic dogs, withspecies-specific transgenes.©2003, Ferrata Storti FoundationKey words: hemophilia; factor VIII; coagulation; genetherapy; inhibitory antibodies.Though factor VIII (FVIII) substitution therapyhas greatly improved the lives of patientssuffering from hemophilia A, there are stilllimitations to the current treatment that havetriggered interest in alternative treatments bygene therapy. Significant progress has recentlybeen made in the development of gene therapyfor the treatment of hemophilia A that serves asan ideal trailblazer for the treatment of other diseasesby gene therapy. These advances parallelthe technical improvements of existing viral vectorsystems and the development of new deliverymethods. 1-3 Both viral vectors as well as non-viralvectors have been considered for the developmentof hemophilia gene therapy. In general,viral vector-mediated gene transfer is far moreefficient than non-viral gene transfer and hastherefore been the method of choice.What is the ideal gene therapy vector forhemophilia A? The ideal gene therapy vectorwould need to have the following properties:haematologica vol. 88(supplement n. 12):september 2003

116T. VandenDriessche et al.• the vector should have the potential for longterm,preferably life-long FVIII gene expression;• the vector should be capable of efficient andpreferably selective gene transfer into theappropriate target cells, particularly hepatocytes,which normally express FVIII;• lack of de novo expression of viral genes in thetransduced target cells is required to avoidimmune rejection of transduced target cells;• the vector should not trigger any toxic oradverse side-effects;• inadvertent germline gene transfer should notoccur following somatic gene therapy;• vector preparations should be devoid of helpervirus or replication-competent virus;• the vector should not trigger inflammatoryimmune responses;• transduction in antigen-presenting cells(APCs) should be avoided to decrease the likelihoodof anti-FVIII antibodies;• vector administration should preferably benon-invasive;• the vector should not integrate into the targetcell genome to avoid insertional mutagenesisof cellular genes, particularly tumor suppressorgenes.There are currently no viral vectors that meetall these requirements. Nevertheless, there are anumber of promising vectors that can be usedfor hemophilia A gene therapy, each with theirown advantages and limitations, in particular:onco-retroviral and lentiviral vectors, highcapacityadenoviral vectors (HC-Ad) and adenoassociatedviral vectors (AAV). These viral vectorsare devoid of viral genes and consequentlycannot replicate in the transduced cells andpropagate in vivo, in contrast to the viruses theyhave been derived from. Furthermore, these replication-deficientvectors cannot express any viralantigens de novo in the transduced cells, herebysafeguarding the transduced cells from possibleimmune rejection by cytotoxic T-lymphocytes(CTL). Nevertheless, FVIII gene transfer mayresult in the presentation of endogenously synthesizedFVIII–derived peptides in the context ofMHC class I molecules and trigger CTL responsesthat could eliminate the FVIII-engineered targetcells. In addition, the viral vector particlesthemselves may induce a humoral immuneresponse that would lead to the formation ofneutralizing antibodies directed at the vector capsidor envelope structures. These antibodiescould potentially interfere with viral transduction,if the patient were to receive a subsequentviral vector challenge.Our work in particular, focuses on the use ofonco-retroviral, lentiviral and HC-Ad vectors forhemophilia A gene therapy which will be summarizedbelow.MoMLV-based retroviral vectorsMoloney murine leukemia virus (MoMLV)-based retroviral vectors can potentially give rise tostable gene expression by virtue of their stablechromosomal integration and lack of viral geneexpression. 4,5 Their integration allows passage ofthe transgene to all progeny cells. However, celldivision is required for MoMLV transduction andintegration, thereby limiting retroviral-mediatedgene therapy to actively dividing target cells. Genetherapy for hemophilia with MoMLV-based vectorsrequires either direct in vivo transduction ofcells that are naturally proliferating or induced toproliferate using growth factors, or ex vivo expansionand transduction of target cells followed bytheir readministration. Stable packaging cell linesare available that facilitate the large scale productionand characterization of recombinantvirus for potential use in clinical trials. No acutetoxicity or adverse effects have been reported inmore than 100 clinical trials with retroviral vectors.6 Furthermore, retroviral vector preparationsdevoid of replication-competent retroviruses(RCR) have not been shown to cause malignanttransformation in animals or in patients. 7,8 Rareclonal transformation events have only beenobserved in severely immunocompromized primatesreceiving a high dose of contaminatingreplication competent retroviruses (RCR). 9 Retroviralvector-mediated gene transfer is thereforerelatively safe for gene therapy of hemophilia.Retroviral vectors containing the FVIII geneshave been developed. However, first-generationFVIII-retroviral vectors were characterized by lowtiters and low FVIII expression levels which initiallyhampered their usefulness for gene therapy.The presence of an intact FVIII cDNA intoMoMLV retroviral vectors resulted in a 100 to1000-fold decrease in vector titer in comparisonwith the parental vector or vectors carrying othersimilarly sized cDNAs. 10-16 The inhibition ofFVIII expression and the low viral titer were dueto the presence of sequences within the FVIIIcDNA that inhibited RNA accumulation byinterfering with transcriptional initiation orelongation (see above). 12,14 Conservative mutagenesisof the entire 1.2 kb INS element failed toincrease FVIII expression or vector titer. 13 However,we and others showed that this impedimentcould be circumvented by including an intronupstream of the FVIII cDNA which led to a significantincrease in FVIII expression and restoredretroviral titer to normal levels (105 transducingunits/mL) (TU/mL). 13,17 Although otherpost-transcriptional mechanisms may also beinvolved 18, this design was based on the observationthat splicing can improve mRNA accumulationby increasing transcriptional efficiency,stabilizing RNA and/or increasing transport fromthe nucleus to the cytoplasm. 19,20 These intronbasedFVIII onco-retroviral vectors could be concentratedto even higher titers (109-1010TU/mL) following pseudotyping with the vesicularstomatitis virus G-protein (VSV-G). 21,22 Thisconstitutes a 107-fold improvement comparedto non-concentrated early generation FVIII retro-haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 117viral vectors. 10,11,13In vivo gene therapy for hemophilia A withMoMLV-based vectors requires direct transductionof cells that are either naturally proliferatingin vivo or that are induced to proliferate. High,stable levels of functional human FVIII could beachieved in newborn, FVIII-deficient mice injectedintravenously with high-titer VSV-G pseudotypedonco-retroviral vectors containing the B-domain deleted FVIII cDNA. 21 High-levels (>200mU FVIII/mL) of functional human FVIII productioncould be detected in about 50% of theanimals, some of which expressed physiologic orhigher levels (up to 12500 mU/mL). Most highexpresserssurvived an otherwise lethal tail-clipping,unequivocally demonstrating phenotypiccorrection of the bleeding disorder. 21 To ourknowledge, this is the first demonstration thathemophilia A can be cured by gene therapy in aclinically relevant animal model. Efficient genetransfer occurred into liver, spleen and lungs butnot in other organs including testes, with predominantFVIII mRNA expression in the liver. 21Long-term correction of hemophilia has beenachieved in about 50% of the FVIII-deficient animals,whereas the other half was not corrected. 21This differential response is due to a specificimmune mechanism since the latter mice developedinhibitory antibodies to human FVIII (rangingbetween 7 and 350 Bethesda units/mL),whereas none of the corrected mice did. 21 Furthermore,in the transient FVIII expressing mice,induction of antibodies was causally related to thedecrease in FVIII expression. 21 Finally, in theabsence of a specific immune response in FVIII-KO/SCID mice, human FVIII expression could bedetected in all recipient animals. 21 The differentialresponse is reminiscent of the lack of therapeuticefficacy in 10-20% of the hemophilia A patientsthat develop inhibitory antibodies to FVIII followingprotein replacement therapy. The higher proportionof non-expressers in FVIII-deficient micecould be due to the inherently higher immunogenicityof xenogenic human FVIII in non-humanspecies. Since we observed a lower level of FVIIIgene transfer in the transient and non-expressorrecipient mice compared to the long-term expressors,a mechanism whereby transduced cells wereeliminated by a cellular immune response cannotbe ruled out. It is therefore likely that the FVIIItransducedhepatocytes are recognized and eliminatedby FVIII-specific CTLs.The cause of the heterogeneous antibodyresponse and FVIII levels among the differentrecipients is not clear. Since the FVIII-deficientmouse is not an inbred strain, genetic differencesmay have contributed to this variability (Figure1). Alternatively, high levels of FVIII may havebeen required to induce immune tolerance toFVIII in the FVIII-deficient mice and preventinduction of inhibitory antibodies (Figure 2).This would be reminiscent of tolerance-inductionby repeatedly injecting FVIII proteins at highconcentrations in hemophilia A patients. 23 However,expression of a xenoprotein in neonatalmice is quite different from giving a species-specificprotein to a subject with a mature immunesystem. In particular, neonatal mice can becometolerant to soluble xenoantigens because neonatalsplenic B-cells are tolerance susceptible forseveral days after birth during which B-cell clonesmay be expanding. The proportion of tolerancesusceptiblesplenic B-cells decreases to less than10% after the first week after birth. It is indeedpossible that exposure of neonatal FVIII-deficientanimals to sufficiently high levels of human FVIIIprotein following FVIII gene transfer resulted inthe induction of tolerance to human FVIII, atleast in some of the recipients (Figure 2). Thishypothesis would be consistent with the observedcorrelation between high FVIII gene transfer andexpression and lack of inhibitory antibodies inthese mice and with the induction of tolerance ofFVIII in neonatal mice receiving high doses ofhuman FVIII. 24 Conversely, when gene transferefficiency was low, FVIII expression may havebeen below a critical threshold during the neonatalphase to induce tolerance, resulting in theinduction of inhibitory antibodies and/or CTLs(Figure 2). The potential immunogenicity ofhuman FVIII in some of the recipient FVIII-deficientmice is consistent with previous reportsindicating that repeated administration ofhuman FVIII protein in adult FVIII-deficient miceled to the induction of inhibitory antibodies. 25The efficient hepatic gene transfer in neonatalmice could be due primarily to the higher hepatocyteturn-over rate in newborn versus adultanimals. In adult mice most hepatocytes arerefractory to gene transfer using onco-retroviralvectors since disassembly of the nuclear membraneduring cell division is required to allow theonco-retroviral preintegration complex to enterthe nucleus. 26 Hence, gene delivery strategiesthat could overcome the requirement for hepaticcell division and be used in adults, would bedesirable.Lentiviral vectorsIn order to overcome the need for active celldivision, retroviral vectors derived from lentiviruseshave been developed, with HIV as the prototype.27 The lentiviral life cycle differs from thatof onco-retroviruses in that the preintegrationcomplex is readily transported into the nucleus,thus enabling efficient transduction of nondividingcells, including neurons. Gene transferwith lentiviral vectors in dividing cells is moreefficient than with MoMLV-based retroviral vectors.28 While present efforts are aimed at the generationof stable packaging cell lines 29 and safe 30and clinically acceptable vectors, the attributesof the lentiviral vector system may be well suitedfor the treatment of hemophilia A. The latestgeneration lentiviral vector packaging systemrelies on a self-inactivating HIV-derived vectorbackbone which does not contain any HIV-1genes. Instead, only the cis-acting elements nec-haematologica vol. 88(supplement n. 12):september 2003

118T. VandenDriessche et al.Figure 1. Genetic heterogeneity model. Hemophilia A mice are not genetically identical. As a consequence, some of the micemay be prone to develop inhibitory antibodies to FVIII and/or CTLs to the FVIII-transduced cells following in vivo gene therapywith FVIII-vectors. This would result in the neutralization of circulating FVIII proteins and/or the immune rejection of the FVIIIexpressingcells in an MHC class I-restricted fashion. Conversely, some of the mice may have a distinct genotype allowingimmune tolerance to be established to FVIII proteins and FVIII-transduced cells. This genetic heterogeneity hypothesis mayexplain the heterogeneous immune response in the neonatal hemophilic mice that received FVIII-retroviral vectors 21 but mayapply to other vector systems as well in adult recipient mice.essary for integration, reverse transcription andexpression are incorporated into the vector backbone,along with the gene of interest, in casu, theFVIII gene.Although many different types of non-dividingcells in different tissues can be transduced withlentiviral vectors, not all cells are permissive andin particular, there may be unknown limitingfactors that preclude efficient gene transfer intonon-dividing hepatocytes. 28,31,32 In anticipationof using lentiviral vectors for FVIII gene deliveryinto the liver, it was therefore important to firstdemonstrate that lentiviral vectors can transducenon-dividing hepatocytes. An improved lentiviralvector expressing the green fluorescent reporterprotein (GFP) was constructed. The vector containedthe central DNA flap which facilitatesintra-nuclear transport of the lentiviral pre-integrationcomplex and enhances transduction efficiency.33Lentiviral transduction was analysed afterintravenous injection of 109 transducing units oflentiviral-GFP vectors into adult mice (Vanden-Driessche et al., in press). Confocal microscopyrevealed intense GFP fluorescence in hepatocytes(5-10% GFP + ) and non-parenchymal cells in theliver of immuno-deficient Scid recipient mice,mostly around the vasculature. In addition,splenocytes were efficiently transduced (20-30%GFP + ) whereas no transduced GFP + cells couldbe detected in other organs. Transduction resultedin stable genomic integration of the lentiviralvector leading to long-term transgene expressionin liver and spleen (> 3 months). To assess thedependence of transduction on cell division,BrdU was continuously administered. Analysis ofGFP and BrdU colocalisation showed that >95%of transduced hepatocytes and splenocytes werenon-dividing. Liver toxicity was assessed by measuringserum-transaminases which were moderatelyelevated 24 hr post-injection but returnedto normal levels the next day. This confirms thatlentiviral vectors can be used to transduce nondividinghepatocytes in adult mice, obviating theneed to artificially induce hepatocyte proliferation.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 119Figure 2. Tolerance model. Exposure of neonatal hemophilic animals to sufficiently high levels of human FVIII protein followingFVIII gene transfer may have resulted in the induction of tolerance to human FVIII protein and the FVIII-transduced cells in someof the recipient mice. Conversely, when gene transfer efficiency was low, FVIII expression may have been below a critical thresholdduring the neonatal phase to induce tolerance, resulting in the induction of inhibitory antibodies and/or CTLs directed againstthe FVIII protein and the FVIII-transduced cells, respectively. This tolerance model may account for the heterogeneous immuneresponse in the neonatal hemophilic mice that received FVIII-retroviral vectors. 21 Similarly, prolonged expression of FVIII at highenough levels above a critical threshold may also play a role in establishing post-natal tolerance to the FVIII protein and to theFVIII-transduced cells. 37To further characterize the transduced splenocytesand non-parenchymal liver cells, FACSanalysis was performed on isolated single-cellsuspensions. Most transduced GFP + splenocytesformed typical phagocytic pseudopodia andexpressed high levels of MHC-II, co-stimulatorymolecules CD80 (B7-1) and CD86 (B7-2), celladhesion markers CD106 (VCAM-1), CD31(PECAM-1) and CD54 (ICAM-1) and othermarkers such as CD11b, CD11c and F4/80 characteristicfor antigen-presenting cells (APC).Similarly, these APC-specific markers were alsohighly expressed on some of the transduced GFP +liver cells, possibly Kupffer cells. Splenic andhepatic APCs were transduced in Scid and Balb/cmice, but there were fewer APCs in the GFP +fraction of Balb/c mice compared to Scid, consistentwith an immune rejection of the transducedcells. In addition, efficient transduction ofB lymphocytes was apparent in Balb/c mice,whereas T cells were refractory to lentiviral transduction.Endothelial cell specific markers(CD62E and CD105) were barely expressed onthe GFP + cells, indicating that endothelial cellswere refractory to lentiviral transduction.The relatively efficient lentiviral transductionof APCs should be taken into account whendesigning vectors for gene therapy. Inadvertenttransgene expression in transduced APCs may beundesirable in circumstances where an antibodyresponse against the transgene product shouldbe avoided, 34 particularly in the case of hemophiliagene therapy. It has been shown previouslythat when a ubiquitously expressed promoteris used to express a secretable transgeneproduct, it increases the likelihood of neutralizingantibody formation against this protein.Hence, the use of liver-specific promoters may bewarranted to restrict transgene expression inhepatocytes and to circumvent inadvertent transgeneexpression in APCs following systemiclentiviral vector administration. Follow-up studiesin hemophilic mice and dogs and continuedefforts to improve the vector design are warrant-haematologica vol. 88(supplement n. 12):september 2003

120T. VandenDriessche et al.ed to further explore the full potential of lentiviralvectors for hemophilia A gene therapy.HC-Ad vectorsAdenoviral vectors can transduce dividing andnon-dividing cells and are by far the most efficientvectors for hepatic gene delivery. 35 The adenoviralvector genome remains episomal, implyingthat the risk of neoplastic transformation dueto insertional mutagenesis is low. Whereas early-generationadenoviral vectors still containmost viral genes which contribute to inflammatoryresponses, toxicity and short-term transgeneexpression, HC-Ad vectors retain only the necessarycis-acting elements that are required for generatinginfectious vector particles during vectorproduction and the transgene of interest. 35 HC-Ad vectors expressing B-domain deleted humanor canine FVIII from different liver-specific promoterswere injected intravenously into hemophilicFVIII-deficient mice which resulted inphysiologic levels of canine or human FVIII (VandenDriesscheet al, unpublished observations).These results underscore the potential usefulnessof HC-Ad for hemophilia A gene therapy. However,the induction of neutralizing antibodiesdirected against the human or canine xenoproteinsprecluded stable phenotypic correction.ConclusionsThe ideal gene therapy vector for hemophilia Ais not yet available but significant progress hasbeen made in further improving viral vectortechnology for FVIII gene delivery. Proof-of-concepthas recently been established demonstratingthat hemophilia A could be cured by genetherapy in a clinically relevant animal model thatmimics the cognate human disease. 21 This wasaccomplished by intravenous injection of oncoretroviralvectors expressing FVIII into neonatalhemophilic mice. Although these findings mayultimately pave the way towards potential genetherapy for pediatric hemophilia A, clinical trialsin children can only commence once safety hasbeen established in adults or perhaps in childrensuffering from a lethal disease, for which notreatment is currently available. The limitationof onco-retroviral vector mediated gene transferin dividing hepatocytes, could be overcome byusing either lentiviral or HC-Ad vectors, whichresulted in efficient hepatic gene transfer in adultmice. Using HC-Ad vectors, physiologic FVIIIexpression levels could be achieved in hemophilicmice. Follow-up studies in hemophilic mice anddogs are required to further evaluate the potentialof these vectors for gene therapy.The induction of inhibitory antibodies followinggene therapy may be related to several confoundingvariables including the type of vectorused, the purity of the vector preparation, thepromoter used to drive FVIII expression, the siteof administration, the transduced cell types(APC) or the underlying genetic defect and otherhost factors. Whether gene therapy wouldincrease or decrease the likelihood of inhibitorformation compared to protein replacementtherapy is one of the important questions thatstill needs to be addressed. Continuous productionof high levels FVIII in situ following genetherapy may actually induce immune tolerancereminiscent of current immune tolerizationstrategies by repeated high dose clotting factoradministration. Alternatively, gene transfer tonaive patients could evoke the same or even amore potent immune response to FVIII. A majorconcern is that the use of viral vectors expressingcoagulation factors or that impurities in the vectorpreparations may provide immunologicaldanger signals that may facilitate inhibitor formation.36 In addition, gene transfer may result inthe presentation of endogenously synthesizedFVIII–derived peptides in the context of MHCclass I molecules potentially resulting in CTLresponses that could eliminate the FVIII-engineeredtarget cells. The secreted FVIII protein thatis produced in vivo may also be presented in thecontext of MHC class II as in the case of infusedclotting factors. It is not clear whether gene therapycould break tolerance in patients that are tolerantto FVIII and whether inhibitors can be suppressedonce they occur following gene therapy.Since many adult hemophilia patients have aninfectious or inflammatory disease, complexinteractions can influence the therapeutic efficacyof the gene therapy procedure or the propensityfor inhibitor formation and caution is warrantednot to exacerbate these underlying conditions.Despite the tremendous progress in the fieldover the past few years many questions remainlargely unexplored. Extensive gene therapy studiesin preclinical hemophilia models are neededto anticipate the possible outcome in patientsand to increase the overall efficiency of the variousgene therapy strategies while further improvingtheir safety. The development of hemophiliagene therapy will undoubtedly continue to contributeto a better understanding of vector-hostinteractions that will benefit the entire field ofgene therapy. The results from the preclinicalstudies in hemophilic animal models indicatethat the simultaneous development of differentstrategies is likely to bring a permanent cure forhemophilia one step closer to reality.AcknowledgmentsThe authors wish to acknowledge Dr. Schiedner,Dr. Kochanek, Dr. Follenzi, Dr. Naldini, Dr. Berneman,Dr. Ory, Dr. Mulligan, Dr. Saint-Remy, Dr.Kazazian, Dr. Lillicrap, Dr. De Geest Mr. Thorrez,Mrs. Vanslembrouck, Ms. Gillijns, Mrs. Vanderhaegen,Mr. Lenjou Mrs. Johnston and Mrs. Hertelfor their valuable contributions.References1. Chuah MK, Collen D, VandenDriessche T. Gene therapyfor hemophilia: hopes and hurdles. Crit Rev OncolHematol 1998; 28:153-71.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 1212. Chuah MK, Collen D, VandenDriessche T. Gene therapyfor hemophilia. J Gene Med 2001; 3:3-20.3. VandenDriessche T, Collen D, Chuah MK. Viral vectormediatedgene therapy for hemophilia. Curr Gene Ther2001; 1:301-15.4. Miller AD, Miller DG, Garcia JV, Lynch CM. Use ofretroviral vectors for gene transfer and expression.Methods Enzymol 1993; 217:581-99.5. Miller AD. Retroviral vectors. Curr Top MicrobiolImmunol 1992; 158:1-24.6. Anderson WF. Human gene therapy. Nature 1998; 392:25-30.7. Cornetta K, Morgan RA, Gillio A, Sturm S, Baltrucki L,O'Reilly R, et al. No retroviremia or pathology in longtermfollow-up of monkeys exposed to a murineamphotropic retrovirus. Hum Gene Ther 1991; 2:215-9.8. Anderson WF, McGarrity GJ, Moen RC. Report to theNIH Recombinant DNA Advisory Committee onmurine replication-competent retrovirus (RCR) assays(February 17, 1993). Hum Gene Ther 1993; 4:311-21.9. Donahue RE, Kessler SW, Bodine D, McDonagh K,Dunbar C, Goodman S, et al. Helper virus induced Tcell lymphoma in nonhuman primates after retroviralmediated gene transfer. J Exp Med 1992; 176:1125-35.10. Hoeben RC, van der Jagt RC, Schoute F, van Tilburg NH,Verbeet MP, Briet E, et al. Expression of functional factorVIII in primary human skin fibroblasts after retrovirus-mediatedgene transfer. J Biol Chem 1990; 265:7318-23.11. Israel DI, Kaufman RJ. Retroviral-mediated transfer andamplification of a functional human factor VIII gene.Blood 1990; 75:1074-80.12. Lynch CM, Israel DI, Kaufman RJ, Miller AD. Sequencesin the coding region of clotting factor VIII act as dominantinhibitors of RNA accumulation and protein production.Hum Gene Ther 1993; 4:259-72.13. Chuah MK, Vandendriessche T, Morgan RA. Developmentand analysis of retroviral vectors expressinghuman factor VIII as a potential gene therapy for hemophiliaA. Hum Gene Ther 1995; 6:1363-77.14. Hoeben RC, Fallaux FJ, Cramer SJ, van den WollenbergDJ, van Ormondt H, Briet E, et al. Expression of theblood-clotting factor-VIII cDNA is repressed by a transcriptionalsilencer located in its coding region. Blood1995; 85:2447-54.15. Koeberl DD, Halbert CL, Krumm A, Miller AD.Sequences within the coding regions of clotting factorVIII and CFTR block transcriptional elongation. HumGene Ther 1995; 6:469-79.16. Fallaux FJ, Hoeben RC, Cramer SJ, van den WollenbergDJ, Briet E, van Ormondt H, et al. The human clottingfactor VIII cDNA contains an autonomously replicatingsequence consensus- and matrix attachment region-likesequence that binds a nuclear factor, represses heterologousgene expression, and mediates the transcriptionaleffects of sodium butyrate. Mol Cell Biol 1996; 16:4264-72.17. Dwarki VJ, Belloni P, Nijjar T, Smith J, Couto L, RabierM, Clift S, Berns A, Cohen LK. Gene therapy for hemophiliaA: production of therapeutic levels of human factorVIII in vivo in mice. Proc Natl Acad Sci USA 1995;92:1023-7.18. Krall W, Kohn DB. Expression levels by retroviral vectorsbased upon the N2 and the MFG backbones. Gene Ther1996; 3:365.19. Palmiter RD, Sandgren EP, Avarbock MR, Allen DD,Brinster RL. Heterologous introns can enhance expressionof transgenes in mice. Proc Natl Acad Sci USA1991; 88:478-82.20. Hamer DH, Leder P. Splicing and the formation of stableRNA. Cell 1979; 18:1299-302.21. VandenDriessche T, Vanslembrouck V, Goovaerts I,Zwinnen H, Vanderhaeghen ML, Collen D, et al. Longtermexpression of human coagulation factor VIII andcorrection of hemophilia A after in vivo retroviral genetransfer in factor VIII- deficient mice. Proc Natl Acad SciUSA 1999; 96:10379-84.22. VandenDriessche T, Naldini L, Collen D, Chuah MK.Oncoretroviral and lentiviral vector-mediated gene therapy.Methods Enzymol 2002; 346:573-89.23. Kreuz W, Becker S, Lenz E, Martinez Saguer I, EscuriolaEttingshausen C, Funk M, et al. Factor VIII inhibitors inpatients with hemophilia A: epidemiology of inhibitordevelopment and induction of immune tolerance forfactor VIII. Semin Thromb Hemost 1995; 21:382-9.24. Pittman DD, Alderman EM, Tomkinson KN, Wang JH,Giles AR, Kaufman RJ. Biochemical, immunological,and in vivo functional characterization of B-domaindeletedfactor VIII. Blood 1993; 81:2925-35.25. Qian J, Borovok M, Bi L, Kazazian HH Jr, Hoyer LW.Inhibitor antibody development and T cell response tohuman factor VIII in murine hemophilia A. ThrombHaemost 1999; 81:240-4.26. Lewis PF, Emerman M. Passage through mitosis isrequired for oncoretroviruses but not for the humanimmunodeficiency virus. J Virol 1994; 68:510-6.27. Naldini L. Lentiviruses as gene transfer agents for deliveryto non-dividing cells. Curr Opin Biotechnol 1998;9: 457-63.28. Park F, Ohashi K, Chiu W, Naldini L, Kay MA. Efficientlentiviral transduction of liver requires cell cycling invivo. Nat Genet 2000; 24:49-52.29. Kafri T, van Praag H, Ouyang L, Gage FH, Verma IM. Apackaging cell line for lentivirus vectors. J Virol 1999;73:576-84.30. Zufferey R, Nagy D, Mandel RJ, Naldini L, Trono D.Multiply attenuated lentiviral vector achieves efficientgene delivery in vivo. Nat Biotechnol 1997; 15:871-5.31. Chinnasamy D, Chinnasamy N, Enriquez MJ, Otsu M,Morgan RA, Candotti F. Lentiviral-mediated gene transferinto human lymphocytes: role of HIV-1 accessoryproteins. Blood 2000; 96:1309-16.32. Park F, Ohashi K, Kay MA. Therapeutic levels of humanfactor VIII and IX using HIV-1-based lentiviral vectorsin mouse liver. Blood 2000; 96:1173-6.33. Follenzi A, Ailles LE, Bakovic S, Geuna M, Naldini L.Gene transfer by lentiviral vectors is limited by nucleartranslocation and rescued by HIV-1 pol sequences. NatGenet 2000; 25:217-22.34. Pastore L, Morral N, Zhou H, Garcia R, Parks RJ,Kochanek S, et al. Use of a liver-specific promoterreduces immune response to the transgene in adenoviralvectors. Hum Gene Ther 1999; 10:1773-81.35. Kochanek S. Development of high-capacity adenoviralvectors for gene therapy. Thromb Haemost 1999; 82:547-51.36. Matzinger P. An innate sense of danger. SeminImmunol 1998; 10:399-415.37. Chao H, Walsh CE. Induction of tolerance to humanfactor VIII in mice. Blood 2001; 97:3311-2.haematologica vol. 88(supplement n. 12):september 2003

[Gene Therapy]review paperImmunologic sequelae andpotential for inhibitor developmentin adeno-associated viralgene therapy for hemophilia Bhaematologica 2003; 88(suppl. n. 12):122-126http://www.haematologica.org/free/immunotolerance2001.pdfROLAND W. HERZOGDept. Pediatrics, The Children’s Hospital of Philadelphia andUniversity of Pennsylvania Medical Center, Philadelphia, PA,USAFormation of inhibitory antibodies against a coagulationfactor antigen in a gene therapy setting is currentlybeing characterized for adeno-associated viral(AAV) gene transfer in animal models of hemophilia.Hemophilia B dogs of two different colonies witha missense or null mutation in the factor IX (F.IX)gene as well as hemophilia B mice with a large F.IXgene deletion were subjected to AAV-mediatedtransfer of species-specific F.IX transgenes to skeletalmuscle or liver. Treated animals were analyzed fortransgene expression and F.IX-specific B- and T-lymphocyteresponses. Particular emphasis in thisanalysis was given to the role of the underlying mutation,the route of vector administration, and immunemodulation. All of these were found to have a significantimpact on the risk of anti-F.IX formation,which represents a T helper cell-dependent process.These studies provide the groundwork for a betterunderstanding of F.IX-specific CD4 + T-cell activationin AAV-mediated gene transfer that may lead to neutralizinganti-F.IX IgG formation, and thus help definegene transfer protocols with minimal risk for suchimmune responses.Correspondence: Roland W. Herzog, The Children’s Hospital ofPhiladelphia, Abramson Research Center 310, 34th St. andCivic Center Blvd., Philadelphia, PA 19104, USA. Phone: international+1.215.590-4907. Fax: international +1.215.590-3660. E-mail: rwherzog@mail.med.upenn.eduConventional treatment of the inheritedbleeding disorder hemophilia is based onintravenous (IV) infusion of functionalcoagulation factor VIII (F.III, hemophilia A) orfactor IX (F.IX, hemophilia B). Formation ofinhibitory anti-F.VIII or anti-F.IX currently representsthe most serious complication of this proteinreplacement therapy. 1,2 More recently, genetherapy strategies have been introduced into theclinic in the form of phase I clinical trials. 3-6While inhibitor formation is well documented inprotein therapy with a prevalence of 3-4% inhemophilia B and 20-30% in hemophilia A, therisk of such an immune response in gene-basedtreatment is only beginning to be defined. A historyof inhibitor formation has been an exclusioncriterion for enrollment of patients in genetherapy trials, and inhibitors have indeed notbeen observed in these trials. However, it is notclear yet whether the risks and characteristics ofinhibitor formation are similar or different ingene therapy as compared with protein therapy.Moreover, a particular combination of genetransfer vector, DNA construct, and target tissue/routeof administration is likely to produceits own unique set of signals to the immune system.7 Therefore, the risk and characteristics ofimmune responses caused by gene transfer haveto be defined for each particular protocol. Here,animal experiments designed to define anti-F.IXresponses in adeno-associated virus (AAV)-mediatedgene transfer are summarized.AAV vectors have been shown to transfer a F.IXtransgene efficiently to skeletal muscle fibers orhepatocytes following intramuscular (IM) injectionof the vector or infusion into the portal circulation,resulting in sustained expression andpartial to complete correction of the coagulationdeficiency in hemophilia B mice and dogs. 8-12AAV vectors are based on a single-stranded DNAvirus with a small (4.7-kb) genome flanked byinverted terminal repeats (ITRs). The wild-typevirus is a non-pathogenic, replication defectivemember of the parvovirus family. 13 The vectordoes not contain viral coding sequences (whichhave been replaced by a F.IX expression cassette)and can be produced in the absence of a helpervirus. 14 A phase I trial of IM administration ofthe vector in patients with severe disease (

IV International Workshop on Immune Tolerance in Hemophilia 123Biology of lymphocyte responses in muscledirectedgene transfer – murine models ofgene transferIn early preclinical experiments, we observedthat IM injection of an AAV vector or a first generationadenoviral vector, both expressinghuman F.IX (hF.IX) from the strong viral CMVenhancer/promoter, resulted in rapid inductionof neutralizing anti-hF.IX IgG in immunocompetentmice. 7,15 While these results may not besurprising given the use of a non-species-specifictransgene, this experimental system was usefulin delineating lymphocyte subsets involved inanti-F.IX formation following muscle-directedgene transfer (as summarized in Figure 1). AntihF.IXIgG formation occurred by day 14 after vectoradministration resulting in neutralization ofsystemic hF.IX expression in the context of eithervector system. However, a series of B- and T-cellassays revealed substantial differences in the typeof immune responses for the two vectors. Adenoviralgene transfer resulted in activation of cytotoxicT-lymphocytes specific for hF.IX expressingcells as well as for adenoviral antigens. Musclefibers expressing the transgene were targeted byinflammatory infiltrates including CD8 + (CTL)lymphocytes, thereby causing destruction oftransduced fibers and elimination of transgeneexpression. These results highlight an importantdifference between protein and gene therapy,namely that cells expressing the coagulation factorafter gene transfer are capable of presentingantigen-derived peptides by their MHC class Imolecules. If this occurs in the target cells suchas muscle fibers, these cells become potential targetsfor CTL responses. If MHC class I presentationoccurs in professional antigen-presentingcells (APCs), such as dendritic cells, this maycause activation of CTLs. In the case of adenoviralgene transfer, MHC class II-resticted T helpercells (CD4 + T-cells) of Th1 and Th2 subsets wereactivated resulting in formation of primarilyIgG2a (the murine equivalent of human IgG1)anti-hF.IX. In contrast, AAV-transduced muscletissue continued to express the transgene even inthe presence of antibody formation without evidenceof inflammation or activation of hF.IXspecificCTL. Data from Jooss et al. suggest thatAAV vectors transduce APCs (dendritic cells) relativelyinefficiently, and that this accounts forthe lack of CTL activation against the transgeneproduct. 16,17 Activation of T helper cells by AAVvector encoded hF.IX showed a stronger biastoward the Th2 subset resulting in predominantsynthesis of IgG1 (the murine equivalent ofhuman IgG4). 7,18 These characteristics suggestthat inhibitor formation in AAV-mediated genetherapy is mechanistically similar to antibodyformation in the context of protein infusion, andis probably not similar to the immune responsetypically seen in the context of a viral infection.Immune modulation may prevent inhibitor formationin the context of a F.IX gene deletionIn order to address the risk of inhibitor formationin AAV-mediated gene transfer, we constructedvectors that express species-specifictransgenes. Hemophilia B mice that received anAAV vector expressing murine F.IX (mF.IX)developed inhibitory anti-mF.IX within 4-8weeks after IM injection of the vector. 18 Thesemice had been generated by means of geneticFigure 1. Model for lymphocyte activation resultingin anti-F.IX formation after AAV-mediated gene transferto skeletal muscle. Note that T-cell activation andantibody (inhibitor) formation is more likely to occurin the context of a F.IX null mutation (e.g. gene deletion,early stop codon, etc.) than in the context of amissense mutation. This diagram summarizes resultsobtained in murine models of gene transfer. AAVmediatedgene transfer to skeletal muscle results insecretion of F.IX from transduced myofibers. Uponuptake of the F.IX antigen by professional APCs suchas dendritic cells, peptide fragments are displayed byMHC class II molecules, which may lead to activationof CD4 + T helper cells. In the context of AAV-mediatedgene transfer to skeletal muscle, these are primarilyof the Th2 subset, thus promoting IgG1 secretionby B cells, while limited activation of Th1 cellsmay result in some synthesis of IgG2a or IgG2b. Activationof F.IX-specific, MHC class I restricted CD8 +cytotoxic T lymphocytes was not observed.haematologica vol. 88(supplement n. 12):september 2003

124R.W. Herzogengineering (knock out) technology and containeda large deletion including the promoterand the first three exons of the F.IX gene. 19 Consequently,the animals did not produce endogenousF.IX transcript or protein. Anti-mF.IXresponses were weaker and more delayed as comparedto anti-hF.IX formation. 7,18 Inhibitor titersof 5-20 BU persisted for >12 months. HemophiliaB mice also formed inhibitors after repeatedIV infusion of recombinant mF.IX protein. 18These data are consistent with clinical experience,in which an increased risk of inhibitor formationhas been observed following proteintreatment in hemophilia B patients with substantialloss of F.IX coding information (e.g. genedeletions 20,21 ). This experimental system was usefulin evaluating strategies for avoiding inhibitorformation in a setting with high risk of anti-F.IXsynthesis. We had hypothesized that transientimmune modulation around the time of vectoradministration combined with sustained transgeneexpression from AAV-mediated gene transfermay be sufficient for prolonged systemic F.IXexpression without inhibitor formation. To testthis hypothesis, hemophilia B mice receivedAAV-mF.IX vector by IM injection at a single timepoint combined with IV or intraperitoneal (IP)injection of an immune suppressive drug thatwas given at the time of vector administrationand intermittently at several subsequent timepoints. 18 Only decreased inhibitor titers and partialsuccess was observed with agents that specificallyblock co-stimulatory pathways that areimportant for activation of B- and T-lymphocytes(signal 2 blockers such as anti-CD40L andCTLA4-Ig fusion protein). However, it is conceivablethat optimization of dosing and scheduleof drug administration could improve thesuccess rate. Inhibitor formation was initiallyefficiently blocked by infusing FK506 every otherday for one month, resulting in substantialshortening of the aPTT due to transgene expression.FK506 inhibits an intracellular signalingpathway required for T-cell activation. Nonetheless,correction of the coagulation deficiency waslost once administration of the drug was stopped,and low titer inhibitors were measured by 3-4months. The most promising results wereobtained with cyclophosphamide, a DNA alkylatingagent with a cytotoxic effect on dividingcells such as proliferating B- and T-cell clones.Cyclophosphamide is commonly used in treatmentof autoimmune diseases and in someimmune tolerance regimens in hemophiliapatients who developed inhibitors during proteintherapy. 22,23 This compound was given biweeklyup to week 6 after vector administration at dosesup to 50 mg/kg; such doses do not cause significantchanges in white blood cell counts inmice. Hemophilia B mice treated with this combinationof vector injection and transientimmune suppression showed correction of theaPTT without anti-mF.IX formation for >1 year(duration of the experiment).Scale-up to hemophilia B dogs and sustainedF.IX expression in the context of aF.IX missense mutationHemophilia B dogs represent an excellent largeanimal model for scale-up studies in a novel geneTable 1. Incidence of inhibitor formation in gene therapy for hemophilia B using hemophilia B dogs and mice and an AAV-2 vectorencoding a species-specific transgene. The table lists the animal model/mutation, route of administration (IM vs. hepatic),number of animals that developed inhibitors in the respective study, and the referenced paper.Hemophilia B animal model Route of administration Vector dose Incidence of References(vg/kg) inhibitor formationUNC-Chapel Hill dogs (F.IX missense mutation) IM ≤3x10 12 0/7 Herzog et al. 1999, Chao et al. 1999UNC Chapel Hill dogs IM 8.5x10 12 Transient (1/1) Herzog et al. 1999UNC Chapel Hill dogs Portal vein (for hepatic gene transfer) ≤4x10 12 0/7 Snyder et al. 1999, Wang et al. 2000,Mount et al. 2002Auburn dogs (F.IX null mutation) IM 1x10 12 2/2 Herzog et al. 2001IM + cyclophosphamide 1x10 12 0/1Auburn dogs Mesenteric vein (for hepatic gene transfer) ≤3x10 12 (1x10 12 )* 1/3 (0/2) 1 Mount et al. 2002Knock out mice with F.IX gene deletion IM° ≤4x10 12 10/10 Fields et al. 2001IM + cyclophosphamide 4x10 12 2/6 +IM: intramuscular. *Two dogs treated with 1x10 12 vg/kg did not develop anti-F.IX, while one dog treated with 3x10 12 vg/kg showed inhibitor formation by week 5 after genetransfer. Results from this animal are complicated by the fact that the dog had additionally inherited pyruvate kinase deficiency (associated with chronic hemolytic anemia)and anti-phospholipid. °Sustained expression of canine or human F.IX without inhibitor formation has been achieved by hepatic AAV-mediated gene transfer in hemophilia Bmice on a C57BL/6 genetic background (Snyder et al. 1999, Wang et al. 1999). The risk of inhibitor formation in other strain backgrounds remains to be evaluated. + Those 2mice with inhibitor formation received 25 mg cyclophosphamide/kg, while 0/3 animals treated with 50 mg/kg showed inhibitor formation. vg: vector genomes as titered byquantitative dot blot hybridization.haematologica vol. 88(supplement n. 12):september 2003

IV International Workshop on Immune Tolerance in Hemophilia 125therapy protocol and for safety studies such asassessment of the risk of inhibitor formation.There are two characterized colonies with no circulatingF.IX antigen. The dogs at UNC-ChapelHill have a missense mutation in the F.IX gene.Modeling studies predict that the mutant F.IXmolecule does not fold correctly and therefore islikely destroyed intracellularly. 24,25 IM administrationof an AAV vector (expressing canine F.IX)at multiple sites at a single time point directedsustained systemic expression of vector dosedependentlevels of biologically active canine F.IX(>4 years) without vector-related toxicity. 8,12 Furthermore,these experiments illustrated the simplicityand safety of this non-invasive method ofgene transfer. Expression was demonstrated byELISA and by partial correction of the whole bloodclotting time (WBCT) and, at higher vector doses,of the aPTT. Dose escalation showed that~5×10 12 -1×10 13 vector genomes [vg]/kg wererequired for therapeutic levels of expression (>1%of normal). At the reported vector doses (up to8.5×10 12 vg/kg), inhibitors were either absent(4/5 dogs) or transient (spontaneous remissionwithin 2 months, 1/5 animals). 8 Inhibitor formationmay be observed at higher vector doses, inparticular at high doses per site of injection (Herzoget al., Hum Gen Ther, in press). Thus, althoughthe risk of inhibitor formation in the context of amissense mutation is greatly reduced, it may benecessary to include transient immune suppressionat high vector doses in muscle-directed genetransfer.Similar success in the UNC hemophilia B dogshas been reported for AAV-mediated hepatic genetransfer following portal vein infusion of the vector,albeit at lower vector doses. 9,11,26 No anticF.IXwas reported in these studies, and expressionwas sustained for at least several years. Similarto IM injections, the AAV vector does notcause heptotoxicity, and vector administration isgenerally well tolerated by the animals. Recentprogress in vector delivery techniques and the useof alternate AAV serotypes (all data discussed herewere obtained with AAV-2 vector) with greatertransduction efficiency of skeletal muscle are likelyto reduce if not eliminate the dose advantageof the liver route over the muscle route (refs. no.27,28 and Arruda et al., unpublished results). However,the risk of inhibitor formation in theseapproaches remains to be defined. In particular,one has to consider that levels of locally producedF.IX may be much higher following gene transferthat is the case for IM injection of the AAV-2 vector,which may significantly influence immuneresponses.Sustained F.IX expression in hemophilia Bdogs with a F.IX null mutation followinghepatic gene transferAnalogous to results in the hemophilia B mice,dogs with a F.IX null mutation (an early stopcodon and unstable mRNA, hemophilia B dogcolony at Auburn University), IM administrationof an AAV vector expressing cF.IX induced persistenthigh titer inhibitors. 29,30 This immuneresponse was observed at vector doses that did notcause inhibitor formation but rather resulted insustained cF.IX expression in the model characterizedby a missense mutation. Inhibitor formationwas not observed in a null mutation dog thatadditionally received cyclophosphamide asdescribed above for treatment of murine hemophiliaB. Inhibitor formation was also observed ina null mutation dog that received IV infusion ofpurified plasma-derived cF.IX. 29 Surprisingly, 2animals of this colony that were treated by hepaticgene transfer did not form anti-cF.IX and continueto express cF.IX at levels of 5-12% of normalwith complete or substantial correction ofthe WBCT and of the aPTT. 26 This was achieved bymesenteric vein infusion of an AAV vector containinga strong hepatocyte-specific enhancer/promotercombination. Similar results hadbeen obtained previously in hemophilia B miceusing hepatic gene transfer. However, these micehave been bred on a C57BL/6 genetic background,which is more promiscuous for transgeneexpression without immune responses followingthis route of administration than otherstrains of inbred mice, thus making definite statementsabout the risk of inhibitor formation moredifficult. 31,32 Nonetheless, other data showed thatliver-directed AAV-mediated gene transfer may becharacterized by a reduced risk of anti-F.IX formationeven in strains other than C57BL/6. 33Our results from the F.IX null mutation dogsillustrate the lower likelihood of inhibitor formationby AAV-mediated liver-directed gene therapyin a large animal model of hemophilia B as comparedto other treatment modalities. 26 Theimmunologic mechanism responsible for theseobservations is currently under active investigationin our laboratory. Table 1 summarizes theincidence of inhibitor formation as reported inseveral studies using muscle- or liver-directedAAV-mediated gene transfer of a species-specificF.IX gene in animal models of hemophilia B.In summary, AAV-mediated gene transfer ischaracterized by a reduced potential for cellularimmune responses and inflammation but mayresult in T helper cell-dependent antibody formationto the F.IX transgene product. The datastrongly suggest a crucial role of the gene transferrecipient’s underlying F.IX mutation on therisk of inhibitor formation. In muscle-directedgene transfer, inhibitor formation is observed inanimals with substantial loss of F.IX codinginformation such as a large gene deletion or anearly stop codon, whereas sustained systemicexpression without or with only transientinhibitor formation was achieved in the contextof a crm- missense mutation. These differencesin the incidence of antibody formation may beexplained by lack of expression of potentiallyimmunodominant T-cell epitopes during devel-haematologica vol. 88(supplement n. 12):september 2003

126R.W. Herzogopment of the immune system in the setting ofa null mutation. The risk of inhibitor formationcan be substantially reduced by transientimmune modulation or by hepatic gene transfer,which may not elicit an anti-F.IX response evenin animal models that are at risk of inhibitor formationin protein-based therapy.AcknowledgmentsRWH is supported by a Career Development awardfrom the National Hemophilia Foundation (USA).The majority of the work described here is supportedby NIH grants P01 HL64190 and R01 HL61921 toK.A. High. The author is grateful for the support andcontributions of Drs. K.A. High, V.R. Arruda, P.A.Fields, C.D. Lothrop Jr., L.B. Couto, and T.C.Nichols.References1. High KA. Factor IX: molecular structure, epitopes, andmutations associated with inhibitor formation. In: AledortLM, Hoyer LW, Lusher JM, Reisner HM, White GC,eds. Advances in Experimental Medicine and Biology.Vol. 386:79-86, New York, Plenum Press, 1995.2. Hoyer LW. The incidence of factor VIII inhibitors inpatients with severe hemophilia A. Adv Exper Med &Biol 1995;386:35-463. High KA. Gene transfer as an approach to treatinghemophilia. Circ Res 2001;88:137-44.4. High KA. Gene therapy: a 2001 perspective. Haemophilia2001; 7 Suppl 1:23-7.5. Kay MA, Manno CS, Ragni MV, Larson PJ, Couto LB,McClelland A, et al. Evidence for gene transfer andexpression of factor IX in haemophilia B patients treatedwith an AAV vector. Nature Genet 2000;24:257-616. Roth DA, O'Brien J, Treco DA, Selden RF. Non-viraltransfer of the gene encoding coagulation factor VIII inpatients with secere hemophilia A. N Engl J Med 2001;344:1738-42.7. Fields PA, Kowalczyk DW, Arruda VR, McCleland ML,Hagstrom JN, KJ P, et al. Choice of vector determines Tcell subsets involved in immune responses against thesecreted transgene product factor IX. Mol Ther 2000;1:225-35.8. Herzog RW, Yang EY, Couto LB, Hagstrom JN, Elwell D,Fields PA, et al. Long-term correction of canine hemophiliaB by gene transfer of blood coagulation factor IXmediated by adeno-associated viral vector. Nat Med1999;5:56-63.9. Snyder RO, Miao C, Meuse L, Tubb J, Donahue BA, LinH-F, et al. Correction of hemophilia B in canine andmurine models using recombinant adeno-associatedviral vectors. Nat Med 1999;5:64-70.10. Wang L, Takabe K, Bidlingmaier SM, Ill CR, Verma IM.Sustained correction of bleeding disorder in hemophiliaB mice by gene therapy. Proc Natl Acad Sci USA 1999;96:3906-1011. Wang L, Nichols TC, Read MS, D.A. B, Verma IM. Sustainedexpression of therapeutic level of factor IX inhemophilia B dogs by AAV-mediated gene therapy inliver. Mol Ther 2000;1:154-812. Chao H, Samulski R, Bellinger D, Monahan P, NicholsT, Walsh C. Persistent expression of canine factor IX inhemophilia B canines. Gene Ther 1999;6:1695-70413. Carter PJ, Samulski RJ. Adeno-associated viral vectorsas gene delivery vehicles. Int J Mol Med 2000; 6:17-2714. Matsushita T, Elliger S, Elliger C, Podsakoff G, VillarrealL, Kurtzman GJ, Iwaki Y, Colosi P. Adeno-associatedvirus vectors can be efficiently produced without helpervirus. Gene Ther 1998;5:938-45.15. Herzog RW, Hagstrom JN, Kung Z-H, Tai SJ, Wilson JM,Fisher KJ, et al. Stable gene transfer and expression ofhuman blood coagulation factor IX after intramuscularinjection of recombinant adeno-associated virus. Proc.Natl Acad Sci USA 1997;94:5804-9.16. Jooss K, Yang Y, Fisher KJ, Wilson JM. Transduction ofdendritic cells by DNA viral vectors directs the immuneresponse to transgene products in muscle fibers. J Virol1998;72:4212-23.17. Sarukhan A, Camugli S, Gjata B, von Boehmer H, DanosO, Jooss K. Successful interference with cellular immuneresponses to immunogenic proteins encoded by recombinantviral vectors. J Virol 2001;75:269-77.18. Fields PA, Arruda VR, Armstrong E, Chu K, Mingozzi F,Hagstrom JN, et al. Risk and prevention of anti-factorIX formation in AAV-mediated gene transfer in contextof a large factor IX gene deletion. Mol Ther 2001;4:201-10.19. Lin HF, Maeda N, Smithies O, Straight DL, StaffordDW.: A coagulation factor IX-deficient mouse modelfor human hemophilia B. Blood 1997;90:3962-620. Giannelli F, Green PM. The molecular basis of hemophiliaA and B. In: Lee CA, editor. Clinical haematology:Haemophilia. Bailliere Tindall; Vol. 9: 1996. p. 211-8.21. Ljung R, Petrini P, Tengborn L, Sjoerin E. HaemophiliaB mutations in Sweden: a population-based study ofmutational heterogeneity. Brit J Haematol 2001; 113:81-6.22. Nilsson IM, Hedner U, Bjorlin G. Suppression of factorIX antibody in hemophilia B by factor IX and cyclophosphamide.Ann Intern Med 1973;78:91-5.23. Nilsson IM, Berntorp E, Zettervall O. Induction ofimmune tolerance in patients with hemophilia andantibodies to factor VIII by combined treatment withintravenous IgG, cyclophosphamide, and factor VIII. NEngl J Med 1988;318:947-50.24. Evans JP, Brinkhous KM, Brayer GD, Reisner HW, HighKA. Canine hemophilia B resulting from a point mutationwith unusual consequences. Proc Natl Acad SciUSA 1989;86:10095-9.25. Herzog RW, Arruda VR, Fisher TH, Read MS, NichiolsTC, High KA. Absence of circulating factor IX antigen inhemophilia B dogs of the UNC-Chapel Hill colony.Thromb Haemostas 2000;84:352-4.26. Mount JD, Herzog RW, Tillson DM, Goodman SA,Robinson N, McCleland ML, et al. Sustained phenotypiccorrection of hemophilia B dogs with a factor IXnull mutation by liver-directed gene therapy. Blood2002;99:2670-6.27. Chao H, Liu Y, Rabinowitz J, Li C, Samulski RJ, WalshCE. Several log increase in therapeutic transgene deliveryby distinct adeno-associated viral serotype vectors.Mol Ther 2000;2:619-23.28. Xiao W, Chirmule N, Berta SC, McCullough B, Gao G,Wilson JM. Gene therapy vectors based on adeno-associatedvirus type 1. J Virol 1999;73:3994-4003.29. Herzog RW, Mount JD, Arruda VR, High KA, LothropCDJ. Muscle-directed gene transfer and transientimmune suppression result in sustained partial correctionof canine hemophilia B caused by a null mutation.Mol Ther 2001;4:192-200.30. Mauser AE, Whitlark J, Whitney KM, Lothrop CD Jr. Adeletion mutation causes hemophilia B in Lhasa Apsodogs. Blood 1996;88:3451-5.31. Fields P, Armstrong E, Hagstrom JN, Arruda VR, MurphyML, Farrell JP, High KA, Herzog RW. Intravenousadministration of an E1/E3-deleted adenoviral vectorinduces tolerance to factor IX in C57BL/6 mice. GeneTherapy 2001;8:354-61.32. Michou A, Santoro L, Christ M, Julliard V, Pavirani A,Mehtali M. Adenovirus-mediated gene transfer: influenceof transgene, mouse strain and type of immuneresponse on persistence of transgene expression. GeneTher 1997;4:473-82.33. Nathwani AC, Davidoff A, Hanawa H, Zhou JF, VaninEF, Nienhuis AW. Factors influencing in vivo transductionby recombinant adeno-associated viral vectorsexpressing the human factor IX cDNA. Blood 2001;97:1258-65.haematologica vol. 88(supplement n. 12):september 2003

Index of authorsAhmad, R.U., 56Albert, T., 78Astermark, J., 30,46,72Baudo, F., 93Berntorp, E., 26,30,46,72Brackmann, H.H., 78Bril, W.S., 18Brito, D., 43Chuah, M.K.L., 115Collen, D., 115de Cataldo, F., 93Effenberger, W., 78Gilles, J.G., 53Goodeve, A., 2Hampel, H., 86Hanfland, P., 78Hausl, Ch., 56Haya, S., 43Herzog, R.W., 122Hess, L., 78Hodes, R.J., 100Huth-Kühne, A., 86Jacquemin, M.G., 65Kaveri, S., 49Kazatchkine, M.D., 49Kobelt, R., 41Lacroix-Desmazes, S., 49Lages, P., 86Lillicrap, D., 75,111Ljung, R.C.R., 4Lollar, P., 11Lopez-Fernandez, MF., 43Maier, E., 56Mauser-Bunschoten, E.P., 35Mostarda, G., 93Nemes, L., 106Nettekoven, W., 78Oldenburg, J., 78Pitlik, E., 106Reipert, B.M., 56Rodriguez-Martorell, F., 43Roosendaal, G., 35Saint-Remy, J.M.R., 8,65,67Sasgary, M., 56Schwaab, R., 78Schwarz, H.P., 56Sosa, R., 43Turecek, P.L., 56Tusell, J., 43Unkrig, Ch., 78Vacchio, M.S., 100van den Berg, H.M., 35van den Brink, E.N., 18VandenDriessche, T., 115Vetter, H., 78Villar, A., 43Voorberg, J., 18Warrier, I., 70Zeitler, H., 78Zimmermann, R., 86

Subscribe now for 2004!haematologica ISSN 0390-6078haematologica was founded in 1920 by Adolfo Ferrata and is the oldest hematology journal in existence in the world today.It is owned by a non-profit organization, the Ferrata Storti Foundation, and its headquarters are in Pavia, Italy.haematologica publishes peer-reviewed papers across all areas of experimental and clinical hematology and makes them immediately,freely accessible online to scientists and to the public around the world, for the benefit of scientific progress and education.haematologica takes advantage of Internet and uses its own efficient online manuscript submission, tracking, peer-review, and publishingsystem. The journal's objective is to review a paper in 3 weeks and immediately communicate the editorial decision to the authors. When amanuscript is accepted without changes, it can be published online within 6-8 weeks from submission. Although original research papersremain the heart of haematologica, the rapid exchange of observations and ideas also includes editorials, review articles, andscientific correspondence.haematologica promotes published papers by electronically supplying MedLine® (National Library of Medicine) with their citationsat the time of online publication.haematologica is the official organ of the Italian Society of Hematology, the Italian Society of Experimental Hematology, the Spanish Associationof Hematology and Hemotherapy, the Italian Society for Hemostasis and Thrombosis and the Italian Association for Pediatric Oncohematology.Subscription Order Card — Fill out this form with your data and send it to: Haematologica Journal Office, Strada Nuova 134, 27100 Pavia, ItalySubscribe to haematologica!Yes! I would like to subscribe to haematologica (12 issues for year 2004)at the rate checked below.Subscription ratesInstitutional (surface mail ❏ €uro 350) (air mail ❏ €uro 400)Personal (surface mail ❏ €uro 150) (air mail ❏ €uro 200)Fill out this form with your data and send it to us and your subscription will start inJanuary 2004.❏ Please charge my credit card for €uro _______________________❏ VISA ❏ American ExpressCard No. Exp. date: _____/_____ Signature: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .City/State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Country/Zip. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Tel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .