Our research is based on the premise that extraordinary insights into the molecular basis of cancer can be obtained by taking global views of the genomes of patient-derived tumor samples. Implicit in this approach is the reliance on naturally arising human tumors as the source of discovery, as opposed to relying exclusively on experimental model systems. In addition, we are committed to taking global views of cancer genomes that are not constrained by prior assumptions about the nature of cancer pathogenesis.
DNA Microarrays and Cancer
To broaden the view of cancer genomes, we have explored the use of DNA microarrays (DNA chips) to monitor the gene expression (gene activity) of thousands of genes simultaneously across the human genome. This technique, pioneered by Patrick Brown (HHMI, Stanford University), involves the extraction of RNA from tumor samples and its subsequent fluorescent labeling and hybridization to an array of DNA probes. Microarrays covering nearly the entire human genome are now available. In a series of experiments, we initially demonstrated that the classification of cancer—specifically two principle forms of acute leukemia—could be achieved by using DNA microarrays to monitor gene expression, without a prior molecular understanding of this distinction. This implied that such methodologies could be applied to the molecular dissection of cancers beyond that which is routinely feasible.
We have since applied this approach to the molecular classification of many tumor types, including lymphoma, prostate cancer, brain tumors, and lung cancer. Similar approaches have demonstrated that patterns of gene expression (a gene expression "signature") may be found across different tumor types. For example, we identified a signature of metastatic propensity across prostate, breast, and lung cancers, suggesting that a genetic test performed at the time of diagnosis might predict the future behavior of some tumors.
MicroRNAs in Cancer
More recently, our laboratory, in collaboration with Tyler Jacks (HHMI, Massachusetts Institute of Technology) and Robert Horvitz (HHMI, also MIT), has developed methods for measuring the abundance of a new type of RNA—microRNAs (miRNAs). Unlike mRNAs, which lead to the production of proteins, miRNAs play important roles in regulating the abundance of certain mRNAs and the rate at which they are translated into protein. We developed a novel bead-based method for measuring miRNA expression patterns in tumors, and discovered that miRNAs undergo a tremendous amount of dysregulation in tumors. Work with Tyler Jacks suggests that miRNA dysregulation may play an important causal role in the pathogenesis of cancer.
We have also begun to ask whether genomic approaches might lead to the discovery of small-molecule compounds capable of modulating the biological state of a cell. Such chemical biology efforts can yield "tool compounds" useful for the dissection of biological mechanisms, and they can also serve as the starting point for drug discovery efforts. Drug discovery typically starts with prior knowledge of a target gene that is biologically relevant to a disease state (e.g., a gene mutated in cancer). That protein is then biochemically purified, and a collection of compounds screened in a test tube (not in cells) for their ability to bind to the protein. Although this approach can be highly effective, in many cases the relevant target protein is not known in advance, and, all too often, small molecules in protein-binding assays do not behave as they do in living cells.
We therefore asked whether we could use the principles of gene expression signatures described above to develop a surrogate of a complex biological state (for example, the normal state versus the cancer state). In principle, one could screen for compounds that turn the cancer signature into the normal signature, and then, by inference, shift the cell into a condition that is closer to normal. This approach, which we termed gene expression-based high-throughput screening (GE-HTS), has now been applied to the discovery of compounds that induce the maturation of acute myeloid leukemia cells, inhibit androgen action in prostate cancer cells, and inhibit the activity of the causal protein in the childhood bone tumor known as Ewing sarcoma.
We have extended this use of gene expression signatures to discover connections between disease states and drugs through our Connectivity Map project. This project involves a large, publicly accessible database of gene expression profiles, coupled with a suite of computational tools that allow users to search the database for compounds that match a query signature. Although we are still in the early stages of the project, the Connectivity Map approach appears to be highly effective and has the potential to serve as a new type of genomics tool for the biomedical research community.
This research is also supported by grants from the National Institutes of Health and the Leukemia and Lymphoma Society.
As of May 14, 2007