Special Report on Cell Biology: Sweetening the pot

Is improved understanding of glycobiology the secret to improved biologics?

Register for free to listen to this article
Listen with Speechify
As the alarm clock jars you from your too-brief slumber, you throw your legs over the side of the bed and face the day’s first decision: What am I going to wear?
Immediately, the follow-up questions fly.
What’s the weather supposed to be like today? Am I walking to work or driving? Did I grab the laundry out of the dryer last night?
Depending on your environment, you might be thinking trousers or a skirt, button-down shirt or sweater, sandals or boots.
And of course, your job will also dictate your choices.
If you have an office meeting, maybe the dress shirt, tie and jacket, rather than the t-shirt. If you’re a firefighter (like so many DDNews readers), then maybe it’s boots, heavy jacket and helmet. New parent, anything easily washed.
So many factors—environment, function, safety—dictating how you decorate your body.
The same is true for the vast majority of proteins floating through and on eukaryotic cells. Only instead of footwear and headgear, these proteins (and many lipids) are decorated with sugars.
The glyco-basis of health
Alongside nucleic acids, proteins and lipids, saccharide polymers known as glycans are a major component of every cell. As such, they are involved in almost every cellular process, as well as processes that occur between cells and within whole organisms.
Glycans, for example, are the basis of the ABO blood group system, where the difference between the oligosaccharide antigens results from variants of a single glycosyltransferase gene. They also facilitate immune recognition of self versus non-self via the glycocalyx, a crust of glycans attached to proteins and lipids that surround cells. And through the binding of selectins, glycans mediate leukocyte recruitment and migration involved in inflammation.
The choice of one glycan over another can also significantly impact the structure and function of immunoglobulins, of which IgG has been the best studied.
Changes in glycoprofile can alter the conformation of the Fc region, for example, which has significant consequences for IgG effector functions. Similarly, more than 95 percent of human IgG molecules carry a core fucose residue that dramatically impacts the antibody’s ability to bind a receptor on the surface of innate immune cells such as natural killer cells and macrophages.
“The presence of a high proportion of IgG which is core-fucosylated therefore represents a ‘safety switch’ which attenuates potentially harmful ADCC [antibody-dependent cellular cytotoxicity] activity,” explained Genos’ Gordan Lauc and colleagues in a recent Biochimica et Biophysica Acta review. “By contrast, ADCC induced by non-fucosylated IgG seems to be one of the primary modes of function of therapeutic anticancer monoclonal antibodies, since IgG molecules lacking core-fucose are over 100 times more effective in initiating ADCC.”
On the flip side, autoimmune diseases such as rheumatoid arthritis, inflammatory bowel disease and systemic lupus erythematosus have been associated with diminished galactosylation of IgG antibodies.
But whereas some of these changes in glycosylation patterns lead to disease, other changes may be more reflective or symptomatic of disease, without necessarily contributing to its pathology.
As Lauc and his colleagues point out, while many glycans may actively participate in cancer, influencing processes like tumor proliferation, invasion and metastasis, others are components of glycoproteins currently being used as biomarkers of disease, such as prostate-specific antigen, alpha-fetoprotein, carcinoembryonic antigen and MUC-1. The glycosylation state of these proteins is significantly altered in cancer.
“I think sugars are really important because they are a very rapid way of a cell changing something fundamental without having to evolve a change in the genome,” explains Pauline Rudd, a principle investigator at the National Institute for Bioprocessing Research and Training (NIBRT) and a visiting principle investigator at the Bioprocessing Technology Institute (BTI) at A*STAR, Singapore’s Agency for Science, Technology and Research. “The environment makes an impact on glycosylation. If you have a fever or cancer, your protein glycosylation changes in response. If you get sunburned, your skin cells change.”
“Glycosylation is determined by the physiological state of the glycosylation machinery and the nature of the proteins undergoing glycosylation,” echoed Macquarie University’s Morten Thaysen-Andersen and colleagues recently in Molecular & Cellular Proteomics. “Jointly, these attributes determine the repertoire of glycans present on synthesized glycoproteins (glycoforms) and create the important features of protein site- and cell-specific glycosylation. Protein glycosylation is therefore a spatiotemporal dynamic modification that cells can utilize to respond to the constantly changing milieu.”
For Rudd, understanding pathways of disease is about taking an omics-wide approach, as exemplified by her efforts to determine whether glycans could be a bellwether of cancer pathogenesis and prognosis.
Working with Agilent Laboratories’ Zohar Yakhini, Oslo University Hospital’s Anne-Lise Børrensen-Dale and others, Rudd’s group performed N-glycan analysis on serum from breast cancer patients. The team then correlated these results with analyses of the transcriptome, circulating tumor cells (CTCs) and clinical outcomes.
The researchers identified a series of statistically significant correlations, including an association of one group of glycans with low expression of transcripts related to DNA-replication and mismatch repair, and another associated with increased energy production. Similarly, one glycan peak was associated with diminished expression of messages linked to cell adhesion, which might ultimately influence tumor cell mobility and therefore metastasis potential.
Just as importantly, the group identified glycan structures that correlated with clinical outcomes, such as survival and primary tumor size, as well as CTCs.
“We have demonstrated how integrating serum glycomics with gene expression from breast carcinomas may be used to identify serum glycans related to breast carcinogenesis and functional processes in the tumour,” the authors wrote. “This approach may improve the search for biologically relevant serum markers of malignant disease.”
Also looking to leverage glycan profiles, Glycotest has developed an assay based on the identification of abnormal levels of fucosylation or aberrant fucosylation patterns in serum proteins to monitor onset and progression of conditions such as hepatocellular carcinoma, cholangiocarcinoma and fatty liver diseases. Based on engineered recombinant lectins, the assay can be used as panels or for single biomarker detection.
And yet, despite the central role played by glycans in all aspects of biology and disease, they are possibly the least well understood and characterized of the biomolecular quartet.
Accessing the glycome
“If you think of all the funding that has gone into genomics and proteomics, glycomics by comparison has had far less funding,” Rudd noted in an NIBRT video. “And I think now the attention has to turn not only to glycomics, but to other post-translational modifications.”
Part of the challenge that has slowed progress in glycan analysis may be due to the inherent complexity of glycobiology. Unlike genomics, which deals with genes that fundamentally all have the same chemistry, and proteomics, which largely deals with linear amino acid sequences that fold into relatively confined structures, glycomics involves simple sugars that can form branched structures and exhibit a certain degree of heterogeneity from glycoprotein to glycoprotein, cell to cell and minute to minute.
“We still make a lot of generalizations when we talk about proteins, sugars and genes too, for that matter,” says Rudd.
“At the protein level, we talk about full characterization,” she continues. “But if we ask what a single copy of a protein looks like, we don’t know. We can say that a percentage of the sample has C-terminal clipping and we can measure the average number of phosphate groups attached to one copy of the protein.”
However, if we take a single copy of a protein with C-terminal lysine clipping, she explains, we cannot say precisely where the phosphates are. Likewise, we cannot yet say exactly what one copy of a glycoprotein looks like in terms of its glycan structures.
“We can only refer to averages and probabilities but our bodies don’t see averages, we see individual molecules,” she presses. “We still need technology development to understand the complexity itself and its relevance to function. This may well be a point where reductionism ceases to be useful.”
In the introduction to a special issue on glycosylation and immunology in Seminars in Immunopathology in 2012, Genentech’s John Lowe tried to encapsulate the inherent complexity issue in a few sentences.
“There are literally hundreds of distinct glycan structures known in any given mammalian species,” he wrote. “For any given glycoprotein, the glycoprotein is typically not homogeneous with respect to the glycan structures that decorate the protein. At any specific glycosylation site on a specific glycoprotein, individual glycoprotein molecules may differ by displaying distinct structural variants of the glycan.”
“Further combinatorial complexity is introduced when a glycoprotein has more than a single glycosylation site,” he continued. “Additional complexity obtains because glycan structures are dictated by the glycosylation machinery expressed in the cell type where the glycoprotein is expressed. This repertoire typically varies, qualitatively and quantitatively, among cell types within a mammalian organism, often in a lineage-specific manner, and can vary with the physiology or pathophysiology of the cell or organism.”
“The whole world is much more complex than either the genomics—or indeed, any of the other omics groups—can deal with at the moment,” explains Rudd, turning that challenge into an advantage for glycobiologists.
“Because of the glycan heterogeneity, which means that one site on a protein can contain a range of glycans, we’re very used to dealing with the complexity, and even if we don’t understand it, we’re not really phased by it,” she says, adding with a laugh: “Maybe we should have more of a voice!”
Fortunately, this complexity has been met with technological improvements on a number of fronts, as Thaysen-Andersen and his colleagues explained.
“Glycoproteomics has been a beneficiary of the continual performance enhancements of modern mass spectrometers including improved speed, sensitivity, resolution and accuracy, most notably implemented on the latest Q-TOF (Sciex, Waters, Agilent, Bruker) and on multiple Orbitrap (Thermo) instrument platforms,” the authors wrote.
“Specifically, key advances have been made in the enrichment of intact glycopeptides from complex peptide mixtures, in LC-MS/MS-based detection of intact glycopeptides through optimized dissociation and acquisition styles of glycopeptides, and in data handling for more automated, yet still confident, glycopeptide identification and quantitation.”
Improvements in sensitivity and throughput are only part of the equation, however. For Rudd, it is also a matter of making the workflow more approachable for scientists who haven’t spent their careers considering glycans.
“My group and others have worked really hard to make an automated workflow that means now people with very little previous understanding of glycan analysis can actually do the more straightforward glyco-analytics with the minimum need to know arcane information,” she says. “The hope is that this will open up the field.”
She gives the example of Waters Corporation’s QDa mass detector, which goes in-line with a liquid chromatography separations system and, she explains, “automatically gives you the mass spec of each glycan in a pool even if you know only the basics of mass spectrometry, because it has already has the variable parameters set.”
For simple or routine glycan analysis, this allows scientists who are not trained in the use of high-end mass spectrometers to obtain good data.
Rudd’s group has worked to improve each step of the analysis from glycoprotein enrichment using newly developed carbohydrate affinity ligands to enzymatic glycan release and fluorescent labeling to developing a database of 350 or more experimentally derived glycan structures that can be probed by LC-MS software.
For some researchers, an inherent challenge of these technologies is that they essentially destroy the product being studied through mechanisms such as digestion and fragmentation, potentially removing some of the vital biological context.
To maintain that context, several groups have turned to lectin arrays as a method to identify the glycan structures of proteins. Lectins are themselves proteins that bind to specific glycans, and can be dotted onto arrays. (See also sidebar How’d they do that? after the end of this main article.)
In a proof-of-principle study they described in mAbs, Baolin Zhang and colleagues at the Center for Drug Evaluation and Research recently analyzed a panel of 15 therapeutic proteins, including eight monoclonal antibodies, with a commercial array of 45 lectins. The goal was to see if the array results matched the known glycan patterns.
Using GlycoTechnica’s evanescent-field fluorescence scanner, the researchers noted that the chip results were largely consistent with the known glycosylation patterns, and that the arrays were particularly sensitive to alterations of the terminal glycan structures. And these results, the authors were quick to note, were achieved on chips that were not designed or optimized for this purpose.
“The lectin microarray, when coupled with a sophisticated detection system, appears to provide a high-throughput platform for rapid screening of glycan profiles of therapeutic proteins,” the same group suggested in a separate review of glyco-analytical methods, also published in mAbs. “However, the lectin-based microarray still faces major challenges before being adopted as a tool for characterization of therapeutic glycoproteins.”
“In this regard, most commercial lectin microarrays utilize lectins from natural sources (e.g., plants) that are expected to display weak binding affinities to a spectrum of glycans,” they continued. “This phenomenon complicates interpretation of the detected binding events with a specific lectin molecule.”
“Nonetheless, the lectin microarray platform appears to allow high-throughput screening of the presence or absence of specific glycan variants in glycoprotein samples.”
SRI International’s Denon Wang and colleagues, meanwhile, took the opposite approach to array technology, spotting different carbohydrates and probing the chip with an antitumor monoclonal antibody first characterized in mice, but found to cross-react with several human cancers.
As they reported in the Journal of Immunology Research, the researchers produced an array of various natural carbohydrate antigens, including several blood grouping antigens and their precursors. Probing the array with HAE3, they identified a glycan that is normally hidden in cells—a cryptic glycan—but becomes exposed during malignant transformation.
They then used fluorescence-activated cell sorting to determine in what tumor types the glycoepitope was prominent, ultimately seeing strong positive signals in several breast cancer cell lines, including hard-to-treat triple-negative cell lines.
“The potential of these types of cryptic glycans in detecting breast cancer and targeting it through immunotherapy warrants further investigation,” Wang enthused in a media announcement of the study. They hope to extend the work to breast cancer patients to see if the marker is associated with metastasis or CTCs.
Rather than worry about glycoform heterogeneity, however, some companies are trying to minimize it and potentially enhance glycoproteins to make them safer and more efficacious.
Drug development
“Currently, the majority of recombinantly produced biotherapeutics are either non-glycosylated proteins like insulin, or antibodies, which have a very simple glycosylation pattern,” explains Nicole Faust, chief scientific officer of CEVEC Pharmaceuticals.
“Proteins with more complex glycosylation patterns like blood coagulation factors or other serum proteins (e.g. alpha-1 antitrypsin or C1 esterase inhibitor) have classically been purified from human blood plasma,” she continues. “This has by default led to products with a desirable glycosylation pattern, i.e., identical to what is found in a healthy human being.”
Unfortunately, she adds, macro- and micro-heterogeneity—variances in the occupancy of glycosylation sites of a protein or differences between glycan structures at the same glycosylation site between individual molecules, respectively—are common and generally undesirable in recombinant products. Such variances lead to a heterogeneous product after the upstream process and make extensive labor-intensive downstream processing, associated with significant reduction in product yield, necessary.
To address this issue, CEVEC has developed a series of cell lines—their CAP-Go platform—engineered to preferentially generate specific glycan structures on proteins. The result is a product that predominantly consists of the favorable glycoform, Faust explains, and requires less downstream processing.
“With CAP cells, macro-heterogeneity is much less pronounced than with other mammalian expression platforms,” she suggests. “Additionally, we can significantly reduce micro-heterogeneity.”
The first products generated on this platform are in the transition from discovery to production, so the company is currently collecting data to determine product consistency during the transition from discovery to production.
“The current cell line portfolio of about 15 cell lines covers what we and our customers currently see as the most important modifications,” she continues, adding that the company is still open to generating novel lines for specific projects with its customers/partners, including other post-translational modifications such as phosphorylation.
“The manufacturing condition can have a huge impact on the glycoprofile,” says Michael Marit, director of drug discovery at PlantForm, a company using tobacco plants to produce biologics (see sidebar Grow ‘em if you’ve got ‘em after the end of this main article).
“There’s evidence in lot-to-lot variability for innovators to have 10 to 20 percent differences in specific glycan species based entirely on manufacturing conditions,” he continues. “So, we’re just starting to understand as a field the impact of our conditions on glycoprofiles, and also the impact of glycoprofile on the function of our molecules.”
For its part, Glycotope has engineered human suspension cell lines as part of its GlycoExpress (GEX) platform, optimizing the glycosylation patterns of a number of biologic products, including follow-ons to cetuximab and trastuzumab. As with the other research, Glycotope has shown that by minimizing fucosylation and maximizing galactosylation and branching, it can significantly increase the ADCC of its antibodies, reporting a 250-fold improvement in preclinical analysis of its CetuGEX.
Not limited to cancers, however, Glycotope signed a partnership agreement last autumn with Octapharma giving the latter exclusive access to a preclinical portfolio of blood coagulation factors developed with the GlycoExpress platform. The deal also sees Octapharma take a stake in its partner.
Her colleagues in BTI are also working on glyco-optimization, using gene-editing techniques such as ZFNs, TALENs and CRISPR-Cas9. Recently, for example, they created a CHO cell line in which they inactivated the GDP-fucose transporter gene to produce fucose-free antibodies, confirmed by the glycoanalytical group.
As they described earlier this year in Biotechnology Journal, Song Zhiwei and colleagues used the system to produce trastuzumab lacking the core fucose and thereby stimulated improved ADCC. Just as importantly, they noted that the inactivation of the transporter had no impact on cell growth or antibody production.
Although many of these groups are using these lines to produce new products, much of the proof-of-principle work is in the development of biosimilars or biobetters, which offers a potential monetary bonanza as many biologics come off patent. Late last year, Allied Market Research suggested the global market for biosimilars could jump from $2.5 billion in 2014 to $26.5 billion by 2020.
At the same time, are these companies simply trying to reinvent the wheel by trying to improve on existing products?
“Herceptin was produced in these early CHO cell lines that had galactose and had fucose, and it has a certain efficacy,” Marit says. “So, the question right now that a lot of companies are asking is: If we produce a version of Herceptin or trastuzumab without the α-1,6-fucose, are we going to see a better ADCC? How much of its efficacy is based on its ADCC function?”
And perhaps more importantly, he presses, “will that make a clinically significant difference, and will it justify putting a new molecule through clinical trials?”
Biobetting on biobetters?
“In the last five years, there’s been a really big push for detailed analysis of glycosylation and its role in the efficacy of the product,” Marit suggests. “Cetuximab contains an α-gal linkage, which produces a hypersensitivity reaction in about 20 percent of patients. These patients can’t take the drug, and it’s because it is produced in mouse cells, which have this unique glycan linkage.
“So there’s been a lot of concern about the role of glycans in safety, and more stringent requirements for biosimilarity to ensure that follow-on or biosimilars mimic as closely as possible the innovators.”
PlantForm is not the only company looking to leverage plants for biologics production. Protalix, for example, uses both tobacco and carrot. Unlike PlantForm, however, which grows entire plants, Protalix only grows plant cells in bioreactors.
Either method offers advantages over mammalian cells, but also comes at the cost of producing glycan signatures different from the more familiar CHO or human cell lines.
“Plants have two specific glycan-linkages that don’t exist in mammals,” Marit explains. “They have a xylose and a fucose linkage, and these two linkages have been shown to be involved in seasonal allergies. So there is reason for us to want to eliminate them in our intravenous drug products.”
By the same token, he points to products from Protalix that contain both xylose and fucose and yet didn’t elicit immune responses.
“That shows that we don’t fully understand the roles of specific glycans and how they may interact with the immune system, and whether or not it is protein-dependent,” he muses.
Not wanting to worry one way or the other about the non-mammalian glycan structures, PlantForm has largely engineered xylose and fucose out of their tobacco plants.
“The system uses an RNAi cassette that targets the enzyme that transfers the problematic fucose and xylose,” Marit explains. “There’s no enzyme produced and therefore no problematic sugar linked to the protein of interest.”
“When we do this, we get a very homogeneous glycoprofile, but it doesn’t have galactose on it, which all of these CHO-produced innovators have,” he continues. “And they don’t have α-1,6-fucose, which all of these CHO-produced innovators also have.”
The company has put its lead product—trastuzumab—through a variety of analytical tests that have shown their product is very similar to the innovator, suggesting the slight differences in glycoprofile have little impact on drug performance.
“To my knowledge, there are no biosimilars available that have been produced in a host different from the innovator,” Marit offers. That may represent a unique challenge to the company, he admits, when they submit their first clinical trial application to regulators more accustomed to mammalian expression systems.
“We do anticipate a challenge here, but we hope that the argument from safety and efficacy will be enough for the biosimilar path,” he says.
Bottom line
In a world of shrinking healthcare budgets, however, regulators and the industry may have little choice in exploring other options for producing follow-on products and next-generation biologics.
“There are outstanding questions about glyco-optimization and the role of glycans, and I think there is a lot of room for improvement there,” Marit admits. “I think it just depends on how much risk companies are willing to tolerate for early trials to see what the difference is.”

Grow ‘em if you’ve got ‘em
In something of a botanical redemption story, small biotech is doing its best to reverse the shame of Big Tobacco by turning the maligned tobacco plant into a health factory. Specifically, Canada’s PlantForm is using the plant to produce glycosylated antibodies and other biologics for a burgeoning biosimilars and biobetters market.
As the company’s director of drug development Michael Marit explains, PlantForm pursued tobacco for a variety of reasons, many of which have to do with large-scale production.
“[The plants are] extremely hardy,” he explains. “So, we have low capital costs, low production costs relative to the bioreactors in the large production facilities that are required for mammalian cell culture.”
And this has a direct impact on scalability as products move from research bench to production floor.
“If you need more product in a CHO system, you need another fermentor; that takes time to grow,” he continues “For us, we just need more controlled growth or greenhouse space, which is easy; we just grow more plants.”
Product safety is also a significant factor in the decision to go with a plant-based system.
“Because this is a non-mammalian system, we don’t have as much concern for bacterial or viral contamination in our system and product,” Marit presses. “CHO cells and mouse cells, for example, share some viral infections with humans.”
Part of the reason for choosing tobacco plants, in particular, was the ability to take advantage of a natural relationship between the plant and its microbial enemy Agrobacterium, which scientists long ago turned into an expression vector.
Agrobacterium will infect the roots of a particular plant species and cause tumors,” Marit explains. Thus, by inserting genes (e.g., encoding the light and heavy chains of an antibody) into the genomes of Agrobacterium species and infecting plants, they can induce the production of proteins (e.g., trastusumab).
But again, the choice of tobacco also has a safety factor attached to it.
“Our plants are transgenic, so they carry this RNAi cassette that controls the plant-specific glycans,” he continues.
When you start manipulating plants that are part of the food chain—such as the carrot cells used by Protalix—that practice can become an issue.
“We don’t have that issue.”

How’d they do that?
Not content to simply use carbohydrate microarrays to identify potential binding partners as drug targets, Injae Shin and colleagues at Yonsei University and Knokuk University wondered if the same arrays could be used to screen for compounds that might stimulate signal transduction in live cells via cell-surface lectins.
“Although carbohydrate microarrays have been employed for studies of cell-glycan interactions,” the researchers wrote earlier this year in Chemical Science, “microarray technology has seldom been used for [this] purpose.”
The researchers probed an array of unmodified glycans with wildtype cells or those expressing SIGN-R1, a mouse protein involved in immune function. Because stimulated SIGN-R1 triggers the generation of reactive oxygen species (ROS), they could monitor binding via an ROS-sensitive fluorescent probe.
The researchers found that not only did the cells bind specific (and previously characterized) glycans, but they also found binding could be attenuated by pre-incubating the cells with free mannan, a known SIGN-R1 binding partner. Furthermore, fluorescence intensity was proportional to the amount of ROS produced, and ROS could be detected from as few as 50 cells.
“We have demonstrated that carbohydrate microarrays with high-density spotting of a number of glycans by using an automatic microarrayer can be employed for the rapid screening of functional glycans that enhance the cell-surface lectin-associated cellular response,” the authors concluded. Thus, they added, “carbohydrate microarray-based technology will be useful for the simultaneous screening of glycans in efforts aimed at the development of novel chemical probes for studies of functional glycomics.”

Subscribe to Newsletter
Subscribe to our eNewsletters

Stay connected with all of the latest from Drug Discovery News.

March 2024 Issue Front Cover

Latest Issue  

• Volume 20 • Issue 2 • March 2024

March 2024

March 2024 Issue