The Role of Immunomodulation in the Management of Factor VIII Inhibitors

With the development of factor VIII (FVIII) concentrates for which the risk of infectious agent transmission is negligible, the principal treatment-related complication in hemophilia A is now the formation of anti-FVIII antibodies (FVIII inhibitors).¹ The occurrence of a FVIII inhibitor in a person with hemophilia significantly influences the clinical care of the patient from two aspects: the requirement of alternative approaches to effect hemostasis through the use of bypassing agents, and the need to initiate immunomodulatory therapy to induce immunological tolerance to FVIII. This review will address recent developments in this latter area of clinical management.

Since their earliest recognition in the 1940s, the pathobiology and treatment of FVIII inhibitors has been an area of intense activity for the scientific and clinical hemophilia communities. The incidence of this adverse immunological event ranges in the literature from 15–50%,^2,–5 thus prompting questions about why such diverse results have been obtained in these various studies. It is clear that a number of variables are likely to have contributed to this wide range of results. These include the frequency of testing for inhibitors, the type of laboratory test being used, the type of FVIII concentrate and its mode of administration, and the mix of patients in the study population.

Best characterized of the patient variables is the FVIII genotype, with a spectrum of inhibitor risks ranging from > 75% for multi-domain deletions through a 20–30% risk with the common intron 22 inversion mutation, to < 10% with missense mutations and small FVIII deletions and insertions.⁶ There are almost certainly immunogenetic influences on the inhibitor risk, but linkage to specific HLA types has been difficult to document.^7,8 Nevertheless, the recently described association with particular IL-10 and TNFα polymorphic alleles confirms the likely complex involvement of genes regulating host immunity.^9,10 Finally, the incidence of inhibitors in people of African-American and Latino ancestry has been documented to be twice that of Caucasian patients,¹¹ again likely due to differences in immunogenetic characteristics.

In terms of FVIII concentrate and product administration variables, rare “outbreaks” of FVIII inhibitors have occurred with concentrates that have likely experienced structural changes during production,^12,13 and there is good evidence to suggest that FVIII delivery into an immunological environment primed by inflammation or tissue trauma will also be more prone to generate an inhibitor response. This type of environment is, of course, more likely to occur at the time of surgery, when many patients will receive continuous high-dose infusions of FVIII, thus raising the question of whether this form of intensive exposure to the protein might be more likely to stimulate inhibitor development.¹⁴ Since the introduction of recombinant FVIII concentrates 15 years ago, there has been ongoing debate about the propensity of these products to elicit FVIII inhibitors. While most studies indicate that these concentrates are no more immunogenic than their plasma-derived predecessors, some evidence continues to question this fact.¹⁵ Related to this debate is the question of whether the presence of von Willebrand factor in FVIII concentrates might be able to mediate an immunoprotective effect¹⁶ or whether a potential benefit might derive from co-purified immunomodulatory proteins in the plasma-derived concentrates.¹⁷ 

Normal, non-hemophilic individuals are naturally tolerant to FVIII because their FVIII-reactive T cells have been deleted during the process of central tolerance induction in the thymus. However, in hemophiliacs, in whom FVIII is absent, T cells with T cell receptors that recognize FVIII peptides may be present. The ability to mount a productive immune response against FVIII relates, in part, to the presence of these FVIII cognate CD4⁺ T cells and the optimal MHC class II receptors to bind to and display FVIII peptides on the surface of antigen-presenting cells (APCs).¹⁸ In hemophiliacs, the FVIII that incites inhibitor development must be internalized and proteolytically processed by APCs before being presented to T cells in lymph nodes and the spleen in the form of immunogenic peptides by class II MHC receptors. The presentation of a FVIII peptide to a cognate T helper cell constitutes “signal 1” of the initial activation process for effector T cells. For the activation process to be successful, “signal 2” must also be present, a phenomenon that requires expression and interaction of various co-stimulatory molecules on the APCs with receptors on the CD4⁺ T cell (e.g., CD80-86 on APCs with CD28 on T cells). In the absence of this interaction between the co-stimulatory molecules and their cognate receptors on T cells, a non-productive immunologic synapse is formed and effector cell activation does not occur. In contrast, this abortive interaction will result in antigen-specific tolerance through either T cell apoptosis or anergy, or through the generation of regulatory (suppressor) T cells. The fact that approximately 75% of hemophiliacs do not develop FVIII inhibitors is presumably influenced by a variety of factors that include a non-optimal HLA genotype for FVIII antigen presentation, the possession of other components of an immunoregulatory genotype (i.e., IL-10 and/or TNFα genotype), the absence of T cells bearing FVIII-specific T cell receptors and/or the presentation of FVIII in a tolerogenic manner that accomplishes long-term peripheral tolerance to the antigen.

Thus, the generation of FVIII inhibitors is a CD4⁺ T cell–dependent process that will subsequently activate FVIII-specific B cells and mediate the formation of FVIII antibody–producing plasma cells. Once an anti-FVIII effector cell response has been generated, long-lived memory T and B cells will also be generated to provide the basis for ongoing immunoreactivity to the antigen. The CD4⁺ T cell dependency of this process was ironically demonstrated by the loss of FVIII inhibitors in some patients with advanced HIV disease.

Studies performed over the past decade have determined that the main T and B cell epitopes on the FVIII protein are located on the C2, A2 and A3 domains.^19,–21 The antibody responses are oligoclonal in nature and comprise both IgG4 and IgG1 isotype responses. For clinical purposes, the magnitude of the antibody response can be quantified through the performance of a functional inhibitor assay from which a Bethesda unit (BU) inhibitor titer can be reported.²² The International Society on Thrombosis and Haemostasis–supported definition of high titer responses is > 5 BUs and low titer between 0.5 and 5 BUs.²³ Although this assay provides a reasonable basis for discriminating between these two clinically distinct types of antibody response, it is well recognized that there is substantial inter-laboratory variability in the performance of the test, some of which, particularly at low titers, has been eliminated by recent assay modifications.^24,25 The inhibitory antibodies present in these patients are thought to interfere with FVIII’s co-factor function through allosteric mechanisms preventing its interaction with FIXa, phospholipid and VWF.^18,19 In addition, there is evidence that some of the antibodies have catalytic activity toward FVIII.²⁶ 

Despite the fact that we have been aware of inhibitory FVIII antibodies for several decades, our knowledge of potential non-neutralizing anti-FVIII immune responses is minimal. No currently available clinical laboratory test evaluates the presence of these antibodies and we have no idea of their incidence, prevalence and natural history in the hemophilia population. In addition, there is also some evidence that the presence of very low titer (< 0.5 BU) neutralizing antibodies might also substantially affect FVIII pharmacokinetics and, at a time when prophylactic treatment regimens are becoming widespread, this type of interference might prove very significant clinically.

While FVIII inhibitors are present, the potential to promote effective hemostasis through FVIII replacement therapy is extremely limited. Thus, the goal of inducing immunological tolerance to FVIII becomes a critical component of inhibitor management.²⁷ While each patient requires individual evaluation with regards to the appropriateness of initiating a trial of immune tolerance induction, this therapeutic intervention will be undertaken in the vast majority of cases.

To date, and since its first description over 30 years ago,²⁸ the principle underlying immune tolerance induction to FVIII is the repeated intravenous administration of FVIII until the inhibitor disappears and FVIII recovery and survival are normalized.²⁹ 

No randomized controlled trials or case control trials of immune tolerance induction (ITI) have yet been reported. Most studies are reports of case series. Approximately 25 studies of ITI have been published that are of variable quality with regards to the details of the investigation and the methodologies that have been used. These studies are further complicated by the reporting of some of the patients several times in later publications.

Registries documenting the outcome of clinical studies of immune tolerance induction have been established internationally,³⁰ in North America³¹ and in Germany.³² Although the patients enrolled in these registries have been treated with a variety of different tolerance induction protocols and the definitions of success have varied, some features of the treatment groups are consistent (Table 1 ). Thus, the most important predictor of success in ITI appears to be the inhibitor titer at the start of the treatment.³³ Inhibitor titers at the start of ITI of < 10 BUs are significantly associated with ITI success. Additional risk predictors for ITI outcome are the historic peak inhibitor titer (a linear variable, but < 200 BUs has been identified as a reasonable threshold indicator of improved outcome from a meta-analysis) and the duration of presence of the inhibitor. Factors that are recognized as predictors of unsuccessful ITI attempts are the coexistence of an infectious or inflammatory process and an interruption of the course of ITI (> 2 weeks in the German Registry data).³² 

The two most contentious factors that may influence ITI outcomes are the dose of FVIII administered and the type of FVIII concentrate used in the ITI protocol. While initial ITI schedules used very high dose FVIII regimens (200 U/kg/d),²⁸ subsequent use of protocols in which reduced (100 U/kg/d)³⁴ and low dose (25 U/kg 3 times/wk)³⁵ schedules have been used have also proven effective. There is some evidence to suggest that while tolerance induction may occur more rapidly with the high-dose regimens, the overall success rate may not be different between the various dosing schedules. In an attempt to resolve the FVIII dosing issue, a prospective international randomized study is currently in progress in which “good risk” pediatric inhibitor patients (pre-ITI inhibitor < 10 BUs: Peak historic inhibitor < 200 BUs and < 2-year history of the inhibitor) are treated with either 200 U FVIII/kg/d or 50 U FVIII/kg 3 times each week.³⁶ 

The second major debate concerning optimal ITI protocols has centered on whether recombinant FVIII concentrates are as effective at inducing tolerance as VWF-containing plasma-derived concentrates.³⁷ While anecdotal reports suggest that the plasma-derived concentrates may be advantageous in this regard,^38,39 there is no prospective systematic evidence to substantiate this claim. If the plasma-derived concentrates are superior for ITI the other question that will need to be addressed is whether the effect is due to an immunomodulatory effect of VWF¹⁶ or to other co-purified plasma proteins that can exert a tolerogenic influence.¹⁷ 

As indicated above, initiation of ITI should proceed as soon after FVIII inhibitor detection as possible, and preferably when the inhibitor titer is < 10 BUs (Table 2 ). If poor risk factors are present (pre-ITI inhibitor > 10 BUs: peak historic inhibitor > 200 BUs or duration of inhibitor > 2 years) consideration may be given to adding immunoadsorption and/or immunosuppression to the initial FVIII ITI schedule. Thus, where the inhibitor titer is > 10 BUs, the Malmo protocol for ITI used extracorporeal immunoadsorption with a Protein A sepharose column, followed by daily FVIII infusions to keep FVIII levels between 40% and100% and the co-administration of intravenous cyclophosphamide. Currently, this intensive ITI protocol is rarely utilized.

In good-risk patients, these more complicated protocols are not used. Initial FVIII dosing should be between 200 units/kg/d and 50 units/kg 3 times a week and, if available, a recombinant FVIII concentrate will usually be the product of choice. Additional factors that will influence decisions about the type of concentrate used and the ITI dosage schedule include the ease of venous access, the feasibility and desirability of central venous line placement, and cost.⁴⁰ In some instances, infectious or thrombotic complications of central venous lines will force the discontinuance of ITI and will be associated with a lower likelihood of tolerance induction.

Initially, the inhibitor titer should be followed weekly to evaluate the peak anamnestic inhibitory activity, but after 3–4 weeks, follow-up at 3-monthly intervals is reasonable. Following disappearance of the FVIII inhibitor in the Bethesda assay, administration of ITI treatment should continue until further evidence of tolerance is confirmed from FVIII recovery and half-life studies. These studies should be performed after a 72-hour period without FVIII administration with a dose of 50 units/kg. A FVIII recovery > 66% of predicted and a half-life > 6 hours is acceptable evidence of durable FVIII tolerance. This state will usually be attained within a year of disappearance of the FVIII inhibitor.

Once evidence of satisfactory FVIII recovery and half-life has been achieved, it is the usual practice to continue the patient on an indefinite prophylaxis regimen to maintain the state of tolerance. However, no prospective studies have been performed to justify this approach.

Using an ITI schedule such as that described above, approximately 80% of patients will achieve tolerance to FVIII. No studies have addressed the issue of when abortive attempts at ITI should be discontinued, but persistence of a high titer inhibitor beyond 12 months of therapy is a poor prognostic sign. In those who continue to show inhibitory activity, several “salvage” strategies can be considered (Table 3 ). If initial attempts at ITI have utilized a recombinant concentrate, consideration might be given to a VWF-containing plasma-derived product, and if initial therapy has been with a low-dose regimen switching to a high-dose protocol should be assessed, possibly with concurrent immunosuppression.

In addition to alternative FVIII infusion regimens, there are now reports of approximately 20 congenital hemophilia A inhibitor patients who have been treated with the mono-clonal mouse-human chimeric anti-CD20 antibody, rituximab.^41,42 This product depletes all forms of B lymphocytes from pre-B cells to mature activated B cells. In the studies reported to date, success rates of inhibitor eradication have ranged from 50–80%, but many questions remain about the optimal use of this agent in this context. Whether FVIII administration should be continued at the time of rituximab delivery and the optimal dosing regimen for rituximab remains unresolved and will require prospective evaluation.

A number of novel approaches to achieving tolerance to FVIII are beginning to be assessed in preclinical models. Some of these strategies involve the reduction of inhibitor risk through the generation of FVIII molecules with reduced immunogenicity. This has been achieved through either the mutation of specific immunodominant residues in the FVIII C2 domain or through the replacement of human FVIII regions with porcine sequence demonstrated to possess reduced immunogenicity.⁴³ Alternative approaches to tolerance induction are also being explored through the application of tolerogenic modes of antigen presentation. These approaches, which include FVIII delivery to the nasal and gastrointestinal mucosal epithelium⁴⁴ and to immature dendritic cells,⁴⁵ likely operate through the generation of FVIII-specific regulatory T cell populations. In addition, FVIII gene transfer approaches may also play a role in future immunomodulatory protocols that could involve combinations of these various therapeutic approaches.⁴⁶