HUMAN CELL(Hum Cell) Copyright 0 2005 by The Japan Human Cell Society

Vol. 18 No. 3 Printed in Japan

- Original Article

Loss of Integrin a3 Expression is Associated with Acquisition of Invasive Potential by Ovarian Clear Cell Adenocarcinoma Cells

Nao SUZUKI1, Atsushi HIGASHIGUCHI’, Yuko HASEGAWA’, Hiroshi MATSUMOT03, Shinji OIE4, Kimiko ORIKAWA’, Sachiko EZAWA2, Nobuyuki SUSUMU’,

Ki-ichi MIYASHITA3 and Daisuke AOKI’

4 s t r a c b Among various types of surface epithebd ovarian carcinoma, clear cell a d e nocarcinoma often has a particularly poor prognosis even when diagnosed in stage I. It is resistant to existing anticancer drugs and appears to have different biological properties to other histological types of ovarian cancer. The present study was conducted using cell lines derived from ovarian clear cell adenocarci- noma in order to identify genes associated with the acquisition of malignant potential by this type of cancer. Two cell lines derived from ovarian clear cell adenocarcinoma (RMGI and RMGV), with different levels of invasive potential in an invasion assay, were used. DNA frag- ments were extracted from the band showing differences in the level of expression by R T PCR with fluorescent differential display. A total of 56 different DNA base sequences were determined by direct sequencing. Primers were established using these base sequences and the level of expression in each cancer cell line was determined by real-time PCR Integrin a3, the gene of which is present on chromosome 17q, was identified. It was also detected by a microarray analysis as one of the molecules showing a different level of expression between the two cell lines. Then the pattern of integrin a3 expression was investigated using immunc- cytochemical staining, and was found to differ between the two cell lines. The findings obtained using these cell lines derived from ovarian clear cell adenocarcinoma indicate that integrin a3 may associated with the acquisition of malignant potential by clear cell adenocar- cinoma.

Key Words: ovarian clear cell adenocarcinoma, integrin a3, RMGI, RMGV [HUMAN CELL 18(3) : 147 - 155,20051

Introduction

1: Department of Obstetrics and Gynecology, St. Marianna University School of Medicine

2: Department of Obsterics and Gynecology, School of

3: Central Research Laboratory, School of Medicine, Medicine, Keio University

Keio University 4: Taiho Pharmaceutical Co.

When prognostic factors were investigated by Cox proportional hazard analysis in 1,185 patients with sur- face epithelial ovarian carcinoma, it was confirmed that clear cell adenocarcinoma had the worst prognosis”. Among the various histological types of ovarian carci- noma, other studies have shown that patients with clear cell adenocarcinoma have the worst prognosis even if the tumor is diagnosed at an early stage and treated by complete resection’), ’). The Gynecologic

147

Oncology Group (GOG) reported that the response rate was 0% for 27 patients with clear cell adenocarcino- ma among 724 patients with FIG0 surgical stage 111 and IV ovarian cancer treated by chemotherapy after primary debulking surgery4’. These reports indicate that clear cell adenocarcinoma is resistant to existing anticancer drugs5’ and has different biological proper- ties from other histological types of epithelial ovarian cancer. In Japan, patients with clear cell adenocarcino- ma have recently increased to account for 23% of those with all types of surface epithelial ovarian carcinoma, which is markedly higher than the rate of approximate- ly 5% in other countries2’. The mechanism of acquisi- tion of malignant potential by ovarian clear cell adeno- carcinoma is currently unknown. The present study was conducted using two ovarian clear cell adenocarci- noma cell lines, RMG16) without in vitro invasive poten- tial and RMG-V7) with invasive potential, in order to idenbfy genes associated with the acquisition of malig- nant potential by this type of carcinoma.

Materials and Methods

1. Cell culture Three human ovarian clear cell adenocarcinoma

cell lines (RMGI”, RMGII”, and RMG-V)) were used. The cells were cultured in a 1:l mixture of Dulbecco’s modified Eagle’s medium and Ham’s F 12 medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal calf serum (Mitsubishi Chemical Co., Tokyo, Japan) and 80 pg/ml of kanamycin sulfate at 3 7 t in an atmosphere of 95% &/5% C02.

2. Invasion assay Serum-free medium (300 p l ) was placed in the well

of the invasion assay chamber (ECM550, Chemicon International, Temecula, CA, USA) and the ECM layer was moisturized at room temperature for 1-2 hours.

Then medium (500 pl) containing 10% FBS was added to the lower chamber. Cell samples (300 pl of a suspension containing 1.0 x lo6 cells/ml) were pre- pared. The serum-free medium in the well of the cham- ber was replaced by a cell sample, which was incubated at 37-C for 2472 hours. Cells that did not invade the EC matrix gel were carefully removed from inside the well using a swab, and this procedure was repeated twice.

1 The harvested cells were placed in a staining solution (500 pl) in a new well and let stand for 20 minutes, after which the well was washed with water for a few min- utes and &-dried.

3. Reverse transcriptase-polymerase chain reac- tion (RT-PCR) with the fluorescent differential display method RNA was extracted from 5 x lo6 RMGI or RMG-V

cells using an RNeasy Mini Kit (Qiagen, Hilden, Germany) and was treated with DNase using an Enzyme Set-FDD (Takara Bio Inc., Shiga, Japan). A reaction mixture (50 pl) containing 26 pl of total RNA

mM MgC12, 0.5 pl of RNase inhibitor (40 U/mI), and 1 pl of DNase I (10 U/ml) was incubated at 37°C for 20 minutes. Then DEPC-treated water (40 pl) with a 1:l mixture (100 pl) of phenol and chloroform were added to stop the reaction. The mixture was centrifuged at 10,000 rpm for five minutes, and aquous phase was transfered to a new tube, after which 0.1 volumes of 3 M NaOAc (PH 5.0) and 2.5 volumes of 99.5% EtOH were added. Next, the mixture was cooled to -80°C for 20 minutes, followed by centrifugation at 15,000 rpm for 10 minutes. The supernatant wqs discarded. The pellet was resuspended in 75% EtOH and centrifuged at 10,000 rpm for five minutes, after which the super- natant was discarded and the pallet was dried. Reverse transcription was done for synthesis of first-strand cDNA using a fluorescence d~erential display kit (fluo- rescein version: Enzyme Set-FDD, Takara Bio Inc., Shiga, Japan). Total RNA (1 pl) and 10 pM fluores- cence-labeled downstream primers (Nos. 1-9) (5 pl) were combined, and the mixture was heated at 70°C for 10 minutes and rapidly cooled on ice. Then 2 pl of 10 x RNA PCR buffer, 4 pl of 25 mM MgCl2,2 pl of dNTP mixture (10 mM), 0.5 pl of RNase inhibitor, 1 p1 of AMV Reverse Transcriptase XL (5 U/ml), and 4.5 pl of RNase free dH2O were added and mixed slowly. After the mixture was reacted at 55°C for 30 minutes, 95°C for five minutes, and 5°C for five minutes, 80 pl of a TE buffer was added, and the mixture was stored at 230°C. The polymerase chain reaction (F’CR) was done using a fluorescence differential display kit (fluorescein ver- sion: Enzyme Set-FDD) from Takara Bio Inc. (Shiga,

(20-50 pg) , 2.5 pl of 1 M Tris-HC1 (PH 7.5), 20 pl of 25

HUMAN CELLVol. 18 No. 3 (2005)

Japan). A reaction mixture (20 jd) was prepared with 2 pl of 10 x LA PCR buffer, 1 pl of 25 mM MgCl2,0.65 pl

of dNTP mixture (2.5 mM), 0.1 pl of TaKaRa La Taq (5

tion, 0.5 pl of fluorescencelabeled downstream primers (10 pM, Nos. 1-9), and 0.5 pl of upstream primers (2 @I, Nos. 1-24). This mixture was reacted at 94°C for two minutes, 38°C for five minutes, and 7 2 t for five minutes, followed by 34 cycles of 94°C for 30 seconds, 38°C for two minutes, and 72°C for one minute.

Then electrophoresis of the products was conduct- ed on denaturing 5% polyacrylamide gel containing 7 M urea. The PCR reaction solution and a 95% formamide - 20 mM EDTA solution were mixed in equal volumes. The mixture was denatured by heating at 94°C for three minutes and rapidly cooled to obtain single- stranded cDNA, which was subjected to electrophore- sis at 1,500 V for two hours on denaturing 5% polyacry- lamide gel containing 7 M urea. Bands were analyzed using a fluorescence image analyzer (Molecular Dynamics Fluoroimager 595, Amersham Biosciences Corp., Piscataway, NJ, USA).

Sections of the gels were cut out irrespective of the presence or absence of target bands with large d ~ e r - ences in the level of expression and their symmetric bands. Water (50 pl) was added to each gel piece, which was let stand at room temperature for 30 min- utes. Then DNA fragments were extracted from the gel by heating at 100°C for 10 minutes. The extracted frag- ments were amplilied using the upstream primer and fluorescence-labeled downstream primer. A reaction mixture (50 pl) was prepared with 1 p1 of an extract containing DNA fragments, 5 p.l of 10 x LA PCR buffer, 4 pl of 25 mM MgCl2,l pl of dNTP mixture (10 mM), 1 pl of fluorescence-labeled downstream primers (10 pM, Nos. 1-9), 5 pl of upstream primers (2 @I, Nos. 1-24), 0.25 p1 of TaKaRa LA Taq (5 U/pl), and 32.75 pL of dHzO. This mixture was reacted at 94°C for two min- utes, followed by 15 cycles of 94°C for 30 seconds, 40°C for two minutes, and 72°C for one minute.

Electrophoresis of the amplified reaction mixtures was conducted at 50 V on 3% agarose gel containing 1 U/ml of H A Yellow. The target and symmetric bands were detected using a fluorescence image analyzer (Molecular Dynamics FluoroImager 595, Amersham

U/pl), 8.75 pl Of dHsO, 2 pl of iirst-strand cDNA soh-

Biosciences Corp, Piscataway, NJ, USA). Then the bands were cut out, and DNA fragments were eluted. Next, the DNA was amplified using the extended upstream primer and downstream primer from the Cloning-Sequencing Primer Set for FDD flakara Bio Inc., Shiga, Japan) under the following conditions.

A reaction mixture (100 pl) was prepared that con- tained 2 pl of a gel extract containing DNA fragments, 10 pl of 10 x LA PCR buffer, 8 pl of 25 mM MgCl2,2 pl of dNTP mixture (10 mM), 1 pl of extended down- stream primers, 2 pl of extended upstream primers, 0.5 pl of TaKaRa LA Taq (5 U/pl), and 74.5 pL of dHzO. This was initially reacted at 94°C for two minutes, fol- lowed by 2 cycles of 94°C for 30 seconds, 40°C for 30 seconds, slope for two minutes, and 72°C for 30 sec- onds, by 40 cycles of 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, and then 72°C for five minutes. The reaction products were placed on 3% agarose gel1 containing 1 U/ml of H A Red and elec- trophoresis was done at 50 V. Then the gel was stained with ethidium bromide and the target band was cut out. DNA fragments were eluted, and direct sequence analysis was performed using direct sequencing primer 1 (Cloning-Sequencing Primer Set for FDD, Takara Bio Inc., Shiga, Japan).

4. Real-time PCR cDNA (5 pl) was diluted 50afold and added to the

PCR master mix (15 pl), which contained 0.1 pl of DNA polymerase (5 U/p l AmpliTaq Gold, Applied Biosystems, Foster City, CA, USA), 1.2 pl of MgCl2 (25 mM), 1.6 pl of dNTP mixture (each at 2 mM), 2 pl of 10 x PCR Gold buffer, 1 pl of SYBR Green (Takara Bio Inc., Shiga, Japan), 0.4 p1 of the forward primer (10 pM) ,0.4 pf of the reverse primer (10 pM), and distilled water (DW). Taqman GAPDH Control Reagent (Applied Biosystems, Foster City, CA, USA) was used as the housekeeping gene. Integrin a3 gene was ampli- fied by using a spectrofluorometric thermal cycler (Applied Biosystem 7700, PE-Biosystem, Foster City, CA, USA), which was initially run at 50°C for two min- utes and 95°C for 10 minutes, followed by 50 cycles of 95°C for 15 seconds and 55°C for one minute. The primers for integrin a 3 were CCTGCATCTCTGT- GAAGCCTC (forward primer) and CCAGAATTG-

1 49

GTCCCCTCCTC (reverse primer), while those for GAPDH were GAAGGTGAAGGTCGGAGTC (forward primer) and GAAGATGGTGATGGGATITC (reverse primer).

5. Microarray analysis We performed microarray analysis using the

Human Cancer Chip version 3, which detects 425 human cancer-related genes. Total RNA was prepared with an RNeasy mini kit (Qiagen, Hilden, Germany). Fluorescence probes were synthesized by incorporat- ing Cy3-dUTP or Cy5-dUTP (Amersham Biosciences Corp., Piscataway, NJ, USA) using 1 pg of total mRNA as a template and 50 U of AMV reverse transcriptase (Takara Bio Inc., Shiga, Japan). As an internal control, 50 pg of lambda A (Takara Bio Inc., Shiga, Japan) was added to the reaction mixture. Cy3- and CySlabeled probes were prepared using mRNA isolated from con- trol cells and the cancer cell lines, respectively. Both mRNAs were mixed in the reaction buffer (6 x SSC/O.2% SDS and 5 x Denhardt's solution with 1 mg/ml denatured salmon sperm DNA), which was hybridized to the cDNA chip at 60°C for 14 hours. Then the chip was washed with 2 x SSC/O.2% SDS at 35°C for 6 min and with 0.2 x SSC/O.2% SDS at 35C for 6 min.

Finally, the chip was washed with 0.2 x SSC at 35°C for 5 min. Visualization and quantitation were done with ImaGene software (Biodiscovery, El Segundo, CA, USA).

6. Immunocytochemical stainiig Cells were hxed with 4% paraformaldehyde for 15

minutes and washed with phosphate-buffered saline (PBS), followed by incubation in 0.05% saponin-PBS for 30 minutes and further washing with PBS. After block- ing with 1% BWPBS, the cells were reacted with pri- mary antibodies mentioned below for one hour at room temperature, and then with the secondary antibody (mouse IgG from the Vectastain ABC Kit, Vector, Burlingame, CA, USA) for 30 minutes at room tempera- ture. Cells were stained with a 100-fold dilution of strep tavidin-Texas Red (Amersham Pharmacia Biotech, Piscataway, NJ, USA) for 30 minutes at room tempera- ture and washed with water. The cells were cleared in 4', &diamidino-2-phenylindole @MI), with which slide

glasses were prepared. Observation was conducted using a fluorescence microscope (Olympus Corporation, Tokyo, Japan). The following primary anti- bodies were used mouse anti-human integrin a3 mon- oclonal antibody (Chemicon International, Temecula, CA, USA), CD29 (integrin bl) antibody (Clone 4B7R Neomarkers, Fremont, CA, USA), and mouse anti- human integrin a3pl (VIA-3) monoclonal antibody (Chemicon International, Temecula, CA, USA).

Results

1. Invasion assay In the invasion assay, no RMG-I and RMG-I1 cells

invaded the ECM during observation for 24-72 hours (Figs. 1-A-D). In contrast, RMG-V cells invaded the ECM after 48 hours (Figs. 1-E, F).

2. RT-PCR with the fluorescent merent display method DNA was extracted from the band in which the

level of expression differed by three fold between RMGI and RMGV (Fig. 2) and a total of 51 different DNA base sequences were identified by direct sequencing.

These sequences were subjected to a BLASTIN search and integrin a 3 (BC015344: Homo sapiens, clone INAGE 4446368, m W ) , the gene for which is present on chromosome 17q, was detected.

3. Real-time PCR Using the primers for integrin a3, the level of

expression of integrin a 3 mRNA in RMG-I cells and RMG-V cells was determined by real-time PCR and it was found to be increased in RMGI cells (Fig. 3).

4. Microarray analysis The results of microarray analysis are shown in

Table 1. The level of expression of integrin a 3 was higher in RMGI cells than in RMGV cells, as was also shown by RT-PCR with fluorescent differential display.

5. Immunocytochemical staining Immunocytochemical staining showed that integrin

a 3 was expressed on the membranes of RMGI cells and that expression was particularly intense at cell

150

adhesion sites Fig. 4) . In contrast. punctate expression of integrin a3 was observed in the cqloplasm of RhqGi' cells (Fig. 4) . The panern of integrin b l expression was co~nparable with that of integrin a3. lmmunwytochemical

qtaining was also conducted with an antibody that rec- ognized laminins binding to integrins a3 or pl. This also showed expression of integrins a3 and pl on the membranes of RMG-I cells and in the cytoplasm of R M G V cells (Fig. 4) .

Discussion

Among various types of surface epithelial ovarian carcinoma, clear cell adenocarcinoma often shows a poor prognosis even when detected in stage I . It is resistant to current anticancer drugs and also appears

to be differ from the other histcilogical types with respect to its biological properties.

The present study was conducted using cell lines derived from ovarian clear cell adenocarcinoma in order to identify molecules related to the biological properties of this cancer. .An invasion assay was con-

ducted using three ct.11 lines and invasive potential was only confirmed for RMG-V cells. Therefore. further assessment was done using RMGV cells with invasive potential and IUvlG-I cells without it. There were no dif- ferences ot growth rate between the two cell lines (data not shown).

D N A fragments were extracted from the band where differences in the level of expression were observed by RT-PCR with the fluorescent differential

display method. Amplification. electrophoresis. and harvesting of the fragments was repeated several times. after which the base sequence of the target DNA was de termined by direct sequencing . Then a RLASTIN search identified integrin a3 sequence. so primers for integrin a3 were used to investigate the expression of integrin a3 mRNA in RMG-I cells and

RMG-V cells by real-time PCR As a result, the level of expression was increased in RMG-I cells. In addition. differences in the expression of various molecules by the two cell lines were investigated using microarray analysis Again. the expression of integrin a3 was

151

Fig. 2: RT-PCR with the fluorescent differential display method DNA was extracted from the band where the level of expression was at least threefold different between RMGl and RMGV cells. and the base sequence was determined.

found to be higher in RMGl cells than in RMGV cells. These findings indicated that the level of integrin a3 expression differed between RMG-I and Rh4GV cells with diflerent invasive potentials, both of which are cell lines derived from ovarian clear cell adenocarcinoma.

The integrin a3 gene is present on chromosome 17q. We previously investigated chromosome copy numbers using frozen section samples from 20 patients with ovarian clear cell adenocarcinoma by a compara- tive genomic hybridization method, which detects changes of DNA copy numbers at specific regions of the chromosomes, and we found that genes on l7q21-

24 were amplified in clear cell adenocarcinoma". The survival rate of patients with a high 17q copy number (1 7q+) was significantly lower than that of patients with a normal 17q copy number (17q-) @c0.05), and l7q+ patients also had a significantly higher death rate from primary cancer. were significantly more positive on peritoneal cytological examination. and were more often aged 260 years old (.p<0.05)"'. We also investigat- ed the presence or absence of gene amplification at l7q21-24 in six cell lines derived from ovarian clear cell adenocarcinoma which our study group established using the CGH method. Amplification was observed at 17q21-24 in RMG-I cells, but not in RMG-V cells (unpublished data). These findings obtained by using the fluorescent differential display and microarray methods to investigate the expression of various mole- cules by cell lines with different invasive potentials derived from clear cell adenocarcinoma indicate that integrin a3 may be one of the molecules associated with the malignant behavior of this type of ovarian can- cer.

Fig. 3: Expression of intek6n os3 mRYA measured by real-time PCR Expression of intepin a3 mRVA was higher in KhlGl crlls than in K M G V cells

152

Table 1: Results 01 microarray analysis

Rank RMG-V/ RMG-I Molecule

1 2 3 3 5 6 7 8 9 1 0 11 12 13 13 15 16 17 18 19

82.989 46.100 25.942 24.848 15.517 13.535 13.319 9.129 8.552 7.793 6.05 1 5.972 5.960 5.654 5.50s - 5 .- 393 5.192 5.055 3.229

fibroblast growth factor receptor 3 alutathione S- t ransferase ? interferon. gamma-inducible protein 30 caspase 1 Human DNA sequence from clone RP11 560A15 mitogen-activated protein kinase integrin. beta5 neuregulin 1 nuclear protein, ataxia-telangiectasia locus cell adhesion molecule with homology to LlCAM mucin 1, transmembrane cyclin El v-yes- 1 Yamaguchi sarcoma viral oncogene homologl cavcolin 1, caveolae protein cadherin 3, typel. P-cadherin glutathione S-transferase M 4 transforming growth factor, beta receptor I1 diap horase platelet-derived growth factor alpha polypeptide

20 3.229 DeriDheral mvelin Drotein 22

Rank RMG-V/ RMG-I Molecule 1 0.006 5' nucleotidase 2 0.006 hexabrac hion 3 0.007 tumor necrosis factor (ligand) superfamily 4 0.00s interleukin 8 5 0.01 1 karyopherin alpha 3 6 0.013 plasminogen activator 7 0.0 16 RAB32. mcmbcr RAS oncogene family 8 0.01 7 amphiregulin 9 0.0 18 epidermal growth factor receptor pathway substrate 8 1 0 11 0.023 cadherin 2, type 1, N-cadherin 12 0.025 CD44 antigen 13 0.028 chromogranin B 13 ( 1.028 f i h n e c t in 1 15 16 0.040 transforming growth factor. alpha 17 ().045 intcgrin. alpha 3 18 0.045 integrin. alpha 1 19 ~1.050 B-cell CLL/lymphoma 2 2 1

0.019 major h is t oco m pa t i bi I it y co m p I e x

( 1 .( )3( 1 collagcn. type VI1, alpha 1

0.05 1 t ra 11s 10 r m i n o (7 row t h factor. bet a- i n d u c t'd

RMG-VIRMG-I: ratio o t thc expression o t RMG-V relative to that of RMG-I

153

Integrins are cell-substrate and cell-cell adhesion molecules that bind extracellular ligands to cytoskele- tal proteins. Integrins influence a wide range of cellular functions through modulation of cell adhesion, such as growth. differentiation, motility, morphogenesis, and invasion of cancer"" ' I ) . Our immunocytochemical study showed that integrin a3 was expressed on the cell membrane by RMGI cells, and the most intense expression was at cell adhesion sites. In contrast, punc- tate expression of integrin a3 was observed in the cyte

plasm of R M G V cells. A similar pattern of expression was observed for laminin, a Iigand for integrin a3 sub units, integrin pl and integrin a3pl. The expression of integrin a?, integrin pl, and laminin was not detected on the cell membranes of RMG-V cells with invasive. while these all showed positive expression, particularly

at cell adhesion sites. on KI14G-I cells without invasive potential. These findings suggest that loss of the expression of these molecules may be associated with the acquisition of malignant behavior, namely invasive potential, by clear cell adenocarcinoma cells. Bartolazzi et al.'" reported that the expression of integrin a3pl was observed in 26 out of 31 ovarian carcinoma. exclud- ing clear cell adenocarcinonia, but disappeared with a decrease in the differentiation of the cancer cells. Further studies should be done to determine whether integrin n3, integrin pl. and laminin are important mol- ecules associated with the acquisition of malignant potential by clear cell adenocarcinoma. such as investi- gating their expression in tumor samples.

Fig. 4: Immunocytochemiral staining I'pprr: Inkgrin tr3. integrin PI. and laminin CCr-?BI) are drtected on the ccll membranes of RMG-I crlls, parlicularly at cell adhesicin sites

1riwc.r. Punctatr rxprrssitrn of thrw moltrules is obscwed in the c-y!oplasm of K M G V crlls.

154

HUMAN CELL Vol. 18 No. 3 (2005)

References

1) Sugawa T, Umesaki N, Yajima A, et al.: A group study on the prognosis of ovarian cancer in Japan. Acta Obst Gynec Jpn, 44: 827-832,1992

2) Enomoto T, Kuragaki C, Yamasaki M, et al.: Is clear cell carcinoma and mucinous carcinoma of the ovary sensitive to combination chemotherapy with paclitaxel and carboplatin? Proc Am SOC Clin Oncol22: 447,2003

3) Monga M, Carmichael JA and Shelly WE: Surgery without adjuvant chemotherapy for early epithelial ovarian carcinoma after comprehensive surgical staging. Gynecol Oncol, 4 3 195197,1991

4) Omura GA, Brady MF, Homesley HD, et al.: Long- term follow-up and prognostic factor analysis in advanced ovarian carcinoma: The Gynecologic Oncology Group Experience. J Clin Oncol, 9: 1138 1150,1991

5) Sugiyama T, Kamura T, Kigawa J, et al.: Clinical characteristics of clear cell carcinoma of the ovary: a distinct histologic type with poor prognosis and resistance to platinum-based chemotherapy. Cancer, 88 2584-9,2000

6) Nozawa S, Tsukazaki K, Sakayori M, e t a].:

Establishment of a human ovarian clear cell carci- noma cell line (RMGI) and its single-cell cloning. Human Cell, 1: 426-435,1988

7) Nozawa S, Yajiia M, Sasaki H, et al.: A new CA- 125like antigen (CA602) recognized by two mono- clonal antibodies against a newly established ovari- an clear cell carcinoma cell line (RMG-ID. Jpn J Cancer Res, 82: 854-861,1991

8) Aoki D, et al.: Establishment and characterization of RMGV derived from human ovarian clear cell adenocarcinoma. Human Cell (in press)

9) Hirasawa A, Saito-Ohara F, Inoue J, e t al.: Association of 17q21-q24 gain in ovarian clear cell adenocarcinomas with poor prognosis and identifi- cation of PPMlD and APPBP2 as likely amplifica- tion targets. Clin Cancer Res, 9: 19952004,2003

10) Hynes RO: Integrins: versatility, modulation and signaling in cell adhesion. Cell, 69: 11-25, 1992

11) Rouslati E: Anchorage dependence, integrins, and apoptosis. Cell, 77: 477-478, 1994

12) Bartolazzi A, Kaczmarek J , Nicolo G, et al.: Localization of the a3bl integrin in some common epithelial tumors of the ovary and in normal equiv- alents. Anticancer Res, 13: 1-12,1993

Received 2005.8.29, Accepted 2005. 10.8 Correspondence : Nao Suzuki, M.D., Ph.D. Department of Obstetrics and Gynecology, St. Marianna University School of Medicine. 2-161 Sugao, Miyamaeku, Kawasaki kanagawa 2168511, Japan. Tel: +8144-977- 8111 Ext.3332 Fax: 81-44-977-2944 E-mail: nao@marianna-u.ac.jp

155