By 2020, the U.S. patent will expire on an estimated 12 patented biologics worth approximately $67 billion. This creates tremendous opportunity for drug companies developing biosimilars as well as patients seeking more affordable biologic drugs. With the introduction of the Patient Protection and Affordable Care Act, the U.S. federal government outlined an approval pathway that will help ensure biosimilar products have the necessary level of efficacy and safety as reference biologic drugs.

Among its many provisions, the new health care law specifically requires clinical studies that include a comparative assessment of pharmacokinetics (PK) and immunogenicity to the original reference biologic drug. This comparison makes it possible to determine structural differences that may not be detected via traditional analytical means. 

Even with updated guidance on the biosimilar regulatory pathway, determining the structural and functional similarity between a biosimilar and its reference drug can be challenging. Unlike a small molecule whose structure can be precisely reproduced, a biosimilar’s structure is unlikely to be identical to that of the reference drug due to the inherently complex nature of proteins. These structural differences arise primarily from the differences in manufacturing processes—different cloning systems, cell expression systems, post translation modification differences—used in the reference biologic drug, where the processes are proprietary and not publicly shared. Therefore, characterizing the structural and functional differences between the biosimilar and the reference drug are of the utmost importance, as such differences can affect the protein’s safety or efficacy. However, current limitations in analytical technologies inhibit the ability to definitively determine the structural comparability between the reference and biosimilar drugs.

To establish biosimilarity, FDA guidelines mandate that PK and immunogenicity comparability between the reference biologic drug and the biosimilar be established. Client and industry focus has been on the clinical studies to demonstrate this comparability. However, non-clinical studies can bring to light unanticipated structural differences and sometimes reveal information that guides future biosimilarity clinical trials. Our data support the recommendation of conducting a non-clinical study that compares both PK and immunogenicity responses to comprehensively assess biosimilarity prior to initiating a costly and lengthy clinical study.

Importance of PK and Immunogenicity Assays
PK is sometimes described as what the body does to the drug. It is a measure of how the drug is absorbed, how long it remains in the body and in what concentrations, where the drug goes in the body, and how the drug is removed from the body.
Immunogenicity is the ability of a biologic drug to induce an immune response in the body of an animal or human. There are two kinds of immunogenicity: wanted immunogenicity, in the case of vaccines and cancer immunotherapy, and unwanted immunogenicity against a therapeutic biologic, which results in anti-drug antibodies (ADA). The production of ADA can render the biologic drug inactive, thus decreasing efficacy and PK, and in rare cases result in adverse events in the human or animal.

How should CROs properly develop pharmacokinetic and immunogenicity assays to compare the reference product and the biosimilar within clinical and non-clinical samples? Namely, should researchers use two separate, product-specific assays or a single assay using one entity as reference?

The one-assay approach eliminates the concern of variability across data from the reference to the biosimilar as the samples are not analyzed across two assays (Liu, et al. 2015). The alternative—one assay for the reference drug and a separate assay for the biosimilar—gives no such benefit. Added benefits of the one-assay approach are the cost savings associated with method development, validation and sample analysis that only require one assay format instead of two, plus streamlined sample analysis.
We have outlined points for consideration when developing an immunogenicity assay in Table 1. For full descriptions of the technical advantages and disadvantages of the one or two assay approach for PK and ADA, please refer to Marini et al. 2014 and/or Liu, et al. 2015, respectively.

Below are two examples of how AIT Bioscience has used the one-assay approach: first applied to a PK and immunogenicity method for one biosimilar, and the second applied to a non-clinical immunogenicity method for a different drug program.

Case 1: PK and Immunogenicity in a Standalone Study
AIT Bioscience used a one-assay approach to develop both the PK and an ADA assay for a biosimilar drug program in a standalone study. Lacking historical information about the reference drug’s assay format as well as platform and reagents used in its development, we selected an electrochemiluminescence base (ECL) format using the Meso Scale Discovery (MSD) platform for both the PK and ADA assay formats as this technique offers a large dynamic range.

We developed a sandwich ECL ELISA method to quantitate the drug. We optimized this assay with the reference material before the biosimilar drug was available and held off on evaluating the specificity of critical reagents to the biosimilar until the end of method development, thereby saving method development time.

Once the biosimilar was available, it demonstrated close agreement with the reference drug in their respective dose-response relationships in the ligand binding assay.Comparative statistics, as recommended by Marini et al. 2014 to establish bioanalytical similarity, were not applied to determine how closely they agreed as the biosimilar PK assay was not developed to support regulated clinical or non-clinical bioanalytical sample analysis; however, the biosimilar and reference drug curves demonstrate nearly overlapping dose responses (Figure 1).

A one-assay approach was also used to develop a bridging ECL ELISA ADA assay format where the biosimilar drug was conjugated with ruthenium or biotin to generate detection and capture critical reagents, respectively (Figure 2). We evaluated a set number of drug-naïve individuals with either the reference drug or the biosimilar added in excess. We determined the parametric 0.1% false positive error rate confirmation cut point for both the reference drug and the biosimilar to be 71.4% and 67.5%, respectively. These results suggest the reference drug and the biosimilar drug have similar reactivity. A potential disadvantage of using the two assay approach would have been the time and cost to generate an independent surrogate control antibody to the biosimilar, as the positive control antibody described in the one assay approach had reactivity to both the reference and biosimilar drug.

Case 2: Immunogenicity in a Non-clinical Study
In the second example, AIT Bioscience again applied a one-assay approach to a different drug program to support a non-clinical analysis for the presence of ADA. We again developed a bridging ECL ELISA on the MSD platform for this method. The biosimilar and the reference drug concurrently underwent non-clinical analysis for ADA responses while utilizing an anti-reference antibody to generate the positive controls.  Upon evaluation of the drug, naïve animals in both the presence and absence of the reference material and the biosimilar material, a parametric 0.1% false positive error rate confirmation cut point was determined for both the reference drug and the biosimilar to be 25.5% and 35.2%, respectively.

Upon analysis of 30 animal samples (10 control animals, 10 animals dosed with the biosimilar drug and 10 animals dosed with the reference drug), one animal in the biosimilar group had a response above the tier 1 screening cut point with a 5.0% false positive error rate. We initiated confirmation testing and evaluated this sample twice within the same assay: 1.) In the presence of excess biosimilar drug and 2.) In the presence of excess reference drug on the same plate.  The putative positive sample continued to elicit ADA responses above the biosimilar and the reference drug’s confirmation cut points with a 0.1% false positive error rate. Therefore the ADA present in this animal, which received the biosimilar, were reactive to both the biosimilar and reference drug suggesting structural similarity.

Discussion
When differences in immunogenicity results arise, structural differences may be gleaned from animal studies. AIT Bioscience typically recommends this step since differences in immunogenicity results might indicate structural or functional differences between the two products otherwise undetectable by other analytical methods (April 2015 FDA Guidance).  Therefore, conducting an ADA assessment in a non-clinical study of a biosimilar can indicate early on potential differences between the biosimilar and the reference drug, which may result in reduced safety or efficacy. In the case of the second example presented above, the animal dosed with the biosimilar had an ADA response to both the biosimilar and the reference drug. This result supports a structural and functional similarity between the biosimilar and reference drugs.

Minor structural differences, including certain post-translation modifications, can dramatically affect a biotherapeutics’ efficacy or safety. Therefore, a clear difference in preclinical immunogenicity could be a sensitive indicator of an unanticipated structural difference. Thus, rather than a ‘check-box’ activity, animal studies have the potential to reveal vital information about the biosimilarity of a biotherapeutic and guide future clinical studies. 

For references and a table showing consideration points for one vs. two-assay approach for immunogenicity assessment, visit the online version of this story at www.contractpharma.com