Gene expression profiling, genetic networks, and cellular states: an integrating concept for tumorigenesis and drug discovery.

Genome-wide expression monitoring, a novel tool of functional genomics, is currently used mainly to identify groups of coregulated genes and to discover genes expressed differentially in distinct situations that could serve as drug targets. This descriptive approach. however, fails to extract "distributed" information embedded in the genomic regulatory network and manifested in distinct gene activation profiles. A model based on the formalism of boolean genetic networks in which cellular states are represented by attractors in a discrete dynamic system can serve as a conceptual framework for an integrative interpretation of gene expression profiles. Such a global (genome-wide) view of "gene function" in the regulation of the dynamic relationship between proliferation, differentiation, and apoptosis can provide new insights into cellular homeostasis and the origins of neoplasia. Implications for a rational approach to the identification of new drug targets for cancer treatment are discussed.