Regulatory challenges for biosimilar mAbs

The development of hybridoma technology by Kohler and Milstein(1) in 1975 is a major landmark in the generation and production of monoclonal antibodies (mAbs). Even though this was initially used mainly for research purposes, it soon became clear that the properties of mAbs opened up new, unexplored therapeutic possibilities. First, mAbs are directed against one particular epitope in one particular target molecule, and therefore are highly specific (‘magic bullet’); second, mAbs exhibit particular effector functions through the Fc region; and third, mAbs can be raised and selected against virtually any putative target.

The first approved therapeutic monoclonal antibody was Muromonab-CD3 (Orthoclone Okt3®, anti-CD-3) authorised for the reversal of kidney transplant rejection.(2) Intrinsically associated with the procedure of the hybridoma technology, Muromonab-CD3 was of murine origin. Despite their high specificity towards the human target against which they are raised, mAbs of murine origin suffer from a high degree of immunogenicity in humans and lack adequate Fc-mediated effects in humans. To date, rDNA technology allows cloning of the monoclonal antibody encoding sequences from hybridomas as well as from any other origin, including human.

Subsequently, sequences encoding the antigen-binding portion of the (mouse) monoclonal can be recombined with cloned sequences encoding the human Fc-portion, resulting in the production of chimeric antibodies (for example, abciximab, rituximab, infliximab). More advanced strategies allow the generation of ‘humanised’ antibodies in which the CDR-regions of the murine monoclonal are grafted into a human framework (for example, palivizumab, trastuzumab, alemtuzumab, pertuzumab).

Developments in transgenic technologies led to the generation of transgenic mice containing the corresponding human antibody genes. Combination of the latter with hybridoma technology allows the direct generation of fully human antibodies (for example, panitumumab). Alternatively, phage-displayed human antibody fragment libraries, combined with cloning technology, also allow the construction of fully human mAbs (for example, adalimumab).

mAbs exert their pharmacological and therapeutic effects through a variety of mechanisms. They can neutralise the action of, and sequester, soluble targets (for example, anti-tumour necrosis factor alpha: infliximab, adalimumab, certolizumab, golimumab; anti-interleukin (IL)1b: canakinumab; anti-vascular endothelial growth factor: bevacizumab, ranibizumab). When targeting cell-bound receptors, mAbs can be used to deliver a toxin or radioactive label (for example, tositumomab-I131, ibritumomab-tiuxetan) to block the function of a receptor (for example, anti-IL2R: basiliximab, daclizumab; anti-RANK-L: denosumab) or to induce apoptosis (for example, anti-CD20: rituximab, ofatumumab; anti-HER2: trastuzumab, pertuzumab). In addition, therapeutic effects of mAbs may be mediated by antibody-dependent cell cytotoxicity (ADCC) and/or complement-dependent cytotoxicity (CDC).

Patent expiry of approved biopharmaceuticals opens the possibility for other biotech companies to ‘copy’ and market these biologicals, which might possibly, as with generics, reduce costs to patients and healthcare systems. However, biopharmaceuticals are made by living cells and, because of their intrinsic complexity and because no two cell lines, developed independently, can be considered identical, biopharmaceuticals cannot be fully copied.

The final biopharmaceutical product is influenced by many variables, including expression systems, growth conditions, purification process and formulation. Post-translational modifications, such as glycosylation, phosphorylation, sulphation, methylation, N-acetylation, hydroxylation, glycation and oxidation, may affect biological activity and result in an intrinsic molecular heterogeneity. Importantly, and in contrast to chemical drugs, biopharmaceuticals are potentially immunogenic.

In this respect, it is important to note that subtle structural differences (for example, consequent to small differences in the number or type of product variants) may significantly affect the immunogenic potential of the drug product.(3) Additionally, product- or process-related impurities can provoke an immune response.

The European regulatory authorities have introduced the term ‘biosimilar’ in recognition of the fact that biosimilar products are similar to the original product, but never exactly the same.(4,5) Therefore, the European Medicines Agency (EMA) has issued a number of general and product-specific guidelines to be taken into account when developing biosimilars.(6)

The concept of biosimilarity is based upon comparability studies: an extensive full comparison (relative to an authorised reference product) at the level of structural biochemical and functional properties and a reduced non-clinical and clinical evaluation. To date, only biosimilars of relatively small (20-30kDa) biologicals have been authorised: somatropin, epoetin alfa and filgrastim. No biosimilar mAbs (size 150kDa) have been approved yet but two biosimilar mAbs (reference product: Infliximab, Remicade®) are currently under evaluation at the EMA.

When considering the development of biosimilar mAbs, it is important to realise that they comprise multiple domains that contribute to their mode of action and affect their clinical properties. The Fab-region contains the variable domains responsible for the specific interaction with the target. The Fc-region plays an important role in ADCC and in CDC and can exert other general regulatory effects on the cell cycle by triggering signalling pathways. Importantly, the Fc-region is glycosylated, and both type and extent of glycosylation play an important role in the effector function and on the clearance. The glycosylation of the Fab region should also not be ignored.(7) Therefore, evaluation of biosimilar mAbs (including comparability exercises of quality attributes) should not only include antigen binding (Fab) but also Fc-mediated functions (for example, binding to FcγR, FcRn, complement).

Fab-associated functions should not only be restricted to antigen binding but also include the expected functional effects on the target (for example, neutralisation, receptor blocking, receptor activation). Because of this complexity, biosimilarity of mAbs should not be demonstrated merely on an in vitro biochemical evaluation but also on an extensive in vivo functional evaluation. The latter, however, is very much limited because of the lack of appropriate animal models (that is, species specificity of antigen and of Fc-binding partners hampers a full evaluation). Taking all these factors into account, it is clear that the development of a biosimilar mAb is much more demanding than the development of a simple biological as has been the case to date. Not surprisingly, the EMA has issued, very recently, specific guidelines on biosimilar mAbs.(8)

According to this guideline, a first step is the evaluation of particular quality attributes with respect to binding and functional characteristics. Therefore, in vitro studies are required in which the biosimilar and the reference compound are compared to each other with respect to: (a) binding to the target antigen; (b) binding to representative isoforms of the relevant three Fc gamma receptors (FcγRI, FcγRII and FcγRIII), FcRn and complement (C1q); (c) Fab-associated functions (neutralisation of a soluble ligand, receptor activation or blockade); and (d), Fc-associated functions (ADCC, CDC, complement activation).

It is important to note that the guideline specifies that the functional assays should be designed to allow the detection of minute differences in the concentration-activity relationship between biosimilar and reference. In view of the fact that animal models may not be adequate, these extensive in vitro characterisation assays (using target and Fc receptors of human origin) are of crucial importance because they are usually more specific, more sensitive and more representative for the ‘human situation’. The guideline explicitly states that: “If the comparability exercise using the above strategy indicates that the test mAb and the reference mAb cannot be considered biosimilar, it may be more appropriate to consider developing the product as a stand-alone.”(8)

It needs to be considered whether there is a requirement for in vivo non-clinical testing. This evaluation is based upon the presence of relevant quality attributes that have not been detected in the reference product (for example, other post-translational modifications), presence of quality attributes in significantly different amounts than those measured in the reference product, or relevant differences in formulation.

If no concerns are identified, in vivo animal studies may be deemed unnecessary. If critical elements have been detected in the in vitro comparability exercises, then relevant in vivo studies should be designed. A major hurdle is the search for a suitable species because of the high (species) specificity of mAbs, both at the level of the Fab and Fc domain. In most cases, non-human primates, transgenic animals or (human) transplant models may be appropriate, even though they still include limitations.

One of the major concerns of adverse reactions is immunogenicity. On the one hand, the immunogenicity is influenced significantly by the clinical context; on the other, the clinical impact of immunogenicity very much depends on the binding site, titre, affinity and duration of immune response. To date, no appropriate animal model to predict immunogenicity in humans is available. Therefore, initial immunogenicity assessment is mainly based on a risk-based approach and implies an important post-marketing vigilance plan.(9)

Clinical evaluation of the biosimilar mAb requires a comparative analysis between the biosimilar and the reference product. The study design for clinical pharmacokinetic analysis should take numerous factors into account (for example, long half-life, immunogenicity, disease and patient characteristics, and pharmacokinetics of reference). Dosing should be selected on the basis of the sensitivity to detect possible differences and preferably all routes of administration should be investigated. Pharmacodynamic markers might be sensitive to detect differences and a set of markers should be evaluated to provide solid evidence of comparability.

A dose-response or time-response may provide a pivotal proof of comparability. Ultimately, clinical efficacy needs to be evaluated. Importantly, the guiding principle is to demonstrate similarity of the biosimilar to the reference, and not to evaluate the patient benefit. Extrapolation of clinical safety and efficacy data to other indications might be considered, but only if scientifically justified (taking into account the molecular mechanism as well as the relative contribution of Fab and Fc).

In general, when copies of chemical drugs have been approved, approval was based on demonstrated bioequivalence compared with the reference product. Having an identical structure and a proven bioequivalence implies, in most cases, that a generic and reference product, as well as two generics, are interchangeable. For biopharmaceuticals, however, the situation is completely different because two independently developed biopharmaceuticals demonstrated to be bioequivalent will not have identical pharmaceutical quality attributes and therefore cannot be considered interchangeable in the absence of evidence gathered from adequately designed clinical studies.

Indeed, potential differences in immunogenicity can only be observed in large study populations and switching between biopharmaceuticals from different origins may increase the risk to antibody development (for example, epitope spreading). It should also be realised that, in contrast to various generics from the same reference product, which can be considered identical, two biosimilars, independently developed and compared with the same reference product, cannot be considered biosimilar to each other.

In addition, claiming a general interchangeability between biosimilar and reference product based merely on biosimilarity would, in practice, also result in an extrapolation to a possible interchangeability between biosimilars. Thus, from a scientific point of view as well as for the sake of patient safety, biopharmaceuticals (irrespective of their regulatory status as biosimilar or reference) should not be considered interchangeable in the absence of solid clinical data. This is also enforced in the US Health Care Reform Bill, clearly stating that more data are required for a product to be labelled interchangeable rather than the mere fact of being biosimilar.(10,11)

It must also be stressed that, if interchangeability has been proven between two biopharmaceuticals (for example, between two biosimilars or between a biosimilar and its reference), this remains strictly valid only for the two specific products that have been evaluated. Importantly, within this context the life cycle of a biopharmaceutical (reference as well as biosimilar) should also be taken into account. Indeed, shifts in quality attributes may occur consequent to changes in the manufacturing procedure. From a practical point of view, such changes in a biopharmaceutical will result only in a unidirectional switch for the patients (old to new) with virtually no risk for the patient. However, shifts being introduced independently in the reference and biosimilar implies that, over time, both may divert from each other to such an extent that biosimilarity disappears.

In practice, it is impossible to apply the concept of interchangeability to biopharmaceuticals, even in the event that, at a given time point, two products would have been shown to be interchangeable. It should be stressed that lack of interchangeability in this context does not originate from any concern with respect to absence of comparable efficacy or safety in general. Lack of interchangeability originates from the uncertainty of whether or not the risk associated with switching back and forth (the consequence in real-life of being interchangeable) between two biopharmaceuticals during a (chronic) treatment is greater than the risk of using one biopharmaceutical without such switching.(10)

It is noteworthy that this concern is even more pronounced in situations where different biosimilars relative to the same reference product are available. Therefore, stimulation of market introduction of biosimilars should be restricted to new patients. This implies that uptake of biosimilar mAbs can be expected to be much slower in pathologies that require chronic treatment compared to applications in acute treatments.

• A biosimilar, an approved copy of an authorised reference biopharmaceutical, is similar, but not identical, to the reference product.
• To date, only biosimilars of relatively small biologicals have been authorised in Europe.
• Two biosimilar monoclonal antibodies (mAbs) of infliximab are currently under evaluation at the European Medicines Agency.
• Pharmacological and clinical properties of mAbs are determined by multiple functional domains (antigen binding domain and Fc domain). Demonstration of biosimilarity of mAbs requires an extensive comparative evaluation of these different domains.
• The pronounced species-specificity of mAbs precludes reliable in vivo non-clinical evaluation and implies the requirement of adequately designed clinical studies to demonstrate biosimilarity.
• Biosimilarity does not imply interchangeability.

1. Köhler G, Milstein C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975;256:495-7.
2. Reichert JM. Marketed therapeutic antibodies compendium. MAbs 2012;4:413-5.
3. Sauerborn M et al. Immunological mechanism underlying the immune response to recombinant human protein therapeutics. Trends Pharmacol Sci 2010;31:53-9.
4. European Medicines Agency. Guideline on similar biological medicinal products. www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003517.pdf; currently under revision www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2013/05/WC500142978.pdf (accessed 30 May 2013).
5. Declerck PJ. Biotherapeutics in the era of biosimilars: what really matters is patient safety. Drug Saf 2007;30:1087-92.
6. Reichert JM. Next generation and biosimilar monoclonal antibodies: essential considerations towards regulatory acceptance in Europe. MAbs 2011;3:223-40.
7. Jefferis R. Isotype and glycoform selection for antibody therapeutics. Arch Biochem Biophys 2012;526:159-66.
8. European Medicines Agency. Guideline on similar biological medicinal products containing monoclonal antibodies – non-clinical and clinical issues. www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/06/WC500128686.pdf (accessed 30 May 2013).
9. European Medicines Agency. Guideline on immunogenicity assessment of monoclonal antibodies intended for in vivo clinical use. www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/06/WC500128688.pdf (accessed 30 May 2013).
10. US legislation HR 1548. Pathway for Biosimilars Act. www.govtrack.us/congress/bill.xpd?bill=h111-1548 (accessed 30 May 2013)
11. Chow SC et al. Scientific considerations for assessing biosimilar products. Statist Med 2013;32:370-81.