Engineering Biopharmaceuticals

November 1, 2007
Gary Walsh, PhD

Associate Professor in the Industrial Biochemistry Program at the University of Limerick. He is also a member of BioPharm International's Editorial Advisory Board.

Volume 20, Issue 11

An overall increasing proportion of future product approvals will be engineered in some way, either directly or indirectly.


One of the most prominent technical trends of the industry has been the continued increase in the proportion of engineered products coming to the market. The vast majority of biopharmaceuticals approved during the 1980s and early 1990s were either first generation murine monoclonals or unmodified replacement proteins [e.g., human growth hormone (hGH), interferons, blood factor VIII, and erythropoietins—all identical in amino acid sequence to the native human protein and administered in order to replace or augment natural levels of that protein]. This article focuses on more recent approvals and trends in engineering approaches for biopharmaceutical production.

The biopharmaceutical sector continues to grow steadily, with an average of 8–10 new products entering the market each year. By the end of 2006, almost 170 recombinant therapeutic proteins or antibody-based products had gained approval in either the US or the EU, commanding an estimated global market value of $40 billion. One of the most prominent technical trends of the industry has been the continued increase in the proportion of engineered products coming to the market. The vast majority of biopharmaceuticals approved during the 1980s and early 1990s were either first generation murine monoclonals or unmodified replacement proteins [e.g., human growth hormone (hGH), interferons, blood factor VIII, and erythropoietins—all identical in amino acid sequence to the native human protein and administered in order to replace or augment natural levels of that protein].

Advances in protein science and bioinformatics, along with the development of increasingly sensitive and sophisticated analytical methodologies, continue to underscore a greater understanding of the links between protein structure and function. This method allows knowledge-based modification of protein structure to achieve some predefined alteration of functionality.


The focus of many initial engineering experiments entailed alteration of the target protein's native amino acid sequence by molecular techniques, such as site-directed mutagenesis. Some of the alterations included the removal or replacement of large stretches of protein backbone (e.g., chimeric and humanized antibodies); others entailed the addition, removal, or replacement of a single amino acid or, at most, a few amino acids. Examples of the latter approach include several engineered insulin products such as Humalog (insulin lispro, Eli Lilly, Indianapolis, IN).

An alternative engineering approach focuses on the covalent attachment of a chemical moiety to the protein's backbone (e.g., the PEGylation of interferons or the acylation of insulin), or the alteration of natural post-translational modifications that may be present, as in the case of Cerezyme (Genzyme, Cambridge, MA) a recombinant glucocerebrosidase enzyme, whose glycocoponent's sialic acid caps are enzymatically removed to expose mannose residues, promoting macrophage-selective product uptake.

Engineering has been undertaken to achieve various therapeutic objectives, with the most common being:

  • Reduction of product immunogenicity (e.g., engineered antibodies)

  • Generation of a faster or slower acting product (e.g., engineered insulins)

  • Increasing a protein's biological half-life (e.g., PEGylated interferons)

  • Generation of a novel protein product, e.g., fusion products such as Amevive (Biogen, Cambridge, MA) or Enbrel (Immunex, Thousand Oaks, CA).


The remainder of this article focuses on more recent approvals and trends in engineering approaches because engineered products approved throughout the 1990s and over the earlier part of this decade have been reviewed elsewhere.1 The 14 engineered products approved since 2002, summarized in Table 1, represent 36% of the 39 new products to come on the market for the first time in either the EU or US in that time period.

Table 1. Engineered therapeutic proteins that gained approval in the EU or US (2002–2006 inclusive)

Engineered antibodies remain the most common engineered product category to gain approval, with one chimeric and five humanized products entering the market over the last five years (Table 1). The most novel engineering approach witnessed in latter years is the acylation of insulin (Levemir, Novo Nordisk, Bagsvaerd, Denmark). The principle engineering feature of Levemir (insulin detemir) is the covalent attachment (acylation) of the C14 fatty acid (myristic acid) to the side chain of the lysine residue found at position 29 of insulin's B chain. Human serum albumin (HSA) harbors three high affinity fatty acid binding sites and the engineering rationale was to facilitate tight but reversible binding of the insulin to HSA, both at the site of injection and in the blood. This occurs in practice, ensuring a constant and prolonged release of free insulin, which gives the product an activity duration of 24 hours. Levemir was the first and, so far, only approved biopharmaceutical to be engineered in this way.

The development of another engineered product, Somavert (pegvisomant, Pfizer, New York, NY), is notable because it involves PEGylation combined with amino acid substitutions (nine in all). Somavert is indicated for the treatment of acromegaly, a rare endocrine disorder characterized by elevated blood hGH concentrations. The product is a hGH analogue capable of binding to the hGH cell surface receptor without triggering an intracellular response. Somavert acts as a hGH antagonist, reducing the effects of endogenous hGH while PEGylation increases the product's serum half-life.

Several disease conditions are triggered or acerbated by the inappropriate over expression of a gene product. Several biopharmaceuticals are now approved, which treat such conditions by inhibiting the activity of the overexpressed protein. In addition to Somavert, antibody-based products such as Kineret (Amgen, Thousand Oaks, CA) and Erbitux (ImClone Systems, New York, NY) function by direct binding and inactivation of the target protein. On the other hand, products such as Humira (Abbott, Chicago, IL) and Enbrel (discussed before) bring about their effect when the biopharmaceutical binds to the protein's target receptor, acting as an antagonist.


Products engineered by amino acid sequence modification or by attachment of polyethylene glycol or some other chemical moiety will continue to gain approval. Because of increasing appreciation of the importance of post-translational modifications (PTMs) on the therapeutic characteristics of many proteins, knowledge-based alteration of native PTM profiles is now receiving more attention.2

The majority of protein-based biopharmaceuticals bear some form of PTM, with glycosylation being the most complex and widespread modification. It is estimated that up to 50% of all native human proteins are glycosylated and that 1–2% of the human genome encodes proteins that contribute to glycosylation. Approximately one third of all approved biopharmaceuticals are glycosylated, with antibodies as well as blood factors and some hormones (e.g., gonadotropins and erythropoietin) representing the most prominent categories. A protein's glycocomponent can have many significant influences on its therapeutic characteristics, including influencing stability, ligand recognition or binding, serum half-life, and immunogenicity.3

Cerezyne (as discussed earlier) and Aranesp (Amgen, Thousand Oaks, CA) are two products for which glycocomponent has been engineered and are already on the market. Aranesp is a recombinant human erythropoietin (EPO) indicated for the treatment of anaemia. Produced in a Chinese hamster ovary (CHO) cell line, the product displays an increased carbohydrate content when compared to native EPO. The native molecule contains three carbohydrate side chains whereas Aranesp contains five. This increased carbohydrate content extends the product's half-life, facilitating once weekly or once fortnightly administration schedules.

Protein glycosylation occurs naturally in the endoplasmic reticulum (ER) and the Golgi, and is undertaken by a multienzyme pathway comprising 2–3 dozen glycosyltransferase and glucosidase enzymes.4 The characteristic heterogeneity associated with a glycocomprotein's glycocomponent is mainly a reflection of incomplete glycosyl processing. Furthermore, different cell types vary in terms of the exact glycosylation pathways they harbor and this can have profound implications on the suitability of any given cell type as a production platform for therapeutic proteins. For example, yeast-derived glycosylation patterns tend to be of a high mannose type, which generally confers a short circulating half-life on the protein in humans. Plant-based glycosylation tends to be more complex and extensive when compared with mammalian-derived patterns and contains sugar motifs that are immunogenic in humans.

One avenue of current research relates to engineering the glycosylation pathways in alternative potential production systems with a view to humanize the glycosylation patterns characteristic of the proteins they produce. Such engineering is not an insubstantial task, because the exact glycosylation pathways present may not be completely characterized, and the glycosylation reactions themselves occur in a specific predetermined sequence spanning both the ER and Golgi. This necessitates the targeting of human glycosylation reactions to specific organelles.

Despite the difficulties, substantial progress has been recorded, particularly in the case of yeast-based systems. For example, engineered strains of Pichia pastoris have been developed which undertake specific human glycosylation reactions, and there are yeast cell lines now in existence that can actually produce more uniform glycoform profiles than do currently used mammalian cell lines.5,6

Engineering the PTM capacity of a producer cell line to tailor the PTM characteristics of therapeutic proteins produced can also be done for other reasons, e.g., in order to optimize a particular biological activity of the target protein, as exemplified by some recent antibody research.


Overall, the oligosaccharide component of immunoglobulin G (IgG) accounts only for 2–3% of its mass. Intact antibodies display a characteristic oligosaccharide side chain attached by asparagine 297, present in the antibody Fc region (Figure 1).7 This oligosaccharide plays an indirect but vital role in triggering antibody effector functions, such as antibody dependent cell-mediated cytotoxicity (ADCC), which is believed to be a primary mechanism by which several therapeutic antibodies, e.g., Herceptin (Genentech, South San Francisco, CA), bring about their therapeutic effect. Recent results illustrate that remodeling of the antibody glycocomponent can have a substantial effect on the exact profile and strength of effector functions triggered. It has been demonstrated that the removal of fucose residues from the antibody sugar side chain can increase ADCC activity8 and an engineered CHO cell line has been developed in which the fucosylating enzyme has been knocked out. This facilitates the production of completely defucosylated IgG which displays up to a 100-fold increase in ADCC activity.

Figure 1. The structure of IgG highlighting the oligosaccharide component. Refer to text for detail. Reproduced from reference 7.


An understanding of protein structure–function relationship coupled with the continued development and refinement of molecular techniques allowing the alteration of protein amino acid sequence or PTM complement provides increasing scope for the engineering of biopharmaceuticals to optimize their therapeutic usefulness. Therefore, an overall increasing proportion of future approvals will be engineered in some way, either directly by site directed mutagenesis, or indirectly via the development of engineered producer cell lines capable of producing therapeutic proteins displaying an optimized PTM profile.

Gary Walsh, PhD, is an associate professor in the Industrial Biochemistry Program at the University of Limerick, Limerick City, Ireland, +353.61.202664,


1. Walsh, G. Second-generation biopharmaceuticals. Eur. J Pharm. Biopharm. 2004;58:185–196

2. Walsh, G. and Jefferis, R. Post-translational modifications in the context of therapeutic proteins. Nature Biotechnol. 2006;24:1241–1252

3. Kobata, A. Structure and function of the sugar chains of glycoproteins. Eur. J Biochem. 1992;209:483–501

4. Dwek, R., Butters, T. Platt, F., Zitzmann, N. Targeting glycosylation as a therapeutic approach. Nature Rev. Drug Discov. 2002;1:65–75

5. Wildt, S. and Gerngross, T. The humanization of N-glycosylation pathways in yeast. Nat. Rev. Microbiol. 2005;3:119–128.

6. Li, H., Sethuraman, N., Stadheim, T., Zha, D., Prinz, B. et al. Optimization of humanized IgGs in glycoengineered Pichia pastoris. Nat. biotechnol. 2006;24:210–215.

7. Jefferis, R. Glycosylation of recombinant antibody therapeutics. Biotechnol. Progress, 2005;21:11–16.

8. Satoh, M., Iida, S. and Shitara, K. Non-fucosylated therapeutic antibodies as next generation therapeutic antibodies. Exp. Op. Biol. Ther. 2006;6:1161–1173.