Document jmEyxj1orJao5Qajq9LrD9oRk

Toxicology 229 (2007) 101113 Microarray analysis of gene expression in peripheral blood mononuclear cells from dioxin-exposed human subjects Cliona M. McHale a,, Luoping Zhang a, Alan E. Hubbard a, Xin Zhao a, Andrea Baccarelli b,c, Angela C. Pesatori b,c, Martyn T. Smith a, Maria Teresa Landi d a School of Public Health, University of California, Berkeley, CA 94720-7360, United States b Centro di Ricerca di Epidemiologia Occupazionale, clinica e Ambientale (EPOCA), University of Milan, Milan, Italy c Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, Milan, Italy d Genetic Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, DHHS, Bethesda, MD 20892-7236, United States Received 26 June 2006; received in revised form 29 September 2006; accepted 10 October 2006 Available online 17 October 2006 Abstract Tetrachlorodibenzo-p-dioxin (TCDD) is classified as a human carcinogen and exerts toxic effects on the skin (chloracne). Effects on reproductive, immunological, and endocrine systems have also been observed in animal models. TCDD acts through the aryl hydrocarbon receptor (AhR) pathway influencing largely unknown gene networks. An industrial accident in Seveso, Italy in 1976 exposed thousands of people to substantial quantities of TCDD. Twenty years after the exposure, this study examines global gene expression in the mononuclear cells of 26 Seveso female never smokers, with similar age, alcohol consumption, use of medications, and background plasma levels of 22 dioxin congeners unrelated to the Seveso accident. Plasma dioxin levels were still elevated in the exposed subjects. We performed analyses in two different comparison groups. The first included high-exposed study subjects compared with individuals with background TCDD levels (average plasma levels 99.4 and 6.7 ppt, respectively); the second compared subjects who developed chloracne after the accident, and those who did not develop this disease. Overall, we observed a modest alteration of gene expression based on dioxin levels or on chloracne status. In the comparison between high levels and background levels of TCDD, four histone genes were up-regulated and modified expression of HIST1H3H was confirmed by real-time PCR. In the comparison between chloracne casecontrol subjects, five hemoglobin genes were upregulated. Pathway analysis revealed two major networks for each comparison, involving cell proliferation, apoptosis, immunological and hematological disease, and other pathways. Further examination of the role of these genes in dioxin induced-toxicity is warranted. 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Tetrachlorodibenzo-p-dioxin (TCDD); Microarray; Gene expression; Biomarkers; Leukemia; Blood Corresponding author at: 140 Warren Hall, Division of Environmental Health Sciences, School of Public Health, University of California at Berkeley, Berkeley, CA 94720-7360, United States. Tel.: +1 510 643 5349; fax: +1 510 642 0427. E-mail address: cmchale@berkeley.edu (C.M. McHale). 0300-483X/$ see front matter 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.tox.2006.10.004 102 C.M. McHale et al. / Toxicology 229 (2007) 101113 1. Introduction 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD), a major environmental contaminant, has a long biological half-life (7 years) and was listed as an established human carcinogen by the International Agency for Research on Cancer (IARC) in 1997 (IARC, 1997). TCDD exerts varying toxic effects in experimental animals, including developmental, endocrinological, immunological, and reproductive effects (Birnbaum and Tuomisto, 2000). TCDD alters the expression of a wide spectrum of genes via binding to the aryl hydrocarbon receptor (AhR) through recognition of xenobiotic response elements (XRE) (Mimura and Fujii-Kuriyama, 2003) and XRE-II (Boutros et al., 2004; Sogawa et al., 2004) in gene promoters. These genes include phase I and II xenobiotic metabolizing enzymes, as well as genes involved in cell proliferation, cell cycle regulation and apoptosis. However, downstream gene targets are largely uncharacterized and the role of gene expression changes in TCDD-associated toxicity and diseases such as chloracne and cancer are largely unknown. In 1976, an industrial accident exposed several thousand people to substantial quantities of TCDD in Seveso, Italy (Bertazzi et al., 2001). Within several months of the accident, a large outbreak of chloracne was observed (Baccarelli et al., 2005). Follow-up of the exposed population through 1996 revealed an excess of hemopoietic neoplasms in both genders (RR 1.7, 95% CI: 1.2, 2.5) with the highest increase for non-Hodgkin's lymphoma (RR = 2.8, 95% CI: 1.1, 7.0) and myeloid leukemia (RR = 3.8, 95% CI: 1.2, 12.5), occurring after 15 years (Bertazzi et al., 2001). The increased risk of hemopoietic neoplasms in the Seveso population underlines the importance of studying long-term dioxin effects in exposed subjects. Approximately 20 years after the exposure, high TCDD plasma levels were still present in the exposed individuals with significantly higher levels in women (Landi et al., 1997, 1998). We hypothesized that longterm presence of dioxin in these subjects could result in alteration of gene expression and that examination of global gene expression by microarray would help us to elucidate the genetic pathways involved in hemopoietic neoplasm development. We therefore chose peripheral blood mononuclear cells as a relevant target for these studies. We report here results of global gene expression analysis in a sample from the Seveso cohort comparing individuals with high levels of dioxin with those with background levels. We also compared individuals who developed chloracne immediately after the accident with those who were exposed but did not develop chloracne, to investigate gene expression changes underlying the development of chloracne. 2. Materials and methods 2.1. Study subjects We selected 26 subjects from the Seveso cohort for this study. We obtained approvals of the institutional review boards and written informed consent from each study subject. The selection aimed at comparing subjects with extreme levels of plasma dioxin 20 years after the accident, and no major differences with regard to potential confounders. Thus, we selected only never smoker, one gender (female), and similar ages (2449 years), alcohol consumption and medication use between groups. In addition, we previously measured plasma levels of 22 congeners in all subjects, and they did not substantially differ within this study sample. Thirteen subjects were exposed to background levels of TCDD (mean 6.7 ppt, range 4.57.9 ppt) and 13 exposed to high levels of TCDD (mean 99.4 ppt, range 23.9268 ppt). The mean age was 32 and 28 years, in the high and background exposure groups, respectively. The 26 individuals could also be equally divided into two casecontrol groups based on development of chloracne in response to TCDD exposure. TCDD levels ranged from 74.3 to 268 ppt and from 25.2 to 90.2 ppt, in chloracne cases and subjects who did not develop chloracne, respectively. Mean age was 28 and 32 years, in cases and controls, respectively. Based on this selection, a power calculation demonstrated that 13 matched pairs would be sufficient to have an 80% probability of selecting a true gene if the ratio of expression were 2.0 and the family-wise error rate (FWER) were controlled at 5%. 2.2. Sample collection, RNA isolation, amplification and hybridization The increased risk of hemopoietic neoplasms in the Seveso population, in the absence of differences in the complete blood count of the exposed individuals, suggested that peripheral blood mononuclear cells (PBMC) could represent a good target for these studies. Details of sample collection and PBMC isolation have been described in (Landi et al., 2003). Briefly, PBMC were separated from whole blood (550 ml collected in tubes treated with sodium heparin) on a ficoll (Histopaque 1077, Sigma Chemical Co., St. Louis, MO) hypaque density gradient. The cells were washed twice and cryopreserved at a concentration of 2 107 cells/ml with an equal volume of freeze medium (RPMI 1640 (Life Technologies) and with 7.5% cell culture grade dimethylsulfoxide (American Type Culture Collection, Rockville, MD), 20% FBS, 2 mM glutamine, 100 U/ml penicillin and 100 g/ml amphotericin (Life Technologies). A one ml aliquot of cells was frozen at a rate of 1 C per minute and then stored in the vapor phase of liquid nitrogen. C.M. McHale et al. / Toxicology 229 (2007) 101113 103 Total RNA was isolated from cryopreserved mononuclear cells using RNeasy mini kits (Qiagen, Valencia, CA) according to the manufacturer's instructions and quantified using a SmartSpecTM3000 (Bio-Rad, Hercules, CA). Each of the 26 samples (100 ng) was labeled separately according to the GeneChip Eukaryotic Small Sample Target Labeling Assay Version II (http://www.affymetrix. com/support/technical/technotesmain.affx), with the exception that the GeneChip Sample Cleanup Module (Affymetrix, Santa Clara, CA) was used instead of ethanol precipitation. The protocol consists of two rounds comprising oligo-dT-primed cDNA synthesis, second-strand cDNA synthesis, and in vitro transcription (IVT) RNA amplification steps, with intermediate clean-up protocols after each step. During the second round of IVT biotinylated UTP is incorporated into the cRNA. The biotin-labeled RNA is hybridized to the GeneChips, washed, detected with streptavidin phycoerythrin conjugate and scanned by the GeneChip Scanner. The amount of light emitted at 570 nm is proportional to the bound target at each location on the probe array. Hybridization and scanning were performed as described in the GeneChip Expression Manual. Samples were hybridized singly to GeneChips. 2.3. Chip normalization and quality controls Redwood City, CA, http://www.ingenuity.com) a web-based application, which queries the ingenuity pathway knowledge base (IPKB) for genetic interactions. The information contained in the ingenuity pathways knowledge base is derived from the scientific literature and each connection in a network is supported by one or more publications. There exist in the literature over 139 peer-reviewed publications including IPA data, e.g. (Challen et al., 2005). Genes which can be mapped to genetic networks available in the Ingenuity database, known as "focus genes", are used to build networks and a score for each network is calculated according to the fit of the user's set of genes. Score is displayed as the negative log of the p-value, indicating the likelihood of the focus genes in a network being found together by random chance. Thus score of 2 have at least a 99% chance of not being generated by chance alone. In the current study a score of 10 or higher was used to select highly significant biological networks. To evaluate the significance of the association of a particular gene set with the relevant canonical pathway within Ingenuity, a ratio of the number of genes from the data set that map to the pathway divided by the total number of genes that map to the canonical pathway is displayed and Fischer's exact test is used to calculate the corresponding p-value. Quality control indicators including initial sample quality (A260/A280 ratios between 1.7 and 2.2), integrity by gel analysis (28S ribosomal bands approximately twice the intensity of the 18S bands), amplification yields for the first and second rounds (exceeding 25 g/ml and 1000 g/ml, respectively) and quality of the second round amplified RNAs (A260/A280 > 1.85) were all satisfactory. Quality control metrics of the microarray raw data including noise, background, % probe sets present/absent, and 3 /5 ratios for internal control genes (-actin and GAPDH) were satisfactory and consistent. To allow comparisons, all chips were scaled to a target intensity of 500 based on all probe sets on each chip. Samples were run blind so that exposure status was unknown. 2.4. Statistical analysis to identify differentially expressed genes Robust multi-array analysis (RMA) (Irizarry et al., 2003) was used to analyze the data produced by the chips. Genes whose expression was significantly different between high- and background exposed individuals, as well as between chloracne cases versus control individuals were identified using a standard two-sample t-test (allowing for unequal variances in highand low-exposed groups) and two-sided p-values. Expression ratios were based on the normalized mean data of all members of the group. 2.5. Pathway analysis Gene Refseq accession numbers were imported into Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems, 2.6. Quantitative real-time PCR analysis using TaqMan gene expression assays Total cDNA (equivalent to 20 ng input RNA), generated with the SuperScriptTM First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA) was used to confirm GeneChip findings by TaqMan Gene Expression Assays (TMGEA) (Applied Biosystems, Foster City, CA), which were run in quadruplicate under standard assay conditions. Because of its abundance and low variability between different types of lymphoid cells, the TATA box binding protein (TBP) was used as a normalization gene for real-time PCR. The mean baseline cycle threshold (Ct) for TATA box binding protein (TBP; GenBank ID NM 003194), was subtracted from the mean Ct for the other six assays to normalize results. These were then compared between high/background exposed and chloracne case/control sample groups. As PCR amplification is exponential, each cycle represents a doubling of the PCR product, so a cycle difference in relative Ct represents a two-fold difference in starting cDNA quantity. TMGEA used were TATA box binding protein (TBP), Hs99999910 m1; hemoglobin beta (HBB), Hs00758889 s1; histone 1, H3h (HIST1H3H), Hs00818527 s1; interleukin 8 (IL8), Hs00174103 m1; cytochrome p450, family 1, subfamily B, polypeptide 1 (CYP1B1), Hs00164383 m1; cAMP responsive element modulator (CREM), Hs00181804 m1; glutathione S-transferase M3 (GSTM3); Hs00168307 m1, kruppel-like factor 4 (KLF4), Hs00358836 m1; microsomal glutathione S-transferase 1 (MGST1), Hs00220393 m1; nuclear factor, interleukin 3 regulated (NFIL3), Hs00705412 s1; transducer of ERBB2, 1 (TOB1), Hs00271739 s1. 104 C.M. McHale et al. / Toxicology 229 (2007) 101113 Table 1 Genes significantly up-regulated by dioxin exposure ranked by ratio (N = 22 probes/genes) Affy ID Ratio p-Value Gene title 244181 at 229740 at 222067 x at 215528 at 209889 at 216176 at 208527 x at 224489 at 240823 at 222139 at 215235 at 217152 at 204014 at 220918 at 208523 x at 214481 at 207373 at 232613 at 225929 s at 201853 s at 234848 at 206813 at 1.497 1.479 1.391 1.354 1.350 1.337 1.326 1.315 1.301 1.298 1.275 1.270 1.268 1.262 1.239 1.230 1.230 1.226 1.216 1.210 1.205 1.201 0.0429 0.0168 0.0096 0.0371 0.0180 0.0374 0.0317 0.0179 0.0008 0.0209 0.0465 0.0398 0.0133 0.0479 0.0069 0.0067 0.0004 0.0027 0.0233 0.0044 0.0001 0.0171 Phosphoinositide-3-kinase, regulatory subunit 1 (p85 alpha) PP12104 Histone 1, H2bd Mannosyl (alpha-1,6-)-glycoprotein SEC31-like 2 (S. cerevisiae) Hepatocellular carcinoma-related Histone 1, H2be KIAA1267 Homo sapiens, clone IMAGE:5730164, mRNA KIAA1466 gene Spectrin, alpha, non-erythrocytic 1 (alpha-fodrin) CDNA FLJ14074 fis, clone HEMBB1001869 Dual specificity phosphatase 4 Chromosome 21 open reading frame 96 Histone 1, H2bi Histone 1, H2am Homeo box D10 Polybromo 1 Chromosome 17 open reading frame 27 Cell division cycle 25B Human TCRAV5.1a mRNA for T cell receptor alpha-chain. Cardiotrophin 1 Gene symbol PIK3R1 LOC643008 HIST1H2BD MGAT5 SEC31L2 HCRP1 HIST1H2BE KIAA1267 KIAA1466 SPTAN1 DUSP4 C21orf96 HIST1H2BI HIST1H2AM HOXD10 PB1 C17orf27 CDC25B LOC650815 CTF1 RefSeq ID NM 181504 XM 928053 NM 021063 NM 002410 NM 015490 NM 003523 NM 015443 NM 003127 NM 001394 NM 003525 NM 003514 NM 002148 NM 018165 NM 020914 NM 004358 XM 939906 NM 001330 3. Results 3.1. Differential gene expression in the high-low dioxin-exposed groups by microarray Peripheral blood mononuclear cell RNAs from high/background exposed groups (two groups of 13 individuals each) were analyzed by Affymetrix GeneChip array. Among the top 50 most significantly altered genes (p-value 0.05) the majority of genes showed low levels (<20%) of up- or down-regulation of expression by dioxin exposure (data not shown). Aldehyde dehydrogenase 6 family, member A1 (ALDH6A1) exhibited the greatest downregulation with a ratio of 0.646 (p = 0.0005), while homeobox D10 (HOXD10) showed the greatest degree of up-regulation with a ratio of 1.23 (p = 0.0004), among the top 50 most significant genes. Given the low ratios of differential expression overall, we considered a cut-off ratio of 1.2 as having potential biological relevance. Using this cut-off, 22 genes (22 probes) were significantly up-regulated (Table 1) and 113 genes (118 probes) were down-regulated by dioxin exposure (Table 2). Five down-regulated genes CD86, DNAJC10, IGFBP7, PCSK5 and PTEN were each identified by two different probes (Table 2). The gene with the greatest magnitude of downregulation by exposure was HLA-DRB4 (ratio 0.30, p = 0.0157). The gene with the greatest degree of up-regulation by dioxin exposure was PIK3R1 (ratio 1.5, p = 0.0429), Table 1. Four histone genes (HIST1H2BD, HIST1H2BE, HIST1H2BI and HIST1H2AM) were significantly up-regulated by exposure (Table 1). 3.2. Differential gene expression in the chloracne casecontrol groups by microarray Gene expression in the study population was also examined on the basis of chloracne with cases having developed chloracne after exposure and controls remaining asymptomatic for the disease. Low ratios of differential expression were observed for the top 50 chloracne-associated genes with lowest p-values (data not shown). Of note however was the fact that glutathione S-transferase M3 (GSTM3) was up-regulated as evidenced by two different probe sets (202554 s at and 235867 at ratios 1.37 and 1.22; p = 0.0004 and 0.00001). C.M. McHale et al. / Toxicology 229 (2007) 101113 Table 2 Genes significantly downregulated by dioxin exposure ranked by ratio (N = 118 probes/113 genes) Affy ID Ratio p-Value Gene title 209728 at 228949 at 201163 s at 225207 at 209555 s at 220646 s at 230413 s at 226668 at 227787 s at 221841 s at 223087 at 221589 s at 201218 at 222453 at 203574 at 238002 at 225114 at 228170 at 236487 at 244187 at 239346 at 226329 s at 235158 at 223090 x at 203973 s at 226665 at 205922 at 224436 s at 235670 at 204286 s at 203765 at 226383 at 205559 s at 217427 s at 225796 at 222637 at 235346 at 235615 at 225367 at 0.301 0.504 0.571 0.576 0.604 0.611 0.623 0.627 0.629 0.629 0.640 0.646 0.648 0.650 0.651 0.651 0.662 0.663 0.664 0.665 0.666 0.668 0.669 0.669 0.674 0.675 0.675 0.676 0.677 0.680 0.680 0.680 0.680 0.681 0.684 0.688 0.688 0.690 0.691 0.0157 0.0304 0.0218 0.0196 0.0288 0.0065 0.0342 0.0011 0.0098 0.0361 0.0111 0.0005 0.0102 0.0118 0.0258 0.0311 0.0081 0.0354 0.0010 0.0042 0.0068 0.0024 0.0095 0.0008 0.0245 0.0074 0.0236 0.0055 0.0467 0.0196 0.0284 0.0030 0.0059 0.0101 0.0027 0.0095 0.0091 0.0287 0.0065 Major histocompatibility complex, class II, DR beta 4 G protein-coupled receptor 177 Insulin-like growth factor binding protein 7 Pyruvate dehydrogenase kinase, isozyme 4 CD36 molecule (thrombospondin receptor) Killer cell lectin-like receptor subfamily F, member 1 Adaptor-related protein complex 1, sigma 2 subunit WD repeat, sterile alpha motif and U-box domain containing 1 Thyroid hormone receptor associated protein 6 Kruppel-like factor 4 (gut) Enoyl coenzyme A hydratase domain containing 1 Aldehyde dehydrogenase 6 family, member A1 C-terminal binding protein 2 Cytochrome b reductase 1 Nuclear factor, interleukin 3 regulated Golgi phosphoprotein 4 Alkylglycerone phosphate synthase Oligodendrocyte transcription factor 1 Hypothetical protein FLJ30655 Chromosome X open reading frame 33 Chromosome 12 open reading frame 38 Hypothetical protein BC018453 Hypothetical protein FLJ14803 Vezatin, adherens junctions transmembrane protein CCAAT/enhancer binding protein (C/EBP), delta AHA1, activator of heat shock 90 kDa protein ATPase homolog 2 Vanin 2 Nipsnap homolog 3A (C. elegans) Syntaxin 11 Phorbol-12-myristate-13-acetateinduced protein 1 Grancalcin, EF-hand calcium binding protein Chromosome 11 open reading frame 46 Proprotein convertase subtilisin/kexin type 5 HIR histone cell cycle regulation defective homolog A PX domain containing serine/threonine kinase COMM domain containing 10 FUN14 domain containing 1 Protein geranylgeranyltransferase type I, beta subunit Phosphoglucomutase 2 Gene symbol HLA-DRB4 GPR177 IGFBP7 PDK4 CD36 KLRF1 AP1S2 WDSUB1 THRAP6 KLF4 ECHDC1 ALDH6A1 CTBP2 CYBRD1 NFIL3 GOLPH4 AGPS OLIG1 FLJ30655 CXorf33 C12orf38 LOC129531 FLJ14803 VEZT CEBPD AHSA2 VNN2 NIPSNAP3A STX11 PMAIP1 GCA C11orf46 PCSK5 HIRA PXK COMMD10 FUNDC1 PGGT1B PGM2 105 RefSeq ID NM 021983 NM 001002292 NM 001553 NM 002612 NM 000072 NM 016523 NM 003916 NM 152528 NM 080651 NM 004235 NM 018479 NM 005589 NM 001329 NM 024843 NM 005384 NM 014498 NM 003659 NM 138983 NM 144643 NM 198450 NM 024809 NM 138798 NM 032842 NM 017599 NM 005195 NM 152392 NM 004665 NM 015469 NM 003764 NM 021127 NM 012198 NM 152316 NM 006200 NM 003325 NM 017771 NM 016144 NM 173794 NM 005023 NM 018290 106 Table 2 (Continued ) Affy ID Ratio 223065 s at 242648 at 202085 at 0.691 0.692 0.692 223423 at 235463 s at 0.694 0.694 228153 at 225769 at 0.694 0.697 226276 at 222235 s at 222714 s at 204194 at 231736 x at 227268 at 226142 at 234915 s at 238465 at 204160 s at 0.697 0.697 0.698 0.698 0.698 0.699 0.699 0.700 0.700 0.705 228155 at 239328 at 213222 at 0.705 0.706 0.712 209686 at 0.713 210895 s at 219147 s at 201888 s at 219859 at 0.717 0.718 0.719 0.719 211711 s at 213005 s at 204053 x at 238581 at 201653 at 225174 at 0.724 0.731 0.732 0.734 0.735 0.740 214084 x at 0.741 227680 at 222562 s at 218519 at 242245 at 0.742 0.745 0.748 0.749 201889 at 0.749 210176 at 217955 at 224967 at 0.750 0.751 0.753 201162 at 0.753 202388 at 0.755 209814 at 216652 s at 0.758 0.759 C.M. McHale et al. / Toxicology 229 (2007) 101113 p-Value 0.0035 0.0323 0.0124 0.0328 0.0034 0.0220 0.0079 0.0336 0.0110 0.0319 0.0236 0.0311 0.0018 0.0115 0.0062 0.0065 0.0271 0.0039 0.0394 0.0071 0.0438 0.0372 0.0006 0.0045 0.0142 0.0249 0.0149 0.0021 0.0220 0.0347 0.0010 0.0097 0.0009 0.0267 0.0005 0.0005 0.0382 0.0358 0.0058 0.0106 0.0451 0.0301 0.0318 0.0151 Gene title STARD3 N-terminal like Kelch-like 8 (Drosophila) Tight junction protein 2 (zona occludens 2) G protein-coupled receptor 160 LAG1 longevity assurance homolog 6 (S. cerevisiae) IBR domain containing 2 Component of oligomeric Golgi complex 6 Hypothetical protein MGC23909 Chondroitin sulfate GalNAcT-2 Lactamase, beta 2 BTB and CNC homology 1 Microsomal glutathione S-transferase 1 PTD016 protein GLI pathogenesis-related 1 (glioma) Density-regulated protein Hypothetical protein MGC33648 Ectonucleotide pyrophosphatase/phosphodiesterase 4 Chromosome 10 open reading frame 58 RCSD domain containing 1 Phospholipase C, beta 1 (phosphoinositide-specific) S100 calcium binding protein, beta (neural) CD86 molecule Chromosome 9 open reading frame 95 Interleukin 13 receptor, alpha 1 C-type lectin domain family 4, member E Phosphatase and tensin homolog Ankyrin repeat domain 15 Phosphatase and tensin homolog Guanylate binding protein 5 Cornichon homolog (Drosophila) DnaJ (Hsp40) homolog, subfamily C, member 10 Similar to neutrophil cytosol factor 1 (NCF-1) Zinc finger protein 326 Tankyrase Solute carrier family 35, member A5 Synapse defective 1, Rho GTPase, homolog 2 (C. elegans) Family with sequence similarity 3, member C Toll-like receptor 1 BCL2-like 13 (apoptosis facilitator) UDP-glucose ceramide glucosyltransferase Insulin-like growth factor binding protein 7 Regulator of G-protein signalling 2, 24kDa Zinc finger protein 330 Down-regulator of transcription 1, TBP-binding Gene symbol STARD3NL KLHL8 TJP2 GPR160 LASS6 IBRDC2 COG6 MGC23909 GALNACT-2 LACTB2 BACH1 MGST1 LOC51136 GLIPR1 DENR MGC33648 ENPP4 C10orf58 RCSD1 PLCB1 S100B CD86 C9orf95 IL13RA1 CLEC4E PTEN ANKRD15 PTEN GBP5 CNIH DNAJC10 LOC648998 ZNF326 TNKS2 SLC35A5 SYDE2 FAM3C TLR1 BCL2L13 UGCG IGFBP7 RGS2 ZNF330 DR1 RefSeq ID NM 032016 NM 020803 NM 004817 NM 014373 NM 203463 NM 182757 NM 020751 NM 174909 NM 018590 NM 016027 NM 001011545 NM 020300 NM 016125 NM 006851 NM 003677 NM 153706 NM 014936 NM 032333 NM 052862 NM 015192 NM 006272 NM 006889 NM 017881 NM 001560 NM 014358 NM 000314 NM 015158 NM 000314 NM 052942 NM 001009551 NM 018981 XM 927922 NM 181781 NM 025235 NM 017945 XM 086186 NM 001040020 NM 003263 NM 015367 NM 003358 NM 001553 NM 002923 NM 014487 NM 001938 Table 2 (Continued ) Affy ID Ratio 226366 at 0.760 204774 at 205715 at 202539 s at 0.760 0.761 0.762 226493 at 209096 at 215933 s at 229588 at 225612 s at 205560 at 222752 s at 226283 at 201487 at 229533 x at 202651 at 204809 at 206877 at 207654 x at 201636 at 202026 at 218398 at 201589 at 227322 s at 228670 at 204634 at 236609 at 210093 s at 203177 x at 205686 s at 228585 at 218616 at 227593 at 200977 s at 226962 at 208127 s at 0.762 0.765 0.768 0.770 0.771 0.773 0.773 0.774 0.779 0.781 0.783 0.783 0.785 0.786 0.786 0.787 0.788 0.789 0.790 0.793 0.795 0.795 0.795 0.796 0.797 0.798 0.798 0.798 0.799 0.799 0.799 C.M. McHale et al. / Toxicology 229 (2007) 101113 p-Value 0.0203 0.0112 0.0302 0.0451 0.0221 0.0140 0.0479 0.0123 0.0362 0.0008 0.0001 0.0436 0.0312 0.0045 0.0187 0.0084 0.0036 0.0158 0.0362 0.0191 0.0058 0.0078 0.0174 0.0040 0.0154 0.0096 0.0392 0.0347 0.0250 0.0021 0.0332 0.0051 0.0004 0.0138 0.0041 Gene title SNF2 histone linker PHD RING helicase Ecotropic viral integration site 2A Bone marrow stromal cell antigen 1 3-Hydroxy-3-methylglutarylCoenzyme A reductase Potassium channel tetramerisation domain containing 18 Ubiquitin-conjugating enzyme E2 variant 2 Homeobox, hematopoietically expressed DnaJ (Hsp40) homolog, subfamily C, member 10 UDP-GlcNAc: betaGal beta-1,3-Nacetylglucosaminyltransferase 5 Proprotein convertase subtilisin/kexin type 5 Chromosome 1 open reading frame 75 WD repeat domain 51B Cathepsin C Zinc finger protein 680 Lysophosphatidylglycerol acyltransferase 1 ClpX caseinolytic peptidase X homolog (E. coli) MAX dimerization protein 1 Down-regulator of transcription 1, TBP-binding Fragile X mental retardation, autosomal homolog 1 Succinate dehydrogenase complex, subunit D Mitochondrial ribosomal protein S30 SMC1 structural maintenance of chromosomes 1-like 1 (yeast) BRCA2 and CDKN1A interacting protein Telomerase-associated protein 1 NIMA (never in mitosis gene a)-related kinase 4 PMS1 postmeiotic segregation increased 1 (S. cerevisiae) Mago-nashi homolog, proliferation-associated (Drosophila) Transcription factor A, mitochondrial CD86 molecule Ectonucleoside triphosphate diphosphohydrolase 1 Integrator complex subunit 12 Hypothetical protein LOC645580 Tax1 (human T cell leukemia virus type I) binding protein 1 Zinc finger and BTB domain containing 41 Suppressor of cytokine signaling 5 Gene symbol SHPRH EVI2A BST1 HMGCR KCTD18 UBE2V2 HHEX DNAJC10 B3GNT5 PCSK5 C1orf75 WDR51B CTSC ZNF680 LPGAT1 CLPX MXD1 DR1 FXR1 SDHD MRPS30 SMC1L1 BCCIP TEP1 NEK4 PMS1 MAGOH TFAM CD86 ENTPD1 INTS12 FLJ37453 TAX1BP1 ZBTB41 SOCS5 107 RefSeq ID NM 173082 NM 001003927 NM 004334 NM 000859 NM 152387 NM 003350 NM 002729 NM 018981 NM 032047 NM 006200 NM 018252 NM 172240 NM 001814 NM 178558 NM 014873 NM 006660 NM 002357 NM 001938 NM 001013438 NM 003002 NM 016640 NM 006306 NM 016567 NM 007110 NM 003157 NM 000534 NM 002370 NM 003201 NM 006889 NM 001776 NM 020395 XM 928597 NM 006024 NM 194314 NM 014011 108 C.M. McHale et al. / Toxicology 229 (2007) 101113 Table 3 Genes significantly up-regulated in chloracne cases ranked by ratio (N = 33 probes/23 genes) Affy ID Ratio p-Value Gene title 214414 x at 209116 x at 211696 x at 213515 x at 217232 x at 211745 x at 209458 x at 217414 x at 204018 x at 211699 x at 204848 x at 204419 x at 209967 s at 230645 at 223298 s at 209723 at 230511 at 202704 at 200864 s at 202554 s at 240456 at 206834 at 204897 at 232164 s at 201980 s at 206545 at 201528 at 201236 s at 230170 at 203725 at 235867 at 209112 at 204550 x at 3.11 3.04 2.58 2.46 2.39 2.26 2.06 2.04 1.79 1.66 1.60 1.59 1.55 1.49 1.45 1.44 1.43 1.41 1.40 1.37 1.33 1.33 1.32 1.32 1.28 1.27 1.27 1.27 1.25 1.23 1.22 1.22 1.21 0.02158 0.0222 0.01197 0.03382 0.02357 0.02166 0.02117 0.02068 0.01021 0.02107 0.02618 0.04907 0.02096 0.01369 0.04984 0.03159 0.02038 0.04464 0.00839 0.00044 0.03415 0.03594 0.02418 0.01637 0.04271 0.00219 0.02876 0.02303 0.02054 0.02686 0.00001 0.03005 0.04847 Hemoglobin, alpha 2 Hemoglobin, beta Hemoglobin, beta Hemoglobin, gamma A Hemoglobin, beta Hemoglobin, alpha 1 Hemoglobin, alpha 1 Hemoglobin, alpha 1 Hemoglobin, alpha 1 Hemoglobin, alpha 1 Hemoglobin, gamma A Hemoglobin, gamma A cAMP responsive element modulator FERM domain containing 3 5 -Nucleotidase, cytosolic III Serpin peptidase inhibitor, clade B (ovalbumin), member 9 cAMP responsive element modulator Transducer of ERBB2, 1 RAB11A, member RAS oncogene family Glutathione S-transferase M3 (brain) FLJ11795 protein Hemoglobin, delta Prostaglandin E receptor 4 (subtype EP4) Epiplakin 1 Ras suppressor protein 1 CD28 molecule Replication protein A1, 70kDa BTG family, member 2 oncostatin M Growth arrest and DNA-damage-inducible, alpha Glutathione S-transferase M3 (brain) Cyclin-dependent kinase inhibitor 1B (p27, Kip1) Glutathione S-transferase M1 Gene symbol HBA2 HBB HBB HBG1 HBB HBA1 HBA1 HBA1 HBA1 HBA1 HBG1 HBG1 CREM FRMD3 NT5C3 SERPINB9 CREM TOB1 RAB11A GSTM3 FLJ11795 HBD PTGER4 EPPK1 RSU1 CD28 RPA1 BTG2 OSM GADD45A GSTM3 CDKN1B GSTM1 RefSeq ID NM 000517 NM 000518 NM 000518 NM 000184 NM 000518 NM 000558 NM 000517 NM 000517 NM 000517 NM 000517 NM 000184 NM 000184 NM 001881 NM 174938 NM 001002009 NM 004155 NM 001881 NM 005749 NM 004663 NM 000849 NM 001039935 NM 000519 NM 000958 NM 031308 NM 012425 NM 006139 NM 002945 NM 006763 NM 020530 NM 001924 NM 000849 NM 004064 NM 000561 As with the exposure data in the previous section, we considered a cut-off ratio of 1.2 as having potential biological relevance and using this cut-off, 23 genes represented by 33 probes were significantly up-regulated in association with chloracne (Table 3). CREM was identified by two different probes. Five hemoglobin genes: HBA2 and HBD (each identified by a single probe); HBB and HBG1 (identified by three different probes); HBA1 (identified by five different probes) were significantly up-regulated. As shown in Table 4, nine genes were down-regulated. 3.3. Identification of biological networks affected by dioxin exposure and chloracne status We investigated biological interactions among the genes associated with dioxin exposure and chloracne status using the Ingenuity Pathway Analysis (IPA) tool. Analysis of the top 200 genes with the greatest magnitude of differential expression associated with dioxin exposure and chloracne, with a p-value cut-off of 0.05 and a ratio cut-off of 1.25 (up- or down-regulation), showed two significant networks each (score 10). Network genes are listed in Table 5. 3.3.1. Biological networks affected by dioxin exposure The top-scoring network (network 1, score = 27) identified for dioxin exposure includes genes involved in cellular growth and proliferation (KLF4, CD36, C-terminal binding protein 2 (CTBP2)), glucose metabolism (CD36, pyruvate dehydrogenase kinase, isozyme 4 (PDK4)), and cell death (NFIL3, CD36, CTBP2, KLF4). Network 2 (score = 22, Table 5) asso- C.M. McHale et al. / Toxicology 229 (2007) 101113 Table 4 Genes significantly downregulated in chloracne cases ranked by ratio (N = 9 probes/genes) Affy ID Ratio p-Value Gene title 244546 at 230756 at 214850 at 224851 at 229040 at 229748 x at 221652 s at 205898 at 211986 at 0.68 0.74 0.75 0.76 0.76 0.76 0.77 0.77 0.78 0.03902 0.03511 0.03679 0.04771 0.0293 0.03728 0.02217 0.04669 0.00469 Cytochrome c, somatic Zinc finger protein 683 Hypothetical protein LOC153561 Cyclin-dependent kinase 6 Integrin, beta 2 Similar to Tektin-3 Chromosome 12 open reading frame 11 Chemokine (C-X3-C motif) receptor 1 AHNAK nucleoprotein (desmoyokin) Gene symbol CYCS ZNF683 LOC153561 CDK6 ITGB2 LOC389830 C12orf11 CX3CR1 AHNAK 109 RefSeq ID NM 018947 NM 173574 NM 207331 NM 001259 NM 000211 NM 001033515 NM 018164 NM 001337 NM 001620 ciated with dioxin exposure includes genes involved in cell death/apoptosis (CCAAT/enhancer binding protein (C/EBP), delta (CEBPD), GLI pathogenesis-related 1 (glioma) (GLIPR1), insulin-like growth factor binding protein 7 (IGFBP7), KLF4, phosphoinositide-3-kinase, regulatory subunit 1 (PIK3R1), phorbol-12-myristate13-acetate-induced protein 1 (PMAIP1), DNA replication, recombination, and repair (PMAIP1, PIK3R1), and cancer (KLF4, CEBPD, IGFBP7, PIK3R1). While networks consider all possible interactions, canonical pathway analysis queries genes in pre-defined and well-characterized biological pathways. With the canonical pathway analysis, two pathways were significantly associated with dioxin exposure but differential expression of only two genes was observed in each case. In the insulin-like growth factor-1 (IGF-1) signaling pathway (p = 0.0276) insulin-like growth factor binding protein 7 (IGFBP7) was decreased 1.75-fold, Table 5 High-scoring networks (score 10) identified by Ingenuity pathway analysis Network ID Network Association 1 AP1S2, CD36, CLDN1, COMMD10, Dioxin exposure CTBP2, DUSP4, ELK3, FLJ30655, FNTA, HAS2, HLA-DRB4, HRAS, IL15, IL18RAP, IL1A, KLF3, KLF4, LIPE, MAP3K7IP2, MAPK1, NCOR1, NFIL3, NFKB1, PCSK5, PDK4, PGGT1B, POMC, PPARG, RLN2, SAA1, SPI1, TJP2, TLR4, TNFSF11, UBE2I 2 ACP1, BACH1, BARD1, BRCA1, CASP3, Dioxin exposure CDKN1A, CEBPD, CHRNA7, CRSP6, CTNNB1, DDR1, ELL, GLIPR1, GPS2, HIRA, HOXA5, IBRDC2, IFITM1, IGFBP7, IRF5, KLF4, PDCD8, PDE4B, PIK3R1, PMAIP1, POU4F1, PTPRF, S100A4, S100B, SAA1, SMC2L1, SPTAN1, THRAP6, TP53, VEZATIN 3 AHNAK, BTG2, CCND1, CD28, CDK6, Chloracne CDKN2C, CREM, CX3CR1, CYCS, DAPK1, DMTF1, GSTM3, HAS1, HLA-DQA1, HLA-DQB1, HLA-DRB1, IFNG, IL6, IL8, IL17F, IL18RAP, LGALS7, LY96, PCNA, PTGER1, PTGER4, RFC5, RPA1, RSU1, SERPINB9, TGFB1, TNFRSF6B, TOB1, TRIM21, XCL1 4 EPO, GH1, HBA1, HBA2, HBB, Hbb-ar, Chloracne Hbb-b1, Hbb-b2, Hbb-bh1, HBD, HBE1, HBG1, HBG2, HBQ1, HBZ Focus genes are italicized. Score 27 22 28 10 Focus genes 14 12 13 5 Functions Cellular growth and proliferation, carbohydrate metabolism, cell death Cell death, DNA replication, recombination, and repair, cancer Cell death, immunological disease, cell-to-cell signaling and interaction Hematological disease, genetic disorder, hematological system development and function 110 C.M. McHale et al. / Toxicology 229 (2007) 101113 while PIK3R1 was increased 1.5-fold. PIK3R1 along with dual specificity phosphatase 4 (DUSP4), which was increased by 1.27-fold, implicated the stress-activated protein kinase/c-Jun NH2-terminal kinase (SAPK/JNK) signaling pathway (p = 0.031). 3.3.2. Biological networks affected by chloracne The top-scoring network associated with chloracne (network 3, score = 28), shown in Table 5, is related to cell death/apoptosis (BTG family, member 2 (BTG2), CD28, cAMP responsive element modulator (CREM), prostaglandin E receptor 4 (PTGER), replication protein A1, 70 kDa (RPA1), serpin peptidase inhibitor, clade B (ovalbumin), member 9 (SERPINB9), cyclin-dependent kinase 6 (CDK6), chemokine (C-X3-C motif) receptor 1 (CX3CR1), cytochrome c, somatic (CYCS)), immunological disease (CD28 and PTGER), cellcell signaling and interaction (CD28, CX3CR1 and CYCS). The second chloracne-associated network (network 4, score = 10, Table 5) includes five hemoglobin genes shown earlier in Table 4, which are involved in hematological disease, genetic disorders and hematological development and function. All five genes were up-regulated (in some cases almost three-fold) in individuals who developed chloracne and this was confirmed for HBB using realtime quantitative PCR. This finding was strengthened by the fact that 13 separate probe sets identified five hemoglobin genes. No canonical pathways were identified in the chloracne comparison. 3.4. Confirmation by quantitative real-time PCR Ten genes were chosen for further study and confirmation by real-time PCR, based on ratio of differential expression, p-value, number of probe sets identifying a gene, patterns of expression (such as hemoglobin and histone families described above), and potential biological significance. Fig. 1 illustrates the correlation between expression levels measured by real-time PCR and microarray for genes associated with dioxin exposure and chloracne status, and the correspondent levels of significance are reported in Table 6. For 8 of the 10 genes chosen, real-time PCR analysis confirmed the direction of change (up- or down-regulation). GSTM3, which was highly significantly up-regulated in association with chloracne by microarray (p = 0.0004), was confirmed by real-time PCR (p = 0.007). HBB was upregulated by high exposure (p = 0.04) and approached significant up-regulation in association with chloracne (p = 0.07) groups. Real-time PCR showed significant upregulation of HIST1H3H. Fig. 1. Correlation of microarray and real-time PCR measurements of expression of genes associated with (A) TCDD exposure, and (B) chloracne. 4. Discussion RNA samples obtained from 26 individuals approximately 20 years after accidental exposure to dioxin in Seveso, in 1976, were analyzed for differential gene expression associated with exposure and past chloracne status by Affymetrix GeneChip microarrays. Due to the long half-life of dioxin, current plasma TCDD levels were still substantially elevated in the exposed individuals. Overall, we observed a modest alteration of gene expression based on dioxin levels or on chloracne status, both in terms of numbers of genes altered and expression ratios. In the comparison between high levels C.M. McHale et al. / Toxicology 229 (2007) 101113 111 Table 6 Expression of genes associated with TCDD exposure and chloracne by real-time PCR and microarray analyses Real-time PCR Microarray Fold change p-Value Fold change p-Value Genes associated with exposure Up HIST1H3H 2.11 HBB 2.65 0.03* 0.04* 1.60 1.50 0.08 0.33 Down IL-8 CYP1B1 KLF4 MGST1 NFIL3 0.60 0.71 0.65 0.71 1.13 0.22 0.60 0.14 0.56 0.13 0.63 0.19 0.65 0.41 0.65 0.09 0.08 0.04* 0.06 0.03* Genes associated with chloracne Up HBB 2.81 CREM 1.21 GSTM3 2.52 TOB1 0.91 0.07 0.09 0.007* 0.53 2.68 1.48 1.37 1.4 0.02* 0.02* 0.0004* 0.04* *Indicates significant p-values. and background levels of TCDD, up-regulation of four histone genes was notable and modified expression of HIST1H3H was confirmed by real-time PCR. In the comparison between subjects who developed chloracne after the accident and those who did not develop the disease, up-regulation of five hemoglobin genes was noted and modified expression of HBB approached significance by real-time PCR. Involvement of these genes may offer new mechanistic insight into known effects of dioxin exposure on immune development and function (Kerkvliet, 2002; Luebke et al., 2006), carcinogenesis (Mandal, 2005; Nebert et al., 2000; Schwarz and Appel, 2005) and diabetes (Remillard and Bunce, 2002). In the pathway analysis, two IPA networks associated with dioxin exposure included genes involved in cellular growth and proliferation, glucose metabolism, cell death, DNA replication, recombination and repair, and cancer. A key network gene, KLF4, is a member of the kruppel family of transcription factors, which mediates p53-dependent G1/S cell cycle arrest in response to DNA damage (Yoon et al., 2003) and is a potential tumor suppressor gene in colorectal cancer (Zhao et al., 2004). KLF4 acts as a quiescence maintenance factor in B lymphocytes (Yusuf and Fruman, 2003). Reduced KLF4 expression in our study could represent dysregulation of B cell development or, as a potential tumor suppressor gene, could be a contributory factor in TCDD-associated hematological cancers (Bertazzi et al., 2001). In the Seveso population, plasma IgG levels were decreased with increasing TCDD plasma concentration in a study conducted 20 years after the exposure, showing long-term immunologic effects of dioxin (Baccarelli et al., 2002). Chloracne is a sensitive indicator of dioxin poisoning (Yamamoto and Tokura, 2003), but less than 0.1% of the subjects in the Seveso cohort were diagnosed with chloracne suggesting that susceptibility or environmental factors may have played a critical role (Pesatori et al., 2003) and indeed we observed chloracne-related changes in gene expression distinct from those associated with exposure. Chloracne manifests as a localized dermatological disease and dioxin toxicity in chloracne cases appeared to be confined to acute dermatotoxic effects (Baccarelli et al., 2005). However, we show changes in gene expression in PBMC of individuals with chloracne, suggesting more systemic involvement. The top-scoring network associated with chloracne is related to cell death, cellcell signaling and interaction, and immunological alterations. In particular, a protein encoded by the PTGER gene (which was up-regulated in association with chloracne) is one of four receptors identified for prostaglandin E2 (PGE2). Knockout studies in mice suggest that this receptor may be involved in the initiation of skin immune responses through stimulation of T cells (Narumiya, 2003). Expression of CD28, whose co-stimulation is essential for CD4-positive T cell proliferation, survival, and activation (Keir and Sharpe, 2005), was also up-regulated. Indeed, a number of genes found altered in the chloracne cases are involved in the regulation of T cell activity and function, and may in turn be involved in acne-associated inflammation (Jeremy et al., 2003; Trivedi et al., 2006). The identification of several PBMC genes with potential roles in the TCDDmediated pathogenesis of chloracne is interesting in light of a recent paper, which shows that the peripheral blood transcriptome dynamically reflects system wide biology (Liew et al., 2006). A second chloracne-associated network included five hemoglobin genes up-regulated in individuals who developed the disease following exposure. Several studies have investigated the association between TCDD exposure and hematologic and immune functions in exposed subjects (Baccarelli et al., 2002, 2005) with inconclusive results. TCDD levels were among the highest ever reported, and yet almost all clinical laboratory tests on these individuals were normal soon after the accident; any abnormal test result, including variation of hematocrit or CBC levels, was only transitory in nature (Needham et al., 1997). We did measure hematologic and immune parameters in the subjects at the moment of this study (almost 20 years after the accident); as expected, no 112 C.M. McHale et al. / Toxicology 229 (2007) 101113 significant association were found, with the exception of decreased IgG levels. These observations further suggest that the use of PBMC should not introduce a bias in the analysis of different groups. However, it is unclear what the changes in hemoglobin gene expression represent. Expression of two members of the glutathione Stransferase superfamily of oxidative stress response genes was also modified, with GSTMI (DeJong et al., 1991) and GSTM3 (Campbell et al., 1990) up-regulated by high TCDD in chloracne cases. These findings are in accord with previous studies in animal models, which showed a sustained oxidative stress response in mice following TCDD exposure (Shertzer et al., 1998), and another report on global gene expression (1152 genes) in waste incineration workers occupationally exposed to dioxin and PAHs (Kim et al., 2004), which showed upregulation of five genes related to oxidative stress. Even though the mean exposure between the chloracne case (74.3268 ppt) and control (25.290.2 ppt) groups was different, each group comprised almost equal numbers of high- and low-exposed individuals. Therefore, gene expression could reflect both presence of dioxin and susceptibility to chloracne, and we cannot disentangle these effects in the present study. The gene expression alterations described can provide important clues with regard to dioxin toxicity. However, overall there was a weak association (low ratios among the most significant genes) between gene expression and dioxin exposure or chloracne status. This may reflect an attenuated response to long-term human exposure (Steenland et al., 2004), which could not be adequately detected by microarray analysis. Moreover, even though TCDD is cleared slowly (half-life 7 years), age-related individual variations in metabolism and excretion could have occurred (Eskenazi et al., 2004), particularly in the lowest exposure group, which included slightly younger subjects. Finally, the power calculation used for study design was based on a ratio of expression of 2 or higher, while we found expression ratios below 2.0, and limitations in sample availability of suitable quality for microarray analysis precluded expanding the study. Thus, low statistical power to observe small differences may have played a role, despite selection of subjects with extreme levels of exposure, same sex, similar tobacco consumption and no differences between comparison's groups with regard to alcohol consumption, medication use or background plasma levels of 22 congeners present in the area and unrelated with the accident. This study highlights the challenges of examining global gene expression in human exposed populations. Few studies have been reported previously and it is now clear that smaller magnitude changes in expression may be typical in these studies (Forrest et al., 2005; Wu et al., 2003). Differences in baseline expression of genes among individuals, and/or differences in genotype and metabolic activation of chemicals may underlie some of the expected variability. Among the challenges presented by these data are development of appropriate statistical analysis to discern true differential expression, and development of accurate methods to independently confirm the findings as quantitative PCR may not reliably confirm expression changes less than two-fold. In conclusion, approximately 20 years after the accident, we have identified by microarray and confirmed by real-time PCR modest alteration in several genes potentially associated with dioxin exposure and chloracne status in a sample of subjects from the Seveso cohort. As in any such study, confirmation of the expression of these genes in a larger population is necessary to further elucidate their roles in the dioxin-induced perturbation of blood cell development, immunological response, and chloracne we have observed. Finally, our study underscores some of the technical and study design challenges inherent in these kinds of studies. Acknowledgements We are indebted to the study participants. We thank also Drs. Pier Alberto Bertazzi and Neil Caporaso for their help in designing and managing the study. References Baccarelli, A., Mocarelli, P., Patterson Jr., D.G., Bonzini, M., Pesatori, A.C., Caporaso, N., Landi, M.T., 2002. Immunologic effects of dioxin: new results from Seveso and comparison with other studies. Environ. Health Perspect. 110, 11691173. Baccarelli, A., Pesatori, A.C., Consonni, D., Mocarelli, P., Patterson Jr., D.G., Caporaso, N.E., Bertazzi, P.A., Landi, M.T., 2005. Health status and plasma dioxin levels in chloracne cases 20 years after the Seveso, Italy accident. Br. J. Dermatol. 152, 459465. Bertazzi, P.A., Consonni, D., Bachetti, S., Rubagotti, M., Baccarelli, A., Zocchetti, C., Pesatori, A.C., 2001. Health effects of dioxin exposure: a 20-year mortality study. Am. J. Epidemiol. 153, 10311044. Birnbaum, L.S., Tuomisto, J., 2000. Non-carcinogenic effects of TCDD in animals. Food Addit. Contam. 17, 275288. Boutros, P.C., Moffat, I.D., Franc, M.A., Tijet, N., Tuomisto, J., Pohjanvirta, R., Okey, A.B., 2004. Dioxin-responsive AHRE-II gene battery: identification by phylogenetic footprinting. Biochem. Biophys. Res. Commun. 321, 707715. Campbell, E., Takahashi, Y., Abramovitz, M., Peretz, M., Listowsky, I., 1990. A distinct human testis and brain mu-class glutathione S-transferase. Molecular cloning and characterization of a form present even in individuals lacking hepatic type mu isoenzymes. J. Biol. Chem. 265, 91889193. Challen, G., Gardiner, B., Caruana, G., Kostoulias, X., Martinez, G., Crowe, M., Taylor, D.F., Bertram, J., Little, M., Grimmond, S.M., C.M. McHale et al. / Toxicology 229 (2007) 101113 113 2005. Temporal and spatial transcriptional programs in murine kidney development. Physiol. Genomics 23, 159171. DeJong, J.L., Mohandas, T., Tu, C.P., 1991. The human Hb (mu) class glutathione S-transferases are encoded by a dispersed gene family. Biochem. Biophys. Res. Commun. 180, 1522. Eskenazi, B., Mocarelli, P., Warner, M., Needham, L., Patterson Jr., D.G., Samuels, S., Turner, W., Gerthoux, P.M., Brambilla, P., 2004. Relationship of serum TCDD concentrations and age at exposure of female residents of Seveso, Italy. Environ. Health Perspect. 112, 2227. Forrest, M.S., Lan, Q., Hubbard, A.E., Zhang, L., Vermeulen, R., Zhao, X., Li, G., Wu, Y.Y., Shen, M., Yin, S., Chanock, S.J., Rothman, N., Smith, M.T., 2005. Discovery of novel biomarkers by microarray analysis of peripheral blood mononuclear cell gene expression in benzene-exposed workers. Environ. Health Perspect. 113, 801807. IARC, 1997. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans: Polychlorinated Dibenzo-para-Dioxins and Polychlorinated Dibenzofurans. Lyon, France, February 411, 1997. IARC Monogr. Eval. Carcin. Risks Hum. 69, 1 631. Irizarry, R.A., Hobbs, B., Collin, F., Beazer-Barclay, Y.D., Antonellis, K.J., Scherf, U., Speed, T.P., 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249264. Jeremy, A.H., Holland, D.B., Roberts, S.G., Thomson, K.F., Cunliffe, W.J., 2003. Inflammatory events are involved in acne lesion initiation. J. Invest. Dermatol. 121, 2027. Keir, M.E., Sharpe, A.H., 2005. The B7/CD28 costimulatory family in autoimmunity. Immunol. Rev. 204, 128143. Kerkvliet, N.I., 2002. Recent advances in understanding the mechanisms of TCDD immunotoxicity. Int. Immunopharmacol. 2, 277291. Kim, M.K., Oh, S., Lee, J.H., Im, H., Ryu, Y.M., Oh, E., Lee, J., Lee, E., Sul, D., 2004. Evaluation of biological monitoring markers using genomic and proteomic analysis for automobile emission inspectors and waste incinerating workers exposed to polycyclic aromatic hydrocarbons or 2,3,7,8-tetracholrodedibenzo-p-dioxins. Exp. Mol. Med. 36, 396410. Landi, M.T., Needham, L.L., Lucier, G., Mocarelli, P., Bertazzi, P.A., Caporaso, N., 1997. Concentrations of dioxin 20 years after Seveso. Lancet 349, 1811. Landi, M.T., Consonni, D., Patterson Jr., D.G., Needham, L.L., Lucier, G., Brambilla, P., Cazzaniga, M.A., Mocarelli, P., Pesatori, A.C., Bertazzi, P.A., Caporaso, N.E., 1998. 2,3,7,8-Tetrachlorodibenzop-dioxin plasma levels in Seveso 20 years after the accident. Environ. Health Perspect. 106, 273277. Landi, M.T., Bertazzi, P.A., Baccarelli, A., Consonni, D., Masten, S., Lucier, G., Mocarelli, P., Needham, L., Caporaso, N., Grassman, J., 2003. TCDD-mediated alterations in the AhR-dependent pathway in Seveso, Italy, 20 years after the accident. Carcinogenesis 24, 673680. Liew, C.C., Ma, J., Tang, H.C., Zheng, R., Dempsey, A.A., 2006. The peripheral blood transcriptome dynamically reflects system wide biology: a potential diagnostic tool. J. Lab. Clin. Med. 147, 126132. Luebke, R.W., Chen, D.H., Dietert, R., Yang, Y., King, M., Luster, M.I., 2006. The comparative immunotoxicity of five selected compounds following developmental or adult exposure. J. Toxicol. Environ. Health B: Crit. Rev. 9, 126. Mandal, P.K., 2005. Dioxin: a review of its environmental effects and its aryl hydrocarbon receptor biology. J. Comp. Physiol. [B] 175, 221230. Mimura, J., Fujii-Kuriyama, Y., 2003. Functional role of AhR in the expression of toxic effects by TCDD. Biochim. Biophys. Acta 1619, 263268. Narumiya, S., 2003. Prostanoids in immunity: roles revealed by mice deficient in their receptors. Life Sci. 74., 391395. Nebert, D.W., Roe, A.L., Dieter, M.Z., Solis, W.A., Yang, Y., Dalton, T.P., 2000. Role of the aromatic hydrocarbon receptor and [Ah] gene battery in the oxidative stress response, cell cycle control, and apoptosis. Biochem. Pharmacol. 59, 6585. Needham, L.L., Gerthoux, P.M., Patterson Jr., D.G., Brambilla, P., Turner, W.E., Beretta, C., Pirkle, J.L., Colombo, L., Sampson, E.J., Tramacere, P.L., Signorini, S., Meazza, L., Carreri, V., Jackson, R.J., Mocarelli, P., 1997. Serum dioxin levels in Seveso, Italy, population in 1976. Teratogen. Carcin. Mutagen. 17, 225240. Pesatori, A.C., Consonni, D., Bachetti, S., Zocchetti, C., Bonzini, M., Baccarelli, A., Bertazzi, P.A., 2003. Short- and long-term morbidity and mortality in the population exposed to dioxin after the "Seveso accident". Ind. Health 41, 127138. Remillard, R.B., Bunce, N.J., 2002. Linking dioxins to diabetes: epidemiology and biologic plausibility. Environ. Health Perspect. 110, 853858. Schwarz, M., Appel, K.E., 2005. Carcinogenic risks of dioxin: mechanistic considerations. Regul. Toxicol. Pharmacol. 43, 1934. Shertzer, H.G., Nebert, D.W., Puga, A., Ary, M., Sonntag, D., Dixon, K., Robinson, L.J., Cianciolo, E., Dalton, T.P., 1998. Dioxin causes a sustained oxidative stress response in the mouse. Biochem. Biophys. Res. Commun. 253, 4448. Sogawa, K., Numayama-Tsuruta, K., Takahashi, T., Matsushita, N., Miura, C., Nikawa, J., Gotoh, O., Kikuchi, Y., Fujii-Kuriyama, Y., 2004. A novel induction mechanism of the rat CYP1A2 gene mediated by Ah receptorArnt heterodimer. Biochem. Biophys. Res. Commun. 318, 746755. Steenland, K., Bertazzi, P., Baccarelli, A., Kogevinas, M., 2004. Dioxin revisited: developments since the 1997 IARC classification of dioxin as a human carcinogen. Environ. Health Perspect. 112, 12651268. Trivedi, N.R., Gilliland, K.L., Zhao, W., Liu, W., Thiboutot, D.M., 2006. Gene array expression profiling in acne lesions reveals marked upregulation of genes involved in inflammation and matrix remodeling. J. Invest. Dermatol.. Wu, M.M., Chiou, H.Y., Ho, I.C., Chen, C.J., Lee, T.C., 2003. Gene expression of inflammatory molecules in circulating lymphocytes from arsenic-exposed human subjects. Environ. Health Perspect. 111, 14291438. Yamamoto, O., Tokura, Y., 2003. Photocontact dermatitis and chloracne: two major occupational and environmental skin diseases induced by different actions of halogenated chemicals. J. Dermatol. Sci. 32, 8594. Yoon, H.S., Chen, X., Yang, V.W., 2003. Kruppel-like factor 4 mediates p53-dependent G1/S cell cycle arrest in response to DNA damage. J. Biol. Chem. 278, 21012105. Yusuf, I., Fruman, D.A., 2003. Regulation of quiescence in lymphocytes. Trends Immunol. 24, 380386. Zhao, W., Hisamuddin, I.M., Nandan, M.O., Babbin, B.A., Lamb, N.E., Yang, V.W., 2004. Identification of Kruppel-like factor 4 as a potential tumor suppressor gene in colorectal cancer. Oncogene 23, 395402.