ANTICANCER RESEARCH 26: 4195-4202 (2006) Gefitinib-related Gene Signature in Bladder Cancer Cells Identified by a cDNA Microarray RYO INOUE1, HIDEYASU MATSUYAMA1, SEIJI YANO1, YOSHIAKI YAMAMOTO1, NORIO IIZUKA2 and KATSUSUKE NAITO1 Departments of 1Urology and 2Surgery 2, Yamaguchi University School of Medicine, 1-1-1 Minami-Kogushi, Ube, Yamaguchi 755-8505, Japan Abstract. Background: The aim of this study was to identify key (EGFR-TKI), an enzyme involved in certain signaling genes linked to the molecular action of gefitinib, a promising pathways that drives proliferation, invasion and survival of anticancer agent on human bladder cancer cell lines. Materials cancer cells (2). Experimental studies have shown gefitinib to and Methods: cDNA microarrays were used to profile feature have antitumor activity against several human cancer cell lines genes in 5637 and T24 cells before and after treatment with expressing EGFR, including ovarian, breast, colon and gefitinib. PCR-based direct sequencing and Western blot analysis bladder cancer (3-6). A variety of bladder cancer cell lines were performed to examine the mutation status and protein levels have been reported to show a dose-dependent inhibition of of EGFR in the cell lines. Results: Gefitinib significantly inhibited cell proliferation when treated with gefitinib, which correlated the proliferation of 5637 cells, while showing little inhibitory effect with the EGFR protein levels (7), while no clear correlation on T24 cells. Theses effects were independent of the mutation was found between the amount of phospho-EGFR or EGFR status and protein levels of EGFR. cDNA microarray analysis and the sensitivity to gefitinib in other experimental studies identified 15 feature genes classified as a cell cycle, apoptotic (8), or pre-clinical trials (9). Several in vitro studies showed pathway and transcription. Notably, levels of expression of the cell that EGFR-TKI was capable of inhibiting the growth of invasion-related genes, YY1 and E-cadherin, were increased in bladder cancer cells (10, 11). Despite much effort in this 5637 cells sensitive to gefitinib. Conclusion: Unique genes involved research filed, the mainstays of the anti-proliferative effect by in the action of gefitinib were identified. Particularly, the up- gefitinib remain to be elucidated because of its complicated regulation of YY1 and E-cadherin may account for the efficacy of mechanism, as shown in a targeting study (9). gefitinib in bladder cancer. In this study, the in vitro anti-proliferative effects of gefitinib were evaluated on a panel of EGFR-expressing Bladder cancer is the second most common malignancy human bladder cancer cell lines and a cDNA microarray among urological malignancies, approximately 30% of strategy was then used to profile several key genes that which are clinically detected as muscle invasive tumors. could discriminate cell lines sensitive for gefitinib treatment Despite aggressive treatment, such as a radical cystectomy from insensitive lines. To our knowledge, this is the first or radiotherapy, muscle-invasive transitional cell comprehensive study to analyze the molecular pattern of carcinoma (TCC) continues to show a poor prognosis with bladder cancer cells before and after gefitinib exposure. an overall 5-year survival of only 50% (1). New therapeutic strategies are prerequisite to improving the poor outcome Materials and Methods of this-type cancer. Drugs. Gefitinib was kindly provided by AstraZeneca Gefitinib (Iressa®, ZD1839) is an orally administered Pharmaceuticals (Macclesfield, UK). Serial dilutions of a stock epidermal growth factor receptor-tyrosine kinase inhibitor solution of gefitinib were made in dimethyl sulfoxide (DMSO) and stored at –20ÆC until use. The drug was diluted in fresh media before the experiment. Epidermal growth factor (EGF) was purchased from BD Biosciences (Bedford, MA, USA). Correspondence to: Katsusuke Naito, Department of Urology, Yamaguchi University School of Medicine, 1-1-1 Minami-Kogushi, Cell culture. Twelve human bladder cancer cell lines (T24, Ube, Yamaguchi 755-8505, Japan. Tel: +81 836 22 2275, Fax: +81 TCCsup, EJ-1, J82, 5637, KK47, HT1197, HT1376, RT4 SW780, 836 22 2276, e-mail: [email protected] UMUC3 and SCaBER) were employed. The cell lines were cultured as a monolayer in "routine medium"; comprised 90% Key Words: Gefitinib, microarray, bladder cancer, EGFR, YY1, RPMI 1640 supplemented with 10% heat-inactivated fetal bovine E-cadherin. serum (FBS) (Gibco BRL, Gland Island, NY, USA) and 50 IU 0250-7005/2006 $2.00+.40 4195 ANTICANCER RESEARCH 26: 4195-4202 (2006) penicillin and 50 Ìg/ml streptomycin. The cell lines were grown in d) Gene expression analysis. The fluorescence intensity measurements a humidified incubator, supplemented with 5% CO2 at 37ÆC. from each array element were compared to the local background, and background subtraction was performed. To normalize the data, the PCR and direct sequence of the EGFR gene. DNA was extracted Cy3 to Cy5 ratio for each gene was adjusted to a median ratio value from the 12 cell lines using QIAamp DNA mini kits (QIAGEN, of all spots on the array. The spots with signal intensity less than the Hilden, Germany) according to the supplier’s protocols. PCR and mean value of negative control spots, a peculiarity in this array, were direct sequence were demonstrated by Lynch's protocols (12). excluded from the following analysis. Poor hybridization spots or areas of the array with obvious defects were flagged manually. Our Assessment of cell growth. A classic growth curve analysis was used to preliminary study with RT-PCR showed that levels of ‚-actin gene determine the effects of gefitinib on cellular proliferation in the were unchanged before and after exposure to gefitinib; therefore, the presence of FBS (same lot number). The cells were plated at 3x104 hybridization signal levels were normalized for each experiment in per 35-mm diameter plate. After 24 h (day 2), the baseline cell order to ensure that levels of ‚-actin before exposure were equal to number of three dishes was counted by using tripan blue dye those after exposure. This examination was independently performed exclusion test. The remaining dishes received gefitinib at in triplicate in each cell line. Genes, whose expression was increased concentrations from 0.1, 1.0 and 10.0 ÌM, all of which were clinically or decreased in all three experiments of one cell line, were selected. achievable concentrations in the tissue. The viable cell number of individual dishes was counted in triplicate at each drug concentration Statistical analysis. The statistical analyses were performed using JMP every 24 h for 8 days. The mean counts were plotted against the time Version 4.02 (SAS Institute). ANOVA with Fisher’s PLSD test were and the logarithmic tracts of the growth curves were made. used to compare differences in the number of viable cells among the six cell lines. A p-value of less than 0.05 was considered significant. Western blot analysis. Western blotting was performed in order to investigate the protein levels of EGFR and the effect of gefitinib on EGFR phosphorylation, as described previously (7, 8). Results EGFR (Anti-EGFR, Cell Signaling Technology), phospho- EGFR (Anti-phospho-EGFR; pY1068, Cell Signaling Technology) PCR and direct sequence of EGFR gene. Neither a point and ‚-actin (Anti-‚-actin, Cell Signaling Technology) antibodies mutation nor a deletion in exons 19 and 21 of the EGFR were used to measure the protein levels. The blots were visualized gene was detected in the 12 bladder cancer cell lines. Based by chemiluminescence (ECL, Amersham Biosciences, Piscataway, on the growth curve, the percentage growth inhibition by NJ, USA). gefitinib was analyzed in six selected cell lines. cDNA microarray. Growth curve analysis. Human bladder cancer cell lines were a) RNA extraction. Total RNA was extracted from T24 and 5637 exposed to continuous treatment with gefitinib at with and without 4-h incubation of 10 ÌM gefitinib using RNeasy concentrations 0.1-10 ÌM. The results are shown in Figure 1. A mini kits (Qiagen, Valencia, CA, USA) according to the supplier’s significant dose-dependent growth-inhibitory effect was protocol. RNA was dissolved in RNase-free H2O. The resulting RNA was measured spectrophotometrically (BECKMAN; observed in 5637 and TCCsup after gefitinib treatment, while Beckman Instruments, Inc.). no effect was found in T24. Mild to moderate growth inhibitory b) cDNA probe preparation. Reverse transcription and fluorescence effects were observed in EJ-1, J82 and KK47. The proliferations labeling were performed to synthesize fluorochrome-labeled cDNA of T24, KK47, J82, EJ-1, TCCsup and 5637 were inhibited by probes with an RNA fluorescence labeling kit (BD Atlas 13.9%, 51.0%, 65.9%, 73.6%, 82.5% and 99.7%, respectively, 6 PowerScript Fluorescent Labeling Kit, BD Biosciences, Palo Alto, days after 10 ÌM gefitinib administration (Figure 1, *p<0.05, CA, USA) for cells with and without gefitinib treatment with Cy3- dUTP and Cy5-dUTP (Amersham Biosciences), respectively, ANOVA with Fisher’s PLSD test). Among the 6 cell lines, the according to manufacturer’s recommendation. most sensitive to gefitinib was 5637, while the most insensitive c) Hybridization and estimation of expression levels on cDNA was T24. The proliferation of 5637 was significantly inhibited 1 microarrays. BD Atlas Glass 1.0 Microarrays containing cDNA for day after 10 ÌM-gefitinib exposure. Based on this, it was 1090 genes (Atlas Glass Human 1.0 microarray,