Microarray Analysis of Bleomycin-Exposed Lymphoblastoid Cells for Identifying Cancer Susceptibility Genes

Jacqueline Cloos,1,2 Wim P.H. de Boer,3 Mireille H.J. Snel,1 Paul van den IJssel,4 Bauke Ylstra,4 C. Rene´Leemans,1 Ruud H. Brakenhoff,1 and Boudewijn J.M. Braakhuis1

1Section Tumor Biology, Department of Otolaryngology/Head-Neck Surgery, 2Department of Pediatric Oncology/Hematology, 3Department of Medical Oncology, and 4Microarray Core Facility, VU University Medical Center, Amsterdam, the Netherlands

Abstract Introduction The uncovering of genes involved in susceptibility to The development of head and neck squamous cell carcinoma the sporadic cancer types is a great challenge. It is well (HNSCC) is strongly associated with excessive tobacco established that the way in which an individual deals smoking and alcohol intake (1). During the last decade, much with DNA damage is related to the chance to develop emphasis has been laid on the genetic factors that act in concert cancer. Mutagen sensitivity is a phenotype that reflects with environmental features to determine an individual’s risk an individual’s susceptibility to the major sporadic for developing HNSCC. These so-called molecular epidemiol- cancer types, including colon, lung, and head and ogy studies are based on the knowledge that there is a variation neck cancer. A standard test for mutagen sensitivity in the population with respect to the capability of cells to is measuring the number of chromatid breaks in deal with DNA damage. This variation has been addressed by lymphocytes after exposure to bleomycin. The aim of analyzing polymorphisms in genes involved in DNA damage the present study was to search for the pathways processing or detoxification pathways (2). In addition, involved in mutagen sensitivity. Lymphoblastoid cell functional tests have been developed that measure how cells lines of seven individuals with low mutagen sensitivity respond to induced DNA damage. Of these latter tests, the were compared with seven individuals with a ‘‘mutagen sensitivity test’’ (3) has particularly often been used. high score. RNA was isolated from cells exposed to Mutagen sensitivity is determined in peripheral blood lympho- bleomycin (4 hours) and from unexposed cells. cytes as the mean number of chromatid breaks per cell (b/c) Microarray analysis (19K) was used to compare gene at metaphase induced by bleomycin exposure in the late S-G2 expression of insensitive and sensitive cells. The phase of the cell cycle. A high mutagen-sensitive phenotype profile of most altered genes after bleomycin exposure, (having more than a mean of 1.0 b/c) was found at higher analyzed in all 14 cell lines, included relatively many frequency in patients not only with HNSCC but also with lung genes involved in biological processes, such as cell and colon cancer compared with the control population without growth and/or maintenance, proliferation, and regulation cancer (4-9). Mutagen sensitivity is not influenced by gender, of cell cycle, as well as some genes involved in DNA tumor stage, or smoking and alcohol intake of the subjects and repair. When comparing the insensitive and sensitive only slightly increases with age (4, 10, 11); it is considered to individuals, other differentially expressed genes were be an intrinsic factor, reflecting how an individual copes with found that are involved in signal transduction and DNA damage (12). In combination with carcinogenic exposure cell growth and/or maintenance (e.g., BUB1 and DUSP4). (smoking and alcohol intake), a mutagen hypersensitivity This difference in expression profiles between phenotype is associated with a greatly increased risk of mutagen-sensitive and mutagen-insensitive individuals developing HNSCC (13). Moreover, it was shown in a justifies further studies aimed at elucidating the genes prospective patient study that mutagen hypersensitivity is responsible for the development of sporadic cancers. related to the development of second primary tumors (5). The (Mol Cancer Res 2006;4(2):71–7) findings of all mutagen sensitivity studies support the notion that a common genetic susceptibility to DNA damage, and thereby a susceptibility to cancer, exists in the general Received 10/6/05; revised 12/21/05; accepted 12/21/05. population (9). In a twin study, we showed that mutagen The costs of publication of this article were defrayed in part by the payment of sensitivity has a high heritability estimate, indicating a clear page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. genetic basis (12). It is interesting to note that, in line with our Note: Supplementary data for this article are available at Molecular Cancer finding, a Mendelian inheritance pattern of radiation-induced Research Online (http://mcr.aacrjournals.org/). Requests for reprints: Boudewijn J.M. Braakhuis, Section Tumor Biology, b/c score was found when investigating cancer-prone families Department of Otolaryngology/Head-Neck Surgery, VU University Medical (14), again indicative for a clear genetic basis of mutagen Center, P.O. Box 7057, 1007 MB Amsterdam, the Netherlands. Phone: 31-20- sensitivity. In that study, it was calculated that mutagen 444-0905; Fax: 31-20-444-3688. E-mail: [email protected] Copyright D 2006 American Association for Cancer Research. sensitivity may be explained by one or at the most two genes. doi:10.1158/1541-7786.MCR-05-0196 The question arose whether bleomycin-induced chromatid

break sensitivity is linked with cancer predisposition in the cell proliferation, and regulation of cell cycle were the most same way as radiation-induced sensitivity. Although bleomycin prominently altered processes (P < 0.001 in all cases, Fisher’s is called a ‘‘radiomimetic’’ agent, it differs from radiation in the exact test) with 96, 41, and 20 of the 206 genes involved, way the damage is induced (15). Despite the differences, a good respectively. Two genes involved in DNA repair, POLH and correlation was found when bleomycin- and radiation-induced XPC, were found to have a significantly increased expression b/c values were compared in the mutagen sensitivity test of the as a result of the exposure to bleomycin. same cells (16), suggesting that a kind of similar mechanism is underlying these damage-sensitive phenotypes. Mutagen Sensitivity-Related Genes The aim of the current study was to reveal what pathway(s) is Two groups of seven lymphoblastoid cell lines were selected, involved in the mutagen-sensitive phenotype to facilitate the one group with a relatively high level of bleomycin-induced understanding of susceptibility to several of the most common chromatid breaks and one with a low level of such breaks. cancer types. There is some indication that cell cycle regulation Table 2 shows that the insensitive cell lines had an average is aberrant in mutagen-sensitive individuals (17). However, mutagen sensitivity score of 0.6 b/c, whereas this score for the specific information concerning which pathway and which sensitive group reached the value of 1.04 b/c (significantly molecules are involved is lacking. In the current study, we have higher; P = 0.00145, Student’s t test). The microarray used an integrative genomic approach to screen for possible expression analysis was done on samples that were untreated pathways by comparing gene expression of cells from individ- and on those exposed to bleomycin for 4 hours, and for each uals with a high and a low mutagen sensitivity by mRNA gene, the expression ratios were compared. For 101 genes with a expression microarray analysis. There is evidence that mutagen known function, a statistically significant different expression sensitivity has a genetic basis determined by one or at the most ratio was determined (P < 0.02): 46 genes with higher and two genes and that it is functionally related to cell cycle control. 55 genes with lower expression in the sensitive group Therefore, we hypothesized that mutagen sensitivity would be (Supplementary Information 2). The 4- over 0-hour Z ratios reflected in a specific pattern of mRNA expression that may give are shown in Supplementary Information 3. Cluster analysis of the opportunity to reveal the responsible pathways. Another these ratios (Fig. 1) showed that (a) the sensitive group is clearly intention of our study was to examine global bleomycin-induced separated from the insensitive group and (b) each cell line has its expression profile changes that should be comparable with unique pattern of expression across the genes. The genes that already published profiles of genes that are differentially most significantly differed between insensitive and sensitive expressed after exposure to radiation. groups, with a P < 0.01, are shown in Table 3: 13 genes with higher and 24 genes with lower expression in the sensitive group. Based on the analysis with Expression Analysis Results Systematic Explorer, some specific cellular processes (catego- Expression Arraying rized according to the Gene Ontology database) were found to By using spotted oligonucleotide arrays, we hybridized 28 be more represented in this group of 101 genes compared with samples (14 untreated and 14 after 4-hour bleomycin exposure) the total human genome. Genes involved in signal transduction against our common reference sample and found expression in and cell growth and/or maintenance were most prominent 80% of the 19K genes. The 4-hour time point was chosen (P < 0.01, Fisher’s exact test). because preliminary results with filter arrays have shown that it was optimal for showing differences in expression levels due to bleomycin exposure. Longer incubation periods were not Discussion considered, because that may result in effects not related to Our quest was to identify a pathway involved in cancer break induction, which is typically measured in the mutagen susceptibility. We approached this by analyzing the expression sensitivity assay after 2-hour bleomycin exposure. of genes involved in mutagen sensitivity. The rationale for this study was based on the notion that mutagen sensitivity is a Bleomycin Exposure-Related Genes well-established marker of cancer susceptibility (9). The 14 lymphoblastoid cell lines were exposed to bleomycin Bleomycin was able to induce clear changes in the for 4 hours and the change in mRNA expression was measured expression profile of all cell lines. Differently expressed genes by microarray analysis. Taking P < 0.01 as the criterion of were found to belong to a considerable number of cellular significance, 204 genes with a known function were found to processes. Many of them have been reported to be involved in have a treatment-induced alteration of the expression level of control of chromosomal damage and some were similar to those mRNA. An increase for 53 genes and a decrease for 151 genes that have been described in relation to ionizing radiation– in expression were observed (Supplementary Information 1). induced damage (18, 19). With respect to up-regulation, three The most significant genes (P V 0.002) are shown in Table 1: genes draw attention. First, CDKN1A, the gene encoding for 16 genes with an increased expression and 36 genes with a p21, is a downstream effector of TP53 and an important decreased expression as a result of bleomycin exposure. Table 1 regulator in cell cycle progression (20). In line with what is also depicts the type of molecular function and cellular process expected, expression of this gene is increased by bleomycin regarding the gene product. A wide array of Gene Ontology treatment to induce cell cycle arrest. The expression of this gene processes were found to be overrepresented in the panel of is also increased after ionizing radiation (19). Second, XPC and significantly different genes as determined with Expression POLH are two genes involved in DNA repair, particularly in Analysis Systematic Explorer. Cell growth and/or maintenance, nucleotide excision repair (21). XPC forms a complex with

Table 1. Genes with Differential Expression as a Result of Exposure to Bleomycin Analyzed in 14 Lymphoblastoid Cell Lines

Gene Description Molecular Function Biological Process

Increased expression after bleomycin CAMK1 Calcium/calmodulin-dependent protein kinase I Protein serine/threonine kinase activity Signal transduction CDKN1A Cyclin-dependent kinase inhibitor 1A (p21, Cip1) Cyclin-dependent protein Cell cycle arrest kinase inhibitor activity XPC Xeroderma pigmentosum, complementation group C Damaged DNA binding Nucleotide excision repair H2AFZ H2A histone family, member Z Histone Unknown POLH Polymerase (DNA directed), D Damaged DNA binding DNA repair regulation FDXR Ferredoxin reductase Ferredoxin-NADP(+) reductase Electron transport RGS16 Regulator of G-protein signaling 16 GTPase activator Signal transduction FRDA Frataxin Inositol/phosphatidylinositol kinase Electron transport TNFSF7 Tumor necrosis factor (ligand) superfamily, member 7 Tumor necrosis factor receptor binding Apoptosis TNFSF9 Tumor necrosis factor (ligand) superfamily, member 9 Tumor necrosis factor receptor binding Apoptosis UBR1 Ubiquitin protein ligase E3 component n-recognin 1 Ligase activity Ubiquitin cycle BBC3 (PUMA) BCL2-binding component 3 Protein binding Apoptosis CD37 CD37 antigen Unknown N-linked glycosylation ANKFY1 Ankyrin repeat and FYVE domain containing 1 Protein binding Endocytosis CX3CL1 Chemokine (C-C motif) ligand 22 Chemokine activity Immune response PPM1D Protein phosphatase 1D magnesium-dependent, y isoform Catalytic activity Cell proliferation, negative regulation of

Decreased expression after bleomycin PIP5K1A Phosphatidylinositol-4-phosphate 5-kinase, type I, a 1-Phosphatidylinositol-4-phosphate kinase Signal transduction ACSL1 Acyl-CoA synthetase long-chain family member 1 Ligase activity Lipid metabolism TPX2 TPX2, microtubule-associated protein ATP binding Cell proliferation homologue (Xenopus laevis) BST2 Bone marrow stromal cell antigen 2 Unknown Cell proliferation PPP3CB Protein phosphatase 3 (formerly 2B), Calcium ion binding Cell cycle regulation catalytic subunit CSNK2A1 Casein kinase 2, a1 polypeptide Protein serine/threonine kinase activity Amino acid phosphorylation CD53 CD53 antigen Unknown Signal transduction CENPA Centromere protein A, 17 kDa DNA binding Chromosome organization and biogenesis CCNB2 Cyclin B2 Unknown Cell cycle regulation ZNF410 Zinc finger protein 410 DNA binding Transcription regulation DCTN2 Dynactin 2 (p50) Motor activity Cell proliferation TLP19 Endoplasmic reticulum thioredoxin superfamily Electron transporter activity Electron transport member, 18 kDa LTA4H Leukotriene A4 hydrolase Epoxide hydrolase activity Inflammatory response CCNA2 Cyclin A2 Unknown Cytokinesis TUBA6 Tubulin a6 GTP binding Microtubule organization CECR5 Cat eye syndrome chromosome region, candidate 5 Hydrolase activity Metabolism FKBP4 FK506 binding protein 4, 59 kDa Isomerase activity Protein processing KNTC2 Kinetochore associated 2 Unknown Mitotic sister chromatid segregation LAPTM4A Lysosomal-associated protein transmembrane 4a Unknown Transport TM9SF1 Transmembrane 9 superfamily member 1 Unknown Transport MORF4L2 Mortality factor 4 like 2 Unknown Cell growth regulation ELAVL1 ELAV (embryonic lethal, mRNA binding, 3V untranslated region RNA catabolism abnormal vision, Drosophila) – like 1 MRPS17 Mitochondrial ribosomal protein S17 Nucleotide binding Protein processing PLK1 Polo-like kinase 1 (Drosophila) Transferase activity Mitosis CDC25B Cell division cycle 25B Hydrolase activity Cell cycle RAC1 Ras-related C3 botulinum toxin substrate 1 GTP binding Protein transport SSR2 Signal sequence receptor, h Signal sequence receptor Cotranslational protein membrane targeting (translocon-associated protein h) TM9SF2 Transmembrane 9 superfamily member 2 Transporter activity Transport SF1 Splicing factor 1 RNA binding Transcription regulation SIAH2 Seven in absentia homologue 2 (Drosophila) Ligase activity Apoptosis XBP1 X-box binding protein 1 Transcription factor activity Immune response EIF4G2 Eukaryotic translation initiation factor 4g,2 RNA binding Cell cycle arrest ADAMTS9 A disintegrin-like and metalloprotease (reprolysin type) Metalloendopeptidase activity Glycoprotein degradation NEDD8 Neural precursor cell expressed, developmentally Ubiquitin-conjugating enzyme activity Proteolysis and peptidolysis down-regulated 8 UBE2G1 Ubiquitin-conjugating enzyme E2G 1 Ubiquitin-conjugating enzyme activity Ubiquitin cycle (UBC7 homologue, Caenorhabditis elegans) TIMM10 Translocase of inner mitochondrial membrane Unknown Protein transport 10 homologue (yeast)

NOTE: The 52 most significantly different genes are shown (P V 0.002, two-sided tested).

HR23B and this complex is the main early damage detector expression of certain genes known to be involved in various (22). POLH (also known as XPV) is involved in high-fidelity processes, including cell cycle control and cell proliferation DNA replication in case of large adducts (23). Mutations in (among others, two cyclins). In general, the picture emerges XPC and POLH have been described in patients suffering from that bleomycin induces the alteration of cellular processes, xeroderma pigmentosum, a disorder of nucleotide excision resulting in a retardation of the cell cycle and an induction of repair (21). Bleomycin exposure also resulted in decreased DNA repair analogous to radiation.

These results indicated the validity of our test system and we responsible for the difference in mutagen sensitivity. This can be further analyzed the data to identify the genes that are responsible explained in several ways. It is possible that various relatively for differences in mutagen sensitivity. We approached this by sensitive individuals differ with respect to the pathway that is comparing bleomycin-induced genes between highly mutagen- aberrant. Mixing those expression profiles within one group sensitive and mutagen-insensitive cell lines. Interestingly, there dilutes the effect and makes it difficult to measure. Nevertheless, was no difference between mutagen-sensitive and mutagen- the present approach could still be successful in case of common insensitive cells with respect to the up-regulation of expression downstream effectors likely involved in double strand break of CDKNIA and the DNA repair genes, XPC and POLH. repair. The reason why the present study did not identify such Nevertheless, we observed a most pronounced difference in a common effect may be related to the limited number of expression of genes involved in signal transduction and cell individuals that was investigated in each experimental group, proliferation, such as, for instance, BUB1 and DUSP4. The although large effects should have been detected in this BUB1 gene encodes a kinase involved in spindle checkpoint approach. On the other hand, analyzing mRNA expression has function. Mutations in this gene have been associated with an intrinsic limitation, as a cell may respond to stress by post- aneuploidy and several forms of cancer (24). DUSP4 encodes a transcriptional or post-translational mechanisms. This indicates dual-specificity phosphatase of the activated mitogen-activated that further research is warranted to find possible differences in protein kinases ERK1 and ERK2, which play an essential role in protein profiles of cells varying in the level of mutagen mitogen-regulated growth factor signal transduction (25). These sensitivity. findings seem to be in line with our earlier findings that the main The findings of this study elucidated part of the network of functional difference between insensitive and sensitive cells is genes that are involved in the response to bleomycin and play a the lack of cell cycle arrest in relatively sensitive cells (17). Only potential role in cancer susceptibility. The observed difference in one gene, which we found to be significantly different, was expression of several genes between mutagen-sensitive and related to DNA repair (PNKP). PNKP gene function was shown mutagen-insensitive individuals justifies further studies aiming to restore termini suitable for DNA polymerase, consistent with to elucidate the genes responsible for the development of in vivo removal of 3Vphosphate groups, facilitating DNA repair sporadic cancers. (26). The PNKP protein was shown to have a direct, specific role in double strand break repair within the context of the Materials and Methods nonhomologous end joining apparatus (27). The mRNA Cell Lines expression of PNKP was down-regulated in the sensitive cells, Mechanisms underlying mutagen sensitivity were studied in which may indicate that DNA polymerase activity is not per se lymphoblastoid cell lines that are generated by immortalization related to the variation in the level of mutagen sensitivity but that of blood lymphocytes with EBV. Cell lines made in this way the explanation should be sought in the facilitation of repair of show a very good concordance in mutagen sensitivity score already synthesized DNA. with the original lymphocytes (17) and are proposed to reflect We made an effort to decrease the possible influence of the capacity of an individual to deal with bleomycin-induced the individual heterogeneity by calculating the ratio of the DNA damage (11). Our laboratory has prepared various expression data of treated and untreated cells of each individual lymphoblastoid cell lines from HNSCC patients and noncancer to let each individual serve as his or her own control. individuals according to a previously published protocol (17). Nevertheless, no specific pathway could be singled out as being Fourteen lymphoblastoid cell lines were selected based on the

Table 2. Characteristics of Lymphoblastoid Cell Lines and Their Donors

Cell Line Donor HNSCC Patient Family Member of HNSCC Patients No. Breaks per Cell* Age Gender

Insensitive 1 No No 0.68 29 F 2 No No 0.60 46 F 3 No No 0.70 39 F 4 No Yes 0.57 14 F 5 Yes No 0.47 59 M 6 No No 0.68 40 F 7 No No 0.51 23 M Average 0.60 36 Very sensitive 8 Yes No 1.28 79 M 9 Yes No 1.21 69 M 10 No Yes 0.82 64 M 11 No Yes 0.86 61 M 12 No No 1.00 36 M 13 Yes No 0.93 66 F 14 Yes No 1.15 67 M Average 1.04 63

NOTE: The two test groups are shown, one with high and one with low breaks per cell score. These groups were significantly different with respect to the breaks per cell score (P = 0.00145, Student’s t test). *This is the average value of three independent experiments. The coefficient of variation is <25% for all cell lines.

FIGURE 1. Clus- stering analysis of the 101 genes that were differently expressed between sensitive and insensitive per- sons. The 4- over 0- hour Z ratios are shown. The sensitive group clearly forms a cluster separate from the insensitive group, and each cell line has its unique pattern of expression across the genes. mutagen sensitivity score: a group of 7 highly sensitive cell 0.1 Ag/mL. Cells were pelleted, resuspended in 5 mL of 0.6 lines and a group of 7 relatively insensitive cell lines (Table 2). mol/L KCl, and incubated for 12 minutes at room temperature. The cells were cultured in RPMI containing 15% FCS Fixative (2 mL; 3:1 methanol/acetic acid) was added to the (BioWhittaker, Vervier, Belgium), 1% penicillin/streptomycin suspension and carefully mixed. The cells were then washed (Invitrogen, Breda, the Netherlands), and 0.1% of 1 mol/L twice in fixative before dropping them on wet slides. After air pyruvic acid (Sigma, St. Louis, MO). One day before testing the drying, the slides were stained in 5% Giemsa solution, coded, effect of bleomycin, trypan blue–excluding (viable) cells were and evaluated for breaks at a Â1,250 magnification (16). Each counted in a hematocytometer, centrifuged, and resuspended to cell line was tested several times (Table 2). a concentration of 4 Â 105 cells/mL. mRNA Microarray Expression Analysis Mutagen Sensitivity Test The experimental procedures and data are available at http:// This procedure was done as has been published earlier (12). www.ncbi.nlm.nih.gov/geo/ according to the Minimum Infor- In short, the cell culture was checked for an approximate mation About a Microarray Experiment standards (accession concentration of 4 Â 105 cells/mL. The cells were exposed to 5 code no. GSE3598; supplementary data can also be found at milliunits/mL bleomycin for 2 hours and at the last hour also in this location). Cells were exposed to 5 milliunits/mL bleomycin the presence of colcemid (Sigma) at a final concentration of for 4 hours. Untreated cells of each cell line were also harvested

at the same time. Pellets of 50 Â 106 cells were washed with Dye, Amersham, Freiburg, Germany) or Cy5 (Fluorolink Cy5) PBS, snap frozen in liquid N2, and stored at À80jC. RNA for the common reference. Samples were hybridized was isolated using RNAzol according to the manufacturer’s at 37jC for 14 hours and slides scanned as described protocol (Campro Scientific BV, Veenendaal, the Netherlands). previously (28). ImaGene (Biodiscovery Ltd., Marina del Quality control of total RNA samples was done with the RNA Rey, CA) feature extraction was used to record spot intensities. 6000 Pico LabChip kit (Agilent Technologies, Palo Alto, CA) The signal mean was taken to represent the actual spot and analyzed on the Agilent 2100 bioanalyzer. High-quality intensity after subtraction of the mean background value. The RNA should show clearly visible 18S/28S rRNA peaks, and expression platform we used has been described previously in the 28S peak was not allowed to be lower than the 18S peak. full detail (28), and these authors show a good correlation As a common reference to be used on each slide, we made between array and Taqman results regarding the level of a large stock of RNA from normal keratinocytes, several expression of several genes. cancer cell lines, and lymphoblastoid cell lines either exposed or unexposed to bleomycin. Statistical Analysis The Human Release 1.0 oligonucleotide library containing First, all spot intensities of all arrays were log2 transformed, 18,861 60-mer oligonucleotides representing 17,260 unique and intensities lower than 10, typical for the empty spots, genes as designed by Compugen (San Jose, CA) was obtained were classified as missing. Next, for each microarray, the spot from Sigma-Genosys (Zwijndrecht, the Netherlands). Cell intensities were Z score normalized (29) to achieve data samples were labeled with Cy3 (Fluorolink Cy3 Monofunctional standardization across the 14 cell lines. Missing values (some

Table 3. Genes with a Differential Expression between Mutagen-Sensitive and Mutagen-Insensitive Lymphoblastoid Cell Lines

Gene Description Molecular Function Biological Process

Higher expression in sensitive lines CDCA8 Cell division cycle associated 8 Unknown Cytokinesis PPAT Phosphoribosyl pyrophosphate amidotransferase Magnesium ion binding Nucleoside metabolism APRIN Androgen-induced proliferation inhibitor DNA binding Cell proliferation, negative regulation of KCNJ14 Potassium inwardly-rectifying channel, Inward rectifier potassium channel Ion transport subfamily J, member 14 BUB1 Budding uninhibited by benzimidazoles 1 ATP binding Cell cycle homologue (yeast) MYO7A Myosin VIIA ATP binding Perception EBNA1BP2 EBNA1 binding protein 2 Unknown Ribosome biogenesis DUSP4 Dual-specificity phosphatase 4 Hydrolase activity Cell cycle regulation GNAT1 Guanine nucleotide binding protein (G protein) GTP binding Signal transduction PHOX2B Paired-like homeobox 2b Transcription factor activity Regulation of transcription TRIM29 Tripartite motif containing 29 Transcription factor activity Transcription from RNA polymerase II promoter CRSP6 Cofactor required for Sp1 transcriptional activation, RNA polymerase II transcription Androgen receptor signaling pathway subunit 6, 77 kDa mediator activity USP47 Ubiquitin-specific protease 47 Ubiquitin thiolesterase activity Ubiquitin cycle

Lower expression in sensitive lines PRPSAP1 Phosphoribosyl pyrophosphate Nucleotide biosynthesis synthetase-associated protein 1 RASA2 RAS p21 protein activator 2 Ras GTPase activator activity Intracellular signaling cascade RBP1 Retinol binding protein 1, cellular Retinol binding Vitamin A metabolism CD36 CD36 antigen (collagen type I receptor, Protein binding Cell adhesion thrombospondin receptor) DLG5 Discs, large homologue 5 (Drosophila) Protein binding Cell proliferation, negative regulation of VSNL1 Visinin-like 1 Calcium ion binding Unknown NEFL Neurofilament, light polypeptide, 68 kDa Protein binding Unknown BAI3 Brain-specific angiogenesis inhibitor 3 Receptor activity Signal transduction PNKP Polynucleotide kinase 3V-phosphatase Hydrolase activity DNA repair LSM4 LSM4 homologue RNA binding RNA splicing SIX4 Sine oculis homeobox homologue 4 Transcription factor activity Transcription regulation PSEN2 Presenilin 2 (Alzheimer disease 4) Protein binding Apoptotic program CDK11 Cell division cycle 2-like 6 (CDK8-like) Protein tyrosine kinase activity Protein amino acid phosphorylation PLG Plasminogen Peptidase activity Blood coagulation TNFSF9 Tumor necrosis factor (ligand) superfamily, Tumor necrosis factor receptor binding Apoptosis member 9 SAS Sarcoma amplified sequence Unknown Cell proliferation, positive regulation of TCF19 Transcription factor 19 (SC1) Transcription factor activity Cell proliferation CDY1 Chromodomain protein, Y-linked, 1 Chromatin binding Spermatogenesis SETMAR SET domain and mariner transposase fusion gene Zinc ion binding DNA transposition TFF2 Trefoil factor 2 (spasmolytic protein 1) Unknown Digestion RHOB Ras homologue gene family, member B Protein binding Cell cycle, negative regulation of OR5I1 Olfactory receptor, family 5, subfamily I, member 1 Receptor activity Signal transduction CREB5 Cyclic AMP – responsive element binding protein 5 DNA binding Transcription, positive regulation of RELB V-rel reticuloendotheliosis viral oncogene homologue B Protein binding Transcription regulation

NOTE: The 37 most significantly different genes are shown (P < 0.01, two-sided tested).

6% overall) were set equal to the mean of the microarray biomarker for second primary tumors after head and neck squamous cell distribution after Z score normalization, thus equal to zero. carcinoma. Cancer Epidemiol Biomarkers Prev 2000;9:713 – 7. 6. Zhang ZF, Morgenstern H, Spitz MR, et al. Environmental tobacco smoking, Therefore, for each microarray, Z scores were computed for mutagen sensitivity, and head and neck squamous cell carcinoma. Cancer each green (Cy3)–labeled spot intensity according to the Epidemiol Biomarkers Prev 2000;9:1043 – 9. 2 formula: Z score = [log (intensity) À l]/r, where l is the 7. Wu X, Lippman SM, Wei Q, et al. Mutagen sensitivity and risk of oral mean and r is the SD of all log2-transformed green-labeled spot premalignant lesions (OPL). Cancer Epidemiol Biomarkers Prev 2001;10:156. intensities of the microarray. Likewise, Z scores were computed 8. Zheng YL, Loffredo CA, Yu ZP, et al. Bleomycin-induced chromosome breaks as a risk marker for lung cancer: a case-control study with population and hospital for each red (Cy5)–labeled spot intensity of the microarray and controls. Carcinogenesis 2003;24:269 – 74. then subtracted from the ‘‘green’’ Z scores for each spot on 9. Au WW. Mutagen sensitivity assays in population studies. Mutat Res 2003; the microarray. These differences were then again Z score 544:273 – 7. normalized to obtain the so-called Z ratio for each spot: a 10. Cloos J, Steen I, Joenje H, et al. Association between bleomycin genotoxicity standardized expression value of the gene in question relative to and nonconstitutional risk-factors for head and neck-cancer. Cancer Lett 1993; 74:161 – 5. the common reference. 11. Cloos J, Reid CB, van der Sterre ML, et al. A comparison of bleomycin- Two types of transcriptome comparisons were made. First, induced damage in lymphocytes and primary oral fibroblasts and keratinocytes in from all 14 cell lines, the expression profiles of unexposed 30 subjects. Mutagenesis 1999;14:87 – 93. cells were compared with those of bleomycin-exposed cells 12. Cloos J, Nieuwenhuis EJ, Boomsma DI, et al. Inherited susceptibility to (‘‘bleomycin effect’’). Second, the response to bleomycin of bleomycin-induced chromatid breaks in cultured peripheral blood lymphocytes. J Natl Cancer Inst 1999;91:1125 – 30. the seven insensitive cell lines was compared with that of the 13. Cloos J, Spitz MR, Schantz SP, et al. Genetic susceptibility to head and neck seven sensitive cell lines (‘‘insensitive versus sensitive’’). A squamous cell carcinoma. J Natl Cancer Inst 1996;88:530 – 5. nonparametric analysis was warranted, because the Z ratio 14. Roberts SA, Spreadborough AR, Bulman B, et al. Heritability of cellular data of some samples were not normally distributed. In radiosensitivity: a marker of low-penetrance predisposition genes in breast addition, genes with measurable expressions in less than four cancer? Am J Hum Genet 1999;65:784 – 94. 15. Cloos J, Gille JJ, Steen I, et al. Influence of the antioxidant N-acetylcysteine cell lines per group were excluded from further analysis. and its metabolites on damage induced by bleomycin in PM2 bacteriophage Therefore, regarding the first comparison, we applied the DNA. Carcinogenesis 1996;17:327 – 31. Wilcoxon signed rank test (30) onto all 28 samples taken 16. Adema AD, Cloos J, Verheijen RHM, et al. Comparison of bleomycin together. Regarding the second comparison, we computed and radiation in the g2 assay of chromatid breaks. Int J Radiat Biol 2003;79: 655 – 61. the differential expression of each gene using the Wilcoxon- 17. Cloos J, Temmink O, Ceelen M, et al. Involvement of cell cycle control Mann-Whitney rank sum test (31). For this, we applied a Z in bleomycin-induced mutagen sensitivity. Environ Mol Mutagen 2002;40: score normalization on all differences of the 4- and 0-hour 79 – 84. Z ratios to arrive at one expression ratio (the 4- over 0-hour Z 18. Amundson SA, Do KT, Shahab S, et al. Identification of potential mRNA ratio) for each gene in each cell line. With this approach, we biomarkers in peripheral blood lymphocytes for human exposure to ionizing radiation. Radiat Res 2000;154:342 – 6. attempted to decrease interindividual heterogeneity, as each 19. Jen KY, Cheung VG. Transcriptional response of lymphoblastoid cells to individual served as his/her own control. The calculation of all ionizing radiation. Genome Res 2003;13:2092 – 100. Ps was based on two-sided testing. 20. Liu G, Lozano G. p21 stability: linking chaperones to a cell cycle checkpoint. Cluster analysis of the latter comparison was done with the Cancer Cell 2005;7:113 – 4. software program Spotfire DecisionSite (Spotfire, Somerville, 21. Cleaver JE. Cancer in xeroderma pigmentosum and related disorders of DNA MA). Variable settings were standard, with no filter or data repair. Nat Rev Cancer 2005;5:564 – 73. 22. Shivji MK, Eker AP, Wood RD. DNA repair defect in xeroderma adjustment, and the hierarchical unsupervised clustering was pigmentosum group C and complementing factor from HeLa cells. J Biol Chem executed for genes and samples with Pearson correlation and 1994;269:22749 – 57. Complete Linkage selected. Expression Analysis Systematic 23. Masutani C, Kusumoto R, Yamada A, et al. The XPV (xeroderma Explorer (available from http://david.niaid.nih.gov/david/ pigmentosum variant) gene encodes human DNA polymerase D. Nature 1999; 399:700 – 4. ease.htm) was used to determine the relative contribution of 24. Yu H, Tang Z. Bub1 multitasking in mitosis. Cell Cycle 2005;4:262 – 5. the affected genes to certain pathways. For each category of the 25. Armes JE, Hammet F, de Silva M, et al. Candidate tumor-suppressor biological process of the Gene Ontology database (http:// genes on chromosome arm 8p in early-onset and high-grade breast cancers. www.godatabase.org), the frequency of the affected genes was Oncogene 2004;23:5697 – 702. compared with that of all characterized genes of the total human 26. Jilani A, Ramotar D, Slack C, et al. Molecular cloning of the human gene, genome. The Fisher’s exact test was used to determine the PNKP, encoding a polynucleotide kinase 3V-phosphatase and evidence for its role in repair of DNA strand breaks caused by oxidative damage. J Biol Chem 1999; statistical significance of the results. 274:24176 – 86. 27. Chappell C, Hanakahi LA, Karimi-Busheri F, et al. Involvement of human References polynucleotide kinase in double-strand break repair by non-homologous end 1. Goldenberg D, Lee J, Koch WM, et al. Habitual risk factors for head and neck joining. EMBO J 2002;21:2827 – 32. cancer. Otolaryngol Head Neck Surg 2004;131:986 – 93. 28. Buermans HP, Redout EM, Schiel AE, et al. Microarray analysis reveals 2. Hashibe M, Brennan P, Strange RC, et al. Meta- and pooled analyses of gstm1, pivotal divergent mRNA expression profiles early in the development of either gstt1, gstp1, anti cyp1a1 genotypes and risk of head and neck cancer. Cancer compensated ventricular hypertrophy or heart failure. Physiol Genomics 2005;21: Epidemiol Biomarkers Prev 2003;12:1509 – 17. 314 – 23. 3. Hsu TC. Genetic predisposition to cancer with special reference to mutagen 29. Cheadle C, Vawter MP, Freed WJ, et al. Analysis of microarray data using sensitivity. In Vitro Cell Dev Biol 1987;23:591 – 603. Z score transformation. J Mol Diagn 2003;5:73 – 81. 4. Cloos J, Braakhuis BJ, Steen I, et al. Increased mutagen sensitivity in head- 30. Conover WJ. Wilcoxon signed rank test. In: Practical nonparametric and-neck squamous-cell carcinoma patients, particularly those with multiple statistics. 2nd ed. New York: John Wiley & Sons; 1980. p. 278 – 94. primary tumors. Int J Cancer 1994;56:816 – 9. 31. Conover WJ. Wilcoxon-Mann-Whitney test. In: Practical nonparametric 5. Cloos J, Leemans CR, Van Der Sterre MLT, et al. Mutagen sensitivity as a statistics. 2nd ed. New York: John Wiley & Sons; 1980. p. 215 – 28.

Mol Cancer Res 2006;4(2). February 2006 Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2006 American Association for Cancer Research. Microarray Analysis of Bleomycin-Exposed Lymphoblastoid Cells for Identifying Cancer Susceptibility Genes

Jacqueline Cloos, Wim P.H. de Boer, Mireille H.J. Snel, et al.

Mol Cancer Res 2006;4:71-77.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/4/2/71

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2006/03/09/4.2.71.DC1

Cited articles This article cites 29 articles, 7 of which you can access for free at: http://mcr.aacrjournals.org/content/4/2/71.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/4/2/71.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/4/2/71. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.