報告書番号

U03029

タイトル（和文）

遺伝子発現を指標にした生物影響評価（その１） --正常ヒト乳腺上皮細胞(MCF10A)での各種発がん物質に対する細胞毒性と遺伝子発現--

タイトル（英文）

Health Risk Assessment based on Global Gene Expression Analysis (Part 1)- Analysis of Global Gene Expression and Cytotoxicity in Normal Human Breast Epitherial Cell Line (MCF10A) under Exposure to Known Carcinogens

概要 （図表や脚注は「報告書全文」に掲載しております）

本研究では、ゲノムスケールの遺伝子発現データに基づいた、生物影響の迅速な評価法の開発を目指している。まず、ヒト乳腺上皮細胞に、既知の発がん性物質および変異原を11種類を用い、細胞毒性について検討した。その結果、TCDDを除く、10種類の物質に対し毒性を示す濃度を明らかにできた。また、このうち砒素、水銀、カドミウムについては、毒性が見られる1/100程度の濃度で、細胞内の代謝が活性化されることが明らかとなった。砒素について、毒性がみられる濃度、代謝が活性化される濃度、どちらも見られない濃度で、DNAマイクロアレイを用い約１万の遺伝子（全遺伝子の1/3）の発現パターンを解析したところ、対照に対し発現量が二倍以上の遺伝子が73個見つかった。また、73遺伝子のクラスター解析により、発現パターンから、砒素の暴露量を分類・推定できることが示された。

概要 （英文）

Public concern has been increasing on health effect of environmental perturbations, such as electromagnetic fields and chemicals. Animal toxicity studies are expensive due to the long time frames required, are politically challenged due to the use of animals, and the applicability of the results obtained to human exposure is open to question. In this study, we have developed a rapid and accurate estimation method for carcinogens based on global gene expression analysis of human cells. Human non-tumorigenic breast epithelial cells (MCF10A) were exposed to a variety of known carcinogens and mutagens including metals (arsenic, mercury, chromium (VI), cadmium), dioxins (2,3,7,8-TCDD, 3,3'DiPB), genotoxicans (MNNG, 4NQO, MMS, t butyl hydroperoxide) and promoter (TPA). All of these chemicals, except TCDD, showed toxicity, as shown by growth inhibition for MCF10A. TCDD showed no toxicity at concentrations up to 1 micro-M (solubility limit in DMSO) possibly due to low expression of the Ah receptor in MCF10A. Some metals (arsenic, mercury, cadmium) have not only shown cytotoxicity, but also enhancement of cell activity based on NADH accumulation. For arsenic, global gene expression analysis by DNA microarray (ca. 10,000 genes, almost 1/3 of the total human genes) was carried out under the condition of high dose (cytotoxic) exposure (5h, 12h, 24h) medium dose (activation) exposure (12h, 24h) or low dose (no phenotypic effect) exposure (24h). Only 73 genes were found to have a 2-fold differential in expression vs. unexposed controls. Some of the affected genes have code for intriguing functions, such thioredoxin, heat shock protein, transcription factor and cell signaling. Clustering the 73 genes based on cosine correlation coefficient, showed a clear colleration to exposure dose, but not exposure period. Using the k-nearest neighbor algorithm, the chosen genes from 14 to 73 could predict 5 of 6 exposure conditions, clearly demonstrating the utility of global gene expression in detecting arsenic's effect on MCF10A. In future work, we expect to identify gene clusters to predict exposure to other toxic agents.

報告者

キーワード