Immunological Characterization of Factor VIII Autoantibodies in Patients with Acquired Hemophilia A in the Presence or Absence of Underlying Disease

Thrombosis Research 104 (2001) 381 – 388

REGULAR ARTICLE

Immunological Characterization of Factor VIII Autoantibodies in Patients with Acquired Hemophilia A in the Presence or Absence of Underlying Disease

Tomoko Matsumoto1, Midori Shima1, Kazuyoshi Fukuda1, Keiji Nogami1, John C. Giddings2, Tomohiko Murakami1, Ichiro Tanaka1 and Akira Yoshioka1 1 Department of Pediatrics, Nara Medical University, Kashihara, Japan and 2Department of Haematology, University of Wales College of Medicine, Cardiff, UK (Received 14 June 2001 by Editor A. Takada; revised/accepted 28 September 2001)

Abstract The development of a factor VIII autoantibody results in a severe hemorrhagic diathesis known as acquired hemophilia A. Underlying pathologies, such as autoimmune disease or chronic inflammatory disease, are observed in about half of the patients. We have investigated a total of 16 cases with acquired hemophilia A and divided the patients into two groups according to the presence or absence of other clinical conditions. Group A comprised nine cases with no detectable associated pathology. Group B consisted of seven cases with other clinical diagnoses. Significant levels of factor VIII activity (FVIII:C) and factor VIII antigen (FVIII:Ag) were detected in Group A and the pattern of FVIII:C inactivation was characteristic of Type 2 inhibitors. In contrast, no FVIII:C was detected in Group B and, in five of seven cases, the inhibitory pattern was Type 1. IgG4 antibody subclass specificity was dominant in both groups. IgG1 antibody reactivity was higher in Group B than in Group A. Our results suggested a close relationship between the presence of underlying disease and immunoCorresponding author: Midori Shima, MD, PhD, Department of Pediatrics, Nara Medical University, 840 Shijo-cho, Kashihara, Nara 634-8522, Japan. Tel: +81 (744) 29 8881x3431; Fax: +81 (744) 24 9222; E-mail: .

logical and coagulation characteristics in acquired hemophilia A. D 2001 Elsevier Science Ltd. All rights reserved. Key Words: Acquired hemophilia; Autoantibody; Factor VIII activity; Factor VIII antigen

A

cquired hemophilia A presents acutely as a severe hemorrhagic diathesis in nonhemophiliac patients and is secondary to the development of inhibitory autoantibodies to factor VIII [1]. Underlying diseases, including autoimmune and chronic inflammatory diseases, complications of pregnancy and childbirth, or concurrent medication, are observed in approximately half of the patients. In addition, the incidence of acquired hemophilia A appears to be higher in elderly individuals. The hemorrhagic defect is comparable to that of severe congenital hemophilia A and can be intractable and progressive such that the presence of the autoantibody can lead to a broad range of symptoms such as large ecchymoses, intramuscular hematomas, and intracranial bleeding and may even be life-threatening. The immunological and biochemical characteristics of factor VIII autoantibodies in acquired hemophilia A are very similar to the inhibitory alloantibodies to factor VIII that may arise as a result of treatment in patients with severe congenital hemophilia A. Thus, the antibodies are

0049-3848/01/$ – see front matter D 2001 Elsevier Science Ltd. All rights reserved. PII S0049-3848(01)00385-1

382

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

mainly IgG of the IgG4 subclass, while the main factor VIII binding site is present in the A2 domain of the heavy chain and the C2 domain of the light chain [2]. Factor VIII alloantibodies have been classified mostly as Type 1, in that factor VIII is completely inactivated in the presence of excess inhibitor. In contrast, factor VIII autoantibodies appear to be mainly Type 2, and factor VIII activity persists independently of antibody concentration [3]. In addition, several reports have shown that most factor VIII alloantibodies recognize both the A2 and C2 domains of factor VIII, whereas a number of factor VIII autoantibodies that recognized the C2 domain alone have been identified [4]. The mechanisms whereby factor VIII autoantibodies develop remain unclear. In the present study, we have investigated a total of 16 cases of acquired hemophilia A and have classified the patients into two groups depending upon the presence of absence of underlying pathology. We have examined the immunological and coagulation properties of the factor VIII autoantibodies in an attempt to identify pathogenetic characteristics.

1. Materials and Methods 1.1. Patients A total of 16 subjects with acquired hemophilia A were studied. Group A consisted of nine cases with no detectable underlying disease (six men,

three women; mean age 52.3 years; range 28–79 years). Group B comprised seven cases with a variety of associated disorders (four men, three women; mean age 51.9 years; range 22 – 78 years). The patient characteristics are shown in Tables 1 and 2. In Group B, Cases 10 and 15 were postpartum. Case 11 was undergoing oral treatment with the psychotropic agent Depixol, for schizophrenia. Case 12 had diabetes mellitus and Case 13 was being treated for both hypertension and a gastric ulcer. Case 14 had diabetes mellitus and a bullous pemphigoid, while case 16 had chronic idiopathic thrombocytopenic purpura (ITP). All plasma samples were obtained with informed consent following local ethical guidelines. 1.2. Measurement of Factor VIII Factor VIII activity (FVIII:C) was measured using a one-stage coagulation method with factor VIII-deficient plasma as the substrate [5]. Factor VIII antigen (FVIII:Ag) was measured by a sandwich enzyme-linked immunosorbent assay (ELISA) [6]. Briefly, serial dilutions of test plasma were added to microtitre wells coated with an antifactor VIII alloantibody purified from the plasma of a severe hemophilia A patient with a high titer factor VIII inhibitor. After washing, peroxidase-conjugated fractions of the same human alloantibody were added, followed by o-phenylenediamine diluted in 0.1 M phosphate–citrate buffer (pH 5.0) at a final concentration of 0.04% (w/v) and containing H2O2 at 0.04% (v/v).

Table 1. Profiles of autoantibodies to factor VIII for Group A Case number

1 2 3 4 5 6 7 8 9 * Negative.

Age

35 30 72 78 69 29 51 28 79

Sex

M M M F M F F M M

FVIII:C (U/dl)

1.2 1.6 1.4 1.0 1.2 8.0 2.3 2.0 2.3

FVIII:Ag (U/dl)

1.0 2.0 2.8 3.8 10.4 24.0 54.0 90.0 100.0

Titer (BU/ml)

33.0 2.6 860.0 570.0 7.7 2.0 5.2 14.5 3.5

Specificity Immunoblot

C2 neutralization (%)

L * H, L L * * * * *

93 50 76 80 > 95 48 5 10 9

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

383

Table 2. Profiles of autoantibodies to factor VIII for Group B Case

Age Sex FVIII:C FVIII:Ag

number

(U/dl)

(U/dl)

Titer

Specificity

Underlying diseases

(BU/ml) Immunoblot C2 neutralization (%)

10 11 12

39 45 71

F F M

<1 <1 <1

<1 <1 <1

309.0 68.0 6.8

H, L L L

54 >95 77

13

68

M

<1

<1

145.0

L

>95

14

78

M

<1

1.2

125.0

H, L

83

15 16

22 40

F M

<1 <1

<1 1.8

34.0 3.0

H, L *

43 >95

Postpartum Schizophrenia Diabetes mellitus (nephropathy and retinopathy) Hypertension, anemia and gastric ulcer Diabetes mellitus, pemphigoid and cerebral infraction Postpartum Chronic ITP

* Negative.

1.3. Measurement of Factor VIII Inhibitor Activity Factor VIII inhibitor activity was measured using a one-stage coagulation method based on the method of Kasper et al. [7]. The results were recorded as Bethesda units (BU)/ml. In order to assess concentration-dependent patterns of FVIII:C inhibition, plasma samples were adjusted initially to 1–2 BU/ml and then serially diluted and mixed with normal plasma at a ratio of 1:1. Residual FVIII:C was measured after incubation at 37 °C for 2 h. 1.4. Measurement of IgG Subclasses of Antibodies to Factor VIII Specific IgG subclasses of the antibodies were determined by ELISA. In outline, purified factor VIII was immobilized to ELISA plates. Following incubation with the antibody-containing plasma, the plates were washed, blocked with human albumin and incubated with peroxidase-conjugated mouse monoclonal antibodies specific for individual human IgG subclasses (Binding Site, Birmingham, UK). Chromogenic substrate (Peroxidase System TMB:KPL, Gaithersburg, MD) was added and colorimetric measurements recorded at 450 nm [8]. Plasma samples from 20 normal healthy subjects were used as controls; the average optical density (OD ± 2 S.D.) for controls was 0.067 ± 0.015 for IgG, 0.052 ± 0.012 for IgG1, 0.046 ± 0.009 for IgG2, 0.045 ± 0.008 for IgG3, 0.047 ± 0.011 for IgG4 and 0.045 ± 0.010 for IgM. Test samples were consid-

ered positive when the OD was above the control average OD + 3 S.D., and also when the addition of 2.5 U of factor VIII blocked the reaction by 50% or more. 1.5. Immunoblot Analysis The binding of specific factor VIII antibodies to factor VIII fragments was analyzed by Western blotting. Purified factor VIII and thrombincleaved factor VIII (final thrombin concentration 1 U/ml) were transferred to nitrocellulose film (Bio Rad, Hercules, CA) after electrophoresis in 7.5% polyacrylamide gel. Nonspecific binding was blocked with 5% skimmed milk, and the patient plasma was added followed by mouse monoclonal antibody to subclass-specific human

Fig. 1. Immunoblot analysis of Group A autoantibodies. In each panel, the left lane demonstrates reactivity with purified factor VIII (FVIII). The right lane illustrates reactivity with thrombin-catalyzed factor VIII (Th).

384

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

2. Results 2.1. Profiles of Autoantibodies to Factor VIII in Group A

Fig. 2. Dose – response relationship for the activity of autoantibodies to factor VIII in Group A. (^) Case 1 (1:10). (&) Case 2. (&&) Case 3 (1:1000). (
) Case 4 (1:100). (Xl) Case 5. ( ) Case 6. (4) Case 7. (r) Case 8. (+) Case 9.

.

IgG. Bound antibody was detected using a chemiluminescent substrate (Western Blot Chemiluminescence Reagent; NEN Life Science Products, Boston, MA) [9]. 1.6. C2 Neutralization Experiments Inhibitor plasma was adjusted to 2–3 BU/ml and mixed with chemically synthesized C2 peptides [10]. The mixture was incubated at 37 °C for 2 h after which an equal volume of normal plasma was added. Residual FVIII:C was measured following a further incubation at 37 °C for 2 h. Neutralization of antibody by C2 fragments was considered positive when residual FVIII:C activity was 20% higher than that in samples to which no C2 was added.

The average inhibitor activity of factor VIII autoantibodies in Group A was 183 BU/ml (range 2–860 BU/ml). There were five cases that could be regarded as ‘‘low responders’’ with values at or below 10 BU/ml (mean 4.2 BU/ ml) and four cases that could be considered as ‘‘high responders’’ with values at or above 10 BU/ml (mean 369 BU/ml; Table 1). In this group, FVIII:C was detected in all nine cases with a mean value of 2.3 U/dl (range 1–8 U/dl). Similarly, FVIII:Ag was detected in all cases with a mean value of 32 U/dl (range 1–100 U/ dl). Cases 5–9 had markedly higher values of detectable FVIII:Ag in relation to the respective level of FVIII:C (Table 1). In particular, the levels of FVIII:Ag in Cases 7, 8 and 9 were within the normal range, corresponding to the well-documented, cross-reacting, material-positive (CRM + ) subgroup of congenital hemophilia A. Analysis of the antibody binding site on the factor VIII molecule by immunoblotting was possible in Cases 1, 3 and 4, which exhibited high titer inhibitors. Case 3 illustrated multiple bands against purified factor VIII, reflecting reactivity with both the heavy chain (92–200 kDa) and the light chain (80 kDa) (Fig. 1). In experiments utilizing thrombin-treated factor VIII, there was reactivity to both the 72-kDa fragment of the light chain and the 44-kDa fragment of the heavy chain. Cases 1 and 4 exhibited reactivity to the light chain of purified factor VIII (80 kDa)

Fig. 3. Immunoblot analysis of Group B autoantibodies. As for Fig. 1, in each panel, the left lane demonstrates reactivity with purified factor VIII (FVIII). The right lane illustrates reactivity with thrombin-catalyzed factor VIII (Th).

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

385

pattern was characteristic of Type 2 inhibitors (Fig. 2). 2.2. Profiles of Autoantibodies to Factor VIII in Group B

Fig. 4. Dose – response relationship for the activity of autoantibodies to factor VIII in Group B. (^) Case 10 (1:10). (&) Case 11 (1:10). (4) Case 12. (
) Case 13 (1:10). (Xl) Case 14 (1:100). (+) Case 15. ( ) Case 16.

.

and thrombin-catalyzed factor VIII (72 kDa) but did not react with the heavy chain. Neutralization of inhibitor activity by C2 peptides was observed in six cases in this group, but was negative in Cases 7, 8 and 9 (Table 1). In experiments to examine the kinetics of FVIII:C inhibition, residual FVIII:C was detected in the presence of excess antibody in all cases. This

In Group B, the mean inhibitor activity of the factor VIII autoantibodies was 98.7 BU/ml (range 3–309 BU/ml). There were two low responders and five high responders. FVIII:C was not detected ( < 1%) in any of the seven cases. FVIII:Ag was below the level of detection in five cases although trace amounts were found at 1.2% in Case 14 and 1.8% in Case 16. Analysis of binding sites by immunoblotting was possible in all subjects except Case 16. Both heavy and light chain binding sites were identified in three cases and light chain sites alone (72 kDa) were detected in three cases (Fig. 3). C2 peptides neutralized inhibitor activity in all seven cases (Table 2). In the experiments to examine inhibitor kinetics, there was a linear relationship between antibody concentration and residual FVIII:C activity. In the presence of undiluted inhibitor plasma, the residual FVIII:C was below the detection threshold in five of seven cases. These results were consistent with a Type 1 inhibitor pattern (Fig. 4).

Table 3. Antifactor VIII IgG, IgM and IgG subclass measured by ELISA in the 16 acquired hemophilia A patients Group

Case

FVIII:C

Immunoglobulin class

number

inhibitor (BU/ml)

IgG

IgM

IgG1

IgG2

IgG3

IgG4

A

1 2 3 4 5 6 7 8 9

33.0 2.6 860.0 570.0 7.7 2.0 5.2 14.5 3.5

0.212 0.144 1.036 0.480 0.532 * 0.108 * 0.326

* * * * * * * * *

0.248 * * * 0.774 * * * *

* * * 0.136 * * * * *

* * * * * * * * *

0.164 1.256 2.510 2.890 0.800 * 1.378 * 2.826

B

10 11 12 13 14 15 16

309.0 68.0 6.8 145.0 125.0 34.0 3.0

1.272 0.478 0.574 0.892 1.210 1.385 0.422

* * * * * * *

* * 0.696 0.934 1.666 * 0.336

* * * 0.100 0.118 0.100 *

* * * * * * *

3.248 2.306 0.650 2.228 2.126 2.042 *

* Indicates that each antifactor VIII immunoglobulin or IgG subclass was negative.

IgG subclass

386

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

2.3. IgG Analysis of Autoantibodies to Factor VIII Negative ELISA results for both IgG and IgG subclass antibodies were found in Cases 6 and 8 in Group A (Table 3). The remaining seven cases in this group were all positive for IgG. High levels of IgG4 antibody were detected in all these cases, whilst IgG3 antibody was consistently negative. IgG1 antibody was found in two cases and IgG2 antibody was present in one case. IgG antibody was found in all seven cases in Group B. High levels of IgG4 were observed in six cases, whilst IgG3 was negative in all cases. IgG1 antibody was demonstrated in four cases and IgG2 antibody was seen in three cases.

3. Discussion The development of factor VIII autoantibodies in nonhemophilic patients can be associated with a variety of underlying diseases including autoimmune diseases, malignant tumors and medication allergies. For example, analysis of 215 cases of acquired hemophilia revealed chronic rheumatoid arthritis in 7.9%, postpartum in 7.3%, malignant tumors in 6.7%, systemic lupus erythematosus (SLE) in 5.6%, dermatoses in 4.5%, bronchial asthma in 3.9% and posttransfusion complications in 2.8%. The authors noted, however, that 46.1% of cases exhibited no evidence of underlying pathology [11]. In the present investigation, therefore, we focused on the coagulation and immunological profiles of inhibitors in acquired hemophilia A in the presence or absence of underlying diseases. Many previous studies have identified Type 2 inhibitors in patients with acquired hemophilia where FVIII:C persists in circulating blood in the presence of factor VIII autoantibody. In the current study, this pattern was prominent in patients in Group A. FVIII:C and FVIII:Ag were detected in all cases in this group. In addition, levels of FVIII:Ag were higher than those of FVIII:C except in Case 1. These findings suggest that circulating immune complexes might have retained factor VIII activity and that factor VIII clearance might have been suppressed in these instances. Cases 7, 8 and 9 especially exhibited

normal levels of factor VIII antigen. In this context, Nogami et al. [12] described the presence of circulating factor VIII immune complexes in similar patients and showed that the autoantibodies interfered with activated protein C binding and cleavage of factor VIII. In contrast, FVIII:C was not detected in the majority of cases in the current Group B and the antibodies reflected a Type 1 inhibitor pattern. Hoyer et al. [13] suggested that antibody epitopes governed the differences between Types 1 and 2 inhibitors. In our study, plasma from only three patients, Cases 1, 3 and 4, in Group A reacted with purified factor VIII and thrombin-catalyzed factor VIII in immunoblotting tests whilst samples from the other six individuals were unreactive. In contrast, plasma from six of seven patients in Group B reacted with factor VIII on immunoblotting. In particular, these immunoblotting experiments in Cases 7, 8 and 9 from Group A revealed no reaction to the light chain (80 and 72 kDa). Furthermore, inhibitor neutralization experiments confirmed that the antibodies from Cases 7, 8 and 9 were not inactivated by C2 domain. Unexpectedly, antibodies from Cases 2, 5, 6 and 16 failed to demonstrate reactions in the immunoblotting experiments although positive results for the C2 epitope were demonstrated by inhibitor neutralization. The reasons for the lack of immunoblot reactions in these latter instances remain to be determined but may be related to the precise structural conformation of the factor VIII immune complexes. Nevertheless, our results confirmed that antibodies recognizing different factor VIII epitopes may exhibit different functional characteristics and the findings show that assessment of both inhibitory activity and immunobinding reactions provides useful data for the characterization of factor VIII autoantibodies. Previous purification studies and kinetic data using specific antiimmunoglobulin antisera have shown that both factor VIII alloantibodies and factor VIII autoantibodies may be oligoclonal in origin, with a mixed composition of IgG subtypes, incorporating IgG1, IgG2, IgG4 and, more rarely, IgG3. Most of the reported inhibitors, however, have demonstrated a restricted heterogeneity of immunoglobulin composition, and IgG4 appears to be the dom-

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

inant subtype [14]. This property might be associated with a lack of complement fixation and immunoprecipitation, and could account for the absence of reported cases of renal or vascular involvement in patients with high levels of IgG4 autoantibodies. The current data were in keeping with the earlier reports and demonstrated a high incidence of IgG4 antibody subclass regardless of the presence or absence of coexisting pathology. Two of our cases (Cases 10 and 15) in particular developed high titer IgG4 antibodies postpartum, but despite the high levels of inhibitor (309 and 34 BU/ml, respectively), the clinical outcome was relatively uneventful. As well as the frequent detection of IgG4 antibody in Group B, there appeared to be a greater degree of heterogeneity amongst these patients. IgG1 was detected in four of seven cases, and IgG2 was found in three individuals in Group B. In comparison, only three of nine cases in Group A demonstrated IgG1 or IgG2 antibodies. It is possible, therefore, that the characteristics of acquired hemophilia in group A reflected a greater degree of immunological recruitment and this difference could influence the clinical effectiveness of immunosuppressive therapy in patients with or without underlying disease. Reports of acquired hemophilia have been increasing in recent years, and some cases have presented with intractable complications. Early diagnosis and the establishment of treatment strategies are vital for effective clinical management. Protocols to suppress antibody activity could depend on the characteristics of the factor VIII antibody, and detailed laboratory investigations, including inhibitor kinetics and factor VIII binding properties, could provide valuable diagnostic data in this respect. For example, the management of hemostasis by coagulation factor replacement therapy might prove to be especially difficult in patients with Type 2 inhibitors, but since the immunological response is relatively weak, immunosuppression of antibody activity might be relatively successful. Further studies on the relationship between the nature of factor VIII autoantibodies and clinical outcomes could offer challenging incentives for the management of these difficult cases.

387

References 1. Shapiro SS, Hultin M. Acquired inhibitors to the blood coagulation factors. Semin Thromb Hemostasis 1975;1:336–42. 2. Scandella D, Mahoney SdeG, Mattingly M, Roeder D, Timmons L, Fulcher CA. Epitope mapping of human factor VIII inhibitor antibodies by deletion analysis of factor VIII fragments expressed in Escherichia coli. Proc Natl Acad Sci USA 1988;85:6152–6. 3. Biggs R, Austen DEG, Densen KWE, Borrett R, Rizza CR. The mode of action of antibodies which destroy factor VIII: I. Antibodies which have second-order concentration graphs. Br J Haematol 1972;23:125–35. 4. Prescott R, Nakai H, Saenko EL, Scharrer I, Nilsson IM, Humphries JE, Hurst D, Bray G, Scandella D. The inhibitor antibody response is more complex in hemophilia A patients than in most nonhemophiliacs with factor VIII autoantibodies. Blood 1997;89: 3663–71. 5. Hardisty RM, Macpherson JC. A one-stage factor VIII (antihaemophilic globulin) assay and its use on venous and capillary plasma. Thromb Diath Haemorrh 1962;7:215–28. 6. Kamisue S, Shima M, Nishimura T, Tanaka I, Nakai H, Morichika S, Takata N, Kuramoto A, Yoshioka A. Abnormal factor VIII Hiroshima: defect in crucial proteolytic cleavage by thrombin at Arg1689 detected by a novel ELISA. Br J Haematol 1994;86:106–11. 7. Kasper CK, Aledort LM, Counts RB, Edson, JR, Fratantoni J, Green D, Hampton JW, Hilgartner MW, Lazerson J, Levine PH, McMillan CW, Pool JG, Shapiro SS, Shulman NR, Van Eyes J. A more uniform measurement of factor VIII inhibitors. Thromb Diath Haemorrh 1975;34:869–72. 8. Shima M, Sawamoto Y, Nakai H, Kamisue S, Murakami T, Morichika S, Tanaka I, Hayashi K, Yoshioka A. Kogenate Japanese clinical study group. Measurement of anti-factor VIII IgG, IgG4 and IgM alloantibodies in previously untreated hemophilia A patients treated with recombinant factor VIII. Int J Hematol 1995;62:35–43.

388

T. Matsumoto et al./Thrombosis Research 104 (2001) 381–388

9. Shima M, Scandella D, Yoshioka A, Nakaiet H, Tanaka I, Kamisue S, Terada S, Fukui H. A factor VIII neutralizing monoclonal antibody and a human inhibitor alloantibody recognizing epitopes in the C2 domain inhibit factor VIII binding to von Willebrand factor and to phosphatidylserine. Thromb Haemostasis 1993;69:240–6. 10. Saenko EL, Scandella D. A mechanism for inhibition of factor VIII binding to phospholipid by von Willebrand factor. J Biol Chem 1995;270:13826–33. 11. Green D, Lechner K. A survey of 215 nonhemophiliac patients with inhibitors to factor VIII. Thromb Haemostasis 1981;45:200–3. 12. Nogami K, Shima M, Giddings JC, Hosoka-

wa K, Nagata M, Kamisue S, Suzuki H, Shibata M, Saenko EL, Tanaka I, Yoshioka A. Circulating factor VIII immune complexes in patients with type 2 acquired hemophilia A and protection from activated protein C-mediated proteolysis. Blood 2001; 97:669–77. 13. Hoyer LW, Gawryl MS, dela Fuente B. Immunochemical characterization of factor VIII inhibitors. In: Hoyer LW, editor. Factor VIII inhibitors. New York: Alan R. Liss, 1984;73–85. 14. Sawamoto Y. Analysis of IgG subclass in factor VIII inhibitor by enzyme linked immunosorbent assay. J Nara Med Assoc 1992;43: 511–21.