| Prediction of Toxicant-Specific Gene Expression Signatures after Chemotherapeutic Treatment of Breast Cell Lines Melissa A. Troester,1* Katherine A. Hoadley,2* Joel S. Parker,3 and Charles M. Perou1,4 1Department of Pathology and Laboratory Medicine and 2Curriculum in Genetics and Molecular Biology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA; 3Constella Health Sciences, Durham, North Carolina, USA; 4Department of Genetics and Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA Abstract
Global gene expression profiling has demonstrated that the predominant cellular response to a range of toxicants is a general stress response. This stereotyped environmental stress response commonly includes repression of protein synthesis and cell-cycle-regulated genes and induction of DNA damage and oxidative stress-responsive genes. Our laboratory recently characterized the general stress response of breast cell lines derived from basal-like and luminal epithelium after treatment with doxorubicin (DOX) or 5-fluorouracil (5FU) and showed that each cell type has a distinct response. However, we expected that some of the expression changes induced by DOX and 5FU would be unique to each compound and might reflect the underlying mechanisms of action of these agents. Therefore, we employed supervised analyses (significance analysis of microarrays) to identify genes that showed differential expression between DOX-treated and 5FU-treated cell lines. We then used cross-validation analyses and identified genes that afforded high predictive accuracy in classifying samples into the two treatment classes. To test whether these gene lists had good predictive accuracy in an independent data set, we treated our panel of cell lines with etoposide, a compound mechanistically similar to DOX. We demonstrated that using expression patterns of 100 genes we were able to obtain 100% predictive accuracy in classifying the etoposide samples as being more similar in expression to DOX-treated than to 5FU-treated samples. These analyses also showed that toxicant-specific gene expression patterns, similar to general stress responses, vary according to cell type. Key words: breast cancer, class prediction, doxorubicin, etoposide, 5-fluorouracil, gene expression, microarrays. Environ Health Perspect 112: 1607-1613 (2004) . doi:10.1289/txg.7204 available via http://dx.doi.org/ [Online 14 September 2004]
Address correspondence to C.M. Perou, Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Campus Box 7295, Chapel Hill, NC 27599-7295 USA. Telephone: (919) 843-5740. Fax: (919) 843-5718. E-mail: cperou@med.unc.edu *These authors contributed equally to the work. Supplemental data is available online (http://ehp.niehs.nih.gov/txg/members/2004/7204/ supplemental.pdf) . All microarray raw data tables are available at the UNC Microarray Database (https: //genome.unc.edu/) and have been deposited in the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE1647 (submitted by C. Perou) . This work was supported by National Institute of Environmental Health Sciences (NIEHS) grant (5-U19-ES11391-03) . M.A.T. was supported by NIEHS Individual National Research Service Award (NRSA) 5F32ES012374 and Institutional NRSA in Environmental Pathology 2T32ES07017. K.A.H. was supported by National Institutes of Health Institutional NRSA in Genetics T32GM07092. The authors declare they have no competing financial interest. Received 26 April 2004 ; accepted 14 September 2004.
The full version of this article is available for free in HTML or PDF formats. |
