FVIII proteins with a modified immunodominant T-cell epitope exhibit reduced immunogenicity and normal FVIII activity.

Factor VIII (FVIII)-neutralizing antibodies (inhibitors) are a serious complication in hemophilia A (HA). The peptide FVIII2194-2213 contains an immunodominant HLA-DRA*01-DRB1*01:01 (DRB1*01:01)-restricted epitope recognized by CD4+ T-effector cells from HA subjects. The aim of this study was to identify amino acid substitutions to deimmunize this epitope while retaining procoagulant function and expression levels comparable to those of wild-type (WT) FVIII proteins. The shortest DRB1*01:01-binding peptide was FVIII2194-2205, and residues important for affinity were identified as F2196, M2199, A2201, and S2204. T-cell proliferation experiments with Ala-substituted FVIII2194-2205 peptides identified F2196A as a substitution that abrogated proliferation of clones specific for the WT sequence. T-cell clones that were stimulated by recombinant WT-FVIII-C2 (rWT-FVIII-C2) protein did not proliferate when cultured with rFVIII-C2-F2196A, indicating the immunogenic peptide includes a naturally processed T-cell epitope. Additional amino acid substitutions at F2196 and M2199 were evaluated by peptide-MHC class II (MHCII)-binding assays, T-cell proliferation assays, epitope prediction algorithms, and sequence homologies. Six B-domain-deleted (BDD)-FVIII proteins with substitutions F2196A, F2196L, F2196K, M2199A, M2199W, or M2199R were produced. Proliferation of T-cell clones and polyclonal lines in response to rBDD-FVIII-F2196K and rBDD-FVIII-M2199A was reduced compared with responses to WT-BDD-FVIII. The BDD-FVIII-F2196K sequence modification appears to be the most promising sequence variant tested here, due to its effectiveness at eliminating DRB1*01:01-restricted immunogenicity, low potential immunogenicity in the context of other MHCII alleles, expression level comparable to WT-BDD-FVIII, and retained procoagulant activity. These results provide proof of principle for the design of less immunogenic FVIII proteins targeted to specific subsets of HA patients.

T cells from hemophilia A subjects recognize the same HLA-restricted FVIII epitope with a narrow TCR repertoire.

Factor VIII (FVIII)-neutralizing antibodies ("inhibitors") are a serious problem in hemophilia A (HA). The aim of this study was to characterize HLA-restricted T-cell responses from a severe HA subject with a persistent inhibitor and from 2 previously studied mild HA inhibitor subjects. Major histocompatibility complex II tetramers corresponding to both of the severe HA subject's HLA-DRA-DRB1 alleles were loaded with peptides spanning FVIII-A2, C1, and C2 domains. Interestingly, only 1 epitope was identified, in peptide FVIII2194-2213, and it was identical to the HLA-DRA*01-DRB1*01:01-restricted epitope recognized by the mild HA subjects. Multiple T-cell clones and polyclonal lines having different avidities for the peptide-loaded tetramer were isolated from all subjects. Only high- and medium-avidity T cells proliferated and secreted cytokines when stimulated with FVIII2194-2213 T-cell receptor ß (TCRB) gene sequencing of 15 T-cell clones from the severe HA subject revealed that all high-avidity clones expressed the same TCRB gene. High-throughput immunosequencing of high-, medium-, and low-avidity cells sorted from a severe HA polyclonal line revealed that 94% of the high-avidity cells expressed the same TCRB gene as the high-avidity clones. TCRB sequencing of clones and lines from the mild HA subjects also identified a limited TCRB gene repertoire. These results suggest a limited number of epitopes in FVIII drive inhibitor responses and that the T-cell repertoires of FVIII-responsive T cells can be quite narrow. The limited diversity of both epitopes and TCRB gene usage suggests that targeting of specific epitopes and/or T-cell clones may be a promising approach to achieve tolerance to FVIII.

Six amino acid residues in a 1200 Å2 interface mediate binding of factor VIII to an IgG4κ inhibitory antibody.

The development of neutralizing anti-factor VIII (FVIII) antibodies complicates the treatment of many hemophilia A patients. The C-terminal C2 domain is a particularly antigenic FVIII region. A crystal structure of recombinant FVIII-C2 bound to an Fab fragment of the patient-derived monoclonal antibody BO2C11, which recognizes an immunodominant inhibitor epitope on FVIII and blocks its ability to bind von Willebrand factor (VWF) and phospholipids, revealed that 15 amino acids in FVIII contact this antibody. Forty-three recombinant FVIII-C2 proteins, each with a surface-exposed side chain mutated to alanine or another residue, were generated, and surface plasmon resonance studies were carried out to evaluate effects of these substitutions on BO2C11/FVIII-C2 binding affinity. Thermodynamic analysis of experiments carried out at three temperatures indicated that one beta hairpin turn at the antigen-antibody interface (FVIII-F2196, N2198, M2199 and F2200) plus two non-contiguous arginines (FVIII-R2215 and R2220), contributed appreciably to the affinity. B-domain-deleted (BDD) FVIII-F2196A, FVIII-F2196K and FVIII-M2199A were generated and characterized. Their pro-coagulant activities and binding to VWF were similar to those of WT-BDD-FVIII, and FVIII-F2196K avoided neutralization by BO2C11 and murine inhibitory mAb 1B5. This study suggests specific sites for amino acid substitutions to rationally design FVIII variants capable of evading immunodominant neutralizing anti-FVIII antibodies.

F8 gene mutation type and inhibitor development in patients with severe hemophilia A: systematic review and meta-analysis.

This systematic review was designed to provide more precise effect estimates of inhibitor development for the various types of F8 gene mutations in patients with severe hemophilia A. The primary outcome was inhibitor development and the secondary outcome was high-titer-inhibitor development. A systematic literature search was performed to include cohort studies published in peer-reviewed journals with data on inhibitor incidences in the various F8 gene mutation types and a mutation detection rate of at least 80%. Pooled odds ratios (ORs) of inhibitor development for different types of F8 gene mutations were calculated with intron 22 inversion as the reference. Data were included from 30 studies on 5383 patients, including 1029 inhibitor patients. The inhibitor risk in large deletions and nonsense mutations was higher than in intron 22 inversions (pooled OR = 3.6, 95% confidence interval [95% CI], 2.3-5.7 and OR = 1.4, 95% CI, 1.1-1.8, respectively), the risk in intron 1 inversions and splice-site mutations was equal (pooled OR = 0.9; 95% CI, 0.6-1.5 and OR = 1.0; 95% CI, 0.6-1.5), and the risk in small deletions/insertions and missense mutations was lower (pooled OR = 0.5; 95% CI, 0.4-0.6 and OR = 0.3; 95% CI, 0.2-0.4, respectively). The relative risks for developing high titer inhibitors were similar.

Recombinant factor IX-Fc fusion protein (rFIXFc) demonstrates safety and prolonged activity in a phase 1/2a study in hemophilia B patients.

Current factor IX (FIX) products display a half-life (t(1/2)) of ∼ 18 hours, requiring frequent intravenous infusions for prophylaxis and treatment in patients with hemophilia B. This open-label, dose-escalation trial in previously treated adult subjects with hemophilia B examined the safety and pharmacokinetics of rFIXFc. rFIXFc is a recombinant fusion protein composed of FIX and the Fc domain of human IgG(1), to extend circulating time. Fourteen subjects received a single dose of rFIXFc; 1 subject each received 1, 5, 12.5, or 25 IU/kg, and 5 subjects each received 50 or 100 IU/kg. rFIXFc was well tolerated, and most adverse events were mild or moderate in intensity. No inhibitors were detected in any subject. Dose-proportional increases in rFIXFc activity and Ag exposure were observed. With baseline subtraction, mean activity terminal t(1/2) and mean residence time for rFIXFc were 56.7 and 71.8 hours, respectively. This is ∼ 3-fold longer than that reported for current rFIX products. The incremental recovery of rFIXFc was 0.93 IU/dL per IU/kg, similar to plasma-derived FIX. These results show that rFIXFc may offer a viable therapeutic approach to achieve prolonged hemostatic protection and less frequent dosing in patients with hemophilia B. The trial was registered at www.clinicaltrials.gov as NCT00716716.

CD4+FOXP3+ regulatory T cells confer long-term regulation of factor VIII-specific immune responses in plasmid-mediated gene therapy-treated hemophilia mice.

Gene transfer of a factor VIII (FVIII) plasmid into hemophilia A (HemA) mice achieved supraphysiologic FVIII expression, but triggered production of high-titer FVIII-specific antibodies and loss of functional FVIII activity. To test whether FVIII-specific regulatory T cells (Tregs) can modulate immune responses against FVIII, we developed a HemA mouse model in which all T cells overexpressed Foxp3 (HemA/Foxp3-Tg). FVIII plasmid therapy did not induce antibody production in HemA/Foxp3-Tg mice. CD4(+)Foxp3(+) T cells isolated from plasmid-treated HemA/Foxp3-Tg mice significantly suppressed proliferation of FVIII-stimulated CD4(+) effector T cells. The percentage of CD4(+) T cells expressing CD25, glucocorticoid-induced tumor necrosis factor receptor, and cytotoxic T lymphocyte antigen 4 increased significantly in spleen and peripheral blood for 9 weeks. Mice receiving adoptively transferred Tregs from FVIII-exposed HemA/Foxp3-Tg mice produced significantly reduced antibody titers compared with controls after initial challenge with FVIII plasmid and second challenge 16 weeks after first plasmid treatment. Adoptively transferred Tregs engrafted and distributed at 2% to 4% in the Treg compartment of blood, lymph nodes, and spleens of the recipient mice and induced activation of endogenous Tregs; the engraftment decreased to negligible levels over 8 to 12 weeks. Antigen-specific Tregs can provide long-lasting protection against immune responses in vivo and limit recall responses induced by a second challenge via infectious tolerance.

Lineages of human T-cell clones, including T helper 17/T helper 1 cells, isolated at different stages of anti-factor VIII immune responses.

The development of neutralizing antibodies (inhibitors) after factor VIII (FVIII) infusions is a serious complication that affects approximately one-quarter of hemophilia A patients who have access to replacement therapy. To investigate the differentiation of naive T cells into FVIII-specific helper T cells that promote B-cell activation and antibody secretion, HLA-DRA-DRB1*0101-restricted T-cell clones that respond to a specific epitope in FVIII were isolated from a mild hemophilia A subject (the proband) 19 weeks and 21 months after his development of a high-titer inhibitor. Clones responding to the same epitope were also isolated from his multiply infused brother, who has not developed a clinically significant inhibitor. The 19-week proband clones were T helper (T(H))17/T(H)1- or T(H)1/T(H)2-polarized, whereas all 8 clones isolated 21 months postinhibitor development were T(H)2-polarized cells. In contrast, all 6 clones from the brother who did not develop an inhibitor were T(H)1-polarized, indicating that tolerance to FVIII can be maintained even with circulating T(H)1-polarized cells that respond vigorously to in vitro FVIII stimulation. This is the first evidence that T(H)17/T(H)1-polarized cells play a role in hemophilic immune responses to FVIII. Furthermore, this is the first report of successful isolation and expansion of antigen-specific human T(H)17/T(H)1 clones using standard culture conditions.

Inhibitors of factor VIII in black patients with hemophilia.

BACKGROUND: Black patients with hemophilia A (factor VIII deficiency) are twice as likely as white patients to produce inhibitors against factor VIII proteins given as replacement therapy. There are six wild-type factor VIII proteins, designated H1 through H6, but only two (H1 and H2) match the recombinant factor VIII products used clinically. H1 and H2 are found in all racial groups and are the only factor VIII proteins found in the white population to date. H3, H4, and H5 have been found only in blacks. We hypothesized that mismatched factor VIII transfusions contribute to the high incidence of inhibitors among black patients. METHODS: We sequenced the factor VIII gene (F8) in black patients with hemophilia A to identify causative mutations and the background haplotypes on which they reside. Results from previous Bethesda assays and information on the baseline severity of hemophilia, age at enrollment, and biologic relationships among study patients were obtained from review of the patients' medical charts. We used multivariable logistic regression to control for these potential confounders while testing for associations between F8 haplotype and the development of inhibitors. RESULTS: Of the 78 black patients with hemophilia enrolled, 24% had an H3 or H4 background haplotype. The prevalence of inhibitors was higher among patients with either of these haplotypes than among patients with haplotype H1 or H2 (odds ratio, 3.6; 95% confidence interval, 1.1 to 12.3; P=0.04), despite a similar spectrum of hemophilic mutations and degree of severity of illness in these two subgroups. CONCLUSIONS: These preliminary results suggest that mismatched factor VIII replacement therapy may be a risk factor for the development of anti-factor VIII alloantibodies.

B-cell and T-cell epitopes in anti-factor VIII immune responses.

Adequate hemostasis is achieved for many hemophilia A patients by infusion of plasma-derived or recombinant factor VIII (FVIII), but unfortunately, a significant subset of patients develop an immune response in which anti-FVIII antibodies, referred to clinically as "inhibitors," interfere with its procoagulant activity. Inhibitors are the subset of anti-FVIII antibodies that bind to surfaces on FVIII (B-cell epitopes) that are important for its proper functioning in coagulation. Less antigenic FVIII molecules may be designed by identifying and then modifying the amino acid sequences of inhibitor B-cell epitopes. Conversely, characterization of these epitopes can yield important information regarding functionally important surfaces on FVIII. The production of inhibitor antibodies is driven by T cells. T cells recognize FVIII as foreign when FVIII-derived peptides bind to major histocompatibility complex (MHC) class II molecules on the surface of antigen-presenting cells. The class II-peptide complexes must then be recognized by T-cell receptors (TCRs). T-cell stimulation requires sustained association of antigen-presenting cells and T cells through formation of a class II-peptide-TCR complex, and peptide sequences that mediate this association are termed "T-cell epitopes." MHC class II tetramers that bind FVIII-derived peptides and recognize antigen-specific TCRs are proving useful in the characterization of human leukocyte antigen-restricted T-cell responses to FVIII.

A family with combined mutations of the hemophilia A and X-linked adrenoleukodystrophy genes.

X-linked adrenoleukodystrophy and hemophilia A are two distinct, potentially devastating, genetic diseases whose corresponding genes are located in close proximity on the X chromosome. Here we report a family with members affected with both conditions, only the second such family ever reported. Although a structural genomic rearrangement involving both genes was initially predicted to underlie this extremely rare phenotype, genotyping revealed the unlikely occurrence of two individual point mutations. Given the impact of this result on the heritability of the two disorders within the family, this case illustrates the significance of performing detailed molecular analysis in patients with multiple genetic disorders.

The factor VIII C1 domain contributes to platelet binding.

Activated factor VIII (FVIIIa) forms a procoagulant complex with factor IXa on negatively charged membranes, including activated platelet surfaces. Membrane attachment involves the FVIII C2 domain; involvement of the adjacent C1 domain has not been established. Binding of recombinant FVIII C1C2 and C2 proteins to platelets was detected by flow cytometry using (1) anti-C2 monoclonal antibody ESH8 followed by a phycoerythrin-labeled secondary antibody; (2) biotinylated C1C2 detected by phycoerythrin-labeled streptavidin, and (3) C1C2 and C2 site-specifically labeled with fluorescein. Highest binding and lowest background were obtained using fluorescein-conjugated proteins. More than 90% of activated platelets bound C1C2, compared with approximately 50% for equimolar C2. Estimates using fluorescent microbeads indicated approximately 7,000 C1C2-binding sites per platelet, approximately 1,400 for C2, and approximately 3,000 for fluorescein-labeled FVIIIa. Unlike C2 or FVIII(a), C1C2 bound to approximately 700 sites/platelet before activation. C1C2 binding to activated platelets appeared independent of von Willebrand factor and was competed effectively by FVIII(a), but only partially by excess C2. Fluorescein-labeled FVIIIa was competed much more effectively by C1C2 than C2 for binding to activated platelets. Two monoclonal antibodies that inhibit C2 binding to membranes competed platelet binding of C2 more effectively than C1C2. Thus, the C1 domain of FVIII contributes to platelet-binding affinity.

Severe haemophilia B due to a 6 kb factor IX gene deletion including exon 4: non-homologous recombination associated with a shortened transcript from whole blood.

In genotyping a severe hemophilia B subject, exons 1-3 and 5-8 were normal. Exon 4 did not amplify, suggesting a partial gene deletion. Previously, a French family with an exon 4 deletion had severe haemophilia B with a circulating, dysfunctional factor IX protein missing its first growth factor-like domain; breakpoints were not analyzed. Using a 5' primer for exon 3 and a 3' primer for exon 5 fragments, the subject's factor IX gene amplified a 5 kb fragment whereas 11 kb was predicted, indicating a 6 kb deletion. Restriction endonucleases localized the 3' intron 4 deletion breakpoint to 1.2 kb 5' to exon 5. Sequencing through the breakpoints revealed a 5,969 bp deletion that included exon 4 and was accompanied by a 13 bp duplication inserted near the 3' breakpoint site. Haemophilia was familial; on testing, his mother was confirmed as a heterozygous carrier, whereas his sister was homozygous for the normal, larger fragments. As exons 4 and 5 of the factor IX gene are in frame, this deletion should produce a shortened transcript, missing 114 bp (38 codons from the first growth factor-like domain). Reverse transcription of mRNA prepared from whole blood and PCR identified the shorter cDNA fragment. Western blotting demonstrated a smaller factor IX protein.

Immunomodulation of transgene responses following naked DNA transfer of human factor VIII into hemophilia A mice.

A robust humoral immune response against human factor VIII (hFVIII) following naked DNA transfer into immunocompetent hemophilia A mice completely inhibits circulating FVIII activity despite initial high-level hFVIII gene expression. To prevent this undesirable response, we compared transient immunomodulation strategies. Eight groups of mice (n = 4-9 per group) were treated with naked DNA transfer of pBS-HCRHPI-hFVIIIA simultaneously with immunosuppressive reagents that included cyclosporine A (CSA), rapamycin (RAP), mycophenylate mofetil (MMF), a combination of CSA and MMF, a combination of RAP and MMF, a monoclonal antibody against murine CD40 ligand (MR1), recombinant murine Ctla4Ig, and a combination of MR1 and Ctla4Ig. All animals except those receiving only CSA exhibited delayed or absent immune responses against hFVIII. The most effective immunosuppressive regimen, the combination of Ctla4Ig and MR1, prevented inhibitor formation in 8 of 9 animals; the ninth had transient low-titer antibodies. All 9 mice of this group produced persistent, therapeutic levels of hFVIII for more than 6 months. When challenged with the T-dependent antigen bacteriophage Phix174, tolerized mice exhibited normal primary and secondary antibody responses, suggesting that transient immunomodulation to disrupt B/T-cell interaction at the time of plasmid injection effectively promoted long-term immune tolerance specific for hFVIII.

Naked DNA transfer of Factor VIII induced transgene-specific, species-independent immune response in hemophilia A mice.

The development of antibodies to a previously unexpressed protein product may limit the success of human gene therapy approaches. We inserted B-domain-deleted factor VIII (FVIII) cDNA of human, canine, or murine origin into the multiple cloning site of a liver-specific vector, pBS-HCRHPI-A, to yield plasmids pBS-HCRHPI-FVIIIA, pBS-HCRHPI-cFVIIIA, and pBS-HCRHPI-mFVIIIA, respectively. Fifty micrograms of each plasmid in 2 ml of solution was rapidly injected into the tail vein of three groups of hemophilia A mice. Factor VIII levels ranging from 3 to 12 IU/ml were obtained from all three groups (normal is 1 IU/ml in human plasma) 3 days after treatment. These initial very high levels of functional human, canine, or murine factor VIII, however, fell gradually to undetectable levels within 2-3 weeks, and their disappearance correlated with the generation of high-titer, inhibitory anti-FVIII antibodies. Notably, this immune response occurred independent of the species of origin of the exogenous factor VIII. Antibody titers to factor VIII were detected beginning at 2 weeks, reached a plateau and remained at high levels for over 6 months. The majority of anti-hFVIII IgG was IgG1 isotype specific, suggesting a humoral response mediated by Th2-induced signals. Consistent with this idea, in a separate group of mice treated with pBS-HCRHPI-FVIIIA, transient immunosuppression by cyclophosphamide significantly delayed (5/6) or abolished (1/6) inhibitory antibody formation against the transgene.

High-level factor VIII gene expression in vivo achieved by nonviral liver-specific gene therapy vectors.

Two liver-specific nonviral gene transfer vectors have been developed to accommodate heterologous genes. The expression cassettes contain (1) a hepatic locus control region from the apolipoprotein E (ApoE) gene (HCR), (2) a liver-specific alpha(1)-antitrypsin promoter (HP), (3) a 1.4-kb truncated factor IX first intron (I) or a synthetic minx intron (mI), (4) a multiple cloning site (MCS) for inserting cDNA sequences, and (5) a bovine growth hormone polyadenylation signal (bpA) to make pBS-HCRHPI-A or pBS-HCRHPmI-A. These vectors were first evaluated with reporter genes encoding human factor IX (hFIX) and green fluorescent protein (GFP). hFIX constructs, pBS-HCRHPI-FIXA and control pBS-HCRHP-FIXIA with the hFIX intron in its native position, produced comparable hFIX gene expression levels (0.5-5 microg/ml) 6 months after naked DNA transfer to mice, whereas the factor IX level from pBS-HCRHPmI-FIXA averaged about 50% lower. RT-PCR analysis of the mRNA indicated that introns inserted upstream from the cDNA were correctly processed and spliced. GFP expression was detected in 15-30% of the hepatocytes in pBS-HCRHPI-GFPA-treated mice. Next, a B domain-deleted human factor VIII (hFVIII) cDNA was inserted into the modified vectors. High-level hFVIII expression (up to 750 ng/ml) was achieved initially in both C57BL/6 mice and Rag2 mice. Moreover, therapeutic levels of hFVIII (20-310 ng/ml) circulated in Rag2 mice 6 months after treatment. These liver-specific gene expression cassettes can deliver a large, heterologous gene such as hFVIII cDNA to achieve high-level, persistent transgene expression after in vivo hepatic gene therapy.

Structure and function of the factor VIII gene and protein.

Factor (F) VIII is a large gene located near the terminus of the long arm of the X chromosome. It contains 26 exons that code for a signal peptide and a 2332 amino acid polypeptide with three different types of domains, namely A1-A2-B-A3-C1-C2. The A domains are homologous with each other and those of ceruloplasmin; substitution into the known crystal structure of the copper binding protein produces molecular models. The large, central B domain is highly glycosylated but has a variable sequence, even among FVIIIs from different species. Most of B can be deleted and the resulting recombinant protein has essentially normal survival in circulation and corrects the bleeding tendency in hemophilia A patients. The C domains are similar to each other, and the crystal structure of a recombinant human C2 domain is known, allowing construction of a molecular model of C1. The FVIII protein is secreted as a heterodimer following at least two intracellular cleavages within the B domain. In circulation it is stabilized by binding to von Willebrand factor (vWF) with a plasma half-life of about 10 hours. After specific thrombin cleavages that remove the remainder of the B domain and one of the high-affinity von Willebrand factor binding sites, FVIII becomes heterotrimeric FVIIIa, capable of enhancing intrinsic FX activation by FIXa. Inactivation of FVIIIa occurs by A2 dissociation or by specific cleavages within A1 and A2 by activated protein C. Control of intrinsic FX activation is critical for hemostasis and thrombosis.

Elevated hemostatic factor levels as potential risk factors for thrombosis.

OBJECTIVES: To review the state of the art relating to elevated hemostatic factor levels as a potential risk factor for thrombosis, as reflected by the medical literature and the consensus opinion of recognized experts in the field, and to make recommendations for the use of specific measurements of hemostatic factor levels in the assessment of thrombotic risk in individual patients. DATA SOURCES: Review of the medical literature, primarily from the last 10 years. DATA EXTRACTION AND SYNTHESIS: After an initial assessment of the literature, key points were identified. Experts were assigned to do an in-depth review of the literature and to prepare a summary of their findings and recommendations. A draft manuscript was prepared and circulated to every participant in the College of American Pathologists Conference XXXVI: Diagnostic Issues in Thrombophilia prior to the conference. Each of the key points and associated recommendations was then presented for discussion at the conference. Recommendations were accepted if a consensus of the 27 experts attending the conference was reached. The results of the discussion were used to revise the manuscript into its final form. CONCLUSIONS: Consensus was reached on 8 recommendations concerning the use of hemostatic factor levels in the assessment of thrombotic risk in individual patients. Detailed discussion of the rationale for each of these recommendations is presented in the article. This is an evolving area of research. While routine use of factor level measurements is not recommended, improvements in assay methodology and further clinical studies may change these recommendations in the future.

Non-inversion factor VIII mutations in 80 hemophilia A families including 24 with alloimmune responses.

Heteroduplex screening identified 74 small mutations in the factor VIII genes of 72 families with hemophilia A. In addition, patients from 3 families with high titer inhibitors had partial gene deletions and 5 unrelated families that were negative for heteroduplex formation had a mutation on direct sequencing. The latter had mild hemophilia A with an inhibitor, and sequencing their exon 23 fragments found a transition predicting a recurrent Arg2150 to His. Of 69 distinct mutations (including the 3 partial gene deletions), 47 are novel. Of small mutations, 51 were missense (one possibly a normal variant and two that could also alter splicing) at 39 sites, 13 were small deletions or insertions (3 inframe and one a normal variant in an intron), 13 were nonsense at 12 sites and 2 altered intron splice junctions. In 24 families, at least one affected member had evidence for an alloimmune response to factor VIII: of these, 11 were associated with missense mutations. In 14 families, de novo origin was demonstrated.