An Orthogonal Approach to Biosimilarity

Dana Hursey/Radius Images/Getty Images

There is a resounding consensus among industry experts that multiple bioequivalency approaches and orthogonal methods will be required for the assessment of biosimilarity. Pharmaceutical Technology spoke to experts Daniel Galbraith, PhD, chief scientific officer, BioOutsource; Michael Sadick, PhD, senior manager, large molecule analytical chemistry, Catalent Pharma Solutions; Gary Chambers, business manager for biopharma labs, Europe, SGS Life Science Services; Glenn Petrie, PhD, senior scientific advisor, ABC Laboratories; and Joerg Windisch, chief scientific officer, Sandoz to learn more about best practices for analytical testing of biosimilars.

Current testing methods

PharmTech

: Which methods of analysis to demonstrate biosimilarity are most effective?   

Galbraith (BioOutsource)

: The key assays for each molecule are those that mimic the biological activity of the drug 

in vivo

and are shown to be sufficiently sensitive to specific differences in the physiochemical structure. The reason being is that these are the ones that are most likely to fail in clinical trials. These are mainly the bioassays using specific target cells. The ones that are used the most are the ligand-binding assays; these are used as markers or indicators for the biological functional activity. Ligand-binding assays are fast and cheap to do, so they are used as screening assays on a large number of samples.  

Sadick (Catalent)

: As far as I am aware, there is no one, or even a few, type(s) of assessment(s) that can verify biosimilarity. It really will require an orthogonal approach that combines physicochemical and functional analyses. Even for functional assays, there ought to be an orthogonal array of cell-based and ELISA-based potency assessments, as well as binding kinetic determinations. The array of functional assays may be winnowed with time as more is known.
 

Chambers (SGS Life Science Services)

: In our experience, orthogonality is the key approach because one test may not reflect subtle differences between an innovator and a biosimilar. For example, the paired analysis of Fourier transform infrared spectroscopy (FTIR) and circular dichroism (CD) will provide measurement of secondary structure. While both measure alpha helices and beta-sheets, FTIR is stronger with beta-sheets, while CD is better with alpha helices. This approach also applies to aggregation where size-exclusion chromatography-multi-angle laser light scattering (SECMALS) and sedimentation velocity analytical ultracentrifugation (SVAUC) are utilized. Any variations can then be further investigated using characterization methods such as peptide map tandem mass spectrometry (PMAPMS/MS) and electrospray ionization mass spectrometry (ESIMS).
 

Petrie (ABC Laboratories)

: The critical analysis for demonstrating biosimilarity is the bioassay. Binding, chromatographic, and electrophoretic analyses provide useful and required data, but the activity assay is the only one that truly reflects bioequivalence.
 

Windisch (Sandoz)

: There is no single method that will be the most important one. At the end of the day, you always have to account for the different structural components. You need to start with the primary structure, the amino acid sequence, the higher-order structure, the folding, the heterogeneity, the glycosylation, the impurities, and then you have to move from structural characterization to characterizing all of the different functionalities of the molecule. It’s really an all-encompassing exercise. There are really two principles here: one of them is redundancy, and the second thing is orthogonality. Overall, you could end up with anywhere between 50–60 methods to analyze structure-the chemical analytics-and another probably 15 methods to test function.
 

PharmTech

: What can binding studies tell investigators about similarity between product candidates? 
 

Sadick (Catalent)

: ELISA, or any other type of quantitative binding assay (whether it be designed as a content assay or as a potency assay) will define the steady state binding characteristics of the molecule. If the molecule is a monoclonal antibody (mAb), then an array of binding assays should be used to assess both CDR/ligand interaction and Fc/Fc receptor interaction. Even more detailed information can be derived from binding kinetics testing, using either surface plasmon resonance (SPR; e.g., Biacore) or bio-layer interferometry (BLI, e.g., Octet).
 

Chambers (SGS Life Science Services)

: Once you have established the acceptable variation, these can tell you if the binding between candidates is similar. If it is not, the data can be compared to secondary/tertiary structure analysis methods to assess whether the differences can be attributed to a conformational change in the candidate.
 

Petrie (ABC Laboratories)

: Typical ELISA or electrochemiluminescent assays can provide concentration data and a relative tool for comparison of binding curves. True binding characteristics require the use of surface plasmon resonance techniques which provide the association/disassociation rates and binding constant. This allows quantitative comparison of the binding characteristics of product candidates and the innovator product.
 

Galbraith (BioOutsource)

: Binding studies are an important first step in the biological or functional characterization of a molecule. Simple binding studies can tell us if the molecule binds and how tightly. More in-depth analysis such as SPR will inform on the kinetics of binding and disassociation, which is key in showing the same modes of action.
 

PharmTech

: Has the demand for biosimilarity testing services increased in the past few years?
 

Chambers (SGS Life Science Services)

: The demand has increased, with more clients requesting development partnerships rather than just testing support.
 

Petrie (ABC Laboratories)

: We have had a definite uptick in the number of requests for analysis of biosimilars. Based on the number of biopharmaceuticals going off patent and the FDA’s clarification of the requirements for biosimilarity, I see continued growth for the foreseeable future.
 

Galbraith (BioOutsource)

: We have seen an exponential increase in biosimilarity testing for two reasons: There are more companies involved in biosimilar development today, and there are more molecules being targeted. I would estimate that the industry has doubled every year for the past four years.
 

Sadick (Catalent)

: We are beginning to see a marked increase of interest, by current and potential clients, for provision of biosimilarity testing strategies and services.
 

PharmTech

: Is the most practical way to assess bioequivalency to use one assay for both the originator and follow-on biologic? 
 

Petrie (ABC Laboratories)

: While difficult, the preferred method for comparison of the bioequivalent and innovator drugs is assay by the identical method. Bioassays and ELISAs already have a high degree of variability. Introduction of a second method further complicates the situation and makes comparison of the data extremely difficult. Justification for the use of a second assay and its validity produces additional regulatory challenges.
 

Galbraith (BioOutsource)

: If there were one assay that could cover all of the potential functions a monoclonal antibody is able to perform, this would be ideal. We could then assess the innovator and biosimilar alongside and get an idea of the level of similarity. However, because the human body is a complicated system, we cannot replicate this in the lab and therefore, a single assay is not possible. In the lab we need to assess each function a monoclonal antibody performs in different assays to build up a full picture of all of these activities.
 

Sadick (Catalent)

: Not necessarily. Variability will be based upon the precision, accuracy, and robustness of each of the tests that are combined orthogonally to assess biosimilarity.
 

Chambers (SGS Life Science Services)

: I would say we are being pushed to look at multiple bioequivalancy approaches to support similarity assessment.
 

PharmTech

: Is it possible for a follow-on biologic to be truly equivalent, but test results do not validate biosimilarity? 
 

Galbraith (BioOutsource)

: It is conceivable that some minor effector functions do not show similarity but the overall conclusion is that this will not affect the clinical efficacy of the drug and hence, a conclusion of similarity will be made.
 

Chambers (SGS Life Science Services)

: Yes, a biosimilar may achieve equivalent clinical results, but analytically we may see variations. These differences may not impact the product binding and so do not impact activity.
 

Petrie (ABC Laboratories)

: There is the potential that a biosimilar could have the same clinical results as the innovator without meeting all the analytical requirements. Bioassays and binding assays have enough variability that they could fail bioequivalent criteria. However, convincing regulators that despite these failures the drug is equivalent presents a considerable obstacle to approval.
 

PharmTech

: What are some of the potential barriers to obtaining sufficient amounts of reference product from originator companies for biosimilarity testing? 
 

Sadick (Catalent)

: Under the current system, it can be quite difficult to obtain sufficient amounts and variety of originator material to provide a truly appropriate baseline for comparison. Costs can be very steep to obtain originator material, and not all originators make their material available, despite being paid for their medications. It usually requires a dedicated group within the company (the biosimilar company and/or the contract testing company) to facilitate obtaining originator material.
 

Chambers (SGS Life Science Services)

: One barrier is ensuring that material obtained remains within the expiry date prior to testing. Additionally, obtaining material around the same time as its manufacturing date to put on stability studies is another potential barrier.
 

Galbraith (BioOutsource)

: It can be challenging to obtain sufficient batches of product. At BioOutsource, we link into a global supply network and have good access to batches, but with some drugs, there is a global limit to supply of innovator drugs.
 

Presenting biosimilar data to regulatory authorities

PharmTech

: How should biosimilarity results be presented to FDA?
 

Windisch (Sandoz)

: We have gained quite a bit of experience through our recent filing. FDA wants to look at the molecule and its attributes from the perspective of the clinical relevance of all of these attributes. The agency wants manufacturers to take a complete look at the molecule, look at all of its structural components, make a systematic evaluation as to the clinical relevance of the different parts of the molecules, and then it wants them to do a criticality assessment and rank them. Going from that criticality evaluation, the agency wants manufacturers to focus on those differences that matter or even potentially matter and provide ample analytical and functional data on each one of them.
 

Galbraith (BioOutsource)

: FDA is interested in the totality of evidence, essentially taking all of the information and summing it up into a statement that defends the claim of similarity. Each assay used in this assessment needs to demonstrate its ability to identify where changes in the molecule are key to clinical efficacy. 
 

Chambers (SGS Life Science Services)

: The format should be tabulated with innovator data against biosimilar data. To keep this simple, it is better to have one column for innovator and biosimilar and provide ranges for each parameter tested. In addition, in the same table, a manufacturer should include a column for variation. This table should include specifications that include experimentally determined method and process variation, allowing an assessment of similarity to be made.
 

PharmTech

: What additional tests will likely be required to demonstrate interchangeability? 
 

Galbraith (BioOutsource)

: Interchangeability is really more of a clinical assessment and is not likely to be answered by laboratory analytics.

Chambers (SGS Life Science Services)

: To demonstrate comparable quality, safety, and efficacy, following batch release/stability testing and full ICH Q6B characterization, forced degradation studies should also be included. Ultimately, clinical trial data will also be required.
 

Petrie (ABC Laboratories)

: While not providing the critical data supplied by binding studies and bioassays, biosimilars require the complete set of analytical techniques required for any biopharmaceutical. This may include high-pressure liquid chromatography (HPLC), SEC, liquid chromatography-tandem mass spectrometry (LC–MS/MS), peptide map, capillary electrophoresis (CE), and micro-flow imaging (MFI).
 

Windisch (Sandoz)

: Manufacturers already have to provide complete characterization for a biosimilar. What you could assume is that FDA would look at any potential small differences-such as in the heterogeneity-somewhat more critically when it comes to interchangeability.
 

Testing complex structures

PharmTech

: What specific challenges exist when it comes to testing the biosimilarity of mAbs? 
 

Sadick (Catalent)

: The challenges, as I see them, are that there are multiple levels at which bio-dissimilarity could impact a monoclonal antibody. Certainly, there is the risk that any change in amino acid sequence, glycosylation, secondary, tertiary, or quaternary structure could result in a change in immunogenicity. Additionally, any of those same alterations could impact complementarity determining region/ligand interaction, Fc/Fc receptor interaction, or both.
 

Galbraith (BioOutsource)

: The challenge of testing mAbs is accurately defining the acceptable range of the critical quality attributes. Each batch of the originator drug can vary, and it is possible that changes may be made to the manufacturing process that could result in a change of the profile of the quality attributes. Attempting to define an acceptable range for your biosimilar molecule within these potentially moving goalposts can be challenging.
 

Chambers (SGS Life Science Services)

: Establishing expected variation is a challenge. This is often determined using a statistically suitable number of innovator batches and this also poses sourcing issues, as all material tested should be within the expiry date. Sourcing sufficient material often becomes a rate-limiting factor.
 

Petrie (ABC Laboratories)

: Establishing the biosimilarity of mAbs is challenging due to the complexity of their structure. Multiple subunits, disulfide linkages, post-translational modifications, and glycosylation require a myriad of analytical techniques. The advent of powerful mass spectroscopic techniques has simplified these analyses to an extent, but an enormous amount of effort is still required.
 

Windisch (Sandoz)

: With monoclonal antibodies, you will often hear they are so much more challenging than some of the other molecules that have already been done; this is only partially true. A monoclonal antibody is complex in that it is large, but in other ways, it’s also a fairly robust and relatively simple molecule. The biggest challenge is probably understanding the structural-functional association as it relates to the activity of an antibody.
  In most cases, the binding is not so much of an issue, because typically the binding site doesn’t have sugars and is not glycosylated. What are more challenging are the effector functions, which are in the Fc fragments. These are influenced by glycosylation and include the recruitment of the cellular immune system-the antibody-dependent cellular cytotoxicity-and the recruitment of the molecular immune system--or the complement-dependent cytotoxicity. Both of these can be influenced by the glycans, by the sugars in its Fc fragment, and one needs to fully understand the structures there and how they influence biological activities. This is really an interplay between doing cell line and process development work and then analyzing the variant that you get both from a structural and a functional perspective. 
 

PharmTech

: Will biosimilar products that incorporate fusion proteins or bispecific antibodies be more difficult to test?
 

Galbraith (BioOutsource)

: Enbrel (etanercept) is a fusion protein and is currently undergoing testing; this has not presented any difficulties thus far.
 

Windisch (Sandoz)

: Fusion proteins such as etanercept can be more challenging to test and can be more complex in structure (i.e., they can be highly glycosylated). On the other hand, you have to look at each molecule individually. You can have a more complex structure, but not all of the molecular attributes are clinically relevant. The challenge is developing your cell line in your process to actually create a close match of the molecule. While it can be more challenging to test fusion proteins, for bispecific antibodies, there is no difference from a normal antibody. 
 

Chambers (SGS Life Science Services)

: Yes. In some cases, these are highly potent and require testing at low-level concentrations for which standard mAb testing methods are not designed. Additional tests will also be required to account for differences in degradation pathways.
 

Petrie (ABC Laboratories)

: Fusion proteins, ADCs, and bispecific antibodies present the same challenges as mAbs due to their complexity. Characterization and analysis are required for not only the drug substance, but the linker and the fusion component, adding complications to the demonstration of biosimilarity.
 

In Vivo vs. In Vitro

PharmTech

: Does using a biologically derived technique as an assessment method complicate testing? 
 

Petrie (ABC Laboratories)

: Any bioassay presents special challenges related to cell lines, cell culture, laboratory technique, etc. These challenges are only multiplied for biosimilars. Small differences in the method, even those improving the method, may differentially affect the results generated for the biosimilar and innovator drug.
 

Galbraith (BioOutsource)

: Biological products are much more complicated and variable than traditional small-molecule products; this is a simple fact. The reason for this is that we use living cells or we use materials harvested from living systems such as blood or serum. These materials do not lend themselves to consistency, and therefore, the assays require much more control and larger datasets. The upside is that these assays will show potency, something that chemistry analysis is not able to do.
 

Sadick (Catalent)

: The

in-vitro

biological activity of a therapeutic molecule is not often completely biomimetic to the therapeutic action of that molecule in vivo. Thus, differences in

in-vitro

activity cannot always be directly related to the

in-vivo

activity (in a 1:1 fashion). The

in-vitro

test, however, should be reflective of the molecule’s therapeutic mechanism of action. Differences in

in-vitro

activity should accurately predict differences in

in-vivo

activity, providing vital information. 
 

PharmTech

: Can results of

in-vitro

tests be predictive of biological activity

in vivo

? 
 

Galbraith (BioOutsource)

: There is always the caveat that

in vitro

cannot truly replicate the

in-vivo

world. Activity we see in a test-tube can sometimes be due to the environment. These tests, however, have moved on significantly even in the past couple of years and are significantly better at estimating the

in-vivo

activity.
 

Sadick (Catalent)

: The orthogonal combination of physicochemical and biological analyses of a biosimilar molecule is not proof positive of the biosimilarity of that molecule

in vivo

. However, a combination of

in-vitro

tests should be reflective of a molecule’s critical attributes. Results from these tests should hopefully minimize the extent of any clinical studies needed for verification of biosimilarity, safety, and efficacy. 
 

Petrie (ABC Laboratories)

:

In-vitro

analysis can confirm that the biopharmaceutical has the proper three-dimensional structure, binding characteristics, and mode of action. These assays, however, cannot predict activity

in vivo

due to the effects of bioavailability, clearance rates, etc.
 

PharmTech

: How is the immunogenicity potential of a biosimilar candidate assessed? What testing methods are typically used? 
 

Galbraith (BioOutsource)

: Immunogenicity can be assessed

in vitro

-the cytokine storm assay has been applied to some products. More often, however, this is left to the clinical trials, where anti-drug antibodies are assessed in a screen of the patients.
 

Petrie (ABC Laboratories)

: Immunogenicity for the innovator and biosimilars are determined identically. ADAs are determined in Phase I and II by means of increasingly specific ELISA or ECL assays.
 

Sadick (Catalent)

: While some predictive

in-silica

testing (of amino acid sequences and glycosylation patterns) may be performed, and even may be somewhat useful, these predictive studies usually only minimally reflect the

in-vivo

reality of immunogenicity. That, unfortunately, leaves preclinical assessment of immunogenicity, which, in itself, is not always predictive of immunogenicity in humans. Continued immunogenicity assessment will likely be required in human recipients for a while.  

Article Details

Pharmaceutical Technology

Vol. 39, No. 4 Pages: 64–69

Citation

: When referring to this article, please cite it as R. Hernandez, "An Orthogonal Approach to Biosimilarity,"

Pharmaceutical Technology

39

(4) 2015.