Challenges in Comparability Studies of Glycoprotein Biosimilars

As defined by the US Food and Drug Administration (FDA) “a biosimilar product is a biological product that is approved based on showing that it is highly similar to an FDA-approved biological product, known as a reference product, and has no meaningful differences in terms of safety, purity and effectiveness from the reference product.”¹ In Europe a legal framework for approving biosimilars was established in 2003. The growth hormone Omnitrope (somatropin) was the first biosimilar drug approved by the European Medicines Agency (EMA) in 2006. Twenty-one biosimilar products are currently on the market in Europe. In the U.S, the passage of the Affordable Care Act (ACA) in 2010 created an abbreviated licensure pathway for biological products classified as "biosimilars”. A generic version of Lovenox (enoxaparin sodium), a low molecular weight heparin made by chemically modifying heparin to make it smaller, was actually approved through the existing Abbreviated New Drug Approval (ANDA) process in 2010, and has been considered by some to be the first biosimilar in the US. Enoxaparin is probably the most complex drug on the market, and FDA has used five evaluation criteria in order to ensure that the active ingredient in generic enoxaparin is the same as that in Lovenox.

However, the first biosimilar drug approved under ACA was Sandoz’s Zarxio™, which was released for sale in March of 2015. Zarxio™ has the same amino-acid sequence as an existing protein therapeutic drug called Neupogen, made by Amgen, with a molecular weight of 18.8 kDa. The approval of the first protein biologic Zarxio™ as a biosimilar in the US will surely open the gateway for many more sponsors seeking approval of protein based biosimilars. It is also anticipated that a number of companies that have biosimilar monoclonal antibody (mAB) drugs already approved in the European market will be submitting applications to the US FDA.

While biosimilar products are similar to the original product, they are not their exact copies. In August 2015, the FDA published guidelines for “Scientific Considerations in Demonstrating Biosimilarity to a Reference Product."² In this article, the FDA expects that first a sponsor will first extensively characterize the proposed product and the reference product with state-of-the-art technology, because extensive characterization of both products serves as the foundation for a demonstration of biosimilarity.

Glycoprotein Biosimilars

In the case of glycoproteins such as mABs, the following criteria are specifically outlined in the FDA guidelines for the extensive characterization:

1. Primary structures, such as amino acid sequence
2. Higher order structures, including secondary, tertiary, and quaternary structure (encompassing aggregation)
3. Enzymatic posttranslational modifications, such as glycosylation and phosphorylation
4. Other potential variations, such as protein deamidation and oxidation
5. Intentional chemical modifications, such as PEGylation sites and characteristics.²

Although glycosylation is mentioned only as a subset of one of these criteria, it is probably the most challenging aspect of demonstrating biosimilarity for glycoprotein therapeutics. This is in part due to inherent heterogeneity of the protein glycosylation machinery. The complexity of glycoprotein structural characterization and comparability of glycan-containing biologics is derived from the fact that in contrast to non-glycosylated proteins glycoproteins are not discrete single molecules, but occur as heterogeneous mixtures of varying complexity. This heterogeneity is derived entirely from the glycan fragment and stems from its variable size, branching, linkage and substitution by non-carbohydrate moieties. To fully capture this often-considerable diversity and to characterize a reference product and its biosimilar counterpart, it is necessary to employ the appropriate experimental design and analytical methodologies. If correct experimental protocols are not followed, it is easy to miss the subtle differences that can exist in glycoform sub-structures and could have a significant impact on the glycoprotein, including its bioactivity, solubility, immunogenicity, stability against proteolysis, and clearance rate.

The next wave of biosimilars that are anticipated to be enter the US market are antibodies (ABs). ABs have a specific glycan modification in the Fc region that is critical for its immune effector functions. Glycan diversity can affect the Fc-dependent activities of these ABs, and any increased Fc sialylation can result in decreased binding to immobilized antigens and also decreased antibody-dependent cell-mediated cytotoxicity activity.³ In comparison, increased Fc sialylation adds to the anti-inflammatory activity of AB. Therefore it is important to be able to detect and monitor these glycan moieties that are shown to be important for an AB’s biological activity.

In designing a strategy to tackle comparability of glycoprotein glycosylation, the experimental design needs to incorporate specific protocols in order to be able to observe these differences (Figure 1). For example, below is a ten-step design platform to capture these subtle changes, bearing in mind that each step needs to be carefully considered and modified depending on the therapeutic glycoprotein of choice.

 Figure 1. Experimental designs for glycoprotein biosimilar analysis

• Monosaccharide composition by HPLC and GC-MS
• Glycosyl-linkage analysis by GCMS or exoglycosidase digestion followed by MS
• Complete release and isolation of N- and O-linked glycans
• Sequencing and identifying the type of complex glycans
• Accurate quantitation of released N- and O-linked glycans
• Identifying site(s) of attachment of N- and O-linked glycans
• Determination of glycan heterogeneity at each glycosylation site
• Quantifying the glycan occupancy at each glycosylation site
• Determining anomeric configuration of each residue
• Identifying and determining the points of attachment of non-carbohydrate constituents such as phosphate and sulfate to the glycan

Most FDA approved glycoprotein therapeutics are currently made in Chinese Hamster Ovary (CHO) cell lines and to a lesser extent in baby hamster kidney cells, murine myeloma, and hybridoma cell lines. CHO cell lines express complex type N-glycans capable of carrying α(2,3) linked sialic acids at the outermost positions of the glycan sequence and are known to express two sialic acids, namely N-acetylneuraminic acid (NeuAc) and the non-human N-glycolylneuraminic acid (Neu5Gc).⁴ Challenges become apparent from every aspect of trying to implement the above platform (Table 1), including the ability to identity and quantitate any CHO-derived non-human glycan epitopes that can be present, including terminal Gal-α1-3Gal and Neu5Gc.⁵ Other potential pitfalls are

Table 1. Glycoprotein analysis methods and their pitfalls

1. incomplete release of all glycans, especially the larger sialylated tetraanteannary structures
2. accurate detection and quantitation of α(2,3) linked sialic acids and possibly α(2,6) linked sialic
3. establishing the purity of enzymes used and to ensure there are no contamination from exo and endoglycosidase activity
4. comparable quantitation of high mannose glycan compared to sialylated complex types, since high mannose glycans can often be overestimated
5. accurate detection and quantitation of each glycan with terminal residue such as Gal, GalNAc, GlcNAc and NeuAc
6. being able to distinguish the possibility of having Glc addition to the non-reducing end of glycans
7. designing experiments not to overlook any non-carbohydrate moieties
8. comparable quantitation of glycan microheterogeneity at each site of N/O-glycosylation with overall glycan quantitation.

Each of the above possible pitfalls highlights the challenges of developing high-quality carbohydrate biosimilar candidates and the critical need for both expertise and modern analytical instrumentation. It is essential to recognize the limitations and shortcomings of these methods, which can lack specificity in the context of characterization and comparability studies. Failure to do so can result in significant delays and waste of resources during the project.

The key to overcoming some of these challenges is in twofold. First, building a team of experts with in-depth expertise and insight into glycoprotein characterization and the know-how of applying state-of-theart techniques that include high-resolution mass spectrometry and NMR spectroscopy. NMR is particularly powerful in detecting subtle product changes, which often require a rather involved series of 2D-NMR experiments that have the ability to differentiate glycan moieties that cannot be distinguished by MS or HPLC experiments. Too often high-end NMR experiments are neglected due to lack of access and expertise. Second, implementing a number of critical key experiments in glycan characterization at the start of the project and at regular time-points during the study to monitor reproducibility and stability of glycan production without loss to the protein yield.

An additional important step for the success of biosimilar product characterization is for the sponsor company to become engaged with the larger scientific community, inasmuch as the latter understands the challenges of glycoconjugate structural characterization as a whole. This is most easily accomplished by working in close connection with scientific non-profit organizations involved in characterization of therapeutic proteins and monoclonal antibodies. The University of Georgia’s Complex Carbohydrate Research Center (CCRC) is an organization dedicated to research, service and training in the structural characterization of complex carbohydrates derived from animals, bacteria, fungi, and plants. Activities at the CCRC include carbohydrate biosimilar product characterization, such as low molecular weight heparins and antibodies. The CCRC is a focal point in the glyco-community because it offers several specialized hands-on training courses in carbohydrate characterization using state-of-the art instrumentation and provides research service in complex carbohydrates, including covering glycomics and glycoproteomics. The CCRC works closely with other academic institutes, industry, and government agencies including the US Pharmacopeia (USP) and the National Institute of Standards and Technology (NIST). USP sets standards for the identity, strength, quality, and purity of medicines, and its drug standards are enforceable by the FDA. The NIST is a measurement standards laboratory that supplies users with over 1300 Standard Reference Materials. Together, these organizations have the resources that can help for the development of a highquality carbohydrate biosimilar candidate. The sponsor can work closely with these institutions to build a framework around acceptable data packages for comparability studies between the biosimilar product and the licensed biological product.

In conclusion, in view of the high cost of healthcare in the US, it is in the best interest of patients that more high quality generic biologics become available to consumers. However, it is essential for the safety of the patients that each biosimilar is thoroughly characterized, and specific experiments are designed as safety nets and are implemented as part of the characterization and release tests to ensure high quality biosimilars for the US market.

References

1. U. S. Food & Drug Administration: Biosimilars [http://www.fda.gov/Drugs/DevelopmentApprovalProcess/ HowDrugsareDevelopedandApproved/ ApprovalApplications/TherapeuticBiologicApplications/ Biosimilars/]
2. U. S. Food & Drug Administration: Scientific Considerations in Demonstrating Biosimilarity to a Reference Product [http://www.fda.gov/downloads/ Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/UCM291128.pdf]
3. Naso MF, Tam SH, Scallon BJ, Raju TS: Engineering host cell lines to reduce terminal sialylation of secreted antibodies. MAbs 2010, 2(5):519-527.
4. Ghaderi D, Zhang M, Hurtado-Ziola N, Varki A: Production platforms for biotherapeutic glycoproteins. Occurrence, impact, and challenges of non-human sialylation. Biotechnol Genet Eng Rev 2012, 28:147-175.
5. Ghaderi D, Taylor RE, Padler-Karavani V, Diaz S, Varki A: Implications of the presence of N-glycolylneuraminic acid in recombinant therapeutic glycoproteins. Nat Biotechnol 2010, 28(8):863-867.