Immunogenicity Testing: A Critical Component of Biotherapeutic Drug Development
Immunogenicity is the ability of a compound to induce a specific immune response. For vaccines, this is the desired effect– to provide an immune response to afford protection against a harmful agent to the individual, for example, flu, tetanus, mumps, measles, Rubella and many others.
In the case of biotherapeutic drugs, immunogenicity is unintended, and so any ‘off target’ specific immune response is unwanted. Understandably, avoidance and monitoring for such immunogenicity is a major part of drug development, as the consequences can be extremely severe. For biotherapeutics, immunogenicity is a key arm of safety considerations in the drug development process. Much of the progress in monoclonal antibody drug design has centred on engineering antibodies that ‘look’ like naturally occurring antibodies and are humanized as much as possible, not least to reduce the possibility of the drug being recognized as ‘foreign’ and therefore reduce the incidence of an adverse ADA response.
How do we know whether a developed drug will be safe with respect to its immunogenic potential, and how do we assess whether immunogenic responses have occurred?
For many drugs, the typical immunogenic response would involve an antibody response. Therefore, the aim of detection is to identify specific antibodies against the drug. The ‘catch 22’ is we want to measure actual anti-drug antibodies, but none will exist before an antibody response (unwanted) has been raised. The initial approach to developing an ADA assay, therefore, is to look for the ‘absence of normal’. This is achieved by looking at multiple serum samples from as many drug-naïve subjects as possible to have a reliable estimate of what an antibody negative sample looks like in assay terms. This enables a useful opportunity to establish what level of response may indicate positivity using statistical means, and this level is known as the cut-point.
So far, so straightforward?
With no absolute standard, how do we quantify and control assay reproducibility? Controls can be established using highly specific recombinant monoclonal antibodies and/or polyclonal antibodies raised in other species, such as rabbits. These options are extremely useful, but as they are not a true example of human anti-drug antibodies, this essentially limits the ADA assay to being semi-quantitative / quasi quantitative rather than a fully quantitative assay.
In the first stage of ADA development, what if any of the drug-naïve samples produce what looks like a positive response? This is not an uncommon occurrence. In the cut-point determination above, are such samples included in the calculation of the positive cut-point? The answer is ‘it depends’. The potentially positive samples need to be investigated further to demonstrate that the positive signal is indeed specific for the drug. This leads to the next tier of ADA testing.
Subsequently, samples screening positive are further analysed to ‘confirm’ the positive responses using a modified screening assay where the sample ADA is pre-incubated with the drug, which, if true ADA, will inhibit the ADA response.
If positivity is confirmed, a third analysis, namely the titration assay, is used to estimate the magnitude of the positive response. This is the established three-tiered approach and documented in the FDA Guideline ‘Immunogenicity Testing of Therapeutic Protein products – Developing and Validating Assays for Anti-Drug Antibody Detection (2019).’
How is the system working?
A substantial catalogue of immunogenicity data from biotherapeutic development has accrued over the last 20-plus years. In recent times, a review of the sufficiency of the 3 tier paradigm is being undertaken to answer such questions as: are we doing enough? Have we done too much? Where next? Ultimately, is the analytical approach to immunogenicity serving its purpose well or could it benefit from any modifications? The responses to these questions are varied, although a question that recurs is how useful is the 2nd ADA step, i.e., the confirmation step? This arises because many sponsors have reviewed the vast amount of data available from their immunogenicity studies, many involving large cohorts and have discovered that the confirmation step may not add any further useful information to the screening step. One proposed modification is to simply perform the screening and then assess the magnitude of the response in the titre assay. This may seem slightly controversial. However, a simplified approach may expedite the assessment with equally useful results. Many observations and variations on the approach to confirmatory assays and the limits of their usefulness have been made in articles such as Neyer et al. J. Immunological Methods (2006) 315, 80-87 and Kubiak et al., Bioanalysis (2020), 12(4), 245-256.
One aspect that has recently met with broad agreement is the presentation of the ADA data to the agencies. A harmonized approach has been proposed and published in the AAPS Journal Myler et al. AAPS J 24, 4 (2022). https://doi.org/10.1208/s12248-021-00649-y.
In addition to the ADA testing above, when ADA are present in pivotal clinical studies, an assessment of the functionality of the ADA is required. In other words, do the assays have a neutralising effect on drug activity? This will typically involve a cell-based assay that can demonstrate that the ADA found can inhibit the function of the drug. Examples include antibody dependent cytotoxicity assays (ADCC) or enzyme inhibition assays, depending on the mode of action of the drug.
Immunogenicity is clearly complex, and within drug development, when approaching the regulatory agencies, a broader package of work is included termed the ‘immunogenicity risk assessment’. This contains much more than the assay data and includes the wider clinical considerations on the disease type immunological susceptibility and may consider previously licensed therapeutics that are of similar structure and likely to possess a similar immunogenicity.
With the advent of bispecific, tri specific antibody drugs, oligonucleotides, AAV (Adeno associated virus) based delivery systems, the world of immunogenicity testing is expanding. In many cases, the risk of immunogenicity may be low based on molecular structure, e.g., in the case of many oligonucleotides. With AAV delivery systems, the potential immunogenic moieties include viral proteins, oligonucleotides and the expressed protein product, of which the latter is generally considered the most likely to be immunogenic.
Currently, many biologic drugs have undergone rigorous evaluation for immunogenicity in preclinical studies and within clinical testing. Advancements in understanding the nature of therapeutic proteins and other biopharmaceuticals and specific regions of the structure that are more likely to induce an immune response have been studied, such that more ‘immunogenic’ regions can be engineered out of therapeutics in discovery or early development phases.
Refinements and strategies for immunogenicity assessment are an ongoing process and will no doubt evolve over time, but even with the increase in knowledge gained over several years of biotherapeutic development, Immunogenicity testing is of paramount importance, and there can be no complacency.