"Although DNA is the blueprint, proteins are the workhorses of cells; they actually carry out the functions," says U-M chemistry Prof. David M. Lubman, who licensed the technology to Eprogen Inc., of Darien, Ill. The process, called ProteoSepTM, uses conventional high performance liquid chromatography (HPLC) techniques and could be "a real winner for the pharmaceutical industry," where HPLC already is widely used, says Lubman. The method has the potential to help researchers understand how various drugs affect protein expression in the cell and how cellular proteins change as a cell turns cancerous. It may also eventually help physicians predict the course of a particular patient's disease and decide how to treat it.
Currently, scientists typically use two-dimensional gel electrophoresis, a tricky and time-consuming procedure, to extract individual proteins from the complex assortment of proteins in cells. In the first step, prepared samples are dabbed onto a gel and exposed to an electric field, which separates the various proteins on the basis of their charges. Next, the proteins are sorted in the second dimension by molecular weight, still in a gel matrix. The result is a 2-D "map" of separate protein spots. Such maps can be used, for example, to pinpoint protein changes that are peculiar to specific types of cancer or to identify unique markers for a particular cell type or disease state. While gel electrophoresis is widely used, it has drawbacks: at least two full days are required to complete the separation in two dimensions, and reproducibility of results is often poor. What's more, the technique has its limits as an identification tool because the protein spots must be cut out of the gel and processed before any proteins can be further analyzed and definitively identified.
In the ProteoSep system, proteins remain in a liquid phase throughout the separation process. Like 2-D gel electrophoresis, the new method first separates the proteins by charge, using a technique called chromatofocusing. Protein samples are introduced to a column containing a solution at a particular acidity (pH). A buffer solution is added, little by little, changing the pH of the column in increments. With each change, groups of proteins with the same isoelectric point (the pH at which a protein has no net charge) are extracted from the column and can be collected as separate fractions.
In the second phase of the process, a technique called nonporous reverse-phase HPLC is used to quickly and efficiently separate out the proteins in each fraction collected in the first phase. "The process is easy to run using a standard HPLC and some commercial buffers," says Lubman. "Anyone who has an HPLC---as pharmaceutical labs do---can set this all up themselves, totally automate the separation process, and have results in a very short time. In our work, we have separated out more than a thousand proteins in hours versus days."
Because the proteins remain intact and are in liquid form, they can then go directly into a mass spectrometer for further analysis. The final result is a 2-D map that can be used like those obtained with gel electrophoresis.
"For example, you can look at differences between an untreated colon cancer cell line and one that has been treated with a drug that's known to be effective," says Lubman. "By comparing the two maps, you can determine which proteins are interacting with the drug or how the drug is affecting protein expression." To simplify such comparisons, Lubman and Eprogen have commercialized a software package called ProteoVueTM with one program under development (DeltaVueTM) that overlaps the two maps, subtracts out parts that are identical, and reveals the remaining differences. The ProteoSepTM process and ProteoVueTM software also make it possible to see how the protein profiles of drug-treated cells change over time or with various drug dosages.
"We've also been looking at cancer progression," Lubman adds. "In breast cancer, for example, we're trying to see how cellular proteins change as a cell goes from being a normal cell to a pre-cancerous cell to one that's starting to become cancerous, to a truly cancerous cell, to a totally metastatic cell."
The method eventually might help physicians predict the course of a patient's disease or decide how to treat it. "In many cases, the protein profile can be related to the prognosis of the disease," says Lubman. "It's also known that, among breast and ovarian cancers, some cancers look pathologically very similar but are very different on a molecular level, so they may need different treatments. We're trying to use this technique to better understand and deal with those differences."
The University of Michigan
News Service
412 Maynard
Ann Arbor, MI 48109-1399