Effective In Vitro Bioanalytical Assays for Comparing Pharmacodynamics of Biosimilar Monoclonal Antibodies

Abstract

The rise in the development of biosimilar therapeutics has caused increasing demand for new approaches for precise, sensitive and accurate bioanalysis. In this white paper, we describe how a combination of the biopharmaceutical market, regulatory guidelines and the molecular characteristics of biologics has inspired the development of novel ligand-binding assays (specifically, flow cytometry and surface plasmon resonance) that may streamline evaluation of biosimilars.

Introduction

The coming year faces the expiration of the first patent of a monoclonal antibody (mAb) biotherapeutic—the 6.8 billion-dollar-a-year drug, rituximab (Rituxan® as marketed by Biogen Idec/Genentech), will lose European patent protection in November, 2013.

This expiry represents just the first in a series of upcoming patent expirations for biotherapeutics, including, but not limited to, trastuzumab (breast cancer), cetuximab (colorectal, head/neck cancers) and natalizumab (for multiple sclerosis and Crohn’s disease). In anticipation of markets for these therapeutic areas opening up upon patent expiration, drugmakers have ramped up the development of biosimilars.

The European Medicines Agency (EMA), having approved its first biosimilar in 2006, has already begun reviewing its first application for approval of a biosimilar mAb – a version of infliximab, originally developed by Janssen Biotech, Inc., a subsidiary of Johnson and Johnson. The United States Food and Drug Administration (FDA) has received nine biosimilar investigational new drug applications, and some copy biologics have already been approved in China and India. Specifically, with several antidiabetes drug patents expiring in 2015, some insulin copy biologics and insulin analogues are already available in India.

Investments in biosimilar development are expected to pay off — in Europe, revenue from the sale of biosimilars reached $172 million in 2010, and may be as high as $4 billion by 2020.

However, developing a biosimilar therapeutic can be expensive, with costs reaching 80 per cent of the cost of developing an innovator biologic drug and about 20 times as high as for developing a small molecule generic. Each new biosimilar faces the challenge of proving that any differences in potency and safety from the innovator drug are not clinically significant. Especially in the case of monoclonal antibodies, which are large and complex, chemical differences between biosimilars and innovators may be numerous. Such cases require rigorous demonstration of biosimilarity as a proxy for therapeutic “interchangeability,” the ultimate (though probably unprovable) standard.

Therefore, there is a critical need for increasingly accurate and precise nonclinical, in vitro assays for measuring drug potency, as these are the cornerstone of quality control of manufactured therapeutics. A recent survey showed that 32 per cent of drugmakers declared that innovations in assay technology were required to meet the demands of proving biosimilarity. Such assays can better determine lotto-lot variability in the manufactured product, assess the impact of process changes on drug quality, assess drug stability, and more. Therefore, increasing the precision of an assay improves the assay’s statistical power, facilitating the comparison between biosimilars and innovators.

Although regulatory agencies are considering biosimilars on a case-by-case basis, they have issued some guidelines on what types of in vitro studies should be performed in the evaluation of biosimilarity. Of note is the European Medicines Agency (EMA)’s draft guidance, titled “Guideline on Similar Biological Medicinal Products Containing Monoclonal Antibodies,” published in November 2010. The EMA’s guidance recommends measuring, among other parameters, binding to the target antigen and binding to all Fcγ receptors, FcRn and complement. Binding to Fc receptors and complement can result in cytotoxicity and safety concerns. The recent FDA draft guidance, “Scientific Considerations in Demonstrating Biosimilarity to a Reference Product,” is less specific in its recommendations, but indicates a strategy for evaluating biosimilarity based on “totality of the evidence.” In other words, comparing a biosimilar with an innovator using multiple, orthogonal assays is like matching fingerprints — the more multivariate the fingerprint, the more likely that a match is predictive of clinical biosimilarity. Most recently, the Indian regulatory agencies, the Department of Biotechnology and the Central Drugs Standards Control Organization issued its own guideline on the development of similar biologics, in part to attract investment from global pharmaceutical companies.

Innovations in in vitro assay development are being welcomed by regulatory agencies, who are championing the “risk-based,” or “stepwise,” approach to evaluating biosimilarity, suggesting that the results, very sensitive, highly predictive nonclinical assays, can help shape the direction of further testing [FDA guidance]. For example, appropriate pharmacodynamic (PD) markers can be a very sensitive indication of potential clinical differences between two drugs. Regulatory agencies have identified immunogenicity testing as an area enhanced by in vitro ligand-binding analysis. Regulatory guidance (ICH Q6B, 1999) states: “When an antibody is the desired product, its immunological properties should be fully characterized. Binding assays of the antibody to purified antigens and defined regions of antigens should be performed, as feasible to determine affinity, avidity and immunoreactivity (including cross-reactivity).”

In this white paper, we describe two bioanalytical approaches involving measurements of ligand binding that have recently been applied to comparing PD of biosimilars. Specifically, we show the use of surface plasmon resonance (SPR) and flow cytometry to quantify the binding between therapeutic mAbs, alemtuzumab (and its variants) and infliximab, to molecules mediating cytotoxicity.

Flow Cytometry: How It Works

Flow cytometry is an essential tool for in-depth cell analysis. In a traditional flow cytometer, cells in a liquid stream pass through a laser beam, which excites any fluorescent molecules on the cell. Emitted fluorescence is then measured by detectors tuned to specific wavelengths. With the capacity to simultaneously measure multiple parameters on hundreds of individual cells per second, flow cytometry is a powerful technology with a wide variety of applications in pharmaceutical development.

As shown in Figure 1, flow cytometry can be a sensitive, information-rich method for measuring mAb binding to Fc receptors on the cell surface. When developing a flow cytometry assay for PD assessment of a therapeutic, factors to consider include:

• Sample type (e.g., whole blood, separated PBMCs, serum, cultured cells)
• Stability of marker(s)/analyte(s) being measured
• Appropriate fluorochromes, dyes and conjugates to get the clearest data from samples
• Incubation temperatures and periods most suited to the matrix and/or analytes
• Most suitable controls to monitor assay performance
• Appropriate quality control checks on instrumentation to ensure cytometers are performing reproducibly

Surface Plasmon Resonance: How It Works

Surface plasmon resonance (SPR) is a powerful technique for measuring the binding of any pair of interacting molecules, including drugs and targets, and antibodies and antigens. Interactions are measured in real time, enabling the determination of kinetic parameters. The SPR signal is proportional to the mass of analyte and does not require any type of label. Figure 2 shows the principle by which the SPR signal is generated.

SPR is a very versatile platform with many different applications throughout pharmaceutical development. The measurement of the kinetics of critical molecular interactions between the drug and its target and other key receptors (e.g., Fc receptors in the case of antibodies), can increase the precision and accuracy of comparisons between biosimilars and innovators.

Results

Effect of differential glycosylation of alemtuzumab (Campath®-1H) on binding to FcγRIII: SPR and flow cytometry assays

Although alemtuzumab can be produced in suspension CHO cells with much greater yield than in the traditional rat cells, alemtuzumab from CHO cells is improperly glycosylated compared to that from rat and human cells. Glycoengineered alemtuzumab, made in CHO cells transfected with glycosyltransferases, showed higher antibody-dependent cell-mediated cytotoxicity compared to the wild type drug.

In order to determine the mechanism of cytotoxicity, we used SPR and flow cytometry binding assays to measure the binding of alemtuzumab and its glycosylated forms to several Fc receptors.

To demonstrate that the assays could detect differences in the measured binding, the assays were first qualified by adjusting the concentration of alemtuzumab to 50 per cent, 70 per cent, 80 per cent, 100 per cent, 120 per cent, 130 per cent and 150 per cent of the reference alemtuzumab concentration.8 Determination of the relative potency was performed using a 5-parameter logistic parallel line model using Statlia software. The curves of the different starting concentration samples were all determined to be parallel to the reference curve with p>0.05. The assay had good linearity with a correlation coefficient of 0.993 and measured potencies in the range of 91.4 to 100.6 per cent.

The curves showing binding between the various forms of alemtuzumab and FcγRIII (CD16a) are shown in Figure 3. A negative control antibody, engineered to have no CD16a binding, showed no response in this assay. The glycoengineered antibodies Glyco1 and Glyco 2 showed enhanced CD16a binding compared to the wild type alemtuzumab with relative potencies of 300 per cent and 411 per cent respectively. The glycoengineered forms of alemtuzumab did exhibit small deviations from strict parallelism [p=0.017 and p=0.038 respectively at p>0.05]. These data correlated with the increased ADCC activity previously observed. Precision between replicates at each concentration was generally less than five per cent.

Similar analyses were performed for CD32a and CD64 although the magnitude of the signal obtained was markedly different, reflecting the strength of binding to the different forms of the receptors. Unlike CD16a, CD32a and CD64 bound to wild type alemtuzumab with higher affinity than to its glycosylated forms (data not shown).

The assay was also tested for other therapeutic monoclonals—namely, bevacizumab (Avastin®), rituximab (Rituxan® or MabThera®) and eculizumab (Soliris®). Eculizumab, a monoclonal antibody against C5 which has a hybrid Fc domain of IgG2 and IgG4 and hence no Fc binding, was used as a negative control in the assay. Although no absolute measurements of potency were determined for these therapeutic monoclonals, their relative binding to CD16a was as follows: alemtuzumab bound CD16a with the strongest affinity, followed by rituximab and then bevacizumab. Eculizumab, as expected, exhibited no measurable binding.

To support our SPR observations using recombinant alemtuzumab and Fc receptors, we used flow cytometry to measure the binding of alemtuzumab to Fc receptor-expressing cells. Suspension CHO cells were transfected with vectors directing the expression of various Fc receptors. These cells were incubated with the therapeutic monoclonal antibodies and the bound antibodies were detected using a FITC-conjugated goat Fab2 anti-human kappa light chain antibody. The level of binding was measured by flow cytometry with the labeled cells gated by forward and side scatter, and the median fluorescence intensity was measured.

As with the SPR assays, the assay was qualified for alemtuzumab using the same nominal concentrations to mimic different potencies of 50 to 150 per cent of the reference. The assay performance for CD16a was very similar to that observed using the SPR assay. The correlation coefficient was 0.99 with measured potencies in the range of 97.9 to 108 per cent.

As shown in Figure 5, the glycosylated variants of alemtuzumab exhibited higher binding to CD16a than the wild type alemtuzumab. Relative potency values could not be obtained due to the lack of parallelism due to both Glyco1 and Glyco2 reaching a higher plateau of binding than the wild type reference.

Binding of infliximab to neonatal Fc receptor FcRn and complement component C1q

The EMA draft guidance recommends testing biosimilars for their binding to the Fc receptor FcRn. The neonatal, intracellular FcRn receptor is responsible for transport of IgG across the placenta. FcRn binds to IgG and albumin at low pH but not at high pH. This receptor is responsible for “salvage” of internalized IgG or albumin and therefore endows these proteins with a long half life. Structurally, the molecule is similar to Class I MHC and consists of a specific heavy chain combined with β-2 microglobulin.

Although FcRn is not commercially available, we obtained a supply of recombinant dimeric receptor suitable for SPR assays. We coupled the receptor directly to the chip surface and qualified the assay using the same protocol as for the alemtuzumab assays. Although the magnitude of the signal obtained was very low, the data were consistent and reproducible (Figure 6). In the qualification of the assay, the r2 value was 0.96 and the measured potencies were from 81 to 110.9 per cent.

The final assay that was developed to satisfy the requirements of the EMA biosimilar monoclonal antibody guidance was to measure C1q binding. C1q is the first component of the complement cascade, binds to IgM or IgG which is complexed with antigen. A large hexameric molecule, its binding affinity in solution is extremely low. In fact, it has proved impossible to detect binding of IgG to C1q which was immobilized on an SPR chip. Although it is possible to measure binding of C1q to immobilized IgG, this assay will be difficult to convert for potency or comparability studies.

We therefore developed an ELISA to measure C1q binding. The therapeutic monoclonal antibodies were serially diluted and coated to microtitre plate surface, C1q was added and, following washing, the bound material detected using anti-C1q-HRP conjugate. The assay qualification data showed a correlation coefficient of 0.98 with measured accuracies of between 100.4 and 114.8 per cent (Figure 7).

Conclusion

Our studies of Fc receptor binding by monoclonal antibodies indicate that binding to soluble or cell-surface Fc receptors can be accurately measured by surface plasmon resonance or by flow cytometry. We observed excellent precision for replicates, typically less than 5 per cent CV. The commercial data
were amenable to potency determination by the parallel line method, and we could easily distinguish variant glycoforms and different antibodies. Results showed high correlation with results of antibody-dependent cell-mediated cytotoxicity (ADCC) experiments, without the variability due to biological complexity and statistical noise. ADCC experiments testing the effect of Fc receptor binding to alemtuzumab yielded more variable results, with CVs of individual triplicates in the range of 0.2-18.6 per cent (mean CV=5.4 per cent).

These methods are accurate, robust, reproducible and currently in use within a GMP setting for batch release of drug lots. The assays are sensitive to changes in potency related to glycosylation, aggregation, concentration or protein modification. However, many changes can occur in biologics without affecting a potency readout, so careful physico-chemical characterization is still essential.

Our data also indicate that flow cytometry and SPR might be used to predict potential immunogenicity, induced cell proliferation, cell-mediated lysis, and receptor-mediated events (cAMP, cytokine release, antigen expression). A combination of such assays can provide a comprehensive picture of functional activity which is sensitive to small changes and provides reliable tools for quality control and for comparison of biosimilar with innovator products.

References

1. Mullard A. Can next-generation antibodies offset biosimilar competition? Nat Rev Drug Discov. 2012 Jun 1;11(6):426-8.
2. Nair, A. Pharmaceutical Technology. 2011 Feb; 35(2):18.
3. Frost & Sullivan. Analysis of European Biosimilars Market. Dec 2011, Pages: 223 [http://www.frost.com/prod/servlet/svcg.pag/HCPD]
4. 9th Annual Report and Survey of Biopharmaceutical Manufacturing, BioPlan Associates, Inc, April 2012, www.bioplanassociates.com
5. European Medicines Agency. Guideline on Similar Biological Medicinal Products Containing Monoclonal Antibodies. EMA/CHMP/BMWP/403543/2010. Nov 2010.
6. United States FDA. Scientific Considerations in Demonstrating Biosimilarity to a Reference Product. Feb 2012.
7. India launches ‘similar biologics’ guidelines at BIO2012 http://www.in-pharmatechnologist.com/Regulatory-Safety/Indialaunches-similar-biologics-guidelines-at-BIO2012
8. Harrison A et al. Methods to measure the binding of therapeutic monoclonal antibodies to the human Fc receptor FcγRIII (CD16) using real time kinetic analysis and flow cytometry. J Pharm Biomed Anal. 2012 Apr 7;63:23-8.

James D. Hulse, Ph.D. is Managing Scientific Director, Discovery & Development Solutions, EMD Millipore Corporation. He can be contacted at [email protected].