November 15, 2010 (Vol. 30, No. 20)
Vicki Glaser Writer GEN
Conventional Methods Are Inadequate When It Comes to Quantitative and Structural Analysis
Characterization of the carbohydrate portion of glycoproteins in a high-throughput, cost-efficient manner presents myriad technological challenges. Unlike the linear peptide components of these molecules, carbohydrates can be complex, branched structures containing monomeric units that may link an oligosaccharide to a protein at more than one site.
Furthermore, a single sugar molecule can attach to multiple other sugars, and more than one oligosaccharide can attach to a single glycosylation site on a glycoprotein. Oligosaccharides can also contain a number of different sugar residues, some of which may have an identical mass, complicating molecular weight-based analyses.
“An oligosaccharide composed of three monomeric units can have 64,000 potential isomers,” said Fan Xiang, Ph.D., application development scientist at Shimadzu. All of these factors contribute to the complexity of glycosylation profiling, which encompasses a determination of the carbohydrate composition of a glycoprotein and the identity and structure of the various glycan isoforms.
Conventional HPLC and mass spectrometry (MS) technology have been sufficient for more qualitative analysis of the carbohydrate structures added on during post-translational modification (PTM) of a protein, but they have limitations when it comes to quantitative and structural analysis of glycosylation.
“Most groups doing glycan analysis only have MS2 capability,” explained Dr. Xiang, and they are “still struggling to see glycan structure.” Tandem mass spectrometry (MSn) techniques are increasingly being used to define detailed glycan structures—these include MS2 and MS3 approaches, and up to MS5 for more complex isoforms.
“Tandem mass spectrometric experiments have revealed that oligosaccharides might have characteristic signal-intensity profiles that depend on the glycosidic linkage and branching structures,” added Dr. Xiang. By developing an MS spectral library of structurally defined oligosaccharides, MS profiling strategies can then be used to compare the MSn spectra from an experimental analyte to the glycan spectral library, enabling identification of the carbohydrate structure and isomeric components using comparative peak analysis algorithms.
Dr. Xiang and colleagues at Shimadzu collaborated with researchers in the department of chemistry at the University of Michigan, Ann Arbor, on research presented at the ASMS meeting in Salt Lake City. They described the use of MALDI-QIT-TOF MS technology to determine the N-glycan structure of human CD24, a small glycosylphosphoinositol-anchored protein involved in signal transduction in T cells and immune-system activation. ]
The carbohydrate portion of CD24 and, in particular, the negatively charged sialic acid moieties on the sialylated N-glycan structures are believed to mediate protein-protein interactions involving CD24.
Deglycosylation of CD24 with PNGase F was followed by N-glycan purification on a porous graphitized carbon column. Half of the N-glycan fraction isolated was subjected to in-column permethylation. MALDI-QIT-TOF MS was used to analyze both the natural and permethylated N-glycan fractions, and the untreated fraction was analyzed in both positive and negative mode. Using this technique, the researchers identified seven carbohydrates, including three sialylated N-glycans.
From a drug development and regulatory perspective, glycosylation profiling presents additional challenges, noted Dr. Xiang, as there is no agreement or formal documentation on how to submit glycan structure data as part of a regulatory filing.
Presenting at Informa’s “Comparability for Biologics” meeting in Berlin, Parastoo Azadi, Ph.D., technical director at the Complex Carbohydrate Research Center at the University of Georgia, described the tools and techniques used for comparability studies of glycoproteins to assess batch-to-batch variability in glycosylation.
Analytical methods can differentiate between N-linked and O-linked glycosylation, provide data on the percent and sites of glycosylation, and identify what carbohydrates are present at which sites. They can yield both qualitative and quantitative information, with the demand for quantitation increasing and techniques and technology evolving to meet that demand.
“Quantification is still a black box for carbohydrates,” she said. The field is advancing beyond conventional MALDI-MS technology for generating molecular weight data to the more advanced MS-MS techniques that provide sequence information on glycan fractions.
Whereas HPLC of labeled oligosaccharides remains an analytical standard, new derivatization methods to modify and stabilize the carbohydrates for analysis, coupled with MS-MS techniques, “may enable us to do accurate quantitation in the next five years,” she predicted.
Even as they become more quantitative, analytical techniques for characterizing carbohydrates also need to evolve beyond one-sample-at-a-time analysis and enable high-throughput profiling as well as structural analysis and identification of glycans in complex mixtures.
The nonprofit Complex Carbohydrate Research Center not only performs basic research on the synthesis, structural characterization, and medical aspects of glycosylated compounds and polysaccharides, it also operates an analytical services facility and publishes the results of its analyses.
In a study presented at the Society for Glycobiology’s annual conference, scientists described the glycobiological characterization of human biglycan, a proteoglycan component of the extracellular matrix present in a variety of tissue types. Altered biglycan expression has been associated with medical conditions such as osteoporosis, osteoarthritis, corneal diseases, and atherosclerosis.
Dr. Azadi and colleagues characterized the glycan portion of human recombinant biglycan using SDS-PAGE, HPLC, and MS techniques. LC/MS-MS analysis following 18O labeling revealed that two potential N-linked glycosylation sites in the non-proteoglycan form of biglycan are fully glycosylated. The researchers used ESI-MS-MS analysis to identify all major glycosylation of the biglycan proteoglycan.
In a paper published in Cell recently, Matthias Mann, Ph.D., and colleagues from the Max Planck Institute of Biochemistry and Harvard Medical School, described the use of a filter-aided sample preparation (FASP) method they developed to enrich for the N-glycoproteome component of tissue and plasma samples.
The authors propose that the “830 mouse and 1998 human N-glycosylation sites” that comprise the Swiss-Prot database (“the most comprehensive resource of annotated N-glycosylation sites”) “is likely a drastic underestimate of the true extent of the mammalian N-glycoproteome.”
In the study described, they used the FASP-based N-linked glycopeptide capture method and orbitrap mass spectrometry to analyze the peptide fragments derived from mouse tissue and plasma samples. The resulting data not only included 74% of known mouse N-glycosites, but also allowed them to identify an additional 5,753 glycosylation sites. In a single LC/MS-MS run using 200 micrograms of starting material, the researchers were able to map more than 2,000 sites.
FASP incorporates an SDS solubilization step that enhances the enrichment efficiency for glycosylated membrane proteins. It enriches for these glycoproteins by “at least 10-fold,” said Dr. Mann.
After on-filter protein digestion, a lectin-based peptide affinity reagent is added to the top of the filter. Glycosylated peptides bind to the lectin and are retained on the filter. They are subsequently washed through the filter and deglycosylated using PNGase F. 18O water incorporation followed by MS analysis pinpoints the glycosylation sites on the peptides.
Dr. Mann’s group is using this method to look for plasma biomarkers in tissue samples from patients with cancer. Changes in the glycosylation status of secreted or membrane-bound glycoproteins in circulation may reveal patterns of increased or decreased glycoprotein expression that can serve as biomarkers to aid in cancer diagnosis and prognosis, and for evaluating disease progression and the risk of metastasis.
Critical Quality Attributes
Daryl Fernandes, D.Phil., founder and CEO of Ludger, led a workshop on comparability issues related to glycosylation throughout the life cycle of a biopharmaceutical product at the “Comparability for Biologics” meeting.
Ludger performs glycoprofiling for the biopharmaceutical industry and develops methods to measure and control glycosylation during biotherapeutics production. The company is also developing a framework for glycosylation analysis that combines HPLC and MS technologies for separation, quantification, and identification of carbohydrates.
Dr. Fernandes emphasized the importance of adopting Quality by Design (QbD) strategies that rely on the identification of critical glycosylation parameters early in the product life cycle. “A sensible interpretation of QbD will help determine significant differences in comparability of glycosylation,” he said.
Virtually every batch of a glycoprotein will have variation in the glycan composition, but some variation will have no effect on the safety or efficacy of the product. “QbD offers an opportunity to deal effectively with the complexity and variability of drug glycosylation,” he added. Each biotherapeutic product is different, so each requires an individualized model to define consistency.
Manufacturers of glycoprotein therapeutics should identify glycosylation critical quality attributes (GCQAs) early in the product life cycle, in Dr. Fernandes’ view. GCQAs can be used to determine which glycosylation features should be eliminated or modified to optimize a drug’s safety and efficacy profile. They can help define the ideal glycosylation profile of a biotherapeutic and guide critical decision-making, related for example to selection of an expression system or of a particular cell line or clone that will yield optimal glycosylation.
Identifying and prioritizing GCQAs for process design and small-scale process development, or in scale-down process modeling, can assist in predicting variability in glycosylation that may occur as a result of changes in cell-culture conditions during scale-up. Selection of glycoprofiling methods and analytical tools for monitoring glycosylation during process design and development and throughout production should take into consideration the most current regulatory guidelines.
M-Scan presented workshops on strategies and analytical techniques used for biopharmaceutical characterization and comparability studies at the recent “BioProduction” meeting in Barcelona and at the “BioAnalytical Congress” in Brussels.
A contract services provider that specializes in PTM characterization and protein and carbohydrate structural analysis, M-Scan utilizes MALDI-MS, electrospray (ES)-MS, and high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) technology for glycosylation-site analysis and glycan profiling.
Regulatory requirements for glycosylation are “becoming more stringent,” said Fiona Greer, global director of quality assurance, a result of the growing recognition that the carbohydrate portion of a glycoprotein may affect the molecule’s ability to fold properly, its stability, its ability to communicate with surrounding cells, and various other factors that can impact the safety and efficacy of a biotherapeutic agent including its circulating half-life, function, and immunogenic potential.
As PTMs are highly product-specific, glycosylation analysis is necessarily product-dependent, and a glycoprofiling strategy must be developed for each individual biotherapeutic product. M-Scan designs a step-wise analytical strategy that includes MS-based and non-MS methods to provide an overview of the total carbohydrate structure of a glycoprotein.
Monosaccharide composition is identified and quantified by MS and/or HPAEC-PAD, with particular attention paid to sialic acid analysis. Gas chromatography (GC)-MS with methylation provides information about the linkages of individual sugars, and MALDI-MS and electrospray LC-MS techniques can be used for glycan population analysis, to obtain structural data, and to identify sites of glycosylation. For a more in-depth analysis, the company relies on MS-MS techniques to perform antennal profiling analysis.