Automating Analysis of Glycosylated Proteins

12/29/2004

Glycosylated proteins are those to which one or even a dozen or more sugars, or glycans, are attached. The glycans often alter the biological properties of the proteins. They are key to many biologics — for example, genetically engineered erythropoietin (EPO), used in the treatment of cancer, anemia and AIDS.

Because glycans attach in varied and complex ways, getting a handle on glycoproteins involved in pharmaceutical processes is no easy feat. PerkinElmer, Inc. and Procognia have teamed up for a more advanced, high-throughput analysis of protein glycosylation that will make it easier for firms to monitor and control glycoproteins. Procognia’s U-c fingerprint lectin array-based platform has been adapted for PerkinElmer’s Protein Array Workstation as well as its newer scanner, ProScanArray HT.

“For protein therapeutics, it’s basically a nightmare for analytical means to monitor and control growing these proteins,” says Johanna Griffin, Procognia’s Chief Communications Officer. “The glycosylation is influenced by everything—the cell type, the culture medium, the temperature, the cell density. All of those influence the combination of glycoforms produced by a cell. Consequently, having a tool like this to analyze the glycans without even having to purify the proteins is incredibly powerful.”

Procognia’s U-c fingerprint technology is based on an array of some 30 lectins on a slide. The glycoprotein in its culture medium is applied directly to the slide, Griffin says, and the glycans bind to the lectins which recognize the features on the glycans. A labeled probe is then used to determine where the proteins have bound to the lectins, and a histogram of binding intensities is generated. The histogram is the fingerprint, Griffin adds, which is deconvoluted by algorithms to produce quantitative information for all of the features on the glycans and the proportion of each of those features present. All of this is automated, using PerkinElmer’s workstation and high-throughput scanner.

The process requires no pretreatment or purification of samples, and can analyze up to 20 samples in three hours. Griffin further explains the significance for manufacturing: “During process development, you would be developing a graph or matrix where you know exactly what to expect the glycosylation to look like at each stage in the fermentation scaleup,” she says. “If you then apply this to a fermentation during manufacturing, you know immediately when a batch may not be going right. You can either throw it away, or make an adjustment to correct it.”

Making a business case for this PAT technology becomes easy. “If you assume that it costs you $2 million per fermentation batch on average, and you just save one batch, you’ve saved $2 million,” says Griffin.