A key part of the biosimilar approval process is proving that the new product is indeed structurally and functionally similar to the originator. This is not a trivial activity, as biologic medicines are complex proteins or glycoproteins, and there are significant challenges involved in determining their structures to the level of precision demanded by the regulators.
In the U.S., the 351(k) approval pathway used for biosimilars is an “abbreviated” pathway, requiring step-wise comparison with a reference product approved under the normal 351(a) pathway, which would have involved exhaustive clinical trials. The biosimilar application must initially include information demonstrating that the new product is highly similar in structure to the reference. The U.S. then requires animal studies to be performed, and at their discretion, a clinical study or studies. However, the premise is that these will be less extensive than for a novel molecule. Uniquely, U.S. rules provide for two types of biosimilar: a normal biosimilar, which requires analytical characterization to justify its highly similar status; and an interchangeable biosimilar, which demands additional clinical switching studies to demonstrate that there are no differences in any given patient.
The U.S. regulatory pathway looks at the totality of data submitted, both clinical and non-clinical, and FDA guidance1 suggests the use of a “meaningful fingerprint-like” analysis algorithm, which relies heavily on a statistical approach to demonstrating analytical similarity. This statistical approach has resulted in tiers of biosimilarity, ranking the quality attributes of the molecules according to how important they are.
The first tier includes those quality attributes that are the most critical, where a statistical equivalence test would be required to demonstrate comparability. The FDA advises that the focus should be on a few of the highest-risk critical attributes, for example, product content, or glycosylation in some molecules. For the second, less critical tier, it might be possible to determine quality ranges based on standard deviations of these attributes compared to the reference product. The third tier represents the least important attributes, and these data may be presented in raw form, or as graphical comparisons.
The whole paradigm of the FDA’s accelerated pathway turns the traditional 351(a) pathway around, with the analytical sections becoming the most critical initial step, and the requirement for clinical studies greatly reduced. Biosimilarity is demonstrated in a stepwise manner, where the foundation is the performance of quality and analytical comparisons, followed by in vitro and any necessary in vivo testing ahead of clinical work.
Physicochemical analytical characterization is absolutely essential at all stages of biosimilar development. At the outset, the biosimilar manufacturer has to determine the exact structure and sequence of the originator product. Multiple analytical techniques are applied to this reference molecule to study its primary protein sequence and post-translational modifications, in order to determine the quality target product profile (QTPP). Multiple batches of the originator product should be assessed.
For the second step, during the development of the manufacturing process for the biosimilar, analytical physicochemical characterization can aid in the determination of the most appropriate cell line, for example. The third step is perhaps the most critical part in comparability terms, and is a side-by-side comparison of the originator molecule and the biosimilar. Again, a battery of techniques for the qualitative and quantitative determination of primary and higher order structures is available. It is important to remember the tiered approach for risk assessment- which quality attributes have been deemed the most critical, and how the statistical analysis of the resulting data is going to be approached. Not forgetting that throughout the entire development process for the biosimilar, physicochemical analysis underpins a range of activities, including formulation development, stability testing and quality control.
A number of other factors must be considered when comparing the biosimilar to the originator molecule. The selection of the reference material is important in global biosimilar development programs. Some countries’ regulations allow the use of comparators authorized by certain other regulatory authorities for some studies, if appropriate demonstration is made that it is indeed representative of the authorized product in the country of application. The number of batches of biosimilar and reference product that will be analyzed and compared must be carefully considered—it is important to remember that the originator product will have undergone changes over its lifetime too.
The biosimilarity ‘fingerprint’ can then be built up using multiple orthogonal analytical techniques, with quantitative ranges for the various quality attributes being established. The revised quality guideline from the European Medicines Agency2 states that the comparative characterization studies should be performed using “sufficiently sensitive analytical tools.” The regulatory authorities are well aware of the capabilities of modern analytical instrumentation, and would expect to see such methods used to interrogate both biosimilar and reference product.
Analyzing the structure of a biologic is nowhere near as straightforward as the process of structural determination for a small molecule. The amino acid sequence of a protein is not simply a direct translation from the gene sequence at the cellular level. The expressed protein will differ due to microheterogeneities formed during fermentation and downstream processing in addition to various post-translational events, which cannot be predicted from the oligonucleotide code, and only determined by careful studying of the expressed product. Glycosylation is arguably the most important of these events, but any of these modifications can alter the molecule’s effectiveness and immunogenicity. Furthermore, if a carbohydrate moiety is attached to the protein backbone, this will increase the potential heterogeneity of the product. It is therefore extremely important to be able to determine what these alterations are.
The ICH Topic Q6B guideline3 was first published in the 1990s, and covers the characterization of biotech products, with the aim of setting specifications for quality. It mandates multiple physicochemical analyses, and provides a starting point for designing a strategy to prove biosimilarity. Six structural characterization and six physicochemical property topics are cited as important, as shown in Table 1.
A wide range of analytical tools is available for determining this essential structural and property information. Some of these are classical analytical techniques, and others are newer method or instrumentation developments, such as hydrogen–deuterium exchange (HDX) or ion mobility mass spectrometry.
As an example, Figure 1 shows the use of intact mass measurement via electrospray mass spec to determine the intact molecular weight of an antibody product. This not only determines the molecular weight, but also gives some insight into the glycosylation pattern and the different glycoforms. The differing numbers of galactose and fucose present on this IgG molecule are evident from the mass spec analyses. This provides a simple way of monitoring glycoforms patterns and screening for biosimilarity.
However, if post-translational modifications are to be studied in more detail and quantitative results collected, the molecule can be broken down into smaller constituent parts. One way to do this is to take the antibody, reduce and protect the disulfide bridges, and then cleave it into smaller pieces using a protease enzyme. These fragments can then be analyzed, either via a peptide map, or by using online LC/MS or LC/MS/MS to sequence the individual peptides. This peptide mapping philosophy can even be extended to determine the location of the disulfide bridges.
For biosimilars, the carbohydrate profile may not necessarily be the same as the originator - but any differences observed would certainly have to be shown to have no impact on safety and efficacy. Glycosylation characterization requires information on carbohydrate content, structure and site of attachment, as described in ICH Q6B. Some of the glycomic analytical strategies that can be applied are shown in Figure 2. For example, the intact molecule can be studied via electrospray or Matrix Assisted Laser Desorption Ionization Time-of-Flight (MALDI-ToF) mass spectrometry, GC/MS can be used to determine the quantitative monosaccharide composition, and chromatographic methods such as High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD) used for sialic acid content. The glycoprotein can also be split into smaller glycopeptides ahead of peptide mapping, or the carbohydrates removed and studied either in native form or derivatized. Whichever preparative approach is used, analyses can be via MS and LC methods for both qualitative and quantitative assessment.
Higher order structure
There are many different techniques available for determining higher order structure and aggregation of the molecule. A number of these are detailed in Table 3, and again, the different methods are complementary as each have their own advantages and disadvantages. For example, circular dichroism is commonly used, as it is a quantitative technique that is sensitive to helix content that can give information about secondary protein structure, but formulation buffers can interfere. Fourier transform infrared spectroscopy (FTIR) is less sensitive to buffer content, and provides an insight into secondary structure, including sheet content. A range of potentially useful qualitative techniques includes fluorescence spectroscopies and differential scanning calorimetry. Recently, techniques more commonly used in research, such as 2D protein NMR, are being used to look at biosimilar protein structure in comparison to originator product.
Aggregation is a serious issue which also needs to be assessed. A number of biophysical techniques (Table 4) are available, ranging from simple and cheap sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS PAGE) through to complex quantitative methods such as Sedimentation Velocity Analytical Ultracentrifugation (SV-AUC), based on ultracentrifugation, and size-exclusion chromatography with multi-angle light scattering (SEC-MALS).
Although this article has concentrated on physicochemical assessment, the requirement for comparative functional assays must not be forgotten. ICH Q6B describes both potency and biological activity assessment as essential characterization steps. It is also important to link the physicochemical studies with functional attributes. The choice of biological assay will obviously depend on the nature of the biosimilar molecule and may involve cell-based potency and mechanism of action assays.
The overall picture
Fundamentally, when proving biosimilarity, the strategy has to be to select from the many techniques available that together give a sufficiently detailed picture of a biosimilar’s structure and activity, and how well it correlates to the reference product. This must clearly demonstrate to the regulator that there are no “residual uncertainties” that the biosimilar will behave in similar fashion to the original when administered to patients.
It is interesting to look at the first product to gain FDA approval as an example. Sandoz gained FDA approval for its Zarxio biosimilar to Neupogen on the back of a range of studies. These include Edman degradation, peptide mapping, and mass spectrometry to compare the amino acid sequences, an in-vitro proliferation bioassay to assess potency, surface plasmon resonance to study target binding, and protein concentration content determination. These represent those quality attributes deemed to have “very high” criticality for efficacy, safety and immunogenicity. “High” criticality attributes included subvisible particles, measured by light obscuration; oxidized variants, studied via reverse phase chromatography; higher order structure, assessed by circular dichroism and NMR spectroscopy; and high molecular weight variants / aggregates assessed using size exclusion chromatography. Less critical were truncated variants, norleucine and deamidation, all of which were assessed using some form of chromatography. More than 80 batches of products were analyzed during development, with the final head-to-head comparison carried out using six batches of Zarxio, four of U.S.-licenced and two of EU-authorized Neupogen.
It is important to remember that the development pathway for a biosimilar is not the same as that for a novel molecule. The primary aim is not so much establishing activity and safety; rather, it is proving that the biosimilar is so close in structure to the originator that it will have the same activity and safety attributes as the marketed product, with sufficient certainty to convince the regulatory authorities.
Careful study of the originator is as important as the study of the biosimilar. Multiple orthogonal analytical methods are required to establish this comparison, including strategies for both primary and higher order structure, and there is an increasing focus on linking the structure to its biological activity.
- US-FDA: Biosimilars http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm290967.htm
- EMA Multidisciplinary: Biosimilars http://www.ema.europa.eu/ema/index.jsp?curl=pages/regulation/general/general_content_000408.jsp&mid=WC0b01ac058002958c
- ICH Q6B Guidelines http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q6B/Step4/Q6B_Guideline.pdf
Dr. Greer was a founding director of M-Scan, pioneering new developments, particularly mass spectrometry, for analysis and sequencing of glycoproteins. For over 35 years she has characterized a range of biotechnology products. In 2010, she became global director, biopharma services development, SGS Life Sciences.