You are here

Key considerations in the preclinical development of biosimilars

Drug Discovery Today, May 2015, Pages 3 - 15



  • Preclinical data for the development of biosimilars are different from those obtained in the development of the reference biologic.
  • The principles and processes for demonstrating biosimilarity involve a stepwise process of comparing the proposed biosimilar with its reference biologic product.
  • The goal of biosimilar development is to match the proposed biosimilar quality target product profile (QTPP) to that of the reference biologic product.
  • It is hoped that biosimilars will help increase therapeutic options and patient access to important biologics for the treatment of various diseases, ranging from rheumatoid arthritis to cancer.

Biosimilar development requires several steps: selection of an appropriate reference biologic, understanding the key molecular attributes of that reference biologic and development of a manufacturing process to match these attributes of the reference biologic product. The European Medicines Agency (EMA) and the FDA guidance documents state that, in lieu of conducting extensive preclinical and clinical studies typically required for approval of novel biologics, biosimilars must undergo a rigorous similarity evaluation. The aim of this article is to increase understanding of the preclinical development and evaluation process for biosimilars, as required by the regulatory agencies, that precedes the clinical testing of biosimilars in humans.

Biosimilar development: Overview

The development process for biologics requires that they are produced and purified from living systems such as bacteria, yeast or mammalian cell lines [1] . By contrast, conventional small-molecule drugs, such as aspirin, are typically synthesized using relatively simple chemical processes [2] and [3]. The patent expiration of small-molecule drugs enables pharmaceutical companies to synthesize identical copies of these molecules chemically and eventually make available generic versions of these drugs (i.e. identical copies of the patent-expired small-molecule drugs) [4] . Patent protection for several key biologics will expire in the near future [5] and [6]. The exact manufacturing processes, such as specific cell lines and culture conditions, through the final processing steps, inevitably dictate the resulting unique quality characteristics of a biologic drug [3] . Manufacturing processes for biologics are typically proprietary, and without access to this information other pharmaceutical companies have to produce biosimilars using existing technology [7] . In addition, the manufacturing of biologics involves an inherent degree of variability, which can add to the technical complexity and challenge in producing a highly similar copy for biosimilars and pose a challenge for innovator companies making an internal process change as well (comparability exercise; see Glossary ) [6] and [7]. As such, the term ‘biosimilar’ is used to describe a biologic that is considered ‘highly similar to a reference biologic product notwithstanding minor differences in clinically inactive components’, and for which there are ‘no clinically meaningful differences between the biologic product and the reference product in terms of the safety, purity, and potency of the product’ [6] . Because they are not identical, biosimilars cannot be considered generic versions of their reference biologic products [8] .

This review article is divided into two parts: part 1 briefly reviews the regulatory guidance documents for biosimilar development with a focus on preclinical topics, as well as the overall goals and principles for the development and evaluation of biosimilars. Part 2 provides a detailed review of the preclinical development of biosimilars, in particular monoclonal antibodies (mAbs), up to the point of first-in-human, Phase I clinical trials.

Part 1: biologics and biosimilars – understanding the basics of development

Key differences between biologics and small-molecule drugs

There are numerous differences between biologics (such as proteins, which are defined by the FDA as being greater than 40 amino acids) [9] and small-molecule drugs, as outlined in Table 1 [1], [7], [8], [10], [11], [12], [13], [14], and [15]. For example, biologics are proteins that are typically much larger than chemically synthesized small-molecule drugs, exhibit a high degree of structural complexity, including primary, secondary, tertiary and possibly quaternary structures, and are subject to post-translational modifications [14] . Another important characteristic of these biologics is that they are potentially immunogenic, that is, capable of eliciting an immune response when administered [11] .

Table 1 Key differences between biologics and small-molecule drugs [1], [7], [8], [10], [11], [12], [13], [14], and [15].

Biologics Small-molecule drugs
• Protein is produced in living cell expression systems (typically uses recombinant DNA technology); conditions of biosynthesis and purification cannot be exactly duplicated; cell-based, process-related variability is expected • Product can be reliably and consistently synthesized using standard chemical reactions
• Often large, complex molecules, with primary, secondary, tertiary, quaternary structure • Generally small molecules with simple structures
• Multiple isoforms of the product can exist in a single preparation; not possible to characterize the mixture fully with a single test • Molecular structure is well defined and a preparation can be purified to homogeneity, if required for activity
• Stability of product can be impacted by multiple factors (e.g. light, heat, agitation, etc.); product characteristics can also change (drift) over time as the cell lines evolve in culture and over the manufacturing process lifecycle • Product usually is stable under most conditions
• Not possible for competitor pharmaceutical companies to replicate branded biologics using current technology; production process often is proprietary; exact copy biologics are not possible • Generic versions (exact copies) can be produced more easily by competitor pharmaceutical companies using widely available nonproprietary technology

The manufacturing process for biologics is more complicated than that required for small-molecule drugs. Unlike small-molecule drugs, which are produced through chemical synthesis, biologics are manufactured in living systems, and their development typically involves cloning and expressing the gene sequence of interest into a specific cell type, fermentation and purification of the final product [1] . Manufacturing of biologics can undergo changes during the process (e.g. because of regulatory requirements or a need to improve manufacturing efficiency, product quality and/or yield); such changes could impact the quality, safety and efficacy of the final product [13] and [16]. Although usually without clinical consequence, changes in the manufacturing of a biologic can necessitate a preclinical and clinical re-evaluation of the biologic, termed a comparability exercise [13] and [17]. The International Conference on Harmonization (ICH) Q5E guidelines outline the necessary steps to compare batches of biologics to evaluate whether a manufacturing process change could have an adverse impact on product quality, efficacy or safety, including immunogenicity [13] and [16]. This evaluation can be based on an analysis of product quality attributes and, in some cases, supporting preclinical and/or clinical data [16] and [17]. Finally, the characteristics of biologics themselves, even under the same manufacturing conditions, can change over time. This phenomenon is known as drift [11] .

Current biosimilars environment

The European Medicines Agency (EMA) adopted the first general biosimilars directive in 2005, and the FDA released draft guidelines for the development of biosimilars in 2012 [9] and [18] ( Box 1 ). A general global guidance for biosimilar development was also published by the World Health Organization (WHO) in 2009 [5] , with the aim of providing a harmonized developmental framework for regulatory agencies to use in their national guidelines. Globally, the WHO biosimilar guidelines have been adopted by many countries into their national regulations.

Box 1 Key regulatory guidance documents


 • Guideline on similar biological medicinal products. Issued October 2005.
  ∘ Introduces the concept of biosimilar products and outlines the general scientific approach needed to demonstrate biosimilarity [18] .
 • Guideline on similar biological medicinal products containing monoclonal antibodies – nonclinical and clinical issues. Issued December 2012.
  ∘ Outlines preclinical and clinical requirements for demonstrating that the proposed biosimilar product is similar to another previously authorized mAb product [24] .
 • Other product-specific guidelines are available from the EMA website at:
 • Guidance for industry: scientific considerations in demonstrating biosimilarity to a reference product. Issued February 2012.
  ∘ Provides draft guidance on the FDA's ‘totality of the evidence’ approach toward demonstrating biosimilarity. The stepwise approach as outlined proceeds from structural and functional analysis, animal testing and human PK/PD studies to clinical assessments of immunogenicity, safety and efficacy [9] .
 • Guidance for industry: quality considerations in demonstrating biosimilarity to a reference protein product. Issued February 2012.
  ∘ Provides draft guidance on analytical studies needed to determine whether a proposed biosimilar is highly similar to its reference biologic as part of a biosimilarity assessment [6] .
 • Guidance for industry: biosimilars – questions and answers regarding implementation of the Biologics Price Competition and Innovation Act of 2009. Draft guidance issued February 2012.
  ∘ Provides answers to commonly asked questions from biosimilar developers and others regarding the agency's interpretation of the Biologics Price Competition and Innovation (BPCI) Act of 2009, a part of the Affordable Care Act [36] .
 • Guidelines on evaluation of similar biotherapeutic products (SBPs). Enacted October 2009.
  ∘ Intended to provide a global framework for licensing of biotherapeutic products (i.e. biosimilars) proposed to be similar to an existing biologic (i.e. the reference biologic), which has been licensed based on quality, efficacy and safety data from a full licensing dossier [5] .
 • Q5E: comparability of biotechnological/biological products. Enacted June 2005.
  ∘ Provides guidance on evaluating comparability of biologics before and after changes in the manufacturing process, to demonstrate that such changes in process have no adverse impact on quality, safety and efficacy of the product [16] .

Currently, the EMA has granted approval of more than a dozen biosimilars manufactured by at least ten different companies [19] ( Table 2 ). Although most of the EMA-approved (e.g. authorized) biosimilars to date are growth factors (small proteins with relatively simple molecular structures), the development of more structurally complex biologics is now possible; indeed, the European Union (EU) recently approved a biosimilar version of infliximab, a mAb, under the trade names Remsima™ and Inflectra™ [20] .

Table 2 Status of oncology biosimilar applications submitted to the EMA [19] .

Biosimilar Active substance Status
Abseamed ® Epoetin alfa Authorized 28 August 2007
Binocrit ® Epoetin alfa Authorized 28 August 2007
Epoetin Alfa Hexal ® Epoetin alfa Authorized 28 August 2007
Retacrit ® Epoetin zeta Authorized 18 December 2007
Silapo ® Epoetin zeta Authorized 18 December 2007
Biograstim ® Filgrastim Authorized 15 September 2008
Filgrastim ratiopharm Filgrastim Withdrawn 15 September 2008
Ratiograstim ® Filgrastim Authorized 15 September 2008
Tevagrastim ® Filgrastim Authorized 15 September 2008
Filgrastim Hexal Filgrastim Authorized 06 February 2009
Zarzio ® Filgrastim Authorized 06 February 2009
Nivestim™ Filgrastim Authorized 08 June 2010
Grastofil ® Filgrastim Authorized 18 October 2013
Ovaleap ® Follitropin alfa Authorized 27 September 2013
Inflectra™ Infliximab Authorized 10 September 2013
Remsima™ Infliximab Authorized 10 September 2013
Alpheon Recombinant human interferon alfa-2a Refused 05 September 2006
Omnitrope ® Somatropin Authorized 12 April 2006
Valtropin ® Somatropin Withdrawn 24 April 2006

Preclinical development of biosimilars

The goal of biosimilar development is to produce a biologic that is ‘highly similar’ to, and exhibits no clinically meaningful differences from, the reference biologic product in terms of ‘safety, purity and potency’ [6] . The term ‘preclinical’ as used herein encompasses a range of studies including: comprehensive analytical characterization; structural and in vitro functional assessments; studies of mechanisms of actions (MOA); and any in vivo pharmacokinetic (PK), pharmacodynamic (PD) and/or safety and immunogenicity assessments in animals that might occur prior to the first clinical evaluations in humans. The development of biosimilars of mAbs will be highlighted throughout this article. Note that the preclinical programs designed to evaluate biosimilars that are from other types of branded biologics, such as interferons and erythropoietins, can differ from those of biosimilar mAbs and are not the focus of this article.

Pathways to approval: differences between novel biologics and biosimilars

Whereas the typical pathway for a new biologic heavily emphasizes the endpoints of clinical evaluations relating to the demonstration of efficacy and safety in humans, the biosimilar development pathway, as defined by the EMA, FDA and WHO, is more grounded in head-to-head analytical, in vitro functional and preclinical in vivo comparisons of the biosimilar to its reference biologic product ( Fig. 1 ). This pathway, therefore, includes an extensive analytical and preclinical evaluation, which facilitates an abbreviated clinical evaluation, assuming a sufficient degree of similarity has been demonstrated in earlier steps [10] .


Figure 1 New biologic versus biosimilar drug development. In contrast to the development program for a new biologic, the preclinical phase of the biosimilar development program is more comprehensive. Conversely, fewer data are required for the clinical phase of the biosimilar development program as a result of establishing adequate similarity between the reference biologic product and the biosimilar in the preclinical evaluation phase. Toxicology studies differ between a biosimilar development program and a new biologic development program. In contrast to a new biologic development program, a biosimilar development program might contain at least one toxicology study, assessing safety, pharmacokinetic (PK), immunogenicity and any pharmacodynamic (PD) effects. Safety pharmacology, reproductive toxicity, genetic toxicity and carcinogenicity studies are not routinely required in a biosimilar development program.

Principles for demonstrating biosimilarity

Although the process for demonstrating similarity between a biosimilar and its marketed reference biologic product has some parallels to the comparability exercise undertaken following a manufacturing process change to a biologic, the procedures outlined in ICH Q5E guidance for demonstration of comparability ( Box 1 ) might not be sufficient to demonstrate biosimilarity [9] and [10]. A larger data package comparing the biosimilar with the reference product will generally be required and, for the purposes of this article, the term ‘similarity evaluation’ will be used to refer to the stepwise process of comparing a proposed biosimilar with its reference biologic product [9] and [13]. The similarity evaluation provides an agency-recommended approach to comparing the physicochemical and biologic activities of a biosimilar with its reference biologic product, with the goal that the established safety and efficacy data for the reference biologic product also applies to the biosimilar [21] . It is necessary, therefore, that the assessments, study components and endpoints chosen for the studies within the similarity evaluation are sensitive enough to detect all potentially clinically relevant differences between the biosimilar and the reference biologic product [5] and [21]. The importance and weight of the preclinical development program is highlighted by the EMA position that a comparative clinical efficacy study for relatively simple biosimilars (such as erythropoietin, granulocyte-colony stimulating factor, etc.) might not be necessary if similar biologic activity and potency to the reference biologic product can be documented with analytical and preclinical in vivo data in conjunction with comparative human PK data [21] and [22]. The EMA guidance and the draft FDA guidance on biosimilars outline a stepwise approach for the similarity evaluation, discussed in detail below [9] and [20].

Biosimilar mAbs: specific guidance

The function of therapeutic mAbs is dependent on their antigen-binding regions (Fab) and potential effector function regions (Fc), the activity of which can be dependent on post-translational modifications (e.g. glycosylation) [12] and [23]. Consequently, the EMA has issued specific guidance for the similarity evaluation required for mAb-based biosimilars [24] . As outlined in Table 3 , the preclinical guidance on mAbs states that preclinical in vivo studies might not be an absolute requirement for registration, as long as an extensive analytical characterization and in vitro functional evaluation of the biosimilar (relative to its reference biologic product) yields satisfactory findings of similarity [12], [23], and [25] ( Table 3 ). The EMA outlines a three-step approach toward the design and implementation of preclinical studies, with the first step consisting of in vitro analytical and functional studies, followed by an evaluation and decision regarding whether in vivo studies are required in step 2 and then, if necessary, in vivo studies are conducted in step 3. These in vivo studies are to be designed to be efficient from an animal utilization perspective and focused on addressing all unresolved differences between the biosimilar and the reference product from step 1 ( Table 3 ). Key functional assessments for mAbs include target antigen, Fc gamma receptor and complement binding assays, and the degree of antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent cytotoxicity [24] ( Table 3 ).

Table 3 EMA 2012 guidance for stepwise preclinical assessment of biosimilar mAbs [12] and [24].

Step 1: in vitro studies
• Designed to assess differences in biologic activity between the biosimilar and the reference biologic product
 ∘ Can be more sensitive to detect differences than animal studies
 ∘ Considered an essential component of the preclinical similarity evaluation
 ∘ Assess sufficient batches of product that are intended for use in subsequent clinical trials
 ∘ Should be sufficiently sensitive to assess possible differences in concentration–activity
• Assessments include measurement of:
 ∘ Target antigen binding
 ∘ Binding to all Fc gamma receptors, including the neonatal Fc (FcRn) and complement (C1q)
 ∘ Antigen binding (Fab) functions (i.e. ligand neutralization, receptor activation and/or blockade)
 ∘ Fc-associated functions [i.e. complement activation, antibody-dependent cell-mediated cytotoxicity (ADCC) activity]
• All possible functions of the mAb are covered, even if not necessarily therapeutically relevant
Step 2: considering the need for preclinical in vivo assessment
• Factors to consider when determining the need for additional in vivo testing
 ∘ Presence of biosimilar attributes that are distinct from the reference biologic product (e.g. a new post-translational modification)
 ∘ Presence of any quality attributes that differ in significant amounts between the biosimilar and the reference biologic product
 ∘ Any relevant differences in formulation (e.g. different excipients)
• These factors are considered collectively to evaluate the need for additional in vivo testing; each does not necessarily constitute a need for in vivo testing
 ∘ In vivo testing might not be needed if studies in step 1 are satisfactory and no factors of concern are identified
 ∘ If in vivo studies are considered necessary, availability of a relevant animal species must also be considered
Step 3: (if needed) preclinical in vivo studies
• Depending on the need for additional information, focus could be on pharmacokinetic (PK), pharmacodynamic (PD) or safety; designed to maximize the information obtained
 ∘ Principles of the 3Rs (animal replacement, refinement and reduction) should be applied; euthanasia of study animals might not be needed
 ∘ Justify study duration and observation period based on PK and clinical use of mAb
• If possible, compare PK and PD quantitatively (i.e. concentration dose assessment) at therapeutic dose
• Although not predictive of immunogenicity in humans, blood samples can be collected and stored if needed to interpret animal in vivo data
• If new excipients are used, it might be necessary to evaluate local tolerance
• Safety pharmacology and reproductive toxicology assessments are not considered relevant for biosimilar mAbs

Despite the availability of specific guidance from the EMA, regulatory guidance on the development of biosimilars, including mAbs, can vary across countries or geographic regions. For example, although the EMA considers differences in formulation as a possible area of consideration during the evaluation of biosimilar mAbs (in evaluating the need for in vivo studies; Table 3 ), it is possible that improvements in formulation science and/or excipients could have occurred in the years since the reference biologic product was approved. Differences in formulation could, therefore, be justified with a biosimilar, and the FDA allows for this possibility in its draft guidance [9] and [24]. Thus, understanding the guidance documents for specific countries or geographic regions is a crucial component of designing a biosimilar strategy.

Part 2: a closer look at the development process for biosimilars

The biosimilar development process: the importance of the reference biologic product

Selection of appropriate reference biologic products is an important consideration in the biosimilar development process. To generate an appropriate dataset for a biosimilar, the reference biologic product (typically sourced from a major market, e.g. USA, EU, Japan, etc.) could be purchased for the similarity evaluations [18] . The dosage and route of administration of a biosimilar are defined by the reference biologic product, which serves as the comparator for the similarity evaluations. Although EMA guidance states a biosimilar must have the same pharmaceutical form, strength and route of administration as the reference biologic product, some formulation changes might be permitted with sufficient justification and/or further studies [21] . Any intentional changes to the biosimilar designed to improve efficacy would not be consistent with the biosimilar development pathway; such a biologic would be considered a ‘biobetter’ and would be evaluated as a new biologic product [21] . The WHO guidance recognizes that the reference biologic product must be licensed and have a full registration dossier and data package [5] . For example, a reference biologic product should have been available on the market for a sufficient duration of time to allow the accumulation of a large body of clinical efficacy and safety data [5] . Whereas initial EMA guidance required the reference biologic product to have marketing authorization in the EU, non-EU authorized biologics can be used to avoid unnecessary duplication of clinical trials in global drug development programs [21] . However, because the reference biologic product might need to be purchased and thoroughly characterized by the biosimilar manufacturer, the availability and acquisition of the reference biologic product across multiple national health regulatory jurisdictions could be limiting factors to the development of a biosimilar in some parts of the world.

The biosimilar development process: identifying the quality target product profile (QTPP) of the reference biologic product

Identifying the key target quality attributes, or QTPP, of the reference biologic product, is another consideration in the development of a biosimilar. This involves a thorough characterization of multiple lots of the reference biologic product and the design of a manufacturing process that produces a biosimilar that closely reflects the reference biologic product, specifically those attributes most essential to appropriate clinical performance in terms of efficacy and safety [5] .

Some characteristics of a biologic can change over time, resulting from operational variations within a manufacturing process or upon storage [15] . Changes in post-translational modifications such as glycosylation are frequently observed during biologic drug manufacturing over time [26] . Consequently, testing multiple lots of a reference biologic product over a period of time is necessary to build a complete picture of the product quality profile, especially if there have been changes to the manufacturing process. The reference biologic product could also exhibit drift in QTPP over time [26] . In some cases, distinct and separate reference biologic product supplies exist in different regions, with the potential for differing product profiles. This can be determined through the analytical characterization of reference biologic product from more than one source region (e.g. USA and EU). In the example shown in Fig. 2 , variations are observed in major glycan (G1F) species across EU- and US-produced lots of a reference biologic product. This analysis demonstrates the need for appropriate selection of reference biologic product while developing an understanding of the regional differences (e.g. separate and distinct supply routes) that could exist globally.


Figure 2 Comparing lots of a reference biologic product: major or important terminal galactose N-glycan species from multiple reference biologic lots, single trade name (EU and US, arbitrary lot numbers). Lots of a reference biologic product from the EU and the USA are compared for the % terminal galactose species and % monogalactosylated biantennary complex N-glycoforms (G1F). Variability is seen within lots produced in the USA and EU. Different ranges of N-glycan profiles also are observed between the EU and US reference biologic products. This analysis demonstrates the variability in reference biologic products produced within a geographic region and in between geographic regions, stressing the need for appropriate selection of a reference biologic product in the demonstration of biosimilarity.

One approach to address potential QTPP drift in the reference biologic product relative to its biosimilar has been to define ‘goalposts’; that is, if the proposed biosimilar performs within the ‘established variation’ of the reference biologic product over time, it can be considered to be ‘highly similar’ based on the totality of evidence from physicochemical and functional evaluation, and preclinical and clinical studies during the biosimilar development [17] . Differences among reference biologic product lots can be identified using highly sensitive, state-of-the-art, robust analytical techniques. A comparison of molecular attributes of different lots of an EU-sourced reference biologic product over time is shown in Fig. 3 . In Fig. 4 , analysis using liquid chromatography/mass spectroscopy (LC/MS) is used to compare the masses of the reference biologic product from two different lots (US and EU reference biologic products).


Figure 3 Lot–lot variation: comparison of different lots of reference biologic product using cation exchange/high performance liquid chromatography (CEX/HPLC). Different lots of an EU reference biologic product (reference biologic lot A and reference biologic lot B) over time were compared by CEX/HPLC for key molecular attributes. As shown in the chromatograms, these key molecular attributes (such as the basic species composition) of the reference biologic product can drift (vary or change with manufacturing process changes) over time.


Figure 4 Lot–lot variation: comparison of different lots of reference product using subunit analysis. Different lots of a reference biologic product from the USA and EU were compared with regard to their N-glycan profile. The glycan profile is described as Man5 (high mannose 5), G0 (a galactosylated noncore fucosylated biantennary complex N-glycan), G0F (a galactosylated biantennary complex N-glycoforms), G1F (monogalactosylated biantennary complex N-glycoforms); G2F (digalactosylated biantennary complex N-glycoforms). The glycan profile of the reference biologic product from the EU and the USA are highly similar.

The biosimilar development process: matching the QTPP of the reference biologic product

The goal of biosimilar development is to match the QTPP of the reference biologic product. The draft FDA guidance recommends doing this by using elements of a quality-by-design (QbD) approach [6] . The manufacturing process is developed in accordance with the following [5] :

  • Established guidelines, such as those of the ICH and Good Manufacturing Processes (GMP).
  • Appropriate quality control and assurance procedures.
  • In-process controls with process validation.

The manufacturing process of a biosimilar during development is to be consistent with the overall ICH guidance for biotherapeutic drug development, although the manufacturing steps do not have to be the same as those of the reference biologic product [5] .

The design of the manufacturing process for a biosimilar can take into account the development history of the reference biologic product [5] . A ‘knowledge gap’ might still exist because some proprietary aspects of the manufacturing process for the reference biologic product could be unavailable to the biosimilar manufacturer (e.g. knowing or having the exact clone of the cell line used to express the protein, or details of the purification process) [13] . Given this, a robust and well-characterized cell line and/or expression system could be identified a priori, with cell culture conditions and a protein purification strategy producing a biosimilar closely matching the QTPP. Important QTPP parameters such as the glycosylation profile can be affected by the choice of expression system or particular cell line clone, the media and/or feed strategy used and specific bioreactor conditions. As such, the development of a biosimilar requires reiterative manufacturing process development using elements of QbD that ultimately will produce a product with QTPPs that are highly similar to the reference product [5] and [13].

All available evidence regarding the QTPP of the reference biologic product is assembled, understood and used during biosimilar development [5] . To summarize, the cell line(s) is chosen and engineered to express the proposed biosimilar (protein) of interest. The proposed biosimilar from several clones of this cell line is then evaluated for QTPP and the clone that has the best-performing product in the QTPP evaluation is chosen. At this point, the bioreactor conditions are manipulated to produce a proposed biosimilar that most closely resembles the reference biologic based on the QTPP evaluation. Multiple iterations can be necessary, which when combined with expert process knowledge and experience provide a robust process understanding to achieve the desired QTPP.

The biosimilar development process: importance of choosing an appropriate protein expression system

Advances have been made in understanding how the quality attributes of expressed recombinant proteins are affected by variables including the type of cell line and the cell culturing process used. If possible, the same protein expression system used to produce the reference biologic product can be used for the production of the proposed biosimilar. For example, if the reference biologic product is expressed in a mammalian cell line such as CHO cells then it would not be appropriate to express the biosimilar in a bacterial system, such as Escherichia coli, unless it can be convincingly demonstrated that such a change would not impact the QTPP of the proposed biosimilar. One specific attribute that could be affected by a change in expression systems is glycosylation [5] . Many of the most widely used mAbs are glycosylated proteins [27] . The glycosylation profile is a key consideration when designing the manufacturing process for a biosimilar; for example, certain patterns of glycosylation can modulate important functional attributes such as PK or potency [27] .

The desired glycoforms of a biologic can be obtained by optimizing the following parameters [27] :

  • Cell line used (e.g. using a GnTIII+ mutant or a Fut8−/− mutant cell line).
  • Additives in the culture (e.g. levels of sodium butyrate).
  • Culturing conditions (e.g. oxygenation, shear stress, etc.).

The post-translational modifications of the reference biologic product are key considerations when developing a biosimilar. Changes in type and degree of post-translational modification have the potential to impact the clinical performance and immunogenicity profile of the resultant biosimilar [17] .

The biosimilar development process: scaling up production

Once the QTPP of the reference biologic product has been identified and the manufacturing steps required to achieve it in the proposed biosimilar have been established, the manufacturing process can be scaled up to production level. Procedures for optimizing large-scale cell cultures include adjustments and optimization of the culturing conditions (e.g. optimizing nutrition, controlling cell death, etc.) to obtain the most efficient product yield, and modification of the physical environment to minimize damage to the protein from aeration or agitation [28] . As with the reference biologic product, the potential exists for lot–lot variability in the key quality attributes of the biosimilar, and this is monitored within and across production runs [24] . The stability of the biosimilar versus the reference biologic product should also be assessed under various stress conditions, such as light, heat and agitation [5] .

The similarity evaluation: overall approach to the physicochemical and functional evaluation of biosimilars

EMA and draft FDA guidelines require that a biosimilar is highly similar to the reference biologic product in terms of physicochemical and functional characterization. All observed differences in these parameters need to be justified with respect to the potential impact on efficacy and safety; however, these differences would not preclude a conclusion of biosimilarity [9] and [21]. The first step in a similarity evaluation involves detailed studies of comparative physicochemical and functional properties, including determination of protein structures, assessment of biologic activity and MOA [5] .

According to the EMA and WHO guidelines and draft FDA guidance, in vitro biological and functional assessments then serve to complement the analytical characterization by enabling evaluation of the impact of any observed structural differences between the biosimilar and its reference biologic product [5] and [9]. If the reference biologic product is known to have multiple mechanisms of action, functional assays can be developed that are specifically designed to assess the full range of activities of the biosimilar [5] .

The similarity evaluation: structural assessments

As outlined by major regulatory guidelines, a thorough analytical and structural characterization and comparison between the proposed biosimilar and the reference biologic product occurs before clinical testing of a biosimilar so all meaningful differences in product attributes and potential impact on function evaluated between the biosimilar and its reference biologic product can be identified [17] . Structural characterization of a biosimilar and its reference biologic product is conducted using appropriate robust, state-of-the-art, biochemical, biophysical and biological functional analytical techniques [5] . Modern analytical methods have the capability to provide in-depth comparative analyses of a proposed biosimilar to its reference biologic product; such techniques include ultra-high-resolution MS, orthogonal techniques and appropriate biological and functional assays [5] and [9].

The primary amino acid sequence of the biosimilar, for example, is expected to be identical to the reference biologic product, although small differences as the result of N- or C-terminal modifications might be justified if they are not expected to impact efficacy and/or safety [9] . Highly sensitive analytical methods such as LC/MS peptide mapping along with Edman sequencing could be used to confirm the primary sequence [29] . An example of a peptide mapping chromatogram to confirm an amino acid sequence is shown in Fig. 5 [30] .


Figure 5 Comparison of primary amino acid sequence: a proposed biosimilar peptide map compared with reference biologic product materials sourced from the USA and EU. The primary amino acid composition of a proposed biosimilar and its EU and US reference biologic products is shown. As observed in the peptide mapping, the amino acid sequence is the same for the proposed biosimilar and its reference biologic products [30] .

In addition to primary, secondary, tertiary and quaternary structural characterization, post-translational modifications (such as glycosylation) and assessment for other potential variants (e.g. protein deamidation and oxidation) should be characterized [9] . If sufficiently comprehensive and robust, the comparative analytical data might justify a more targeted and selective approach to subsequent animal (nonclinical) and clinical studies [9] . Similarly, the final biosimilar dosage form is analyzed to determine any physicochemical and functional impact by formulation and excipients on product stability [9] . Recognizing that biologics are technically more difficult to characterize than small-molecule drugs, biosimilar guidelines outline steps to monitor quality aspects of the biosimilar, including three-dimensional structure, post-translational modifications and biological function, all of which can be impacted by differences in manufacturing processes [18] and [31].

The similarity evaluation: functional assessments

In vitro functional studies are used to compare the pharmacologic and/or biologic activity of a biosimilar with its reference biologic product(s) [9] . Depending on the type of biosimilar produced, the number and types of in vitro studies required will vary on a case-by-case basis. Such assessments can include binding and functional assays. Examples of two functional assays comparing a proposed biosimilar with a reference biologic product are shown in Figure 6 and Figure 7.


Figure 6 Example of a functional assay for a biosimilar mAb and its reference biologic products: antigen-dependent cellular cytotoxicity (ADCC) with peripheral blood mononuclear cells (PBMC). A proposed biosimilar and its reference biologic products from the EU and the USA were evaluated for the ability to elicit the effector function of ADCC. As can be interpreted from the curves, the effective concentration (EC) required to kill 50% (EC50) of the PBMCs is very similar between the proposed biosimilar and the EU and US reference biologic products.


Figure 7 Example of a functional assay: inhibition of cell growth with a proposed biosimilar and its reference biologic products. A proposed biosimilar and its reference biologic products from the EU and the USA were evaluated for activity in a cell growth inhibition assay. As can be interpreted from the curves, the effective concentration (EC) required to inhibit growth by 50% (EC50) of the cell population is similar between the proposed biosimilar and the EU and US reference biologic products [30] .

As with the structural assessments, functional data are considered components of the ‘totality-of-the-evidence’ approach to demonstrate biosimilarity and could be used to justify a more targeted and selective approach toward animal and clinical studies required for approval [9] ; for example, functional assessments could demonstrate and justify ‘no clinically meaningful differences’ between the biosimilar and its reference biologic product of a specific quality attribute. Functional assessments can also assess differences in higher order structure and determine the impact of significant structural differences [9] . In addition, functional assays can be used to provide evidence of a similar MOA for the biosimilar and the reference biologic product (if the MOA is known) [9] . Similar to the structural analysis, multiple lots of reference biologic product procured over a period of time are compared with the proposed biosimilar. The limitations, sensitivity and specificity of such functional assays will also need to be considered when assessing the extent of additional animal and clinical data required to demonstrate biosimilarity [9] .

The similarity evaluation: preclinical pharmacokinetics and toxicity

Because clinical experience and a human safety database would have been established for the reference biologic product, the analytical characterization and demonstration of a high degree of similarity between a proposed biosimilar and its reference biologic product reduces the extent of preclinical in vivo studies [5] . The biosimilar manufacturer could develop preclinical in vivo study plans using data from the reference biologic program contained in publicly available documents [e.g. US Summary Basis of Approval, European Public Assessment Report (EPAR), Health Canada Product Monographs, Freedom of Information (FOI) requests, scientific literature, external presentations, meeting and/or poster abstracts] [5] . For example, it is documented that the administration of rituximab, a mAb targeting CD20 expressed on the surface of mature B cells, causes B cell depletion [32] . B cell depletion could be selected as a PD marker in humans and in nonhuman primates (a pharmacologically relevant nonclinical species) in studies designed to evaluate a proposed rituximab biosimilar.

The FDA draft guidance states that preclinical in vivo studies can be used to assess toxicity, PK and PD measures, and immunogenicity [9] . From an ethical standpoint, preclinical in vivo studies using animals are designed to obtain the maximal amount of information possible and can be conducted in a relevant animal species, meaning one in which the reference biologic product has an established toxicologic and/or PD profile [5] and [33]. For example, if a biologic is cross-reactive with human and one or more nonclinical species, then the animal species might be expected to be reflective of results in humans. If there is a high degree of confidence that the in vitro functional assays are reflective of the MOA of the reference biologic product, these assays could be of greater value in demonstrating similarity and could have greater sensitivity than comparative preclinical in vivo studies [5] . Although it might be possible to complete the preclinical in vivo evaluation in a single study, the evaluation could require a comparison of a range of PK, PD and toxicologic assessments, and the amount of data required might be dependent on the class of biosimilar product (e.g. more data might be required when evaluating complex biologics such as mAbs) [5] . In addition, draft FDA guidance suggests that a comparative single-dose study of a biosimilar and its reference biologic product evaluating PK and PD endpoints could contribute to the totality-of-the-evidence for demonstrating biosimilarity and could be incorporated into an animal toxicity study, although this would not necessarily eliminate the need for human PK/PD studies [9] .

Comparative PK assessments can be designed based on studies previously conducted in the same nonclinical species as the reference biologic product. For example, if the PK of a reference biologic product has been studied in mice, a single-dose comparative study could be conducted in mice to compare PK parameters, while making tolerability observations (e.g. animal clinical status, body weights, necropsy and histopathology for any deaths of unknown etiology or any unexpected euthanasia) and potentially evaluating crucial parameters such as immunogenicity [e.g. levels of antidrug antibodies (ADAs)]. Such a study could be useful even in a nonpharmacologically relevant nonclinical species to compare the kinetics of non-target-mediated clearance of the biosimilar with the reference biologic product. An example of a PK comparison of an intended biosimilar mAb and its reference biologic product is shown in Fig. 8 [34] .


Figure 8 Preclinical pharmacokinetic (PK) comparison: serum concentration–time profile for a biosimilar mAb and its reference biologic product (AUC curves). The serum concentration–time profiles for a proposed biosimilar mAb and a reference biologic mAb are shown at three different doses (2 mg/kg, 10 mg/kg, 20 mg/kg) 168 h post-dose. As can be seen from the curves, the PK profiles are comparable between the proposed biosimilar and the reference biologic at all doses tested [34] .

If conducted, preclinical in vivo toxicity studies can be designed to rule out unexpected toxicity associated with the biosimilar and to demonstrate a lack of biologically meaningful differences between the products. Based on the results of in vitro structural and functional assessments, the FDA requires the collection of animal toxicity data when unexplained analytical differences exist prior to initiating studies in humans [9] . If there is pharmacologic activity in a nonclinical species, preclinical toxicity assessments, including repeat dose administration and PK measurements, should be conducted. These assessments could include evaluating antibody responses and measuring the formation of ADAs and should be conducted over a sufficient amount of time to detect differences between the biosimilar and the reference biologic product [5] and [33]. From a responsible animal use perspective, preclinical in vivo studies are designed and conducted for the shortest possible duration of time that allows the known toxicity profile of the reference biologic product to be compared with the biosimilar.

A repeat-dose toxicity evaluation conducted with the formulation intended for clinical use could be used to eliminate residual concern over unexpected toxicity associated with the biosimilar or to qualify any process-related impurities or formulation differences [5] . Immunogenicity assessments during a repeat-dose toxicity study can determine whether these process-related impurities or formulation differences are meaningful [5] . The EMA and WHO regulatory guidelines suggest consideration of at least one repeat-dose toxicity study including PK and immunogenicity measurements that compare the proposed biosimilar and its reference biologic product in a relevant species [5] and [32]. The route of administration should support clinical use of the reference biologic product and doses should be in the most sensitive part of the dose–response curve to detect any difference between the reference biologic product and biosimilar [5] and [9].

The preclinical in vivo program: immunogenicity assessments

Unlike small-molecule drugs, biologics (including biosimilars) are large protein molecules that can be capable of evoking an immune response. Several factors can influence immunogenicity of biologics (including biosimilars) including [33] :

  • nature of the biosimilar (e.g. B cell depletion can decrease ADA production);
  • presence of process-related impurities;
  • route of administration (subcutaneous versus intravenous);
  • patient population in question, patient-specific factors include:
    • disease-related immunosuppression;
    • concomitantly administered immunosuppressive drugs.

All structural differences between a proposed biosimilar and its reference biologic product also have the potential to impact immunogenicity [7] and [35].

Differences in immunogenicity between a biosimilar and its reference biologic product cannot be accurately predicted in vitro; animal studies, however, could provide an overall comparative view of the immunogenic potential and possibly detect minute differences between a biosimilar and its reference biologic product [5] . A key reason for doing these studies is to help interpret the associated toxicology study [5] . It is important to recognize, however, that nonclinical species cannot be used to predict immunogenicity in humans reliably; immunogenic potential can be reliably assessed only in clinical trials and further evaluated through post-marketing pharmacovigilance [29] and [35].

Impact of the overall analytical profile of a biosimilar on clinical program design

The outcomes of analytical studies can be used to assess at an early stage the overall degree of similarity between a proposed biosimilar and its reference biologic product, and the need for additional assessments before undertaking clinical studies in humans [5] and [21]. Some QTPP attributes have a greater relevance to the demonstration of similarity than others; this could be driven by the potential for an attribute to have a clinical impact. For example, it is required that the amino acid sequence is confirmed as the same as that of the reference biologic product, so this attribute has a high relevance to similarity and high clinical relevance. Some other attributes could have little-to-no clinical impact and, therefore, are considered to have little or no relevance to demonstrating similarity.

The observed degree of similarity of target quality attributes are ideally within the range of the reference biologic product for quality attributes with high clinical significance to minimize the possibility of clinical differences; if overlapping with the reference biologic product, then additional assessment might be required for quality attributes with minor, moderate or high clinical significance. Some specific differences are expected to have a high degree of clinical impact (e.g. differences in attributes that are involved in primary mode of clinical action), whereas other differences could be less clinically relevant (e.g. presence of process-related impurities at levels typical of current standard licensed biologic manufacturing processes). Comparative clinical studies can vary in number and type of study, depending on the level of overall similarity in the characterization assessment that has been observed throughout the preclinical evaluation [24] .

Concluding remarks

Biosimilars can provide additional treatment options for diseases including cancer and chronic inflammatory conditions such as rheumatoid arthritis that can lead to increased access to care. Indeed, state-of-the-art techniques are being used to develop biosimilars of complex biologic therapies including mAbs. A similarity evaluation of the biosimilar is conducted prior to clinical studies. These tests, which help demonstrate a high degree of similarity with the reference biologic product, evaluate the following attributes [9] and [10]:

  • Physicochemical properties.
  • In vitro and in vivo biological activity and function.
  • PK and PD properties.
  • Preclinical in vivo safety and toxicity profile.

The QTPP of the reference biologic product must be fully understood before designing a biosimilar development strategy and undertaking the similarity evaluation. If extensive analytical and preclinical assessments demonstrate that a proposed biosimilar has a high degree of similarity to the reference biologic product, the subsequent clinical development pathway could be abbreviated. Figure 9 shows a comparison of the regulatory content for a novel reference biologic product (originator product) versus a biosimilar. The originator requires the standard content across all modules [module 3 = chemistry, manufacturing and control (CMC) studies; module 4 = nonclinical (animal) in vivo studies; module 5 = clinical studies]. By contrast, a biosimilar requires an expanded module 3, the robustness and content of which allows an abbreviated nonclinical (in vivo animal studies) and/or clinical program (that is, reduced module 4 and module 5). In this manner, the module 3 content underpins the subsequent clinical development.


Figure 9 Quality/nonclinical/clinical integration: originator product versus a biosimilar. The regulatory requirements for a reference biologic product (originator) and the proposed biosimilar are compared. For a reference biologic product, the standard content is required across the modules [module 3 = chemistry, manufacturing, and control (CMC) studies, module 4 = nonclinical (animal) in vivo studies, module 5 = clinical studies]. For a proposed biosimilar, modules 4 and 5 are decreased and module 3 is expanded. Module 3 is the basis for the nonclinical and clinical program. Abbreviation: CTD, common technical development.

Clinical studies should be designed in an efficient manner so as not to recreate efficacy and safety data for the reference biologic product but rather to eliminate any residual uncertainties that might exist regarding the degree of similarity [17] . A robust similarity evaluation ultimately allows a more streamlined in vivo preclinical and clinical study and approval process for biosimilars, which could increase patient access.

Conflicts of interest

This supplement was funded by Pfizer Inc. Lynne A. Bui, MD, was a paid consultant to Pfizer Inc for the research and/or authorship of this supplement. Gregory L. Finch, PhD, Susan Hurst, PhD, Beverly Ingram, PhD, Ira A. Jacobs, MD, MBA, FACS, Carol F. Kirchhoff, PhD, Chee-Keng Ng, PhD, and Anne M. Ryan, DVM, PhD, DACVP, are employees of Pfizer Inc. Medical writing and editorial support to prepare this supplement was provided by Jaqueline Egan and Allison M. England, PhD, of QD Healthcare Group and funded by Pfizer Inc. Lynne A. Bui has the following additional conflicts of interest to disclose: consultant for ActOncology, Array Pharmaceuticals, Astex Pharmaceuticals, Clinipace, Medivation Inc, NovellaXencor Inc, Pathway Therapeutics, Puma Biotechnology Inc, RadioRx Inc and Sun Pharmaceuticals Inc; stock and/or stock options in Annam Biosciences, Annam Pharma and Exelixis.




therapeutic agents (e.g. proteins defined by the FDA as greater than 40 amino acids) [9] that are manufactured in living systems, typically through recombinant DNA technology [1] . Because the manufacturing process for biologics is inherently subject to variation, identical copies cannot be produced [1] . Other guidelines, such as those by the EMA and WHO, do not provide a definition for what constitutes a protein [5] and [33].


a biologic product considered ‘highly similar’ but not identical to its reference biologic product. The FDA further defines a biosimilar as highly similar ‘notwithstanding minor differences in clinically inactive components’, and having ‘no clinically meaningful differences between the biologic product and the reference product in terms of the safety, purity, and potency of the product’ [6] .

Comparability exercise

usually an exercise to characterize a biologic product before and after a manufacturing process change, with the goal of ruling out any adverse impact on quality, efficacy or safety. The EMA regards this exercise as broadly similar to a biosimilarity assessment [16] .


usually the biologic product being equivalent in terms of efficacy and safety, including immunogenicity, before and after a manufacturing process change [16] .


the process by which biologic drugs can change over time and exhibit differences between lots or batches of product [11] .

Generic drug

copies of small-molecule drugs (e.g. ibuprofen). Because identical versions of most small-molecule drugs can be produced through chemical synthesis, they can be substituted for the branded product [5] .

Preclinical evaluation

the phase of study occurring before clinical evaluation. This can include analytical and/or physicochemical characterization; chemistry, manufacturing and control (CMC) evaluation in vitro and/or in vivo functional, PK, PD, toxicologic or immunogenicity assessment.

Reference biologic

the existing, licensed and marketed product against which a proposed biosimilar product will be compared [6] . WHO uses the terminology ‘reference biotherapeutic product’, the EMA uses ‘reference medicinal product’ and the FDA uses ‘reference product’ when referring to a reference biologic [5], [9], and [16].

Similarity evaluation

the process of demonstrating biosimilarity between a proposed biosimilar and its reference biologic product.


1 Global Cancer Research Institute, Inc. (GCRI) for Personalized Oncology, San Jose, CA 95124, USA

2 Development Strategies Group in the PDM (Pharmacokinetics, Dynamics, and Metabolism) Department, Pfizer Inc, Groton, CT 06340, USA

3 Drug Safety Research and Development, Pfizer Inc, Groton, CT 06340, USA

4 Pfizer Inc, Andover, MA 01810, USA

5 Pfizer Emerging Markets/Established Products Medicines Development Group, Pfizer Inc, New York, NY 10017, USA

6 Global Technology Services – BioManufacturing Sciences Group, Pfizer Inc, Chesterfield, MO 63017, USA

Corresponding author: