利用微阵列分析恶性胶质瘤对电离辐射的时序基因反应

J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp J. Radiat. Res., 45, 53–60 (2004) Regular Paper Microarray Analysis of Temporal Gene Responses to Ionizing Radiation in Two Glioblastoma Cell Lines: Up-regulation of DNA Repair Genes Takashi OTOMO1*, Makoto HISHII1, Hajime ARAI1, Kiyoshi SATO1 and Keisuke SASAI2 Radiosensitivity/DNA repair/Glioblastoma/DNA microarray. To determine the patterns of gene expression responsible for the radiosensitivity of glioblastoma cells, we analyzed transcriptional changes after ionizing radiation in different cell lines. After completing clonogenic survival assays, we selected two glioblastoma cell lines with different radiosensitivities. Subsequently, they were investigated by using the technique of DNA microarray, and we then categorized the upregulated genes into 10 groups. Between the two cell lines, the difference in the percentage of DNA repair/replication category was the largest, and this category was present at a greater percentage with radioresistant cell line U87MG. Moreover, among the commonly upregulated genes, the DNA repair/replication category was present in the largest percentage. These genes included G22P1 (Ku70) and XRCC5 (Ku80) genes known as important mem- bers of the nonhomologous end-joining (NHEJ) pathway of DNA double strand break (DSB) repair. Further- more, cell line that specifically upregulated genes included the members of major pathways of DNA DSB or single strand damage repair. These pathways were not only NHEJ, but also homologous recombination (HR) and postreplication repair (PRR). In conclusion, the distribution of genes involved in the DNA repair/replica- tion category was most different between two human glioblastoma cell lines of different radiosensitivities. Among commonly upregulated genes, the DNA repair/replication category was present in the largest percent- age. INTRODUCTION Glioblastoma is the most malignant astrocytic tumor, and radiotherapy is an important component of multimodal treat- ment. Although radiotherapy prolonged the survival of patients with this disease,1) the recurrence rate is still very high. Biological modifications during radiotherapy based on genetic information may improve the efficacy of radiotherapy for glioblastoma.2) Many studies have been reported about correlations between gene expression and radiotherapeutic response,3–5) survival time after regrowth,4) and in vitro radiosensitivity.6–9) Furthermore, Rhee et al.10) reported evidence that cell lines of the same pathological origin could be highly heterogeneous in gene expression. Recent technological advances in DNA microarray have made it possible to profile the gene expres- sion responses to ionizing radiation (IR) simultaneously. A variety of genes, including immediate-early and acute phase genes, DNA repair genes, cytokine, and growth factor genes have been proposed as members of the mechanisms by which cells survive after IR.11–13) Although these gene expression levels change quickly, a few studies reported time-course- related gene induction after IR. Here we report time-related changes and cell line specific patterns of gene expression after IR by using two glioblastoma cell lines of different radi- osensitivities. Moreover, our methods can provide valuable information for designing effective radiotherapy of various tumor types. MATERIALS AND METHODS Cell lines, irradiation, and clonogenic assays Human glioblastoma cell lines, U87MG, A172, and U138MG were obtained from the American Type Culture Collection. Cells were maintained in RPMI-1640 medium with 10% fetal calf serum and antibiotics (90 U penicillin/ml, 90 µg streptomycin/ml) at 37°C in an atmosphere of 5% CO2 in air. We used a 150 kVp X-ray machine (MBR-1505R2, Hitachi Medical Corporation, Japan) for irradiation with a 0.5 mm aluminum filter at a dose rate of 2 Gy/min. We measured the radiosensitivities of these cell lines with a standard clono- genic assay as described previously.14) *Corresponding author: Phone: +81-03-3813-311, Fax: +81-03-5689-8343, E-mail: totomo@med.juntendo.ac.jp 1Departments of Neurosurgery and 2Radiology, Juntendo University School of Medicine, Hongo 2-1-1, Bunkyo-ku, Tokyo 113-8421, Japan. T. Otomo et al.54 J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp Direct DNA sequence of exon 7 of the TP53 gene in cell line A172 The total RNA was isolated from unirradiated A172 cells by the acid guanidinium thiocyanate-phenol-chloroform method.15) The first-strand complimentary DNA (cDNA) that served as the polymerase chain reaction (PCR) template was synthesized from total RNA (2 µg) by using reverse tran- scriptase (Super ScriptTM II, Life Technologies, Invitrogen Corp., Carlsbad, CA, USA) and oligo-dT primers. The fol- lowing procedure was used for direct sequencing of PCR- amplified cDNA molecules: A 200-base pairs fragment span- ning exon 7 of the TP53 gene was PCR-amplified from 0.5 µl of cDNA by using 0.4 µmol of each of the following primers: sense (5′-GTGTGGAGTATTTGGATGACAG-3′) and anti- sense primers (5′-CACAAACATGCACCTCAAAGCT-3′), 1.6 µl dNTPs, 1.5 mmol MgCl2, and 0.2 µl of Taq DNA poly- merase (LA Taq, Takara Bio Inc., Japan). Subsequent proce- dures were performed according to Kyritsis et al.16) Preparation of cells, fluorescent cDNA, and hybridization for the DNA microarray We used U87MG and A172 cells for the analysis with the DNA microarray. Cells were grown to confluence and were irradiated with or without a single dose of 6 Gy as described above. Samples were collected at 3 times (i.e., 0.5, 1, and 6 h after irradiation). As a control, unirradiated cells were used and collected at the same times. Total RNAs were extracted from cells as described above. A fluorescent probe was syn- thesized by incorporating Cy3- or Cy5-dUTP (Amersham Pharmacia Biotech) with an RNA Fluorescence Labeling Core Kit (Takara Bio Inc.,Japan), using 25 µg of the total RNA described above and 25 U AMV reverse transcriptase (Takara Bio Inc., Japan). We labeled total RNA from irradi- ated cells with Cy3 and total RNA from unirradiated cells with Cy5 dyes. After purification, Cy3 and Cy5 probes were mixed with human Cot áT DNA, poly(dA), and yeast tRNA (5 × Competitor áT, Takara Bio Inc., Japan), then followed by ethanol precipitation. The mixture was resuspended in 15 µl of a hybridization solution (6× SSC, 0.2% SDS, 5× Den- hardt’s solution, 0.1 mg/ml salmon sperm DNA). Subse- quently, we performed prehybridization, hybridization and washing according to the method of Taniguchi et al.17) Scanning and analysis of DNA microarrays We performed microarray analysis by using Human Cancer CHIP Version 3.0 (Takara Bio Inc., Japan), in which 638 human cancer-related genes and housekeeping genes (listed on the home page of Takara Bio Inc., http://bio.takara.co.jp/ BIO_EN/DNAChip_Download_e.htm) were spotted on a glass plate. Before each use of the Human Cancer CHIP Ver- sion 3.0, we performed microarray analysis as described above by using the TestARRAY Version 3.0 DNA chip (Takara Bio Inc., Japan) to check for labeling probes. Fluores- cence intensities were captured with a laser confocal scanner; GMS 418 Array Scanner (Affymetrix). Subsequently, obtained data were analyzed with software: ImaGene version 4.1 (BioDiscovery Inc). For quality control, automatically marked spots that were not distinguishable from the back- ground were removed. Negative values resulting from a sub- traction of the background were removed. Eight signal spots of beta-actin cDNAs were used for normalization. After nor- malization, 2.7-fold differences in upregulated expression were used to identify the unchanged or altered genes. We arbitrarily chose this ratio of 2.7 as the cutoff value for matches of independent experiments and comparison of A172 and U87MG cells data, considering the manufacturer’s instruction that showed a twofold induction was significant. We used histograms for gene selection and confirmed these signals in scatter plots created by ImaGene software. The selected genes were checked by using the web site of Uni- Gene (http://www.ncbi.nlm.nih.gov/UniGene/index.html). An- notations and functional classification of genes were based on PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi), On- line Mendelian Inheritance in Man (http://www.ncbi.nlm.nih. gov/entrez/), LocusLink (http://www.ncbi.nlm.nih.gov/Locus Link/), GENE ONTOLOGYTM CONSORTIUM (http:// www.geneontology.org/), Proteome database (http://www. proteome.com/), and other database web sites. RESULTS Radiosensitivities of cell lines Figure 1 shows the radiation survival curves derived from colony formation assays of 3 cell lines. The radiosensitivity of A172 cells was the highest among 3 cell lines. On the other hand, the radiosensitivity of U87MG cells was the lowest. So we selected the U87MG and A172 cell lines for the analysis Fig. 1. Radiation cell survival curves for U87MG cells (open cir- cles), A172 cells (closed circles), and U138MG cells (open squares). The colony formation assay was described in the “MATERIALS AND METHODS” section of the “Cell lines, irradiation, and clonogenic assays.” Error bars represent standard deviations for three indepen- dent experiments. Microarray Analysis of Gene Responses to Irradiation in Glioblastoma Cell Lines 55 J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp because of their difference in radiosensitivity. The survival fraction of U87MG cells was slightly higher than for A172 cells at the dose of 3 Gy. In contrast, more apparent differ- ences in the survival fraction were observed among the 3 cell lines at doses of 6 and 9 Gy. Taking into account these dose- dependent differences in the survival fraction and considering clinically relevant doses, we decided on 6 Gy at the irradia- tion dose for microarray analysis. Confirmed sequence of exon 7 of the TP53 gene in cell line A172 In this study, we used 3 cell lines in the clonogenic assay because of their known TP53 gene status, and they were reported as wild-types.8,18,19) On the other hand, in previous studies mutations of theTP53 gene at exon 7 in the A172 cell line were also reported20). Therefore we used direct sequenc- ing to examine a part of the coding region of the TP53 gene for mutations in the region, including exon 7 of the gene of cell line A172. No mutation was found in the region, and the TP53 gene of cell line A172 was used as a wild type. Analyses of gene expression after X-ray irradiation Table 1 shows the number of upregulated genes in the two cell lines with respect to time after irradiation. We used histo- grams for gene selection and confirmed these signals in scat- ter plots created by ImaGene software (Fig. 2) as described in “MATERIALS AND METHODS.” To detect the immediate early and acute phase transcriptional responses, we set time points at 0.5, 1, and 6 h as described in “MATERIALS AND METH- ODS.” In these phases, many genes that affect cell survival are induced,13,21) and DNA repair kinetics were observed.22) The upregulated genes were analyzed further with respect to cell lines. The results suggested that the upregulated genes repre- sented three groups, those that were U87MG cell-specific, those that were A172 cell-specific, and those that were com- mon in both U87MG and A172 cells (Table 1). Time course- dependent changes of gene expression were observed. In these 3 groups, we used a further classification with respect to function, and Table 5 shows the distribution of up-regulated genes. These functional groups (categories) were: (1) DNA repair/replication; (2) transcription/translation factor; (3) stress response; (4) cell communication/signaling; (5) cell structure; (6) metabolism/energy; (7) apoptosis; (8) protein synthesis/modification; (9) cell cycle; (10) miscellaneous. To compare cell line specific upregulated genes of U87MG and A172, their distribution is represented in Fig. 3. The impor- tant information that came out of these analyses is that the percentage of DNA repair/replication and Transcription/ Translation factor categories were larger in U87MG cells. In contrast, cell cycle, protein synthesis/modification, and apop- tosis categories were present in larger percentages with A172 cells. In the group of commonly upregulated genes, the DNA repair/replication category was present in the largest percent- ages (Table 2). DISCUSSION This is the first study describing time course-dependent and cell-line specific patterns of gene expression after IR, using two human glioblastoma cell lines of different radiosensitivi- ties. These cell lines were selected because of their known TP53 status as wild types and their radiosensitivities as a con- sequence of clonogenic assays. Time-course-dependent Fig. 2. Scatter plots of gene expression levels for A172 (upper) and U87MG (lower) cells. The log-transformed fluorescence inten- sity of each spot was compared between irradiated and unirradiated cells. For each gene, the RNA expression level in the irradiated sam- ple (Cy3) is given on the x axis, and the expression level for the same gene in the control sample (Cy5) is plotted on the y axis. Gene induc- tion in response to irradiation is visualized as a shift downward from the diagonal, and the genes suppressed are seen as a shift upward. The dashed lines represent the straight lines of regression after nor- malization, as described in “MATERIALS AND METHODS.” Fig. 3. Representation of the distribution of genes compared between U87MG- and A172-cell line specific genes. Closed and striped bars indicate the distribution of U87MG- and A172-cell line specific genes, respectively. T. Otomo et al.56 J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp Table 1. Upregulated genes. Category Gene name Symbol Induced time after irradiation (h) In A172 In U87MG 0.5 1 6 0.5 1 6 Apoptosis defender against cell death 1 DAD1 1 1 programmed cell death 10 PDCD10 1 baculoviral IAP repeat-containing 3 BIRC3 1 BCL2/adenovirus E1B 19 kDa-interacting protein 3 BNIP3 1 BCL2-related protein A1 BCL2A1 1 caspase 1, apoptosis-related cysteine protease (interleukin 1, beta, convertase) CASP1 1 CD27-binding (Siva) protein SIVA 1 death-associated protein DAP 1 DNA fragmentation factor, 45 kDa, alpha polypeptide DFFA 1 quinone oxidoreductase homolog PIG3 1 p53-induced protein PIG11 1 v-rel avian reticuloendotheliosis viral oncogene homolog A (nuclear factor of kappa light polypeptide gene enhancer in B cells 3 [p65]) RELA 1 growth arrest and DNA-damage-inducible 34 PPP1R15A 1 nuclear factor of kappa light polypeptide gene enhancer in B cells inhibitor, alpha NFKBIA 1 Ras-related GTP-binding protein RAGA 1 cytochrome c HCS 1 Cell Commu-nication/Signaling integrin, alpha 3 (antigen CD49C, alpha 3 subunit of VLA-3 receptor) ITGA3 1 1 interleukin 8 IL8 1 1 epidermal growth factor receptor pathway substrate 8 EPS8 1 active BCR-related gene ABR 1 EphB4 EPHB4 1 glia maturation factor, beta GMFB 1 guanylate binding protein 1, interferon-inducible, 67 kDa GBP1 1 interferon-related developmental regulator 1 IFRD1 1 signal sequence receptor, alpha (translocon-associated protein alpha) SSR1 1 vascular endothelial growth factor C VEGFC 1 fibronectin 1 FN1 1 Rho GTPase activating protein 1 ARHGAP 1 CD44 antigen (homing function and Indian blood group system) CD44 1 guanine nucleotide binding protein (G protein), alpha-inhibiting activity polypeptide 3 GNAI3 1 integrin beta 3-binding protein (beta 3-endonexin) ITGB3BP 1 1 chondroitin sulfate proteoglycan 2 (versican) CSPG2 1 cadherin 13, H-cadherin (heart) CDH13 1 paxillin PXN 1 1 1 tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, zeta polypeptide YWHAZ 1 Cell Cycle CDC37 (cell division cycle 37, S. cerevisiae, homolog) CDC37 1 D123 gene product D123 1 nucleolar phosphoprotein p130 NOLC1 1 ras homolog gene family, member G (rho G) ARHG 1 G1 to S phase transition 1 GSPT1 1 RAP1A, member of RAS oncogene family RAP1A 1 CDC16 (cell division cycle 16, S. cerevisiae, homolog) CDC16 1 v-jun avian sarcoma virus 17 oncogene homolog JUN 1 CDC28 protein kinase 1 CKS1 1 ras homolog gene family, member A ARHA 1 1 Cell Structure laminin, alpha 4 LAMA4 1 1 Microarray Analysis of Gene Responses to Irradiation in Glioblastoma Cell Lines 57 J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp Table 1. (continued) Upregulated genes. Category Gene name Symbol Induced time after irradiation (h) In A172 In U87MG 0.5 1 6 0.5 1 6 collagen, type VIII, alpha 1 COL8A1 1 1 lamin B2 LMNB2 1 RAB2, member of RAS oncogene family RAB2 1 peripheral myelin protein 22 PMP22 1 RAB36, member of RAS oncogene family RAB36 1 laminin, beta 2 (laminin S) LAMB2 1 vimentin VIM 1 vinculin VCL 1 collagen, type I, alpha 2 COL1A2 1 collagen, type VI, alpha 3 COL6A3 1 keratin 10 (epidermolytic hyperkeratosis; keratosis palmaris et plantaris) KRT10 1 caveolin 1, caveolae protein, 22 kD CAV1 1 1 hexabrachion (tenascin C, cytotactin) HXB 1 1 1 DNA Repair/Replication purine-rich element binding protein A PURA 1 ataxia telangiectasia and Rad3 related ATR 1 general transcription factor IIH, polypeptide 2 (44 kD subunit) GTF2H2 1 proliferating cell nuclear antigen PCNA 1 RAD52 (S. cerevisiae) homolog RAD52 1 adenine phosphoribosyltransferase APRT 1 glutaredoxin (thioltransferase) GLRX 1 1 centromere protein F (350/400 kD, mitosin) CENPF 1 topoisomerase (DNA) II alpha (170 kD) TOP2A 1 ubiquitin conjugating enzyme E2A (RAD6 homolog) UBE2A 1 thyroid autoantigen 70 kD (Ku antigen) G22P1 1 1 X-ray repair complementing defective repair in Chinese hamster cells 5 (double- strand-break rejoining; Ku autoantigen, 80 kD) XRCC5 1 1 1 1 guanine monphosphate synthetase GMPS 1 1 ribonucleotide reductase M1 polypeptide RRM1 1 1 Metabolism/Energy phorbol-12-myristate-13-acetate-induced protein 1 PMAIP1 1 aldolase A, fructose-bisphosphate ALDOA 1 1 dihydropyrimidine dehydrogenase DPYD 1 hypoxanthine phosphoribosyltransferase 1 (Lesch-Nyhan syndrome) HPRT1 1 phosphoinositide-3-kinase, class 3 PIK3C3 1 galactosidase, beta 1 GLB1 1 glutathione peroxidase 1 GPX1 1 glyceraldehyde-3-phosphate dehydrogenase GAPD 1 phosphoglycerate kinase 1 PGK1 1 lysophospholipase like MGLL 1 ATP synthase, H+ transporting, mitochondrial F0 complex, subunit b, isoform 1 ATP5F1 1 malate dehydrogenase 1, NAD (soluble) MDH1 1 1 1 1 ornithine decarboxylase 1 ODC1 1 1 Protein Synthesis/Modification matrix metalloproteinase 3 (stromelysin 1, progelatinase) MMP3 1 1 dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 4 DYRK4 1 F-box only protein 9 FBXO9 1 superoxide dismutase 3, extracellular SOD3 1 tissue inhibitor of metalloproteinase 3 (Sorsby fundus dystrophy, pseudoinflammatory) TIMP3 1 cathepsin L CTSL 1 hsp70-interacting protein HSPBP1 1 T. Otomo et al.58 J. Radiat. Res., Vol. 45, No. 1 (2004); http://jrr.jstage.jst.go.jp changes of upregulated genes were observed, and the func- tional classification of the upregulated genes showed interest- ing distribution patterns. The difference in the percentages of genes categorized in DNA repair/replication between the two cell lines was the greatest (Fig. 3). Differences in other func- tional gene groups were not so significant. Furthermore, in commonly upregulated genes, the DNA repair/replication cat- egory was present in the largest percentage. Therefore there is a possibility that genes categorized as DNA repair/replication play a role in determining radiosensitivity in glioblastoma cells. The significance of this observation remains to be deter- mined and may affect the design of future studies. X-ray irradiation induces complex cellular response.6,7) The complex molecular responses to genotoxic stress are medi- Table 1. (continued) Upregulated genes. Category Gene name Symbol Induced time after irradiation (h) In A172 In U87MG 0.5 1 6 0.5 1 6 v-yes-1 Yamaguchi sarcoma v