Development of ADA Against Recombinant Human Interferon Beta in Immune Tolerant Mice Requires Rapid Recruitment of CD4+ T Cells, Induces Formation of Germinal Centers but Lacks Susceptibility for (Most) Adjuvants

Specificity of the T Cell Response to Protein Biopharmaceuticals

The anti-drug antibody (ADA) response is an undesired humoral response raised against protein biopharmaceuticals (BPs) which can dramatically disturb their therapeutic properties. One particularity of the ADA response resides in the nature of the immunogens, which are usually human(ized) proteins and are therefore expected to be tolerated. CD4 T cells initiate, maintain and regulate the ADA response and are therefore key players of this immune response. Over the last decade, advances have been made in characterizing the T cell responses developed by patients treated with BPs. Epitope specificity and phenotypes of BP-specific T cells have been reported and highlight the effector and regulatory roles of T cells in the ADA response. BP-specific T cell responses are assessed in healthy subjects to anticipate the immunogenicity of BP prior to their testing in clinical trials. Immunogenicity prediction, also called preclinical immunogenicity assessment, aims at identifying immunogenic BPs and immunogenic BP sequences before any BP injection in humans. All of the approaches that have been developed to date rely on the detection of BP-specific T cells in donors who have never been exposed to BPs. The number of BP-specific T cells circulating in the blood of these donors is therefore limited. T cell assays using cells collected from healthy donors might reveal the weak tolerance induced by BPs, whose endogenous form is expressed at a low level. These BPs have a complete human sequence, but the level of their endogenous form appears insufficient to promote the negative selection of autoreactive T cell clones. Multiple T cell epitopes have also been identified in therapeutic antibodies and some other BPs. The pattern of identified T cell epitopes differs across the antibodies, notwithstanding their humanized, human or chimeric nature. However, in all antibodies, the non-germline amino acid sequences mainly found in the CDRs appear to be the main driver of immunogenicity, provided they can be presented by HLA class II molecules. Considering the fact that the BP field is expanding to include new formats and gene and cell therapies, we face new challenges in understanding and mastering the immunogenicity of new biological products.

Test models for the evaluation of immunogenicity of protein aggregates

Protein aggregates have been discussed for a long time as a potential risk factor for immunogenicity in patients. Meanwhile, many research groups have investigated the immunogenicity of differently produced aggregates using in vitro or in vivo models. Despite all knowledge gained in these studies still little is known about the mechanisms of immunogenicity and the kind of protein aggregates bearing the greatest risk for immunogenicity. The choice of a suitable test model regarding the predictability of immunogenicity of protein aggregates in humans plays a major role and influences results and conclusions substantially. In this review we will provide an overview of the test models recently used for the evaluation of immunogenicity of protein aggregates; we will discuss advantages and drawbacks regarding their usability and predictive power for immunogenicity in humans.

Treatment with recombinant human bone morphogenetic protein 7 leads to a transient induction of neutralizing autoantibodies in a subset of patients

Background

Recombinant human bone morphogenetic protein 7 (rhBMP7) is applied for treatment of bone fractures, especially tibial non-unions. Its application may induce autoantibodies (aAB) affecting the targeted and endogenous signaling pathways and in turn negatively impact treatment efficacy.

Methods

Novel and sensitive assays for the quantification of BMP7-aAB and BMP2-aAB were established and used to analyze serum samples from healthy controls (n = 100 men, n = 100 women) and patients with long bone fracture (n = 265) treated or not with rhBMP7. Sera from three to nine time points per patient were available and enabled the evaluation of aAB over a time course of up to one year. Functional activity of the BMP-aAB was tested with a BMP-responsive cell-based reporter assay. Consolidation of the fracture was evaluated as clinical outcome potentially affected by BMP7-aAB.

Results

Prevalence of BMP7-aAB and BMP2-aAB was 1–2.5% in non-treated patients or healthy controls. The rhBMP7 treatment induced a transient increase in BMP7-aAB in a subset of patients, returning to non-detectable levels within six months. IgG from BMP7-aAB positive sera inhibited dose dependently the BMP7-reporter gene activity, whereas control sera were without effect. Successful consolidation of the fracture was observed in the majority of both aAB-positive and aAB-negative patients.

General significance

We conclude that BMP7-aAB can be detected as natural aAB in healthy subjects, and are transiently induced by rhBMP7 therapy in a subset of patients. The aAB are capable of antagonizing BMP7 signaling in vitro, but do not preclude treatment success in patients.

•

There are patients with natural autoantibodies recognizing BMP7.

•

In some patients, rhBMP7-therapy induces BMP7 autoantibodies.

•

BMP7 autoantibodies elicit neutralizing effects on BMP signaling.

•

Therapy-induced BMP7 autoantibodies disappear over time.

•

BMP7 autoantibodies seem not to affect therapy success.

Editor's Highlight: Subvisible Aggregates of Immunogenic Proteins Promote a Th1-Type Response

Protein aggregation is associated with enhanced immunogenicity of biotherapeutics. As a result, regulatory guidelines recommend screening for aggregation during bioprocessing. However, the mechanisms underlying the enhanced immunogenicity of aggregates are poorly understood. In the investigations described herein, the immunogenicity in mice of a humanized single chain variable antibody fragment (scFv) purified after expression in Escherichia coli has been examined. Reproducible scFv aggregates were obtained within the subvisible particle size range (mean diameter 2 µm) using thermal and mechanical stresses. Intraperitoneal immunization of BALB/c strain mice with 1 mg/ml of aggregated or monomeric scFv induced similar IgG and IgG1 antibody responses. In contrast, aggregate preparations stimulated significantly higher levels of anti-scFv IgG2a antibody than did the monomer. In comparative studies, aggregates of ovalbumin (OVA) within the subvisible particle size range were prepared by stir stress, and their immunogenicity compared with that of monomeric OVA in mice. Aggregated and monomeric OVA induced similar anti-OVA IgG and IgG1 antibody responses, whereas IgG2a antibody levels were significantly higher in aggregate-immunized mice. Furthermore, cytokine profiles in supernatants taken from splenocyte-dendritic cell co-cultures were consistent with aggregated preparations inducing a T helper (Th) 1-type response. Aggregated proteins within the subvisible range were therefore shown to induce a preferential Th1 type response, whereas monomeric proteins elicited a selective Th2 response. These data indicate that protein aggregation can impact on both the vigor and quality of immune responses.

Mouse Models for Assessing Protein Immunogenicity: Lessons and Challenges

The success of clinical and commercial therapeutic proteins is rapidly increasing, but their potential immunogenicity is an ongoing concern. Most of the studies that have been conducted over the past few years to examine the importance of various product-related attributes (in particular several types of aggregates and particles) and treatment regimen (such as dose, dosing schedule, and route of administration) in the development of unwanted immune responses have utilized one of a variety of mouse models. In this review, we discuss the utility and drawbacks of different mouse models that have been used for this purpose. Moreover, we summarize the lessons these models have taught us and some of the challenges they present. Finally, we provide recommendations for future research utilizing mouse models to improve our understanding of critical factors that may contribute to protein immunogenicity.