A new method that allows researchers to link genomic and proteomic information from cancer cells to information about drug structure will be a powerful tool in the design and discovery of new drugs, researchers report in the August issue of the Nature publication The Pharmacogenomics Journal*.

The new bioinformatic approach, developed as a joint effort between scientists from the National Cancer Institute (NCI) and LeadScope Inc., an informatics company in Columbus, Ohio, relates gene expression patterns to more than 27,000 substructures and chemical features within compounds that have been tested for their effect on tumor cell growth. This new approach improves upon previous methods that correlate molecular data with the anti-cancer activity of compounds but not with the substructures within those compounds. The system was developed using gene expression data but can easily be applied to protein profiles, as well.

This method is important because researchers today are using recent technological innovations to produce vast amounts of information about the genes and proteins active within cancer cells. This knowledge of the molecular pathways associated with cancer is expected to dramatically accelerate the pace of drug development by providing insight into potential targets. Computer software that links molecular information with information about potential drugs can help researchers use genomic and proteomic data much more efficiently.

Incorporating data on drug structure has tremendously enhanced the power of the bioinformatics system, according to the study's lead authors, John Weinstein, Ph.D., of NCI and Paul Blower, Ph.D., of LeadScope Inc. "Someone who is trying to design or perfect cancer drugs would ideally like to relate a gene or protein profile directly to drug structure," Weinstein said. The new system does just that, allowing researchers to understand which chemical features of drugs govern their behavior in various cell types. Researchers expect this new technology to help chemists select the most promising candidate drugs for further screening from large collections of compounds.

The method links three databases of information on cells and chemical compounds. The original "information-intensive" method of analysis developed by Weinstein and colleagues used data from more than 80,000 chemical compounds and 60 human cancer cell lines. Since 1990, NCI's Developmental Therapeutics Program has tested these compounds for their ability to inhibit growth in the cell lines, known collectively as the NCI-60. The NCI-60 are cells derived from nine different types of cancer (melanoma, leukemia, and cancers of the lung, colon, breast, prostate, kidney, ovary, and central nervous system).

The first of the new databases includes data on the inhibitory effect of the compounds against each of the cell lines. Gene expression patterns of each of the cell lines, determined by cDNA microarrays, make up the second database. The LeadScopeTM/LeadMinerTM software relates this information to the final database, which identifies which of 27,000 structural features are present in each of the compounds. Scientists who use this tool in the future will be able to create and explore new datasets based on the compounds, cells, and expression profiles most relevant to their own research.

Databases linking biological activity and gene expression to the 60 NCI cell lines are available at http://discover.nci.nih.gov and http://www.leadscope.com. A full-capability version of the software for the NCI-60 databases will be available from LeadScope Inc. for interested researchers.

Back to Top