Immunogenicity of biologic therapies for migraine: a review of current evidence

Biologic therapies, such as peptide-based therapies and mAbs, are increasingly used to treat a wide range of health conditions. Treatment with these biologics can lead to a variety of immune responses that range in severity [9]. ADAs are produced by B cells through both T cell–dependent and T cell–independent processes. T cell–independent ADA production generally involves a direct interaction between a therapeutic protein and cell surface B cell receptors, leading to internalization and B cell activation. It is also possible for specialized circulating dendritic cells to directly ingest a therapeutic protein, which is then presented to splenic B cells, resulting in ADA production. ADAs produced by T cell–independent processes are typically transient and have a lower affinity for the therapeutic protein; however, subsequent stimulation of B cells by additional antigens or epitopes may lead to longer lasting and/or higher titer ADA [12]. During T cell–dependent ADA production, antigen-presenting cells (APCs) derive peptides from a therapeutic protein, which are then presented to naïve CD4+ T cells, leading to T cell activation in a class II major histocompatibility complex (MHC)–restricted process [12]. Cytokines produced by activated T cells enhance B cell activation and expansion, resulting in longer lasting and higher affinity immunoglobulin G ADAs. Furthermore, B cell activation may also lead to the formation of plasma cells and memory B cells [13]. ADAs are broadly categorized as either NAb or non-neutralizing antibodies (non-NAbs) [14]. Non-NAbs bind to pharmacologically inactive sites on therapeutic proteins. Conversely, NAbs directly bind to therapeutic protein active sites, thereby potentially reducing efficacy (Fig. 1) [13, 15].

Among biologic therapies, a wide range of immunogenicity rates have been reported, from < 1% to > 50% in some cases [11]. The immunogenicity of biologic therapies is influenced by a range of factors, including product-specific or treatment-related factors (eg, dose, route of administration, protein structure, protein stability, and formulation) and patient-related factors (eg, genetic predisposition, age, disease status, concomitant medication use, comorbidities) [10‐12, 15]. Both the type and magnitude of ADAs produced can vary greatly between individuals, further affecting the extent to which ADAs impact each patient. Finally, ADA binding to a target therapeutic protein can lead to the formation of immune complexes that may significantly impact the extent to which ADAs and NAbs affect a given treatment [13]. It has been suggested that therapies administered by subcutaneous injection are more immunogenic than those administered intravenously, and this is thought to be related to the presence of cutaneous dendritic cells serving as APCs leading to T cell–dependent ADA production [11]. Therapeutic proteins containing human or fully humanized sequences tend to be less immunogenic than those containing murine or chimeric sequences; moreover, proteins with a high likelihood of forming immune complexes, containing epitopes recognized by class II MHC, and/or certain posttranslational modifications (eg, glycosylation) are more immunogenic than those lacking these elements [11, 12]. Due to a significant impact of peptide sequence on immunogenicity, a range of in silico and in vitro tools exist to identify and minimize risk when developing a biologic therapy. Biologic therapies can be evaluated to determine the risk of immunogenicity by identifying potential T cell epitopes using in silico analyses (eg, statistical, interference, or structural modeling) or in vitro T cell stimulation [16]. This approach provides an opportunity to engineer biologics, with the goal of increasing the humanness of the therapeutic protein.

NAbs are ADAs that directly interact with pharmacologically active regions of a mAb and thus may directly prevent or reduce mAb target binding, thereby potentially diminishing efficacy (Fig. 1). Although non-NAbs do not directly affect pharmacologic activity, both NAbs and non-NAbs can affect the systemic exposure of a therapeutic protein. For example, ADA/mAb interactions that increase clearance and shorten the elimination half-life may decrease therapeutic exposure [10, 13]. Conversely, non-NAb/mAb interactions that decrease clearance and lengthen the elimination half-life may increase therapeutic exposure [10, 13].

The extent of these effects on the efficacy of a therapeutic protein depends on many factors, including ADA kinetics (eg, onset and duration of immune response), titers, target affinity, and the propensity for immune complex formation [11, 13, 15, 17]. Immune complexes form when ADAs bind to therapeutic proteins, such as mAbs. Immune complexes are cleared through several mechanisms, with larger complexes being cleared more quickly than smaller complexes, which may persist longer [11, 13]. Depending on ADA affinity and titer, among other things, immune complexes can vary greatly in size and stability (Fig. 1) [13, 18]. For example, in cases where the titer and affinity of ADAs are low, immune complexes may be small or may form less readily and thus may have little or no impact on mAb efficacy or safety. Conversely, ADA reactions with high titers and/or high affinity for the therapeutic protein may favor the formation of large ADA therapeutic protein aggregates with more drastic impacts on systemic exposure or pharmacologic activity (eg, interaction between therapeutic mAb and target protein/receptor) [18, 19]. Altogether, these mechanisms may result in loss of efficacy, altered pharmacokinetics, cross-reactivity to endogenous protein, and hypersensitivity reactions, such as infusion reactions to anaphylactic reactions [13, 20, 21].

Immunologic responses to therapeutic proteins can cause adverse reactions, which have a range of different phenotypes (clinical presentations) and endotypes (underlying cellular and molecular mechanisms of response). Hypersensitivity reactions manifest with a range of mild to severe symptoms depending on the underlying mechanisms of immune response (described as types I–IV, according to the revised Gell and Coombs’ classification) and may present at any time from treatment initiation through several months following the abrogation of treatment [11]. Type I reactions are commonly mediated by immunoglobulin E antibodies, although they can also occur independent of immunoglobulin E release through T cells. These reactions usually occur within minutes to hours after treatment and can typically be prevented by prophylactic antihistamine treatment. Symptoms of type I hypersensitivity reactions can include pruritus, flushing, shortness of breath, rash, and hypertension; in severe cases, type I reactions can involve life-threatening anaphylaxis [13, 22, 23]. Therapeutic proteins can bind to cell surface protein targets and attract circulating ADAs, leading to formation of immune complexes on cell membranes in tissues. Type II reactions result from the adverse effects of these ADA/immune complexes on cell membranes [13]. Type III reactions occur when therapeutic proteins bind soluble antigens and aggregate to form immune complexes that are not cleared from the body [23]. These immune complexes may precipitate and deposit, particularly in diffuse capillary networks or tissues high in fenestrated epithelium (eg, kidneys and synovial membrane) resulting in complement activation, inflammation, and local (eg, nephropathy/nephritis) or systemic injury [13, 22, 23]. Type IV reactions, also called delayed type IV hypersensitivity, are thought to be mediated by T cell activation (though other mechanisms may be involved), typically arise 12 h to several weeks following mAb treatment, and may be mild (rash) or severe (Stevens-Johnson syndrome and toxic epidermal necrolysis) [13, 23]. It is noteworthy that type I and type III hypersensitivity reactions are most common with respect to ADAs [22]. With respect to fremanezumab, 2 mutations were introduced into the constant region of the heavy chain to limit normal antibody effector functions. This loss of function mutation was designed to prevent stimulation of antibody-dependent cell-mediated cytotoxicity and triggering of complement-dependent cytotoxicity [24].

There are many challenges to evaluating ADAs and understanding the impact of ADA on the efficacy and safety of biologic therapies. For example, several ADA assay formats are available and, as such, a consensus method for detecting, measuring, and reporting ADAs has not been established [15, 25]. In addition, the reagents required to test for ADAs are specific to each therapeutic protein. This lack of a uniform detection method makes it challenging to interpret similarities and differences in immunogenicity among different biologics, despite similarities in target and/or mechanism of action [10]. Biologic therapies that bind soluble ligands make up a large proportion of antibody therapeutics. Drug targets, when present at sufficiently high circulating concentrations, can potentially interfere with the performance of ADA assays [26]. In this context, eptinezumab, fremanezumab, and galcanezumab bind to the soluble ligand, CGRP, whereas erenumab is a mAb that binds to the CGRP receptor [27]. Nevertheless, in order to provide an overview of the current evidence on ADAs and NAbs in migraine preventive biologics, we have briefly summarized published results from recent migraine prevention clinical trials evaluating mAbs that target the CGRP pathway.

In the HALO EM trial, of the 4 patients receiving fremanezumab who developed ADAs, 2 patients had no adverse events (AEs): 1 patient had 2 transient events of injection-site pain and also experienced 1 event of a mild decrease in hemoglobin with no evidence of hemolysis, and 1 patient who developed ADAs and NAbs experienced an upper respiratory infection. During the HALO CM trial, 2 patients developed ADAs following treatment with fremanezumab: 1 patient had no AEs during the study and the other patient experienced 2 transient events of injection-site pain, 1 event of mild injection-site induration, and 1 event of mild injection-site erythema. Finally, among 43 patients who developed ADAs in the HALO LTS study, including 18 with NAbs, there were no apparent safety consequences nor any impact on efficacy. None of the patients in the HALO studies who developed ADAs had a severe hypersensitivity reaction.

Although patient-level data is not available for other CGRP mAbs described in this review, publications and prescribing information suggest that ADAs formed did not have a significant impact on efficacy or safety [6‐8]. Taken together, no AEs potentially related to ADAs have been identified with this class of drugs.