The applications of toxicogenomics can be characterized into two broad and overlapping classes: mechanistic or investigative research and predictive toxicology.

The biological relevance of the experimental system for transcript profiling is clearly of major importance where a mechanistic understanding of a toxic process or a mode of action is required. In all likelihood, the toxic endpoint is known in advance (at a physiological, histological, and/or biochemical level) and an appropriate test system (in vitro or in vivo) can be designed to model the endpoint as closely as possible. Toxicogenomic applications may help to identify surrogate markers for the development of this phenotype, and indeed the exposure of rodent hepatocytes to the nongenotoxic carcinogen phenobarbital has been studied, using both microarray and gel-based expression technologies. In excess of 300 genes have been identified where expression is modulated by this compound. Many other toxic endpoints could be profiled using these methods, with combinatorial approaches such as transgenic or knockout models, potentially providing insights into the role of specific genes. 

The possibility that a specific group or class of compounds (grouped by toxic endpoint, mechanism, structure, target organ etc.) may induce signature patterns of gene expression changes is the basis for the application of toxic genomics to predictive toxicology. The use of these technologies to analyze genome-wide changes in mRNA expression following treatment of in vitro systems with known reference toxicants may permit the identification of diagnostic gene expression patterns. Pattern recognition may, in turn, allow the design and construction of miniarrays, customized to detect specific toxicity endpoints or pathways. While in vitro systems have practical advantages, there are major drawbacks to consider. Even where appropriate cells in vitro, such as primary hepatocytes, are available, compound-induced changes in transcription may not necessarily reflect accurately the response of the corresponding organ in vivo. Their application in predevelopment toxicology screening would be of substantial benefit in providing an early view of compound safety in advance of traditional studies.

Development of reference data sets to allow a “pattern recognition” approach to toxicology is likely to require the application of complex computer algorithms and statistical approaches. The building of reference data sets, possibly by comparison of microarray output across different laboratories, will require consistency in data analysis and format. A number of resources exist in both the academic and commercial sectors for such purposes.

Regards,

Angelina Matthew,

Managing Editor,

Journal of Genetics and Genomes

Email Id: genetics@scholarlypub.com