RICHARD CUMMINGSLipids and proteins on the surfaces of cells are bedecked with sugar chains, which determine how cells develop, adhere to one another, and communicate. Bacteria and viruses glom onto these complex linear or branched oligosaccharides, called glycans, to infect cells. The immune system fights back, learning to recognize microbes’ sugar coatings and mounting both an innate and an adaptive defense.
Despite glycans’ fundamental importance to biology, they remain poorly understood compared to DNA and proteins. Because of their structural complexity, glycans are arduous to manufacture. And unlike DNA, they are impossible to clone and amplify, so quantities are limited.
But one technology initially developed for understanding genetic material has proven a perfect fit for glycobiology: the microarray. The first glycan microarrays came onto the scene in 2002, just seven years after the advent of microarrays to study gene expression.
Glycan microarrays consist of small quantities of a variety of natural or synthetic oligosaccharides affixed to a surface. Researchers use the arrays to identify proteins, cells, and microbes that bind to the sugars. Because printed microarrays require minuscule quantities of sugars, they made it possible, for the first time, to broadly screen proteins’ glycan affinities.
These days, more than a dozen labs make their own glycan microarrays, and researchers say they will only improve as more glycans are isolated and synthesized. The largest and most widely available arrays are produced by the Consortium for Functional Glycomics (CFG), an international collaboration among glycomics researchers, and by Imperial College London’s Glycosciences Laboratory. Other labs make smaller arrays targeted for use on more-specific projects.
“At this point I don’t think there is a single presentation or single platform that is a clear winner,” says Jeffrey Gildersleeve, head of the Chemical Glycobiology Section at the National Cancer Institute’s Center for Cancer Research in Frederick, Maryland.
And while DNA microarrays’ glory days may be fading as RNA sequencing prices fall, the future of glycan microarrays appears sunny. “I think that [glycan microarrays are] really just about to, in the next five years, come into their own in terms of commercial use, industrial applications, and clinical use,” says Richard Cummings, professor of biochemistry, director of the National Center for Functional Glycomics (NCFG) at the Emory University School of Medicine, and current chair of the CFG.
Here, The Scientist profiles a selection of today’s leading glycan microarrays and some of the discoveries they have enabled.
OUR GLYCANS, OURSELVES
The CFG was first funded in 2001 by the National Institute of General Medical Sciences (NIGMS) to study protein-carbohydrate interactions. An early CFG project resulted in an ELISA-based glycan array to be shared with the entire glycomics community. By 2004, the CFG had switched to a printed microarray including 200 glycans (PNAS, 101:17033-38). Currently, the array consists of 610 mammalian glycans covalently attached to slides via amide linkages.
The CFG has proven to be an invaluable resource to researchers who don’t have voluminous glycan libraries of their own, says CFG chair Cummings. “Most of the glycans of interest are not commercially available,” he says. And even if they were, one would have to purchase them in large quantities.
The CFG’s mammalian glycan array has contributed to discoveries ranging from the receptor affinity of flu viruses, to the lectins bacteria use to invade cells, to the cancer-specific antibodies that bind to glycans present on the surface of tumor cells.
Most recently, Stephan von Gunten of the University of Bern’s Institute of Pharmacology in Switzerland and colleagues used the array to determine glycan binding of immunoglobulins isolated from the blood of healthy donors (Sci Transl Med, 7:269ra1, 2015). They discovered that some immunoglobulins bind to mammalian glycans that bacteria and viruses coopt to invade cells. The researchers hypothesize that human antibodies block these adhesion sites in order to outcompete binding by the pathogens.
Cummings and his colleagues also developed a method called shotgun glycomics, in which they harvest and purify glycans from an array of sources, ranging from breast milk to pig lungs (Nat Meth, 8:85-90, 2011). The researchers then observe binding patterns of known glycan-binding proteins to the arrays and sequence the glycans that they bind. Alternatively, researchers can use the shotgun arrays to identify the glycans to which particular microbes attach.
Although the original NIGMS funding ended in 2011, screening using the CFG array remains available via Emory’s NCFG. Researchers hoping to test the glycan-binding properties of their own samples can apply on the CFG website to send them to Emory for screening, which costs $300 per slide. Any results produced by the CFG array can be accessed at www.functionalglycomics.org. Researchers hoping to use shotgun glycan arrays or other specialized arrays offered by the NCFG can pay to access them. The NCFG will also print custom arrays for a fee.
PALMA ET AL., CURR OPIN CHEM BIOL, 18:87-94, 2014 WITH PERMISSIONResearchers at the Glycosciences Laboratory at Imperial College London produced one of the earliest glycan microarrays (Nat Biotechnol, 20:1011-17, 2002). The group’s microarray system has since evolved to include 830 glycan probes. Each synthetic glycan is affixed to a lipid tag, a combination called a neoglycolipid (NGL), and then printed onto to a glass slide coated with nitrocellulose. The microarray also includes naturally occurring glycolipids.
Imperial College’s NGL-based microarrays stand out from other microarrays because the lipids and the noncovalent bonds linking the sugars to the slide are slightly flexible, allowing the glycans to wiggle around. “We believe that element of mobility is really important,” says Ten Feizi, director of the Glycosciences Laboratory and one of the original creators of the microarray.
The strength of bonds between individual glycans and proteins is weak. Because of this, groups of glycans often interact with proteins or groups of proteins with multiple glycan binding sites on their surfaces. These bonds between proteins and groups of glycans are called multivalent interactions.
All glycan microarrays allow for multivalent bonding, as multiple sugar chains are printed side by side on the array. However, the density of the glycans and their orientation influence whether they are able to slot into their allotted protein binding sites. Feizi says that the NGL-based microarray is more forgiving, as the nonrigid linkers allow the glycans to adjust to the multivalent binding sites on proteins, much as glycans on a fluid cell membrane might adjust during binding.
NGL-based microarrays have in recent years helped researchers understand viruses, innate immunity, and cancer. Feizi recalls that her group’s microarrays really shined in studies of sugar-coated receptors used by the 2009 H1N1 swine flu to infect cells. The researchers employed a small NGL-based array to show that the virus bound not only to a sugar found on the surface of cells in the upper respiratory tract, but also to a different glycan found deep in the lungs (Nat Biotechnol, 27:797-99, 2009). Other glycan microarrays failed to detect the relatively weak binding to the alternate lung receptor. Feizi and her colleagues subsequently showed that a mutated version of the virus had a heightened affinity for the lung receptor, and that this mutant tended to cause more-lethal infections (J Virol, 84:12069-74, 2010).
Last year, the researchers used their microarrays to determine the identity of a glycan structure, F77, bound by an antibody known to stick to prostate cancer cells (J Biol Chem, 289:16462-77, 2014). They first assessed the antibody’s binding against NGLs and glycolipids on their broad screening array. But the array only generally revealed the carbohydrate structures the antibody targeted. After discovering that the antibody also binds heavily glycosylated proteins called mucins, the researchers released glycans from naturally occurring mucins to create “designer arrays,” allowing them to more precisely establish the sugar chain structure the antibody binds.
Researchers hoping to use NGL-based arrays from the Glycosciences Laboratory can apply online at www1.imperial.ac.uk/glycosciences/request/. Glycosciences Laboratory personnel will analyze samples using large screening arrays for £200 per sample. The Wellcome Trust currently funds the facility.
JEFFREY GILDERSLEEVEOver the past decade, the National Cancer Institute’s Jeffrey Gildersleeve has developed a glycan microarray consisting of 503 natural glycoproteins, synthetic “neoglycoproteins,” peptides, and controls. On the surfaces of the neoglycoproteins the researchers vary the arrangements and densities of glycans. This variation allows users to understand how altering the presentation of glycans affects binding of proteins.
Gildersleeve developed his array to screen vast quantities of serum from people with cancer. By presenting glycans in multiple configurations and densities on the surfaces of proteins, he encourages multivalent bonding. Being able to control presentation allows him to understand which configurations of glycans bind which proteins, and allows him to replicate arrangements that work.
“We’re trying to identify unique subpopulations of antibodies in serum that are relevant to cancer care,” he explains. These antibody subfamilies may only be distinguishable by which presentations of glycans they bind.
The Gildersleeve group is probably best known for its success in characterizing antibodies in prostate cancer patients to help predict and then assess how they respond to the Phase 3 cancer vaccine PROSTVAC-VF, in development by Bavarian Nordic.
“The big challenge for most cancer therapies, and especially cancer vaccines and other immunotherapies, is that they can have really, really good results in some patients . . . but lots of patients really see no benefit at all,” Gildersleeve says. Discovering biomarkers to predict how patients will do on therapies aids doctors in choosing the right ones for their patients.
Using their microarrays, the group analyzed the serum of patients prior to their participation in Phase 2 clinical trials of the vaccine and then several months after beginning the therapy. The researchers found that patients who showed elevated immunoglobulin binding to a blood group A antigen before the trial were likely to survive longer than average (Clin Cancer Res, 19:1290-99, 2013). The researchers also found that patients who received the vaccine and then did well tended to show increases in antibodies to the Forssman antigen, a glycan present on viral vectors used to deliver the vaccines (PNAS, 111:E1749-58, 2014).
The Gildersleeve lab is willing to collaborate with other research groups on projects using their microarray. The group is particularly interested in work related to cancer or HIV.
THE SCIENTIST STAFF, BASED ON NAT CHEM BIOL, 10:470-76, 2014, WITH PERMISSION FROM RICHARD CUMMINGSWhile most glycan microarrays focus on the mammalian glycome, some researchers are starting to pay attention to the sugars decorating microbes. Last year, researchers led by Cummings and his frequent collaborator Jim Paulson, professor of cell and molecular biology at the Scripps Research Institute in La Jolla, introduced an array that now includes approximately 350 microbial glycans derived from diverse species.
The CFG began to construct a microbial glycan microarray (MGM) around a decade ago, but halted the project after their funder, the NIGMS, decided the array did not fit into the scope of the consortium’s mission, according to Paulson, who was then CFG director.
Paulson and his colleagues recently began work on the array again and introduced a new and improved version in June 2014 (Nat Chem Biol 10:470-76). Paulson and his colleagues demonstrated that the microarray shows different binding patterns when presented with human versus other mammalian sera. They also showed that rabbit sera’s reactivity to glycans changes after the animals have been exposed to specific microbial strains. These results indicate that the adaptive immune system quickly reacts to the presence of specific microbes, and also that the new microarray, while small, is sufficiently sensitive to pick up these changes.
The researchers further investigated the binding of innate immune glycan-binding proteins called galectins to the array, finding that several of them bind a probe isolated from a strain of the gut bacterium Providencia alcalifaciens. Further investigation showed that binding by some galectins kills this strain and that galectins also bind and kill an assortment of other gram-negative and gram-positive bacterial strains carrying similar glycan antigens.
This type of antigen mimics mammalian glycans, possibly as a way for bacteria to evade the immune system. Indeed, the researchers showed that the same galectins bind analagous antigens on the CFG mammalian array. However, the galectins do not harm mammalian cells when they bind to this antigen. Thus, the MGM was able to uncover an entirely new innate immune mechanism for fighting off diverse bacteria that try to mimic mammals’ own cells.
Many in the glycomics community see the MGM as long overdue, as microbial and mammalian glycans are generally quite distinct. “I think this is a part of the immune system we haven’t mined,” says Laura Kiessling, director of the Keck Center for Chemical Genomics at the University of Wisconsin–Madison. “It’s been opaque to us because we haven’t had any way of looking at it.” Kiessling is currently using the MGM to further explore innate immunity.
Those wishing to use the array can obtain it through the CFG at cost.
Correction: The number of samples that the Glycosciences Laboratory will analyze for £200 per sample using screening arrays was misstated. The laboratory will analyze as many samples as are necessary with existing probes. The article has been updated to reflect this. The Scientist regrets the error.