Clin Exp Nephrol (2007) 11:191–195
DOI 10.1007/s10157-007-0487-2
REVIEW ARTICLE
David Goldsmith · Martin Kuhlmann · Adrian Covic
Through the looking glass: the protein science of biosimilars
Received: January 30, 2007 / Accepted: June 24, 2007
Abstract
Biopharmaceuticals have revolutionized the treatment and
management of many diseases. The advent of recombinant
erythropoietins has greatly benefited patients with anemia
related to chronic kidney disease and cancer, virtually
eliminating the need for blood transfusions. Currently, the
patents for many biopharmaceutical molecules have expired
or are approaching expiration and a number of biosimilars
manufacturers are aiming to claim part of the market share.
Unlike the situation for synthetic “small molecule” drugs,
identical copies of far more complex biopharmaceuticals
cannot be produced. A biopharmaceutical can be 100 to
1000 times larger than a synthetic chemical drug, with
extremely complex three-dimensional structure and bio-
logical functions which are often not completely under-
stood. Due to their nature and complexity, these fascinating
therapeutic molecules are products of highly controlled
biological processes. This review takes a look at how bios-
imilars are fundamentally different from their originator
products by examining the biopharmaceutical production
process and how it can influence the structure and function
of the final drug product.
Key words Biopharmaceuticals · Biosimilars · Assays · Gly-
cosylation · Manufacturing · Pharmacovigilance
Biopharmaceuticals and renal anemia treatment
In chronic kidney disease, erythrocyte function and produc-
tion is impaired, resulting in anemia in a large percentage
D. Goldsmith (*)
Renal Unit, 6th Floor New Guy House, Guy’s Hospital, London, UK
SE1 9RT, United Kingdom
Tel. +44 02071885708 or 5647; Fax +44 02071885692
e-mail: David.Goldsmith@gstt.nhs.uk; goldsmith@london.com
M. Kuhlmann
Vivantes Klinikum im Friedrichshain, Berlin, Germany
A. Covic
C.I. Parhon University Hospital, Iasi, Romania
© Japanese Society of Nephrology 2007
of patients. The application of therapeutic agents to stimu-
late red blood cell production has only been possible in the
last 20 years, with the advent of erythrocyte-stimulating
agents (ESAs).
The patents for many biopharmaceuticals are ending or
coming to an end, and a host of follow-on biologics manu-
facturers are waiting in the wings to launch biosimilar prod-
ucts. Eleven biopharmaceutical products with combined
sales of US$15 billion had lost patent protection by the end
of 2006. That biosimilars are similar, but not identical, to
their originator products is the subject of much discussion.
How similar is similar enough? To what extent can the
safety and efficacy of a follow-on product be demonstrated
without the extensive clinical data required of originator
products? This review aims to answer these questions
through a description of how biopharmaceutical products
are made, and how this can affect the structural and
immunogenic properties of the final product.
Biopharmaceuticals: complex biological entities
A biopharmaceutical (or “biological medicinal product”) is
a pharmaceutical product that contains biotechnology-
derived proteins as active substances.1 Copies of bio-
pharmaceuticals are known as “similar biological medicinal
products” or biosimilars. In discussions of comparisons
between originator and biosimilars, the terms “comparabil-
ity” and “similarity” are used. “Comparability” is used in
the context of biopharmaceuticals produced from a single,
validated production process (i.e., how different batches
from the same manufacturer compare with one another in
terms of physicochemical properties). “Similarity” is used
when comparing products from different manufacturers.
Biosimilars manufacturers must convince regulatory author-
ities that their products are similar to the reference origina-
tor product. A follow-on biopharmaceutical product can
only be called “similar” to the originator drug when the
same safety and efficacy as that of the originator is
established.
192

Table 1. A comparison of the molecular weights of synthetic chemical the current biopharmaceuticals can be divided into two
and biopharmaceutical drugs classes: those that are not post-translationally modified, and
Chemical drug Molecular weight (Da) those that are.5
One of the most important post-translational modifica-
Glucophage 166
Vioxx 314
tions is glycosylation. Many important therapeutic proteins
Prozac 346 are glycosylated, including erythropoietin,6 the colony-
Zantac 351 stimulating factors (CSFs),7 and a host of recombinant
Paxil 375 human immunoglobulins (IgGs).8 Glycosylation can influ-
Zocor 419
Augmentin 420
ence the biological activity of a protein by different mecha-
Crixivan 712 nisms. First of all, glycosylation can affect half-life by
Taxol 854 influencing the active clearance of a protein. The serum
half-life of erythropoietin is dependent on four sialylated
Biopharmaceutical drug Molecular weight (Da)
N-glycans (sialic acid is itself a glycan). Poorly glycosylated
Neupogen 18 800 forms of erythropoietin are rapidly cleared through the
Roferon-A 19 625 kidney, whereas less-sialylated forms of erythropoietin are
Humatrope 22 125 rapidly cleared by hepatocytes and macrophages. Thus,
Avonex 22 500
Epogen 30 400 increasing the degree of sialylation and glycosylation
Pulmozyme 37 000 decreases the renal clearance rate,9 and can increase eryth-
Enbrel 75 000 ropoietin in vivo activity.
Zenapax 144 000 Although the integrity of the primary polypeptide struc-
Rituxan/MabThera 145 000
Factor VIII 264 000 ture may be largely unchanged using different recombinant
Source, EuropaBio2
expression systems and manufacturing conditions, signifi-
cant changes in post-translational processing can and do
occur with different manufacturing processes.8 The follow-
ing section describes the biopharmaceutical manufacturing
Examples of biopharmaceuticals include cytokines, hor- process. Although detailed technical discussion of each step
mones, and monoclonal antibodies.3 In comparison with is beyond the scope of this review, we provide examples of
synthetic small molecules, biopharmaceuticals are 100 to critical stages in the manufacturing process which can affect
1000 times larger in size and are structurally more complex. the final product.
Table 1 gives a list of molecular weights for a selection of
chemical and biopharmaceutical agents. Several successful
biopharmaceuticals are on the market, including recombi-
nant human insulin, growth hormone (GH), erythropoietin, How can protein activity be modified by the
interferon (IFN)-beta, and Factor VIII. These complex production process?
molecules are produced from genetically modified cell lines,
and are extracted using complex purification procedures. The production process: an inside look
The manufacturing protocols used for the production of
biopharmaceuticals are proprietary to the manufacturer, Biopharmaceutical manufacturing is a multistep process,
and are not accessible in the public domain. Because of involving cloning of the appropriate genetic sequence into
the inherent variability in such processes, identical copies a carefully selected expression vector, selection of a suitable
of a protein molecule cannot be produced without prior cell expression system, scale-up and purification, up to
knowledge of the protocols used in its manufacture. formulation of the end product.

Delivery of the gene of interest into host cells

The role of post-translational modifications
Before discussing how the biopharmaceutical production
process can affect the final protein product, it is necessary
to take a brief overview of how proteins can be modified,
with particular emphasis on post-translational modifica-
tions (defined as any alteration that occurs after synthesis
of the protein chain). Post-translational modifications can
have a profound impact on the activities of proteins.4 Exam-
ples of such modifications include enzymatic cleavage
(which can lead to activation, i.e., in hormones), the attach-
ment of lipid moieties (targeting the protein to a cell mem-
brane), or the attachment of glycans (affecting properties
such as serum half-life). The importance of post-
translational modifications is underscored by the fact that
The success of recombinant protein production relies on the
efficient delivery and appropriate expression of recombi-
nant proteins using a cell expression system. One of the first
steps in the process of expressing a recombinant protein is
the cloning of the appropriate genetic sequence into a
plasmid expression vector. Once the gene of interest has
been successfully cloned into the chosen vector, it is then
introduced into a host cell line for expression. A major
factor for consideration is the host cell expression system
itself.10 Today, a wide range of expression systems is avail-
able for recombinant protein production. Some of the most
commonly used systems include Escherichia coli, insect,
yeast, and mammalian cells.11 Currently, the industry main-
stay for biopharmaceutical production is mammalian cells.12

Table 2. Commercial biopharmaceutical products and the expression systems used for their productiona
Expression system Product
Escherichia coli Infergen (rh IFN-alfa)
Humalog (rh insulin)
Neupogen/Neulasta (GM-CSF)
Yeast NovoRapid (rh insulin)
Regranex (recombinant platelet-derived growth factor)
Leukine (GM-CSF)
CHO cells Herceptin (trastuzumab)
Cerezyme (recombinant imiglucerase)
Avastin (bevacizumab)
Rituxan (rituximab)
Murine hybridoma Bexxar (iodine131-labelled tositumomab)
Orthoclone OKT3 (muromonab)
Simulect (basiliximab)
a
Adapted after Thiel (2004)13
One of the most commonly used mammalian cell types for
these purposes are Chinese hamster ovary (CHO) cells.11
The use of mammalian expression systems offers distinct
advantages, including correct post-translational modifica-
tion and protein folding.11 However, this does not preclude
the use of non-mammalian recombinant expression systems;
indeed, a diverse array of alternative expression systems is
under development, including the use of transgenic hosts,
such as chicken eggs, goats, cows, and plants.12 A glance at
a selection of biopharmaceutical molecules from several
manufacturers shows the variety of recombinant expression
systems employed for their production (Table 2).
The expression of the same recombinant gene product
from different sources has a profound impact on the final
structure of the protein. Recombinant human granulocyte-
CSF is available in two forms for clinical use: a non-
glycosylated form expressed in E. coli, and a glycosylated
form expressed in CHO cells.14 A similar heterogeneity in
glycosylation patterns has been observed for human
interferon-gamma (IFN-gamma) produced in different
expression systems.15
Manufacturing scale-up, fermentation, and purification
During the development of a biopharmaceutical, the process
used for the production of sufficient material for preclinical
studies needs to be drastically modified in order to produce
material of sufficient quality and quantity for clinical trials.16
Furthermore, regulatory bodies in Europe and the United
States require that the same manufacturing procedure used
to produce drugs for late-stage clinical trials be used to
produce them for the market.17 Thus, the protocols used for
the manufacture of the final therapeutic product need to
be strictly validated, with built-in controls for monitoring
product quality during each step of the process.
The scaling up of the protein manufacturing process is a
complex procedure. The challenge for the biopharmaceuti-
cal manufacturer is to increase the quantity of protein
produced without compromising its quality and purity.18 The
193
Manufacturer
Intermune
Lilly
Amgen
Novo Nordisk
Johnson & Johnson
Berlex
Genentech
Genzyme
Genentech
Biogen
Corixa
Johnson & Johnson
Novartis
methods for scaling up the host cell system to produce suffi-
cient quantities of biopharmaceutical needs to be adapted to
the final product, the cell line, and the quantities needed. As
an example, adherent host cells (such as those used for the
production of recombinant human erythropoietin) have
specific substrate attachment requirements, and the surface-
to-volume ratio in the bioreactor needs to be adjusted to
obtain maximal cell densities.18 The levels of oxygen19 and
carbon dioxide20 in the cultures are additional factors that
can affect the health of the host cells, and consequently, the
quality of the final product. Cell culture conditions are
known to have a profound influence on post-translational
modifications present in the expressed proteins.21
Another challenge for manufacturers is the presence of
potential contaminants introduced in the manufacturing
process. One such concern is the transfer of viruses or
prions from culture medium containing animal serum.22
Another frequently encountered problem is the formation
of protein aggregates. Certain proteins such as IFN-gamma
have the tendency to aggregate under mild denaturing con-
ditions (low denaturant concentration and pH below 5).23
In the case of IFN-alpha, such protein aggregates have been
shown to be more immunogenic than monomers when
tested in transgenic mice.24 Protein aggregation is a common
problem in the manufacture of biopharmaceuticals, and is
another factor that is influenced by the cell culture condi-
tions and purification processes.25
Glycosylation is also especially sensitive to cell growth
conditions. Changes in culture pH, the availability of pre-
cursors and nutrients, and the presence or absence of
various cytokines and hormones can each affect the extent
of glycosylation in the drug product.26 The accumulation of
enzymes such as sialidases and glycosidases in the cell
culture medium can potentially modify the secreted recom-
binant protein.27 A variety of undesired chemical alterations
to the drug product can also occur during the production
process, including oxidation or deamidation.28 Additional
impurities, such as process-derived chemicals or antibiotics,
can greatly influence the biological activity and safety profile
of the final product.29
194

sites on the native protein backbone.37 In the absence of the

Different process, different products: what are
appropriate glycosylation moieties, antibodies reacting to
the consequences?
the exposed sites on the protein developed in 4 out of
16 patients receiving rhGM-CSF therapy.37 The lack of
Regardless of the scale of the manufacturing protocol, dif- appropriate glycosylation in rhGM-CSF in this case
ferent biological preparations can exhibit marked varia- may stem from the choice of host cell system used for
tions. One study revealed variations in glycosylation pattern its production.37
between different laboratory preparations of erythropoie-
tin.30 An independent comparison of eight different eryth-
ropoietin preparations by mass spectrometry demonstrated How can we ensure the safety of
that each sample contained a mixture of isoforms, with at biosimilar molecules?
least 23 distinct glycan structures.31 In the case of erythro-
poietin (like that of many other biopharmaceuticals),
Clearly, the use of a different manufacturing process will
glycosylation pattern is a critical factor that determines bio-
result in differences in the final product. The challenge then
logical activity, solubility, and half-life of the glycoprotein in
remains to assess and quantify these differences, and deter-
the circulation.26
mine whether the new product is as safe and efficacious as
The use of isoelectric focusing (a protein separation
the originator. Currently, neither analytical methods nor in
technique in which a mixture of protein molecules is resolved
vitro assays are able to fully characterize a recombinant
into its components by subjecting the mixture to an electric
protein, and there are no reliable tests for predicting clinical
field in a supporting gel) to analyze 12 commercial epoetin
efficacy. Animals transgenic for the human endogenous
alfa samples manufactured outside the United States and
gene equivalent to a recombinant protein could provide
Europe demonstrated variations in isoform composition
useful qualitative models to assess the relative immunoge-
between these preparations.32 A separate comparability
nicity between different products, but such data cannot
study of another series of biosimilar products manufactured
be extrapolated to predict the immunogenic response in
outside the United States and the European Union revealed
humans.38 Rigorous clinical testing is the only means of
that the products differed widely in composition, exhibited
demonstrating unequivocally that a biosimilar has the same
batch-to-batch variation, and did not always meet self-
safety and efficacy as the originator product. Because the
declared specifications.33 These results were corroborated
majority of immunological responses occur after prolonged
by a Brazilian study, which, furthermore, revealed unaccept-
exposure to the recombinant protein (generally exceeding
able bacterial endotoxin levels in three products,34 leading
the length of standard phase III clinical trials), pharmaco-
to their withdrawal from the market.
vigilance plans need to be integrated for the surveillance of
What are the consequences of differences in the final
newly launched biosimilar products.39 Although this is true
product? A critical issue faced by all biotechnology derived
of large, complex recombinant proteins (such as monoclo-
medicines is product immunogenicity. The increased
nal antibodies),40 care must also be taken with “simpler,”
number of pure red cell aplasia (PRCA) cases in patients
well-characterized molecules such as insulin and growth
receiving a specific brand of epoetin is a case in point. Upon
hormone. Insulin and growth hormone preparations con-
the European Medicines Agency’s (EMEA’s) request to
taining the same active ingredient display wide variations
remove human-derived proteins from the product formula-
in bioavailability because of differences in product
tion, the manufacturer of this product replaced human
formulations.39
serum albumin (HSA) with sorbitol-80 as a stabilizer.35 The
aim of this change was to minimize the risks from human-
borne contaminants such as HIV and prions. However, a
rise in the number of PRCA cases following this change in Conclusion
product formulation was a red flag to all parties concerned
regarding biosimilar drugs. If this minor change in one com- Biopharmaceuticals are far more complex molecules than
ponent of the formulation of an established product could traditional synthetic chemical drugs. Their properties are
lead to such disastrous consequences, what then is the sig- highly dependent on the manufacturing process, and slight
nificance of more fundamental differences in a biosimilar modifications during any step of the protocol can result in
molecule produced by a completely different process? structural and/or functional changes in the active substance.
Such adverse reactions are rare, but are not limited Current analytical techniques are not capable of identifying
to recombinant human erythropoietin products. Immune all the molecular changes that may have occurred upon
responses can also be triggered by other factors, such as modification of the manufacturing process; therefore, such
variations in amino acid sequence, glycosylation pattern, test results cannot be relied upon to predict drug efficacy
and trace amounts of contaminants in the final product.35 In and safety based purely on physicochemical characteristics.
the case of IFN-beta, lack of the appropriate glycosylation Although European regulatory bodies have established
is thought to promote immunogenicity by uncovering guidelines for the approval of several classes of biosimilar
hydrophobic residues and reducing product solubility.36 products, future advances in analytical techniques may be
Similarly, O-linked glycosylation on recombinant human able to predict how well biosimilars match up to their
granulocyte-macrophage (rhGM)-CSF protects specific originators in terms of safety and efficacy.

References
1. Committee for Medicinal Products for Human Use. Guidelines on
similar biological medicinal products containing biotechnology-
derived proteins as active substance: quality issues. European
Medicines Agency EMEA/CHMP/BWP/49348/2005 2005.
2. EuropaBio. The biotech industry’s views on the pharma review
package. Available at http://europabio.org/positions/posPharma.
ppt Accessed Dec 2006
3. Schellekens H. Biosimilar therapeutic agents: issues with bioequiv-
alence and immunogenicity. Eur J Clin Invest 2004;34:797–9.
4. Walsh G, Jefferis R. Post-translational modifications in the context
of therapeutic proteins. Nat Biotechnol 2006;24:1241–52.
5. Gerngross TU. Advances in the production of human therapeutic
proteins in yeasts and filamentous fungi. Nat Biotechnol 2004;22:
1409–14.
6. Elliott S, Egrie J, Browne J, Lorenzini T, Busse L, Rogers N, et al.
Control of rHuEPO biological activity: the role of carbohydrate.
Exp Hematol 2004;32:1146–55.
7. Sylvester RK. Clinical applications of colony-stimulating factors: a
historical perspective. Am J Health Syst Pharm 2002;59:S6–12.
8. Raju TS, Briggs JB, Borge SM, Jones AJ. Species-specific variation
in glycosylation of IgG: evidence for the species-specific sialylation
and branch-specific galactosylation and importance for engineer-
ing recombinant glycoprotein therapeutics. Glycobiology 2000;10:
477–86.
9. Misaizu T, Matsuki S, Strickland TW, Takeuchi M, Kobata A, Taka-
saki S. Role of antennary structure of N-linked sugar chains in
renal handling of recombinant human erythropoietin. Blood
1995;86:4097–104.
10. Bendig MM. The production of foreign proteins in mammalian
cells. Genet Eng 1988;91–127.
11. Hunt I. From gene to protein: a review of new and enabling
technologies for multi-parallel protein expression. Protein Expr
Purif 2005;40:1–22.
12. Dove A. Betting on biogenerics. Nat Biotechnol 2001;19:117–20.
13. Thiel KA. Biomanufacturing, from bust to boom . . . to bubble?
Nat Biotechnol 2004;22:1365–72.
14. Hoglund M. Glycosylated and non-glycosylated recombinant
human granulocyte colony-stimulating factor (rhG-CSF)–what is
the difference? Med Oncol 1998;15:229–33.
15. Hooker A, James D. The glycosylation heterogeneity of recombi-
nant human IFN-gamma. J Interferon Cytokine Res 1998;18:
287–95.
16. Meuwly F, Weber U, Ziegler T, Gervais A, Mastrangeli R,
Crisci C, et al. Conversion of a CHO cell culture process from
perfusion to fed-batch technology without altering product quality.
J Biotechnol 2006;123:106–16.
17. Dove A. Uncorking the biomanufacturing bottleneck. Nat Biotech-
nol 2002;20:777–9.
18. Chu L, Robinson DK. Industrial choices for protein production by
large-scale cell culture. Curr Opin Biotechnol 2001;12:180–7.
19. Dutton G. Improved tools for large-scale bioprocessing. Genet Eng
News 2000;20:57–63.
20. Pattison RN, Swamy J, Mendenhall B, Hwang C, Frohlich BT. Mea-
surement and control of dissolved carbon dioxide in mammalian
cell culture processes using an in situ fiber optic chemical sensor.
Biotechnol Prog 2000;16:769–74.
21. Sugiura T, Maruyama HB. Factors influencing expression and post-
translational modification of recombinant protein C. J Biotechnol
1992;22:353–60.
195
22. Jayme DW, Smith SR. Media formulation options and manufactur-
ing process controls to safeguard against introduction of animal
origin contaminants in animal cell culture. Cytotechnology 2000;33:
27–36.
23. Kendrick BS, Cleland JL, Lam X, Nguyen T, Randolph TW,
Manning MC, et al. Aggregation of recombinant human interferon
gamma: kinetics and structural transitions. J Pharm Sci 1998;87:
1069–76.
24. Braun A, Kwee L, Labow MA, Alsenz J. Protein aggregates seem
to play a key role among the parameters influencing the antigenic-
ity of interferon alpha (IFN-alpha) in normal and transgenic mice.
Pharm Res 1997;14:1472–8.
25. Cromwell ME, Hilario E, Jacobson F. Protein aggregation and
bioprocessing. AAPS J 2006;8:E572–9.
26. Balaguer E, Demelbauer U, Pelzing M, Sanz-Nebot V, Barbosa J,
Neususs C. Glycoform characterization of erythropoietin combin-
ing glycan and intact protein analysis by capillary electrophoresis
– electrospray – time-of-flight mass spectrometry. Electrophoresis
2006;27:2638–50.
27. Gramer MJ, Goochee CF. Glycosidase activities in Chinese
hamster ovary cell lysate and cell culture supernatant. Biotechnol
Prog 1993;9:366–73.
28. Ahrer K, Jungbauer A. Chromatographic and electrophoretic
characterization of protein variants. J Chromatogr B Analyt
Technol Biomed Life Sci 2006;841:110–22.
29. DiPaolo B, Pennetti A, Nugent L, Venkat K. Monitoring impuri-
ties in biopharmaceuticals produced by recombinant technology.
Pharm Sci Technol Today 1999;2:70–82.
30. Storring PL, Tiplady RJ, Gaines Das RE, Rafferty B, Mistry YG.
Lectin-binding assays for the isoforms of human erythropoietin:
comparison of urinary and four recombinant erythropoietins.
J Endocrinol 1996;150:401–12.
31. Yuen CT, Storring PL, Tiplady RJ, Izquierdo M, Wait R, Gee CK,
et al. Relationships between the N-glycan structures and biological
activities of recombinant human erythropoietins produced using
different culture conditions and purification procedures. Br J
Haematol 2003;121:511–26.
32. Schellekens H. Follow-on biologics: challenges of the “next
generation”. Nephrol Dial Transplant 2005;20(Suppl 4):iv31–
6.
33. Combe C, Tredree RL, Schellekens H. Biosimilar epoetins: an
analysis based on recently implemented European Medicines
Evaluation Agency guidelines on comparability of biopharmaceu-
tical proteins. Pharmacotherapy 2005;25:954–62.
34. Schmidt CA, Ramos AS, da Silva JEP, Fronza M, Dalmora SL.
Activity evaluation and characterization of recombinant human
erythropoietin in pharmaceutical products. Arq Bras Endocrinol
Metabol 2003;47:183–9.
35. Louet S. Lessons from Eprex for biogeneric firms. Nat Biotechnol
2003;21:956–7.
36. Karpusas M, Whitty A, Runkel L, Hochman P. The structure of
human interferon-beta: implications for activity. Cell Mol Life Sci
1998;54:1203–16.
37. Gribben JG, Devereux S, Thomas NS, Keim M, Jones HM,
Goldstone AH, et al. Development of antibodies to unprotected
glycosylation sites on recombinant human GM-CSF. Lancet
1990;335:434–7.
38. Schellekens H. Bioequivalence and the immunogenicity of bio-
pharmaceuticals. Nat Rev Drug Discov 2002;1:457–62.
39. Schellekens H. How similar do “biosimilars” need to be? Nat Bio-
technol 2004;22:1357–9.
40. Schellekens H, Ryff JC. “Biogenerics”: the off-patent biotech prod-
ucts. Trends Pharmacol Sci 2002;23:119–21.