March 15, 2016 (Vol. 36, No. 6)

Glycans Account for Some of the More Ornate Weavings from Nature’s Post-Translational Loom

Through its association with monoclonal antibody quality, safety, and efficacy, glycosylation remains a hot issue for biotech. Accordingly, improving glycosylation is a priority for companies such as Agilent Technologies, which has undertaken several initiatives to improve glycan analysis within the context of the company’s analytical systems.

Based on Agilent’s holistic view of the complete analysis workflow, it has devoted resources toward sample preparation.

2-Aminobenzamide (2AB) fluorescent labeling followed by HPLC or LC/MS remains the gold standard of glycan analysis, but the technique is labor-intensive and introduces the potential for error. “We sought to make glycan analysis as complete, robust, and simple as possible, while focusing on sample preparation,” says Gurmil Gendeh, Ph.D., director of marketing, biopharma market segments, Agilent. “Injection into the column followed by fluorescence detection is the easy part. The bottleneck is getting large numbers of samples ready for analysis. Conventional 2AB labeling can take as long as three days.”

Agilent, in collaboration with ProZyme, automated this sample-prep workflow on its AssayMAP Bravo Automated Liquid Handling Platform. In addition to freeing lab workers from manual pipetting and sample manipulation, automation all but guarantees consistency between runs, between labs, and across geographic locations. Other significant benefits are reduction of human error and the ability to scale throughput to match needs without a proportional increase in labor costs.

Agilent has also introduced AdvanceBio Glycan Mapping, a 2.7 µm, Amide-HILIC column based on a superficially porous particle with reduced diffusion distances. It gives high-resolution separations at lower pressures and enables the use of longer column lengths for increased separation efficiency. Non-UHPLC systems can generally use columns with particle sizes above 2 microns. Sub-2-micron chemistries are also available for analysts seeking higher throughput and resolution.

“The most challenging aspect of analyzing glycans is their heterogeneity and complexity,” Dr. Gendeh maintains. “Proteins consist of only 23 amino acids, and genes of just five bases. But glycans exhibit a wide range of unit structures, which is multiplied by branching and linkage types, which makes them difficult to analyze by mass spec. Plus, they lack chromophores, so unlabeled UV analysis is not possible.”

While single-dimension hydrophilic interaction liquid chromatography (1D HILIC) is nearly always sufficient for monoclonal antibody N-glycans, 1D HILIC columns may not provide the selectivity and sensitivity required for some biopharmaceutical quality and similarity assessments that may have more sialic acids. After collaboration with a customer, Agilent has developed a glycan analysis method based on HILIC in the first dimension, and weak ion exchange in the second.

“Suddenly you see much more resolution and glycan variance that was not apparent with 1D analysis,” Dr. Gendeh explains. HILIC separates based on polarity; ion exchange, based on charge. “Now you can separate based on whether the glycans are neutral or mono- or poly-sialylated.” Sialylation is believed to be responsible for efficacy in therapeutic antibodies.

To resolve the branched glycan structures that occur in samples of (A) fetuin and (B) erythropoietin, Agilent Technologies used a separation approach that combines hydrophilic inter-action chromatography (HILIC) and weak anion-exchange chromatography (WAX). A HILIC/WAX two-dimensional liquid chromatography (2D-LC) run produced this image, which shows highly orthogonal separation. The ion-exchange chromatography in the second dimension reveals the charge pattern of the glycans.

Enzymatic Glycan Release

Releasing glycans from proteins before analysis is still a viable approach, but the method requires serious work. Conventional N-glycan sample preparation depends on reductive amination of aldehyde-terminated saccharides, a process that requires glycans to undergo multiple chemical conversions, including a lengthy high-temperature incubation step. Moreover, reductive amination occurs under anhydrous conditions to minimize desialylation. Sample preparations are therefore burdened with transitioning a sample from aqueous to anhydrous conditions.

“Bioprocessors have explored variations to conventional approaches for N-glycan sample preparation, but have not yet presented a solution that provides not only simplicity within a high-throughput workflow, but also high sensitivity for mass spectrometry,” says Matthew Lauber, Ph.D., principal applications chemist at Waters. Alternative labeling reagents, with functional groups that enhance electrospray ionization efficiency, do not address the cumbersome, time-consuming dependency on reductive amination.

There is therefore great interest in rapid tagging methods that target the precursor glycosylamines of reducing, aldehyde-terminated glycans. One example is a Waters labeling reagent, RapiFluor-MS, which rapidly reacts with glycosylamines upon their release from glycoproteins. The reagent consists of an N-hydroxysuccinimide carbamate tagging group, a quinoline fluorophore, and a highly basic tertiary amine for enhancing ionization. According to the company, the reagent provides rapid labeling kinetics, high fluorescence quantum yield, and significantly enhanced MS detectability.

To accelerate N-glycan preparation, Waters has integrated rapid tagging with enzymatic deglycosylation and HILIC SPE (solid-phase extraction) clean-up for high recovery of the released, labeled glycans.

N-linked glycans are studied by analyzing intact proteins or hydrolysates with MS, or by cleaving the glycans enzymatically and studying them separately. N-linked glycan analysis often suffices for monoclonal antibodies. “But many the emerging therapeutic proteins do not follow this script,” cautions St. John Skilton, Ph.D., senior director of marketing and sales at Protein Metrics. For example, the fusion protein etanercept (Enbrel) has multiple N- and O-glycosylation sites.

O-linked glycans introduce huge complexity and variety into glycan analysis. Protein Metrics is 1 of about 20 third-party software developers that provide capabilities above and beyond vendor-supplied packages for operating MS systems and performing basic spectrum analysis.

Protein Metrics’ programs were developed through collaboration of the company’s glycobiologists and experts in bioinformatics, image analysis, and signal processing algorithms. “Bringing those experts together was critical because they don’t look at these problems the same way mass spectrometrists do,” explains Dr. Skilton. “They were able to dig into what glycosylations actually mean and where they are located on the protein. As a second step, they determined the likely functions of these glycosylations within cellular systems.”

Without software assistance, glycan analysis requires highly skilled experts analyzing raw spectral data by hand. Those individuals are becoming scarce. Protein Metrics asserts that its analysis software can process raw spectral data systematically and in automated fashion. Additionally, Protein Metrics emphasizes that its glycan analysis products are platform-neutral, meaning they can compare results from multiple MS vendors within a unified platform.

Using a Gene Analyzer

In 1994, scientists at Applied Biosystems (currently Thermo Fisher Scientific) demonstrated the use of capillary electrophoresis (CE) and fluorescent dyes for glycan analysis. Two years later, a research group collaborating with the company used a multilane, gel-based DNA sequencer for glycan analysis.

In 2014 and 2015, a research group from Roche published two papers comparing glycan data from an Applied Biosystems multicapillary genetic analyzer with a single-capillary system using laser-induced fluorescence detection and HILIC-based LC, and found data was comparable for the three methods.

The difference is that the Applied Biosystems/Thermo Fisher Scientific multicapillary systems can run 8, 24, 48, or 96 samples in parallel and can detect in several colors. Recently, the company charged Baburaj Kunnummal, Ph.D., product manager for bioproduction and pharma analytics, with developing methods for glycan analysis based on Thermo Fisher’s gene-analysis systems. The end product was GlycanAssure™, a fully integrated glycan analysis system consisting of a multicapillary CE instrument, sample prep kits, and method-specific software. The system was launched in December 2015.

The main advantages of GlycanAssure over existing glycan analysis methods relate to throughput—glycan results from 96 samples in a workday. Other methods provide either high-throughput processing with low-quality data, or low-throughput processing with high-quality. data The single-channel CE and LC methods that are capable of delivering high-quality data can be modified to produce high-throughput processing, provided separation time is shortened—but then these methods produce low-quality data.

“We came up with a complete solution, including sample-prep kits based on magnetic bead separations,” Dr. Kunnummal says. “Traditional glycan sample prep is laborious, with lots of pipetting, vacuum centrifugation, and other manual operations. You’re looking at up to a week to get the data.”

Interestingly, tweaking a multicapillary CE instrument for glycan analysis does not require extensive changes, just finetuning of CE conditions and methods. The capability of the instrument for multiple color detection allows the use of a standard internal label with a second fluorescent dye in every capillary, thus allowing correction of migration anomalies.

The glycan method involves cleaving the sugars from the amino acid backbone; removing the protein; purifying the glycans (by using magnetic beads); labeling the glycans; removing excess fluorescent label (by reusing the magnetic beads); and running on the multicapillary CE instrument. Currently, the method is targeted to analyze N-glycans.

Influencing Glycopatterns

Essential Pharmaceuticals, a supplier of chemically defined media, cell culture supplements, and organ preservation solutions, recently reported data on its Cell-Ess® media supplement demonstrating a 25% improvement in monoclonal antibody yield. Yet the company’s CSO, Adam Elhofy, Ph.D., notes that under certain conditions, supplements and/or feeds that positively influence volumetric productivity can carry an undesirable side effect: changes in glycosylation patterns.

“Whether it’s innovator proteins or biosimilars,” Dr. Elhofy observes, “customers are extremely wary of process changes that affect glycosylation.”

Culture feeds, like peptones, may affect glycosylation patterns, favorably altering a desirable pattern in one region while having a negative effect in another. During development, sponsors settle empirically on optimal glycosylation in the Fab or binding region, but factors that drive beneficial Fab glycosylation may do the opposite in the Fc region. “Developers are forced to find a balance,” Dr. Elhofy declares.

Cell-Ess works in part due to its lipid components, which stabilize the lipid membranes where early- and late-stage glycosylation occurs, mainly the endoplasmic reticulum and Golgi apparatus (where glycan branching occurs).

“The real deal occurs in the Golgi on a lipid membrane, but our supplement stabilizes the entire pathway, including protein secretion,” explains Dr. Elhofy. “A good many post-translational issues, such as protein misfolding, occur when this mechanism is disrupted. Stabilizing the machinery with Cell-Ess allows good things to happen.”

An alternative approach to influencing glycopatterns was unfolding late in 2015, when a startup company called Glycocept obtained from the University of Maryland, Baltimore exclusive licensing rights to a patent and technology for modifying glycosylation on monoclonal antibodies. The technology allows for modifications that alter the antibodies’ effector functions—their ability to kill cancer cells.

One common strategy for modifying antibodies involves increasing their binding to cellular receptors that stimulates killing of tumor cells. Glycocept’s approach, HyGly™, works differently. It changes the antibody’s glycosylation characteristics in a way that decreases its binding to inhibitory receptors. Glycocept expects that this approach will increase the effector function of therapeutic antibodies and will be synergistic with established therapies.

“Countless amino acid and glycan modifications to antibodies have been published, but more often than not they alter affinity to both inhibitory and activating receptors to similar extents and in the same direction,” explains Eric J. Sundberg, Ph.D., CSO at Glycocept and associate professor at the University of Maryland School of Medicine. “It’s a question of balance: You want to decrease or increase binding of the Fc selectively. But if you increase binding by similar amounts to all receptors, you haven’t really achieved anything.”

HyGly modifies antibodies’ Fc region through adding glycosylation sites—achieving “hyperglycosylation”—near the critical glycosylation site on asparagine 297, which is required for antibody activity. By decreasing the relative binding to the inhibitory receptor that diminishes effector function nearly 100-fold, hyperglycosylated antibodies exhibit enhanced effector functions.

Hyperglycosylation has the potential to create biosuperior antibodies, or biobetters. “Preferentially reducing binding to the inhibitory receptor takes the brakes off the effector functions,” Dr. Sundberg tells GEN. “That could result in more potent tumor-killing antibodies.”

Much of the activity of Herceptin, Genentech’s breast cancer antibody treatment, arises from Fc interactions and the induction of effector functions to kill tumor cells. Hyperglocosylation could, for example, create a “super-Herceptin” that is even more effective than the original drug. 

Antibody-modifying technologies typically increase binding of an antibody to a cellular receptor that stimulates the killing of a cancer cell. An alternative technology, Glycocept’s HyGly, alters the glycosylation characteristics of the antibody to decrease binding to a cellular receptor that inhibits killing of a cancer cell. In this image, which depicts Glycocept’s strategy for manipulating antibody effector functions, IgG antibodies are hyperglycosylated at specific sites, resulting in selectively reduced binding to inhibitory versus activating Fc gamma receptors (right), leading to de-repressed effector functions relative to wild-type IgG antibodies (left).

Previous articleSeeing Gene Networks as Social Networks, Analysts Identify Disease Nodes
Next articleClinical Actionability in Cancer