Building a New Paradigm in Drug Discovery

CAMBRIDGE, Mass. — It’s no secret that drug development is a painfully slow and expensive process—a typical new drug takes 15 years and $500 million to come to market, costing the pharmaceutical industry some $20 billion annually. That’s because finding a drug candidate and developing it into a treatment for disease is largely a matter of luck. "Right now, we have no way of determining the interactions of a drug candidate with all the proteins in the human body. To make drug discovery a more targeted science, we need to identify the complete set of all human proteins and develop tools to study them in parallel," says Whitehead Fellow David Sabatini, who was chosen as one of the world’s 100 Top Young Innovators by Technology Review magazine.

Sabatini is addressing this immediate challenge to help streamline the drug discovery process. New technology developed in the Sabatini lab is allowing scientists to glean the function of thousands of proteins simultaneously, in their natural environment inside living cells. This offers an advantage in drug discovery because ultimately, scientists must determine what goes wrong at the protein level in diseases. Proteins are the brick and mortar of cells and the machinery that run the functions of life. As a result, protein malfunctions cause a vast majority of diseases. Thus, virtually every pharmaceutical on the market today is either a protein itself or a type of molecule that alters protein function.

Proteomics

While many private and academic laboratories have realized that the large-scale study of all the proteins in the body—the field of proteomics—holds the key to understanding and treating disease, the technology to study thousands of proteins in parallel has been lagging. "Most laboratories still study proteins one at a time or maybe a couple at a time. If we’re going to really make progress in biology, we need to be able to start studying proteins 1,000 or more at a time," says Joshua LaBaer, director of the Institute of Proteomics at Harvard Medical School.

Approaches using true protein arrays, which are composed of isolated proteins, (without the cells that normally house them) are limited by whether the proteins can be made, purified, and then folded correctly. Other proteins can’t be put on a microarray at all because they are part of a cell membrane and need to be attached to their membranes to function. "Our system takes advantage of the cell to do all the hard work—to make the proteins, to modify them correctly, and to localize them correctly. We can make cell microarrays that contain a very large percentage of the human genome without being limited by the protein’s properties," says Sabatini.

Looking at Proteins on their Home Turf

Sabatini’s innovative approach uses glass slides, about half the size of a typical business card, printed with 10,000 microscopic spots, each representing a different piece of DNA (or gene) coding a protein. Then, live cells are cultured on the slide. Cells growing on the printed spots absorb the DNA in that spot and start over producing the corresponding protein. The result is a cell microarray of 10,000 distinct spots, each acting like a little protein factory. Sabatini envisions eventually having the entire set of human proteins available on handful of slides.

In the basic research arena, these cell microarrays will enable scientists to look at how a protein functions on its own turf inside either healthy or diseased cells. For example, one array could test for the location of 10,000 different proteins inside the cell. Whether a protein is found in the nucleus, cell membrane, or mitochondria can provide clues to the protein’s cellular role. Other cell microarrays could be used to identify the proteins in a cell that are involved with cell migration or cell death.

Alternatively, Sabatini’s cell microarrays could be used in a drug discovery approach, where researchers are looking at the intrinsic properties of proteins. Pharmaceutical companies typically have hundreds of compounds whose protein targets are not known. Cell microarrays can be used as a screening tool to test these compounds and identify their protein targets. In one fell swoop, the researchers could predict the potential diseases each compound might treat and the side effects each compound might produce. A compound that binds to multiple proteins, for instance, may produce side effects when tested in patients and thus may not be good drug candidate. Sabatini is collaborating with Whitehead Fellow Brent Stockwell to do this type of targeted screening in mammalian cells for such diseases as cancer and Lou Gehrig’s disease.