SHORT REPORT Fluorescent cDNA microarray hybridization reveals complexity and heterogeneity of cellular genotoxic stress responses Sally A Amundson* ,1 , Mike Bittner 2 , Yidong Chen 2 , Jerey Trent 2 , Paul Meltzer 2 and Albert J Fornace Jr 1 1 National Institutes of Health, National Cancer Institute, Bethesda, Maryland, MD 20892, USA; 2 National Human Genome Research Institute, Bethesda, Maryland MD 20892, USA The fate of cells exposed to ionizing radiation (IR) may depend greatly on changes in gene expression, so that an improved view of gene induction pro®les is important for understanding mechanisms of checkpoint control, repair and cell death following such exposures. We have used a quantitative ¯uorescent cDNA microarray hybridization approach to identify genes regulated in response to g- irradiation in the p53 wild-type ML-1 human myeloid cell line. Hybridization of the array to ¯uorescently- labeled RNA from treated and untreated cells was followed by computer analysis to derive relative changes in expression levels of the genes present in the array, which agreed well with actual quantitative changes in expression. Forty-eight sequences, 30 not previously identi®ed as IR-responsive, were signi®cantly regulated by IR. Induction by IR and other stresses of a subset of these genes, including the previously characterized CIP1/ WAF1, MDM2 and BAX genes, as well as nine genes not previously reported to be IR-responsive, was examined in a panel of 12 human cell lines. Responses varied widely in cell lines with dierent tissues of origin and dierent genetic backgrounds, highlighting the importance of cellular context to genotoxic stress responses. Two of the newly identi®ed IR-responsive genes, FRA-1 and ATF3, showed a p53-associated component to their IR-induction, and this was con®rmed both in isogenic human cell lines and in mouse thymus. The majority of the IR-responsive genes, however, showed no indication of p53-dependent regulation, representing a potentially important class of stress- responsive genes in leukemic cells. Keywords: cDNA microarray; ionizing radiation; leukemia; p53; genotoxic stress Genotoxic stresses, such as ionizing radiation (IR), can elicit a wide variety of cellular responses, from cell- cycle arrest to mutation, transformation, or cell death. On the molecular level, early responses to IR include activation of proteins such as p53 and NFkB, changes in protein localization, engagement of signal cascades and transcriptional induction of speci®c genes such as CIP1/WAF1 and MDM2. Many factors, both cell-type and stress-speci®c, can interact to determine the ®nal outcome for the cell. IR has proved a useful probe for the study of basic processes such as cell-cycle regulation, apoptotic pathways, and DNA metabolism and the identi®cation of new IR-response genes may also provide novel targets for future experimental approaches in radiotherapy. The human myeloid cell line ML-1 was selected for this study based on our previous experience with stress gene induction in many dierent cell lines. Various factors contribute to the IR-responsiveness of cell lines. For instance, the p53 wild-type status of ML-1 allowed the detection of many genes, such as CIP1/WAF1, GADD45 and MDM2, which require p53 function for optimal IR-induction. Importantly, this cell line contains endogenous wild-type p53, so the results obtained represent cellular responses with physiologi- cal levels of p53, rather than the unnatural and often highly overexpressed levels resulting from arti®cially engineered systems. In addition, the fact that ML-1 is a myeloid cell line, prone to undergo rapid apoptosis following genotoxic stress, also results in the induction of genes speci®cally associated with this process, such as BAX, MCL1, GADD34 and BCL-X L (Zhan et al., 1994b, 1997). The genes we knew to be IR-regulated in ML-1 prior to this study gave us a wide range of target responses we could expect to detect on the microarray, from 450-fold induction for CIP1/WAF1 to approxi- mately tenfold reduction for c-MYC. Since the initial development of cDNA microarray hybridization (Schena et al., 1995), there has been considerable interest in this rapidly emerging technol- ogy. The ability to compare relative levels of thousands of mRNA transcripts simultaneously in a single hybridization has the potential to contribute greatly to our understanding of many dierent ®elds. Development of a quantitative approach for the measurement of relative changes in gene expression would oer the added advantage of expanding this approach beyond simple pair-wise comparisons. Demonstrated applications have included monitoring of dierential gene expression in wild-type and transgenic plants (Schena et al., 1995) and of a human melanoma cell line and its non-tumorigenic variant (DeRisi et al., 1996), expression of heat-shock genes in a human cell line (Schena et al., 1996), metabolic shift in yeast (Lashkari et al., 1997), and the identi®cation of genes related to in¯ammatory disease (Heller et al., 1997). Transcriptional stress-response is an extremely complex area, with the response of many genes dependent on intermeshing signal transduction activ- *Correspondence: SA Amundson, NIH, National Cancer Institute, 37 Convent Dr., Bldg. 37/rm. 5C09, Bethesda, Maryland, MD 20892, USA Received 17 July 1998; revised 14 January 1999; accepted 14 January 1999 Oncogene (1999) 18, 3666 ± 3672 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00 http://www.stockton-press.co.uk/onc