We are applying a combination of experimental and theoretical strategies to the discovery of new agents for treating cancer and AIDS. Since 1990, NCI has tested more than 70,000 chemically defined compounds for their ability to inhibit a panel of 60 different cancer cell lines. Growth inhibition for any single line is simply an index of cytotoxicity or cytostasis, but the patterns of 60 inhibition values encode unexpectedly rich information on mechanisms of drug action and resistance. Each compound's fingerprint pattern of activity is essentially unique among the many billions of distinguishable possibilities. We are mapping these activity patterns into molecular structure descriptors to mine NCI's Drug Information System database of more than 460,000 compounds; we are also correlating them with gene and protein expression patterns that we and our collaborators are assessing in the 60 cell lines -- in part using cDNA microarray, 2D-gel technologies, and chromosomal analyses. Overall, the databases on molecular structure, activity, and targets (>275 million numbers) constitute a major information resource that we are using (1) to generate and test hypotheses related to the cell biology of drug targets; (2) identify new lead molecules and potential targets; and (3) help set priorities for further drug development and individualization of therapy. To cite one instance, we used the information to search for compounds not dependent for activity on intact p53 suppressor gene function. Other highlights of this research include: * Development of mRNA expression profile databases for cells of the NCI cancer drug discovery program. Work has been completed for a set of >8,000 genes (using 2-color EST pin-spotted arrays, in collaboration with the group of Patrick Brown at Stanford) and a set of >6,500 genes (by using oligonucleotide arrays, in collaboration with the group of Eric Lander at the Whitehead Institute). * Development of a protein expression database for cells of the NCI cancer and AIDS drug discovery programs. Thus far, 1,014 spots have been indexed from 2-D polyacrylamide gels from all 60 cell types. Of these, 21 spots have been identified with particular proteins. Included are cytoarchitectural proteins, chaperonins, and molecules directly related to cell-cycle control. This project is tying together our research in molecular pharmacology and """"""""proteome."""""""" * Development of methods for identification and characterization of 2-D gel spots based on matrix-assisted laser desorption ionization (MALDI) mass spectrometry and ion trap electrospray mass spectrometry. * Development and integration ( in progress) of a chromosomal description of the 60 cell lines and others. Included are spectral karyotyping (with L. Kirsch), array-based comparative genomic hybridization (with J. Gray, UCSF), and single nucleotide polymorphism identifications (1,600) (with K. Buetow). * Development of the DISCOVERY program set and Clustered Image Maps, which can integrate information on a tested compound's molecular structure, pattern of activity, and possible molecular target(s). * Application of DISCOVERY to identify candidate anticancer compounds not dependent on intact p53 function (""""""""p53-inverse"""""""" compounds). . * Identification from gene expression profiles of a new possible range of utility for the enzyme-drug L-asparaginase in solid tumors (with D. von Hoff). . * Identification using microarrays of 3 classes of genes that change expression in response to treatment of colon cancer cell lines with the topoisomerase 1 inhibitor camptothecin.
Showing the most recent 10 out of 37 publications
