In Pursuit of Synthetic EPO

Researchers are beginning to close in on the goal of synthesizing erythropoietin-like glycoproteins having defined carbohydrate composition.

The antianemia medication erythropoietin (EPO) had 2005 sales of nearly $9 billion, according to IMS Health. The 165-residue glycoprotein contains several carbohydrate groups, some of which are oligosaccharides of substantial size and complexity. It's produced commercially by fermentation as a mixture of glycoproteins with variable carbohydrate composition.

Homogeneous EPO, in which the carbohydrate composition is constant, has been difficult to isolate in significant quantities. Obtaining homogeneous versions of the drug in substantial amounts is a long-sought goal, as it could ease studies of structure-activity relationships and lead to more-predictable therapeutic formulations.

Chemistry professor Samuel J. Danishefsky of Sloan-Kettering Institute for Cancer Research and Columbia University and coworkers hope to create homogeneous EPO synthetically. They haven't succeeded yet, but in a recent paper they report significant progress toward that goal (Angew. Chem. Int. Ed., published online May 19, dx.doi.org/10.1002/anie.200600538).

In the study, the group developed a revised route to convergent glycopeptide synthesis and demonstrated it by creating some of the largest and most complex homogeneous EPO-like glycopeptides ever made by chemical synthesis. The undertaking has already led to advances in carbohydrate and peptide-ligation chemistry, and Danishefsky believes it will also eventually lead to homogeneous EPO.

EPO-like proteins have been synthesized twice before. In 1999, Nicola Robertson and Robert Ramage of Albachem Ltd. (now Almac Sciences), Edinburgh, Scotland, and the University of Edinburgh reported having synthesized an EPO-like protein by solid-phase synthesis. But it was deglycosylated, lacking carbohydrate groups.

In 2003, R&D Director Gerd G. Kochendoerfer of Gryphon Therapeutics, in South San Francisco, and coworkers, including researchers at the Blood Research Institute, Milwaukee, made four peptides on a synthesizer, added a carbohydrate-mimetic polymer to two of them, and joined the peptides by a technique called native chemical ligation (NCL). The polypeptide chain was then folded and two disulfide bonds were added to produce a bioactive EPO-like protein with enhanced in vivo lifetime.

Synthesizing EPO with its carbohydrate groups, however, has so far proved elusive. One restraint is that conventional NCL requires that a terminal cysteine be present on one of the peptide pieces to be joined, and EPO has few cysteines. Associate professor of cell biology and chemistry Philip E. Dawson of Scripps Research Institute and coworkers had earlier developed a more versatile NCL reaction that doesn't require terminal cysteines, and Danishefsky and coworkers modified this procedure. Their cysteine-free NCL requires the presence of a terminal auxiliary group on one of the peptides, and this is then removed from the product after the ligation.

In their study, Danishefsky and coworkers used one cysteine-based and one non-cysteine NCL to join three glycopeptides that they had preassembled. Their most EPO-like product was a glycopeptide with 24 amino acids and three carbohydrate groups, only one of which even approaches the size and complexity of EPO's largest oligosaccharides. But in a yet-unpublished study, they created a biantennary (two-branch) oligosaccharide that more closely resembles EPO's. And in another study (J. Am. Chem. Soc., 2006, 128, 7460), they determined the molecular mechanism of their version of NCL.

"The glycopeptide synthesis is an impressive piece of work," says chemistry professor Carolyn R. Bertozzi of the University of California, Berkeley, and "a major step toward the ultimate goal of putting together EPO, the most complex natural product to be targeted for total synthesis, as far as I know." Combining glycopeptides "elevates the challenge by orders of magnitude" over peptide ligation alone, she says, "because of the sheer number of functional groups and the demands on chemoselectivity." The noncysteine ligation they use "apparently stands out in that its conditions, yields, and rates are amenable to use with large, glycosylated fragments."

Chemistry professor Stephen B. H. Kent of the University of Chicago is more skeptical. "Where we're going to end up when Danishefsky and coworkers have succeeded—which they undoubtedly will—is with a chemical synthesis of a native glycosylated EPO that will never be reproduced by other groups. It will be one of these heroic syntheses that will be a milestone in the field and will contribute all sorts of useful methodology, but it's unlikely that it will be a practical way for others to make analogs. It's closer to traditional synthetic organic chemistry than to the types of new approaches that will be required to make total chemical protein synthesis routine, useful, and practical for ordinary mortals."

Danishefsky replies that "at present, the effort is really an exciting academic enterprise. However, I am hopeful it will find a practical consequence by allowing medicinal chemistry to be done in the EPO field on pure compounds. The practicality of the technique could change radically when all the building blocks are available in a modular form and can be shuffled around."

Dawson notes that "a major challenge for the future of this synthesis will be to maintain high synthetic yields and optical purity as the peptide fragments get larger and less soluble. Continued refinement of ligation methods will benefit the entire field of synthetic protein chemistry."

Although major challenges remain in the assembly of large glycoproteins, says associate professor of chemistry and chemical engineering Linda Hsieh-Wilson of California Institute of Technology, the studies by Danishefsky's group "significantly extend the reach of NCL and take us one step closer to tailoring glycoproteins with defined oligosaccharides at specific sites. This work has the potential to significantly advance our understanding of the roles of protein glycosylation and has exciting implications for the development of protein and antibody therapeutics."