Carrier testing in hemophilia B with an immunoassay that distinguishes a prevalent factor IX dimorphism

Immunoassays with a monoclonal antibody (A-1) detect a prevalent dimorphism in plasma coagulation factor IX. The antibody was shown to react with a dimorphic segment of the normal factor IX sequence as follows. First, A-1 bound to isolated activation peptide (residues 146 through 180) prepared from activated factor IX from a normal plasma pool. Second, binding of recombinant factor IXs with A-1 or factor IX from normal individuals was strong only when they had Threonine (Thr) at position 148; factor IXs with the Alanine (Ala) allele at that position were far less reactive. Third, immunoblot reactivity of Escherichia coli fusion proteins containing known segments of the factor IX sequence restricted the epitope to residues 147 through 153. In 120 hemophilia B pedigrees, the Ala immunoassay allele frequency was 0.19 and did not differ from the Ala frequency in normal males. In 22 of 49 families, immunoassay testing was informative for classification of obligate or possible carriers. In one large family, 4 obligate carriers were heterozygous for the dimorphism and 3 of their 7 daughters were classified as carriers. In other families, when the affected member had less than 1 nmol/L factor IX antigen, heterozygosity for Thr/Ala alleles excluded the carrier state even when DNA studies were not informative. Strong linkage disequilibrium of Thr/Ala alleles with the common TaqI DNA polymorphism was found. Nineteen of 75 normal and hemophilic factor IX genes had the 1.3-kilobase (kb) fragment and coded for the Ala allele; the rest had the 1.8-kb fragment and coded for Thr. In selected families, the A-1 immunoassay is an inexpensive and rapid method to confirm and supplement restriction fragment length polymorphism analyses of DNA for carrier testing.     

Two allotypes of factor IX present in haemophilia B

Factor IX antigen (IX:Ag) was measured with three different immunoradiometric assays (IRMAs) in 30 healthy people and 43 patients with haemophilia B of varying severity. Two of the IRMAs were based on monoclonal antibodies capable of differentiating between two genetically determined molecular variants of normal factor IX. Most patients with severe hemophilia B lacked demonstrable IX:Ag. The factor IX variant that is undetectable with one of the monoclonal antibodies used was present in 2 out of 6 families with moderate haemophilia B and in 1 out of 6 families with mild haemophilia B. The existence of allotypes of factor IX in hemophilia B may have practical implications for carrier detection and prenatal diagnosis.     

Sensitive solid phase enzyme immunoassay for factor IX antigen and classification of hemophilia B

A sensitive solid phase enzyme immunoassay (EIA) was developed for the measurement of factor IX antigen (IX:AG), using rabbit antihuman factor IX antiserum and beta-D-galactosidase, which enabled us to detect IX:AG as low as 10(-4)U/ml. 37 patients with severe hemophilia B have been investigated by EIA, inhibitor neutralization assay and bovine brain prothrombin time. They could be divided into four genetic variants. 25% had normal levels of IX:AG but decreased levels of factor IX clotting activity. On crossed immunoelectrophoresis of the hemophilia B+ and hemophilia BM, we could not find abnormalities in electrophoretic mobilities compared to normal subjects in the presence of 1 mM Ca++ lactate.     

Factor VIII and IX gene polymorphisms and carrier analysis in Indian population

The efficacy of the three common intra- and extragenic polymorphic sites of the factor VIII and IX genes has been examined in the Indian population, with an aim to develop a strategy that would be accurate and informative, yet economical. The approach for hemophilia A carrier detection includes tests for Bcll, Xbal, and Taql polymorphic sites for introns 18 and 22 and the extragenic locus St 14, respectively, whereas for hemophilia B, tests include detection of Taql, Ddel, and Hhal polymorphic sites for introns 4 and 1, and the 3′ flanking region of the factor IX gene, respectively. In hemophilia A, the cumulative efficiency of these three polymorphisms has been found to be 100%, since all 37 tested families were informative for at least one of these three polymorphisms. It is of interest to note that a case of recombination between St 14 and the factor VIII gene was also observed. Of the 47 unrelated X chromosomes examined (normal = 10, factor VIII:C deficiency = 37), heterozygosity for Bcll, Xbal, and St 14 was found to be 47%, 36%, and 86%, respectively, in the factor VIII gene. However, when 37 unrelated X chromosomes (normal = 10, factor IX:C = 27) were analyzed for polymorphism with Taql, Ddel, and Hhal, it was found that the polymorphism detection rate was only 18% for the Taql site but 45% each for the Ddel and Hhal sites, in the factor IX gene. This indicates a low effectiveness of the Taql restriction site in carrier analysis of hemophilia B families in our population. Am. J. Hematol. 54:271–275, 1997 © 1997 Wiley-Liss, Inc.     

Polymorphism of normal factor IX detected by mouse monoclonal antibodies.

Hemophilia B is an X-chromosomal recessive disease due to deficiency of coagulation factor IX. Three monoclonal antibodies against factor IX were prepared and used to develop immunoradiometric assays (IRMAs) of factor IX antigen (IX-Ag). IX-Ag was measured in 65 normal individuals with one IRMA based on polyclonal anti-IX antibodies and two IRMAs based on three monoclonal anti-IX antibodies. One of the monoclonal antibodies differed in specificity since it neutralized less than 50% of the clotting activity of factor IX (IX-C), whereas the other two monoclonal antibodies neutralized 80-95%. When the former antibody was used as the solid phase in IRMA, two groups of normal individuals were distinguished: group A with measurable IX-Ag, and group B without demonstrable IX-Ag. There were no differences between the groups either in IX-C or in IX-Ag measured with polyclonal antibodies. A subgroup comprising only women could be distinguished in group A, in whom intermediate IX-Ag concentrations were found. Family studies showed the group B variant of normal factor IX to be transmitted according to the pattern of X-linked recessive inheritance. The allelic frequency of group A was 0.66, and that of group B was 0.34.     

Cost-effectiveness in establishing hemophilia carrier detection and prenatal diagnosis services in a developing country with limited health resources

The cost-effectiveness of carrier detection and prenatal diagnosis for hemophilia at the International Hemophilia Training Center, Bangkok, Thailand was studied. From 1991 to 2002, 209 females from 124 families with hemophilia A and B were included. There were 180 hemophilia A carriers and 29 hemophilia B carriers which could be classified into 78 obligate and 131 possible carriers. The phenotypic analysis for possible carriers involved the determination of levels of factor VIII or IX clotting activity (FVIII:C, FIX:C) and the ratio of FVIII:C and von Willebrand factor antigen. The result revealed that 49 females (37.4%) were diagnosed as carriers, 65 females (49.6%) were normal and 17 females (13%) were undetermined. Additional genotypic analysis was provided to 46 families with 74 females with obligate, proven or undetermined carriers within the reproductive life. The polymorphisms associated with factor VIII and IX genes were used including Bcl I for the factor VIII gene and combined use of Mse I, Sal I, Nru I, Hha I and Dde I for the factor IX gene. The informative rate was 59.4% (44/74). Consequently, 12 prenatal diagnoses for fetus at risk were performed. Sex determination was initially determined and followed by the diagnosis of hemophilia through informative gene tracking and/or the measurement of fetal levels of FVIII:C or FIX:C. The result revealed that 3 male fetuses were affected. The total cost of carrier detection and prenatal diagnosis that the families had to pay in the government hospital was 238,600 Baht (US dollars 5,965). It was compared to the estimated cost of minimal replacement therapy using lyophilized cryoprecipitate for the survival time of 30 years in one patient with hemophilia of 1,012,500 Baht (US dollars 25,312.5). The cost of prevention was much less than the replacement therapy. In conclusion, it is cost-effective to establish the service for carrier detection and prenatal diagnosis for hemophilia especially in developing countries with limited health resources.     

An Immunological Investigation of Hemophilia B with a Tentative Classification of the Disease into Five Variants1

Abstract. 23 patients with hemophilia B have been investigated by means of several immunological methods. 16 patients (69.9%) had no detectable factor XI antigen. Five had a normal factor IX antigen and the electrophoretic mobility of this abnormal factor IX was similar to that of its normal counterpart. One of these five patients had hemophilia BW, since ox brain thromboplastin clotting time was severely prolonged. The remaining two patients had reduced or decreased factor IX antigen. Several patients showed a slight prolongation of ox brain thromboplastin time due to an associated slight factor VII deficiency. On the basis of these results, a tentative classification of hemophilia B into five variants is proposed, namely: hemophilia B, or with no factor IX antigen; hemophilia Bt, or with normal factor IX antigen; hemophilia BRA, or with reduced factor IX antigen; hemophilia BM, or with normal factor TX antigen and severely prolonged ox brain thromboplastin; hemophilia B, usually B-, with associated mild factor VII defect. A complete evaluation of the hemophilia B patients is feasible only by means of a battery of tests, namely: factor TX activity assay, factor IX antigen determination, ox brain thromboplastin clotting time, factor VII activity assay.     

Defective propeptide processing and abnormal activation underlie the molecular pathology of factor IX Troed-y-Rhiw

Affected members of a South Wales haemophilia B family (from Troed-y-Rhiw) were shown by Western blotting and immunoperoxidase detection to have a factor IX molecule of higher than normal molecular weight which also shows impaired calcium binding. Gene cloning and DNA sequencing revealed the same arginine to glutamine mutation at position -4 of the propeptide that has been found in two previously described factor IX variants which circulate with the propeptide still attached. The mutation also abolishes a HaeIII restriction enzyme recognition site. A potential carrier was shown to be normal both by Western blotting and DNA studies. The way in which the attached propeptide interferes with normal factor IX function was investigated by activation studies with crude normal and patient factor IX Troed-y-Rhiw preparations using Western blotting and detection with iodinated immunopurified polyclonal antifactor IX serum. We demonstrate that the -4 mutation appears to block cleavage between the Arg145-Ala146 peptide bond in the activation peptide, thus preventing the normal activation of factor IX Troed-y-Rhiw. A small amount of normally processed factor IX is produced, implying that the -4 mutation does not completely prevent propeptide cleavage, thus accounting for the low levels of factor IX activity measured in the plasma of affected family members.     

Comparative efficacy of nonheated and heat-treated factor IX complex concentrate in treatment of hemophiliacs with inhibitors

To determine if heat-treated factor IX complex concentrate is as effective as nonheated factor IX complex concentrate for treatment of subjects with hemophilia A and antifactor VIII antibodies (inhibitor patients), we have retrospectively reviewed consecutive home treatment records of ten inhibitor patients who had been receiving nonheat-treated factor IX complex concentrate (NHT-Konyne) and subsequently converted to heat-treated factor complex concentrate (Konyne-HT) when it was licensed in late 1984. Overall, 162 of 284 (57%) separate bleeding episodes treated with NHT-Konyne and 53 of 80 (66.3%) separate bleeding episodes treated with Konyne-HT required only one treatment course of approximately 60-75 U/kg. The distribution of bleeding sites and the absolute factor IX unitage required per treatment episode were similar for both preparations. These data suggest that the percentage of hemophilic inhibitor patients responding to factor IX complex concentrate remains at least 50%, as was reported several years ago in a controlled study, and that inhibitor bypass activity has not altered by heat treatment.     

Carrier detection in the hemophilias

The subject of carrier detection in the hemophilias has received new impetus in the past several years. Treatment complications arising from clotting factor concentrates have become more evident and earlier prenatal diagnosis and new genetic markers for the clotting factor genes have focused interest on this area. Until now, carrier diagnosis has relied upon standard pedigree analysis and clotting factor assays. The results obtained using these methods are probabilistic, and the coagulation tests are unavoidably influenced by the effects of random X chromosome inactivation and the inherent variability of the methods involved. With the cloning and characterization of both factor IX and factor VIII genes, has come the capability of using gene analysis to diagnose the carrier state. This usually involves the detection of restriction fragment length polymorphisms (RFLPs) and their use as linked markers for the defective clotting factor gene. In hemophilia A, the combined use of three intragenic RFLPs and two closely linked, highly polymorphic extragenic markers will make carrier information available to approximately 90% of kindred. In hemophilia B, phenotypic analysis has been complicated by the more heterogeneous expression of the gene defect. To date, five intragenic and one closely linked RFLP have been reported, as well as two protein polymorphisms detectable by monoclonal antibody immunoassays. With the combined use of these genetic markers it is likely that accurate carrier assignment will be available to more than 80% of hemophilia B families.     

Clinical correlates among 49 families with hemophilia A and factor VIII gene inversions

Inversions between a gene A copy within intron 22 of the factor VIII gene and additional copies outside the factor VIII gene were found in 49 families with hemophilia A. Inversion patterns were that of recombination with a distal gene A copy in 34, a proximal copy in 14, and a third (variant) copy in one. Baseline factor VIII clotting activity levels were <1% of normal in 43 and 1% in 6. No inversion was detected in 61 other families whose affected members had ≤1% activity levels nor in 42 families with moderately severe hemophilia A and 2–5% baseline levels. Both high titer and low level alloantibody inhibitors were found in patients with or without an inversion. Of 13 high titer inhibitors, 8 were persistent and 1 of these patients had an inversion. Of 5 that responded to daily factor VIII infusions, 4 were in patients with gene inversions. Of the 49 families with an inversion, the occurrence of hemophilia was isolated in 30 and the mother was a carrier in the 25 in which additional family members were informative. In three of these families with isolated occurrence, the maternal grandmother was a carrier whereas in three others a de novo mutation occurred in the maternal grandfather's factor VIII gene. Screening for gene inversions in patients with severe (or “borderline” severe) hemophilia A provides a direct marker of the mutation in 45% of families. It is useful even if there is no living affected member and in predicting the likely severity of an infant in which there are no reliable baseline clotting activities, including 70% of families with isolated occurrences of hemophilia A. © 1996 Wiley-Liss, Inc.     

No impact of factor IX Ala-10 mutations in acenocoumarol-treated southern Europeans.

Increased risk of bleeding during oral anticoagulant (OA) treatment may be related to mutations in the precursor of coagulation factor IX. Missense mutations at Ala-10 (Ala-10Thre and Ala-10Val) in the factor IX propeptide lead to impaired carboxylation of factor IX. When patients carrying these mutations are treated with coumarins, functional factor IX levels decrease significantly, leading to an excessively prolonged activated partial thromboplastin time (aPTT) and an increased bleeding risk. Mutations at Ala-10 have been described in northern-European patients, but it is not known whether geographical differences play a role in the prevalence of these mutations. We aimed to analyze the prevalence of mutations at Ala-10 in factor IX in southern-European patients on OA treatment. Patients attending the Oral Anticoagulant Unit of the Hospital Clinic were prospectively included. The aPTT was determined at their normal International Normalized Ratio (INR) control. When the aPTT was excessively prolonged for the INR value, determination of factor IX and genotyping for Ala-10 mutations was performed. A total of 2360 patients were included (1289 men, 1071 women; mean age, 70.5 +/- 12.1 years). Twenty-four patients (16 men, eight women; mean age, 61.0 +/- 16.2 years) had an aPTT over that expected for the INR. The mean aPTT was 69.6 +/- 16.2 s. Only one patient presented with a factor IX level lower than 10%. None of the 24 patients carried mutations at Ala-10. Mutations at Ala-10 in factor IX were non-existent in the southern-European population analyzed, and thus, do not represent a relevant cause of bleeding during OA treatment.     

Characterization of a factor IX variant with a glycine207 to glutamic acid mutation

Factor IXTaipei9 is a factor IX variant from a hemophilia B patient with reduced levels of circulating protein molecules (cross-reacting material reduced, CRM). This variant contained a glycine (Gly) to glutamic acid (Glu) substitution at the 207th codon of mature factor IX. The functional consequences of the Gly-->Glu mutation in factor IXTaipei9 (IXG207E) were characterized in this study. Plasma-derived IXG207E exhibited a mobility similar to that of normal factor IX on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Its specific activity was estimated to be 3.5% that of the purified normal factor IX in a one-stage partial thromboplastin time assay (aPTT). Cleavage of factor IXG207E by factor XIa or factor VIIa-tissue factor complex appeared to be normal. When the calcium-dependent conformational change was examined by monitoring quenching of intrinsic fluorescence, both normal factor IX and IXG207E exhibited equivalent intrinsic fluorescence quenching. Activated factor IXG207E (IXaG207E) also binds antithrombin III equally as well as normal factor IXa. However, aberrant binding of the active site probe p-aminobenzamidine was observed for factor XIa-activated factor IXG207E, indicating that the active site pocket of the heavy chain of factor IXaG207E was abnormal. Moreover, the rate of activation of factor X by factor IXaG207E, as measured in a purified system using chromogenic substrates, was estimated to be 1/40 of that of normal factor IXa. A computer-modeled heavy-chain structure of factor IXa predicts a hydrophobic environment surrounding Gly-207 and this Gly forms a hydrogen bound to the active site serine-365. The molecular mechanism of the Gly-->Glu mutation in factor IXTaipei9 might result in the alteration of the microenvironment of the active site pocket which renders the active site serine-365 inaccessible to its substrate.     

Prenatal diagnosis of hemophilia B by an immunoradiometric assay of factor IX

An immunoradiometric assay of factor IX was developed based on homologous antibodies that arose in a hemophilic patient. With this assay, 11 of 12 patients with severe hemophilia B had factor IX antigen levels below 1 U/dl and 6 patients with mild hemophilia B had various levels. Factor IX antigen in 8 fetuses (16th-20th gestational week) aborted for therapeutic reasons ranged from 1.8 to 10.0 U/dl. Six amniotic fluids contained 0.28-1.2 U/dl factor IX antigen. Using the immunoradiometric assay, we could diagnose hemophilia B prenatally in one fetus at risk. No factor IX antigen (< 0.2 U/dl) was detectable in the fetoscopic sample. After termination of the pregnancy, analysis of blood from the abortus confirmed the diagnosis of severe hemophilia B. We conclude that very sensitive immunologic assays, such as the one described here, will prove useful in prenatal diagnosis of severe hemophilia B, since determination of factor IX activity in fetoscopic samples is unrealiable because of possible contamination with thromboplastic material.     

Monoclonal antibody to an epitope on the heavy chain of factor IX missing in three hemophilia-B patients

A murine hybridoma cell line that produces a monoclonal IgG1 antibody to human factor IX was established to provide a conformational probe for the clotting factor and its genetic variants. The antibody inhibited factor IX procoagulant activity, but did not appreciably interfere with the cleavage of factor IX by factor XIa nor with the binding of antithrombin-III-heparin complex to factor IXa. The antigen- solid-phase-antibody complex could be readily dissociated by relatively low concentrations of guanidine or sodium dodecyl sulfate, but only partially by high concentrations of urea. After gel electrophoresis and blotting of reduced samples of factor IXa, the antibody bound exclusively to the heavy chain. Sensitive immunoradiometric assays were developed using insolubilized monoclonal or polyclonal antibodies. Bovine factor IX had little cross-reactivity with the monoclonal antibody. Of 55 patient samples representing different pedigrees with hemophilia-B, antigen levels by the two assays were in excellent agreement in 49. There were 2 severely affected patients whose levels were too low to quantitate in the monoclonal antibody assay. A third, who had the lowest level of all by polyclonal antibody testing, and 3 less severely affected patients had no detectable antigen in the monoclonal antibody assay system (less than 0.03 U/dl). The latter 3 had at least 100-500 times as much antigen by polyclonal antibody testing. It is proposed that these 3 individuals have structural defects involving the epitope recognized by the monoclonal antibody and that they are due to amino acid substitutions between residues 188 through 359. Furthermore, it is suggested the substitutions lead to abnormal kinetic properties.     

Factor IX New London: substitution of proline for glutamine at position 50 causes severe hemophilia B

We describe a novel point mutation in the fourth exon of human factor IX (encoding the first EGF-like domain) in which cytosine is substituted for adenosine at position 10,401, resulting in the substitution of proline for glutamine at position 50 in the polypeptide chain. Sequence analysis of all eight exons, all exon-intron junctions, 160 base pairs (bp) of DNA 5' to the proposed translation start site, and 60 bp 3' to the translation termination site shows no other difference from the normal factor IX gene, with the exception of a previously described benign polymorphism at position 148 in the protein (Ala----Thr). The affected subject has severe hemophilia B with no detectable factor IX activity despite normal factor IX antigen levels. We purified the abnormal factor IX by immunoaffinity chromatography and demonstrated that its activation by factor Xla is markedly delayed compared with normal factor lX. Once activated, the abnormal factor lX binds antithrombin III in a 1:1 molar ratio, and the activated protein demonstrates catalytic activity, suggesting an intact active site. The mutation creates a new Bst Yl restriction endonuclease cleavage site. Restriction with Bst Yl shows the mutation in maternal DNA and offers the possibility of direct carrier status analysis and prenatal diagnosis in kindreds with this mutation. We designate this new mutation factor lXNew London. This is the only reported mutation in the first EGF-like domain that causes severe hemophilia B.     

Comparison of the behavior of normal factor IX and the factor IX BM variant hilo in the prothrombin time test using tissue factors from bovine,human, and rabbit sources

A subset of hemophilia B patients have a prolonged bovine-brain prothrombin time. These CRM⁺ patients are classified as having hemophilia Bm. The prolongation of the prothrombin time has been reported only with bovine brain (referred to as ox brain in some literature) as the source of thromboplastin; prothrombin times determined with thromboplastin from rabbit brain or human brain are not reported to be prolonged. Factor IX from a hemophilia Bm patient (factor IX Hilo) was isolated. The activity of factor IX Hilo was compared to that of normal factor IX in prothrombin time assays when the thromboplastin source was of bovine, rabbit, or human origin. Factor IX, either normal or Hilo, prolonged a prothrombin time regardless of the tissue factor source. However, unless thromboplastin was from a bovine source, this prolongation required high concentrations of factor IX. Further, factor IX normal was as effective as factor IX Hilo in prolonging the prothrombin time when rabbit or human thromboplastin was used. With bovine thromboplastin, factor IX Hilo was significantly better than factor IX normal at prolonging the prothrombin time. The amount of prolongation was dependent on the amount of factor IX Hilo added. In addition, the prolongation was dependent on the concentration of factor × present in the sample. The prothrombin time changed as much as 20 seconds when the factor × concentration was varied from 50% to 150% to normal (fixed concentration of factor IX Hilo). These results demonstrate the difficulty of classifying the severity of a hemophilia Bm patient based on the bovine brain prothrombin time unless both the factor IX and factor × concentrations are known.     

Diagnostic utility of a mouse monoclonal antibody (5-21-3) employed as a competitive probe in an immunoassay to detect antibody to HIV-1 gp41.

A mouse monoclonal antibody, designated 5-21-3, was raised against HIV-1 gp41 using detergent-disrupted virus as the immunogen. Antibody 5-21-3 was conjugated to horseradish peroxidase (HRP) and employed as a competitive probe against normal and HIV-1 antibody-positive sera in an immunoassay to detect the presence of antibody to HIV-1 gp41. The diagnostic utility of the competitive monoclonal immunoassay was assessed by correlation to a similar assay which employed HRP-labeled polyclonal IgG from a gp41-seropositive donor as the competitive probe. The monoclonal immunoassay was greater than 98% as sensitive and 99% as specific as the polyclonal immunoassay, regardless of the geographic source or disease state of the donor. The monoclonal immunoassay also was nearly as effective as the polyclonal immunoassay in detecting points of seroconversion in individuals enrolled in longitudinal studies. Of particular interest was the finding that the epitope recognized by monoclonal antibody 5-21-3 did not map to the well-characterized gp41 immunodominant region.     

Factor IX variants improve gene therapy efficacy for hemophilia B

Intramuscular injection of adeno-associated viral (AAV) vector to skeletal muscle of humans with hemophilia B is safe, but higher doses are required to achieve therapeutic factor IX (F.IX) levels. The efficacy of this approach is hampered by the retention of F.IX in muscle extracellular spaces and by the limiting capacity of muscle to synthesize fully active F.IX at high expression rates. To overcome these limitations, we constructed AAV vectors encoding F.IX variants for muscle- or liver-directed expression in hemophilia B mice. Circulating F.IX levels following intramuscular injection of AAV-F.IX-K5A/V10K, a variant with low-affinity to extracellular matrix, were 2-5 fold higher compared with wild-type (WT) F.IX, while the protein-specific activities remained similar. Expression of F.IX-R338A generated a protein with 2- or 6-fold higher specific activity than F.IX-WT following vector delivery to skeletal muscle or liver, respectively. F.IX-WT and variant forms provide effective hemostasis in vivo upon challenge by tail-clipping assay. Importantly, intramuscular injection of AAV-F.IX variants did not trigger antibody formation to F.IX in mice tolerant to F.IX-WT. These studies demonstrate that F.IX variants provide a promising strategy to improve the efficacy for a variety of gene-based therapies for hemophilia B.     

Immunochemical characterization of a polyclonal human antibody to factor IX.

Inhibitors of clotting factors occuring in humans are often antibody molecules synthesized in response to exogeneous proteins used in replacement therapy. Extensive studies of inhibitors to factor VIII indicate such antibodies may be monoclonal or polyclonal in nature. To date, only one factor IX inhibitor has been subjected to detailed immunochemical analysis and it appears to be a monoclonal IgGA lambda antibody. We have discovered a second inhibitor of factor IX in a patient with severe hemophilia B and have subjected it to immunochemical analysis. Studies on this second inhibitor have been carried out before and after an anamnestic response. Column chromatography, preparative zone electrophoresis, and specific inhibitor neutralization assays using monospecific heterologous antisera to human immunoglobulin classes, subclasses, and light-chain types indicate that the antibody is of the IgG class and contains both kappa and lambda light chains and probably all four IgG subclasses. Thus, the inhibitor appears to be polyclonal by immunochemical and structural criteria. In addition, preparative isoelectric focusing of pre- and postanamnestic inhibitor samples indicates that recruitment of new clones of IgG antibody occurs as a result of anamnesis. It is conceivable that an antibody initially restricted in immunoglobulin subclass became polyclonal following an anamnestic response.