Immunogenicity and glycosylation: The key issues for biosimilars Huub Schellekens Utrecht University

Immunogenicity and biotech comparability Current analytical methods cannot fully predict biological properties The immune system can detect alterations in products missed by analytical methods Immunogenicity of biopharmaceuticals may have serious clinical consequences

History of the medical use proteins Proteins of animal origin (e.g. equine antisera, porcine/bovine insulin): foreign proteins Human derived proteins (e.g.growth hormone, factor VIII): no immune tolerance Recombinant human proteins(e.g.insulin, interferons, GM-CSF): ??

Most biopharmaceuticals induce antibodies Two mechanisms Reaction to neo-antigens Breakdown of immune tolerance

Types of immune reaction against biopharmaceuticals Breaking of self-tolerance Type of product Human homologues Characteristics of antibody production Slow, after long treatment, binding antibodies, disappear after treatment Cause Mainly impurities and aggregates

Factors influencing immunogenicity Structural properties Sequence variation Glycosylation Other factors Assays Contaminants and impurities Formulation Downstream processing Route of application Dose and length of treatment Patient characteristics Unknown factors

Consequences of antibodies Loss of efficay  Insulin Streptokinase Staphylokinase ADA Salmon calcitonin Factor VIII Interferon alpha 2 Interferon beta IL-2 GnRH TNFR55/IgG1 Denileukin diftitox HCG GM-CSF/IL3 Enhancement of efficacy Growth hormone Neutralization of native protein MDGF EPO General immune effects Allergy Anaphylaxis Serum sickness, etc

Pure red cell aplasia associated with EPO treatment Data from Nicole Casadevall

Bone Marrow Smear Normal Bone Marrow PRCA Bone Marrow

Pure red cell aplasia associated with anti-EPO antibodies Nicole Casadevall - 1996 PRCA case with natural antibodies 2002 13 cases with antibodies associated with epoetin treatment

Why was Eprex implicated? High association between Eprex and PRCA Geographic distribution Association with formulation change

Time course of individual PRCA cases Since Dec 93 t Epo-refractory anemia (diagnosis) l Pure Red Cell Aplasia (diagnosis) Since Feb 93 Epoetin alfa SC Eprex Epoetin alfa IV Eprex Epoetin beta SC NeoRecormon Darbepoetin Since Dec 95

Product formulation Recent concern over use of HSA in Europe because of potential transmission of infectious viruses or BSE prions In 1998, HSA was replaced with polysorbate 80 in prefilled syringes of Eprex® distributed ex-US

Main stabilizers used in the epoetin formulations Epogen®/Procrit® (US) Eprex® (pre 1998) Eprex® (post 1998) NeoRecormon® (1990 launch) HSA HSA Polysorbate 80 Polysorbate 20 Glycine Glycine Complex of 5 other amino acids Calcium chloride Urea

Factors potentially contributing to the immunogenicity of Eprex® Formation of micelles associated with Epo (Hermeling et al, 2003) Silicon droplets in the prefilled syringes Leachates from rubber stoppers Mishandling

What is the role of micelles? Very unstable No biological data Does not explain epidemiological data

Silicon as adjuvant Lot of confusion data in the literature Silicon is inert

The evidence for this leachates hypothesis is based on high-pressure liquid chromatography of formulated EPREX®/ERYPO®. Comparison of the elution profiles of solutions from syringes with uncoated or coated rubber plungers shows the presence of many additional peaks when the plunger is not coated. The compounds indicated by these peaks are low-molecular-weight elastomers, curing agents, antioxidants, and plasticizers, all of which are normal constituents of synthetic rubber. Several of these types of molecule, particularly phthalates (used as plasticizers), have been shown in animal experiments to act as adjuvants for immunogenic responses.1–3 References 1. Larsen ST, et al. Toxicology 2001;169:37-51. 2. Larsen ST, et al. Pharmacol Toxicol 2002;91:264-72. 3. Larsen ST, et al. Food Chem Toxicol 2003;41:439-46.

The leachate theory No biological rationale Adjuvants do not break B cell tolerance No experimental data showing breaking tolerance Does not explain epidemiological data and pathogenesis

Mishandling Mishandling with a slightly less stable product may explain all features of PRCA Biological rationale Fits with data concerning other product Fits the pathogenesis Fits with the epidemiological data

Prediction of immunogenicity Purity of the product Epitope analysis Reaction with patient sera Animal experiments Convential animals (relative immunogenicity) Non-human primates Immune tolerant transgenic mice

Daily i.p.

Reducing immunogenicity Optimizing production,purification and formulation Changing sequence (streptokinase, staphylokinase) Pegylation (ADA)

Also an important issue in the biosimilar discussion Glycosylation Also an important issue in the biosimilar discussion

Glycosylation of biosimilar epoetins can be expected to be different What types of glycosylation are there? What is the biological significance of glycosylation

O-linked glycosylation O-linked to specific serine or threonine but consensus sequence not identified Apparently defined by secondary structural elements like β turn. Start with the attachment of a single monosaccharide normally N-acetylgalactoseamine Then extended by glycosyltransferases

N-linked glycosylation N-linked to Asn-X-Ser/Thr Most consensus sequences non-glycosylated. Depends on secondary structures. Glycosylation before folding Starts with binding of DTP-oligosaccharide: 2 GlcNAc, 9 mannose and 3 glucose molecules. Trimming by removing glucoses and mannoses and possible adding of GlcNAc

The functions of the glycocomponent Protein folding Protein trafficking Protein targeting Ligand recognition Ligand binding Biological activity Stability Pharmacokinetics Immunogenicity

Glycosylation of epoetin 40% sugar Three linked N-glycosylation sites at Asn 24, 38 and 83 One O-linked site at serine 126 Heterogeneity caused by variation in core structures and sialic acid Removal of N-glycosylation sites has no effect on in vitro activity, but greatly reduces the in vivo activity Half life in rodents IV 5-6h but < 2 min if desialylated Adding N-glycosylation sites increases half-life. Single O-linked side chain removal has little effect ?

Aberrant glycosylation biosimilar epoetins Retracrit: > glycoforms without O-glycans. < N-glycolyl and 0-acetyl neuraminic acid Epoetin alpha Hexal: > high mannose

Epoetin lacking O-linked glycosylation About 20% lesser activity Delorme at al. Biochemistry 1992 Explanation for the lower activity of Retracrit?

Conclusion The clinical consequences of immunogenicity may be severe Only clinical trials decisive to reveal immunogenicity The main difference between biosimilars is glycosylation Clinical consequences of differences in glycosylation unknown

What are the unanswered questions? What is biosimilar? Naming Label Safety monitoring Sensitivity Background data Standardization Price Counterfeits