Demonstrate Potential of High-Throughput Screening in Academia
Boston, MA -- October 25, 1999 -- Nine months after the Harvard Medical School Institute for Chemistry and Cell Biology (ICCB) opened its doors, its researchers are reporting their first success. In the October 29 Science, they describe how they used a series of screens to fish out of a library of chemical compounds the first known small-molecule inhibitor to a motor protein involved in cell division.
The scientists describe a molecule that perturbs the normally bipolar mitotic spindle apparatus in such a way that cells arrest midway through mitosis with a characteristically malformed spindle that looks a bit like an exploding star.
Moreover, the scientists, led by Tim Mitchison and Stuart Schrieber, pinpointed the protein through which this chemical exerts its peculiar effect: it is a member of the kinesin family of motor proteins. That makes monastrol the first in a new class of mitosis inhibitors. All others -- ranging from colchicine, a medicine used since ancient Egyptian times, to modern cancer drugs including taxol -- affect microtubules, the long, thin protein tubes that are the main structural element of the mitotic spindle.
The paper demonstrates that it is feasible and efficient for academic scientists to perform high-throughput screening, most of which occurs in industrial research. "We are very pleased with this paper," says Mitchison, professor of cell biology. "For this to come out of an academic lab is unusual, if not unique, at this point."
The project was designed primarily to yield tools for basic research. Indeed, the ICCB scientists are already using monastrol to study the molecular dynamics of the mitotic spindle, a finely tuned structure that separates the two sets of chromosomes during cell division. At the same time, any mitotic inhibitor represents a potential starting point for developing better chemotherapeutics, Mitchison says. With luck, monastrol could lead to a drug as effective as taxol but without its troubling neurotoxicity.
The project began when Steve Haggarty, a graduate student in Schreiber's lab, used a fast, cheap, and sensitive assay to screen compound libraries that he and others in the Schreiber lab had previously developed. Called cytoblot, the assay uses an antibody to detect cells suspended in mitosis.
Haggarty screened a commercial library containing 16,320 compounds synthesized in Russia, traditionally a powerhouse in chemistry. (Russian chemistry laboratories have begun selling their compounds to supplement shrinking research funds.) In three days, he tested each compound on about 3,000 dividing cells per well of a 384-well plate. With new robots the institute has since added, this step would now take one day. Researchers led by ICCB fellow Randall King are perfecting a further minituarization step, testing each compound on 1536-well plates, where each well holds 500 cells in five microliters -- about a tenth of a rain drop.
One hundred thirty-nine compounds proved to halt cell division. From these, Thomas Mayer, a postdoc in Mitchison's lab, first eliminated 53 that target tubulin. Many tubulin inhibitors already exist, both as research tools and as drugs. Mayer then looked under the microscope to see how the remaining 86 compounds affected dividing cells.
One jumped out at him. Cells treated with it did not assemble the bipolar spindle that uses microtubules to line up the duplicated chromosomes and then pulls them apart suddenly just before the two daughter cells pinch off. Instead, they formed a monoaster, a central point from which microtubules radiate out, with chromosomes dotted improbably across their middle. Hence the name monastrol.
The next step -- identifying which protein monastrol disturbed to create this dramatic pattern -- could easily have been a year-long, difficult task, had it not been for a lucky break. The scientists knew from earlier work in Mitchison's and other labs that monoasters also appeared in cells with damage in a particular motor protein. Called Eg5, it is a member of a far-flung family whose founding member kinesin transports vesicles along microtubules along neuronal axons and in other cells. It so happened that Tarun Kapoor, also a postdoc in Mitchison's lab, had already developed an assay for Eg5's ability to move in the presence of its fuel molecule ATP. Testing whether monastrol inhibits Eg5, then, took one day. It does.
The scientists have not formally proven monastrol acts through Eg5, and they cannot say whether it does not also act on additional proteins, Mitchison cautions. But he feels confident that this is a new kind of antimitotic drug that works through this motor, he adds.
With this process, the scientists have reversed the order of drug discovery currently favored in industry. Rather than identifying a protein important in a disease, and then screening compound libraries against that single target, as does industry, the ICCB scientists screened a library against cell division, a "whole, complex piece of biology," says Mitchison, and then identified the target of the most interesting compound.
It is too early so tell whether monastrol will prove valuable for developing better chemotherapy, says Mitchison. A drug that disrupts spindle components other than microtubules might be more specific than current ones, because all body cells need functioning microtubules while only dividing cells are likely to require complete spindles. Monastrol's relatively weak binding to Eg5 -- about one order of magnitude below the level at which companies typically start a development program -- is counterbalanced by the fact that it already has proven to perform the desired function of stopping mitosis, says Mitchison. Moreover, monastrol's chemistry is so well known that the next step of synthesizing and testing modified versions would be straightforward. Chemists at the ICCB are making such variants.
Meanwhile, Mayer and Kapoor are using monastrol in ongoing studies of exactly how Eg5 helps build and maintain the mitotic spindle. Mitchison suspects, however, that there is more to Eg5. For example, spindle microtubules are not static; they constantly slide towards the poles. Kapoor is wondering whether Eg5 acts like a "smart glue" that somehow accommodates this movement while keeping the spindle intact. "We can learn a lot more about Eg5, and monastrol will help with that," Mitchison says. "I just love the fact that the first screen landed smack-dab on one of my favorite proteins. Mother Nature was smiling."
Editors: please note, a color image is available.
Journal
Science