Send to printer »

Feature Articles : Mar 1, 2010 (Vol. 30, No. 5)

PTMs Progress Toward Designer Proteins

Advances in the Analysis and Biosynthesis of Glycans Are Transforming the Field
  • Angelo DePalma, Ph.D.

Of the essential post-translational modifications (PTMs) that proteins undergo before their synthesis is complete, glycosylation is arguably the best known. The medical and economic significance of monoclonal antibodies and other glycoproteins has spurred a resurgence of interest in glycans, particularly in their analysis and ways to direct their biosynthesis in desirable ways.

The U.S. Pharmacopeia (USP), for example, develops both documentary and physical standards for analyzing and characterizing glycoproteins. The standards are critical for assuring the purity and quality of these products.

USP defines standards as “horizontal” if they are not linked to a particular product but to a procedure in a General Chapter, for example, the bacterial endotoxins test in USP <85>. “Vertical” standards are linked to a specific product, i.e., a monograph for a drug substance or product and associated reference materials.

By the time you read this, USP will have published a new standard, “USP Chapter <1084>, Glycoprotein and Glycan Analysis—General Considerations,” in Pharmacopeial Forum (PF36(2)). The standard, which will be subject to a 90-day public comment period, will contain general guidance for glycan and glycoprotein analysis.

Future, procedure-driven tests and chapters will be linked to physical reference standards currently in development. In that regard, Chapter <1084> is an introductory or “umbrella” document. USP has pursued a similar approach with offerings on nucleic acid based techniques (e.g., USP General Information Chapters <1125>-<1130>, now official).

Tina S. Morris, Ph.D., vp for biologics and biotechnology at the UPS, describes the chapter as an “overview on how to approach the analysis of a glycosylation product in light of regulatory expectations.” For example, the chapter will use flow charts or “decision trees” that guide users through analytical options, depending on whether the focus is an intact glycoprotein or cleaved glycans.

According to Dr. Morris, USP enjoys “very good” relations with FDA, which routinely sends liaisons to USP standards committees. Chapter <1084> was written by an ad hoc advisory panel composed of experts from industry, academia, the U.K.’s National Institute for Biological Standards and Control, and USP. “Regulators usually provide very helpful input,” she says.

PTM research is increasingly regulation-driven, says Elizabeth Higgins, Ph.D., CEO of GlycoSolutions, which provides analytic support for developers of glycoproteins. As evidence mounts that glycosylation affects glycoprotein (particularly mAb) safety and efficacy, biopharmaceutical manufacturers use glycosylation as a surrogate, to monitor lot-to-lot consistency. “Companies are increasingly interested in measuring galactosylation or core fucosylation of the Fc glycosylation,” observes Dr. Higgins.

GlycoSolutions’ business is growing, in no small part due to interest in biosimilars. Knowing a priori which glycans confer optimal efficacy and safety has been a goal of glycoprotein research, but hard-and-fast rules that apply to all classes of protein have been difficult to formulate due to a scarcity of applicable data. For biosimilars, however, developers are aiming for glycosylation patterns that are similar to the existing drug. 

GlycoSolutions worked on 20 projects in 2008, 25 in 2009, and expects 30 customers in 2010. “Biosimilars represent our largest growth area right now. We’ve also been expanding our services,” Dr. Higgins says. GlycoSolutions’ workload last year involved mAbs (29%), non-mAb glycoproteins (17%), biosimilars (33%), one gene- therapy project, and one biofuel project. Ninety-four percent of their projects involved development-stage proteins.

Pauline Rudd, Ph.D., a professor at the NIBRT Dublin-Oxford Glycobiology Laboratory, has spent much of the past decade developing chromatographic methods for analyzing the glycosylation of therapeutic proteins. Her technique involves cleaving the sugars enzymatically and attaching fluorescent labels to the glycans. This is followed by analysis by HPLC and fluorescence detection.

Dr. Rudd employs dedicated computer software to compare peaks with entries in a glycan database, which she has constructed by cleaving dozens of glycosylated proteins and analyzing the glycans released. The software/database analyze the exoglycosidase array digestions for final structure assignment, which includes monosaccharide sequence and linkage information.

Recently, Dr. Rudd’s group acquired liquid-handling capability for automating the analysis, which she says takes around eight hours from taking a sample from cell culture to data analysis. The idea is to complete the analysis within the time frame of one biomanufacturing shift.

This standard analysis will not reveal where the glycans were originally located on the protein. That requires site analysis, an extremely time- and labor-intensive process. “Site analysis isn’t easy. You really need a good reason to justify carrying one out,” says Dr. Rudd.

Two interesting applications of Dr. Rudd’s technique are quality by design  and process analytic technology. Due to the complexity and time involved in the analysis (Dr. Rudd claims a five-hour turnaround for immunoglobulins), it is not suitable for real-time quality monitoring. However, it is rapid enough to enable tweaking conditions or for making changes to feeds during the time frame of a several week-long cell culture.

In 2008, Dr. Rudd’s group at NIBRT entered into a research collaboration with Lilly and in 2009, a collaboration with Roche was set up for cell-culture monitoring providing Roche with access to NIBRT’s glycan database and software. Also in 2009, the institute announced a research collaboration with the FDA on the characterization of glycosylation on theraputic enzymes under different bioprocessing conditions.

Designer Proteins?

Just as humanized antibodies seek to duplicate the amino acid sequence produced in humans, researchers are hot on the heels of “designer” glycoproteins with tailor-made glycans. Since glycosylation cannot be genetically encoded, investigators must resort to artificial means to achieve desirable, glycosylation patterns. “Glycan trimming,” a method employed by Heather Desaire, Ph.D., a chemistry professor at the University of Kansas (Lawrence), may be one step toward glycosylation-by-design.

The subtractive method uses glycosidase enzymes, in Dr. Desaire’s case a-mannosidase, to remove mannose monosaccharides selectively from high-mannose glycans and thereby “expose” desirable sugar residues. Enzymes specific for other sugars may also be used.

“The method is designed so that either an individual monosaccharide type is removed, or, if more than one deglycosylating enzyme is used, multiple monosaccharide types could be trimmed from the glycoprotein,” Dr. Desaire says. After the enzymatic reactions, purification and analysis ensures the desired glycosylation profile was obtained, and that the protein folded properly.

Structure-activity relationships for glycosylation are still in an early stage. Most correlations are empirical. Given that limitation, glycan trimming may be employed to improve circulating half-life, in the case of mannose-trimming, by reducing the opportunity for mannose binding lectin (MBL) to clear glycoproteins from circulation.

“We have not shown experimentally that trimmed proteins have longer circulation lifetimes, but the binding of MBL to high mannose glycoproteins is well known,” comments Dr. Desaire. Other potential benefits are reducing immunogenicity by removing sialic acid and other immunogenic residues, and improving stability or solubility.

For years protein glycosylation was believed to be an exclusive trait of eukaryotes. The discovery, in 1999, that the bacterium Campylobacter jejuni added glycans to proteins, was therefore a startling revelation. Research by Christine Syzmanski (National Research Council of Canada), and Brendan Wren and Rebecca Langdon, Ph.D., a scientist at the London School of Hygiene and Tropical Medicine, among others, has firmly established that bacteria conduct widespread N-glycosylation of proteins.

Dr. Langdon’s group has identified genes implicated in protein glycosylation in several different Campylobacter and Helicobacter species. “While we now have a good understanding of how N-glycosylation functions in C. jejuni, we are still unclear of its role in the biology of the bacterium and also how the system works in the other bacteria where the genes have been identified,” Dr. Langdon says.

Using a technique known as protein-glycan-coupling technology, she has expressed protein-glycosylation genes in E. coli, an organism more familiar to bioprocessors than Campylobacter or Helicobacter. Dr. Langdon is currently developing this system to produce protein-glycoconjugate vaccines.

Other groups, particularly Professor Mario Feldman’s at the University of Alberta are working on O-glycosylation as well. However, with O-glycosylation there is less control over the protein acceptor site. “Using N-glycosylation, we can engineer proteins to include the acceptor site for attachment of the sugar substrate. This creates the opportunity to theoretically link any polysaccharide to any protein.”

The implications for (future) glycoprotein manufacture in simple bacterial systems are tremendous, but for now the near-term potential remains in glycoconjugate vaccines. Dr. Langdon has been collaborating with GlycoVaxyn, to commercialize such vaccines.

PTM-Based Medicine

PTM is most often associated with glycosylation, but numerous other modifications occur post-translation. Among the most significant is the formation of disulfide bonds that are critical for protein folding. Researchers at SuppreMol are using a type of chemical oxidation to form disulfide bonds in SM101, an early clinical-stage drug for autoimmune diseases.

SM101 is a recombinant, soluble, non-glycosylated version of FcγRIIb, a human Fcγ receptor that binds to autoantibody/autoantigen complexes and blocks Fcγ receptors on immune cells. When FcγRIIb is working properly the immune response is attenuated, and the inflammation cascade typically of autoimmune diseases is prevented.

Patients with certain autoimmune diseases show levels of soluble Fc receptors in their blood that are significantly lower than in healthy individuals. SM101 essentially supplements these under-represented soluble receptors. The absence of glycans allows SuppreMol to manufacture the drug in E. coli instead of cell culture.

Under a European orphan drug designation, SuppreMol is targeting SM101 to idiopathic thrombocytopenic purpura, an autoimmune disease that depletes platelets (resulting in spontaneous, sometimes life-threatening bleeding), and causes the spleen to swell. The drug has successfully passed through early Phase I; a Phase Ia/IIb patient study is expected to begin in early 2010. Lupus and rheumatoid arthritis are other potential diseases treatable with SM101 that will be addressed at a later stage of development.

Current treatments involve knocking down the immune system or stimulating platelet formation. “But making more platelets also leads to an increase in auto-antigens,” notes Peter Sondermann, Ph.D., CSO. “The right strategy probably involves balancing the immune system against platelet formation. We believe we can modulate the immune system in a way that platelets are not regarded as an antigen, at least for a distinct time.”

The PTM angle here is Dr. Sondermann’s technique for forming disulfide bridges within SM101 after it is released from inclusion bodies in E. coli. During the preclinical and early clinical development phase, this was accomplished using oxidized glutathione in the presence of arginine that supports the folding, but both substances are expensive and must be produced in microorganisms.

“Because we’re making a drug with it, you have to track where the material comes from,” Dr. Sondermann says. The synthetic disulfide-bridge mediating molecule cystamine and the folding-assistant urea are now used in the process. Both compounds are inexpensive, widely available, and reduce safety concerns.