Hypoxia Attenuates the Expression of E-Cadherin via Up-Regulation of SNAIL in Ovarian Carcinoma Cells
2003; Elsevier BV; Volume: 163; Issue: 4 Linguagem: Inglês
10.1016/s0002-9440(10)63501-8
ISSN1525-2191
AutoresTsutomu Imai, Akiko Horiuchi, Cui-ju Wang, Kenji Oka, Satoshi Ohira, Toshio Nikaido, Ikuo Konishi,
Tópico(s)Hemoglobin structure and function
ResumoSince ovarian carcinoma cells detach from the primary lesion and metastasize via peritoneal dissemination, we hypothesized that these cells are exposed to hypoxia, which may affect cell attachment and invasiveness. To address this hypothesis, we first examined in vivo the immunohistochemical expression of hypoxia-inducible factor-1α (HIF-1α) and its topological correlation with E-cadherin expression in ovarian carcinomas. We then examined in vitro the effect of hypoxia on the mRNA and protein expressions of E-cadherin using two ovarian cancer cell lines, SKOV3 and OVCAR3, and normal ovarian surface epithelial (OSE) cells. In addition, hypoxia-induced change in the expression of SNAIL, a transcriptional factor repressing E-cadherin expression, was also analyzed. Finally, we examined the facilitation of invasiveness of ovarian cancer cells under hypoxia using Matrigel invasion assay. Immunohistochemically, nuclear localization of HIF-1α was observed in 32 of the 76 (42%) carcinomas studied, and showed a topological correlation with loss of E-cadherin expression. Northern blotting, real-time PCR and Western blotting demonstrated that E-cadherin expression was remarkably decreased under hypoxia in both SKOV3 and OVCAR3 cells, but not in normal OSE cells. mRNA expression of SNAIL was increased under hypoxia in both ovarian cancer cell lines. Invasion assay revealed that hypoxia increases the invasiveness of ovarian cancer cells. Accordingly, the present study demonstrated that hypoxia induces down-regulation of E-cadherin in ovarian carcinoma cells, via up-regulation of the transcriptional repressor SNAIL. These findings suggest that hypoxia plays an important role in the change in intercellular attachment, which may be involved in the initiation of tumor progression of ovarian cancer cells. Since ovarian carcinoma cells detach from the primary lesion and metastasize via peritoneal dissemination, we hypothesized that these cells are exposed to hypoxia, which may affect cell attachment and invasiveness. To address this hypothesis, we first examined in vivo the immunohistochemical expression of hypoxia-inducible factor-1α (HIF-1α) and its topological correlation with E-cadherin expression in ovarian carcinomas. We then examined in vitro the effect of hypoxia on the mRNA and protein expressions of E-cadherin using two ovarian cancer cell lines, SKOV3 and OVCAR3, and normal ovarian surface epithelial (OSE) cells. In addition, hypoxia-induced change in the expression of SNAIL, a transcriptional factor repressing E-cadherin expression, was also analyzed. Finally, we examined the facilitation of invasiveness of ovarian cancer cells under hypoxia using Matrigel invasion assay. Immunohistochemically, nuclear localization of HIF-1α was observed in 32 of the 76 (42%) carcinomas studied, and showed a topological correlation with loss of E-cadherin expression. Northern blotting, real-time PCR and Western blotting demonstrated that E-cadherin expression was remarkably decreased under hypoxia in both SKOV3 and OVCAR3 cells, but not in normal OSE cells. mRNA expression of SNAIL was increased under hypoxia in both ovarian cancer cell lines. Invasion assay revealed that hypoxia increases the invasiveness of ovarian cancer cells. Accordingly, the present study demonstrated that hypoxia induces down-regulation of E-cadherin in ovarian carcinoma cells, via up-regulation of the transcriptional repressor SNAIL. These findings suggest that hypoxia plays an important role in the change in intercellular attachment, which may be involved in the initiation of tumor progression of ovarian cancer cells. Epithelial ovarian carcinoma is the leading cause of death from female genital malignancies, and more than half of patients are diagnosed at the advanced stage of the disease.1Murdoch WJ Ovarian surface epithelium, ovulation and carcinogenesis.Biol Rev Camb Philos Soc. 1996; 71: 529-543Crossref PubMed Scopus (54) Google Scholar The prognosis of patients with advanced ovarian cancer is most likely related to the degree of peritoneal dissemination. Although the process of cancer metastasis appears to be regulated by a variety of gene products,2Gunthert U Birchmeier W Schlag PM Attempts to understand metastasis formation: II. regulatory factors: introduction.Curr Top Microbiol Immunol. 1996; 213: V-VIIGoogle Scholar little is known about the molecular aspects of peritoneal dissemination of ovarian carcinoma cells. One of the important events in the first step of metastasis is the reduction of cellular adhesion between the tumor cells, facilitating invasion into the surrounding tissues and vascular channels.3Bernstein LR Liotta LA Molecular mediators of interactions with extracellular matrix components in metastasis and angiogenesis.Curr Opin Oncol. 1994; 6: 106-113Crossref PubMed Scopus (142) Google Scholar E-cadherin and the associated catenin complex have been recognized as performing key roles in cell adhesion. Reduced expression of E-cadherin has been reported in various human cancers, being associated with tumor progression.4Shiozaki H Oka H Inoue M Tamura S Monden M E-cadherin mediated adhesion system in cancer cells.Cancer. 1996; 77: 1605-1613PubMed Scopus (0) Google Scholar, 5Bukholm IK Nesland JM Karesen R Jacobsen U Borresen-Dale AL E-cadherin and α-, β-, and γ-catenin protein expression in relation to metastasis in human breast carcinoma.J Pathol. 1998; 185: 262-266Crossref PubMed Scopus (175) Google Scholar, 6Van Aken E De Wever O Correia da Rocha AS Mareel M Defective E-cadherin/catenin complexes in human cancer.Virchows Arch. 2001; 439: 725-751PubMed Google Scholar The search for factors affecting the progression and behavior of tumors has revealed the importance of the microenvironment surrounding the tumor cells. Peritoneal dissemination is a metastatic process in which the cancer cells detach from the primary tumor, attach to the peritoneum, and re-grow at this site. It is therefore hypothesized that ovarian cancer cells leaving the primary tumor experience lower oxygen levels.7Mutch DG Williams S Biology of epithelial ovarian cancer.Clin Obstet Gynecol. 1994; 37: 406-422Crossref PubMed Scopus (16) Google Scholar Hypoxia is known to induce hypoxia-inducible factor-1α (HIF-1α), which binds to the hypoxia-response elements of various target genes and activates the transcription of these genes controlling glucose transport, angiogenesis, erythropoiesis, and vasomotor regulation, and thus may increase the survival of tumor cells. 8Semenza GL Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1.Annu Rev Cell Dev Biol. 1999; 15: 551-578Crossref PubMed Scopus (1637) Google Scholar, 9Ratcliffe PJ O'Rourke JF Maxwell PH Pugh CW Oxygen sensing, hypoxia-inducible factor-1 and the regulation of mammalian gene expression.J Exp Biol. 1998; 201: 1153-1162PubMed Google Scholar, 10Horiuchi A Imai T Shimizu M Oka K Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of HIF-1α, VEGF and cell cycle-related molecules in ovarian cancer cells.Anticancer Res. 2002; 22: 2697-2702PubMed Google Scholar With the progressive and rapid growth of malignant tumors, cancer cells in the ischemic condition are expected to transform to the “metastatic” phenotype through a reduction in intercellular adhesion as well as an increase in cell motility and invasiveness. However, little is known about the hypoxic status in vivo and its functional relationship with intercellular adhesion in malignant tumors. Therefore, we first immunohistochemically examined the expression and localization of HIF-1α with relevance to those of E-cadherin in epithelial ovarian carcinomas. Based on the topological correlation between the two molecules, we hypothesized that hypoxia may affect the cell attachment and invasiveness of ovarian cancer cells. Our preliminary screening using microarray analysis showed possible hypoxia-induced changes in the expression of adhesion molecules including E-cadherin in ovarian cancer cells.11Imai T Horiuchi A Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of cell adhesion molecules in ovarian cancer cells.in: Mok JE Quinn MA Namkoong SE Kim YT 9th Biennial Meeting of the International Gynecologic Cancer Society. Monduzzi Editore, Bologna2002: 99-101Google Scholar In this study, therefore, we examined the effect of hypoxia on the change in expression of E-cadherin at the mRNA and protein levels in ovarian cancer cell lines, as well as in the normal counterpart, ovarian surface epithelial (OSE) cells. Hypoxia-induced change in the expression of SNAIL, a transcriptional factor repressing E-cadherin expression, was also analyzed. Finally, the hypoxic effect on cell invasiveness was examined by in vitro invasion assay. Seventy-six cases of primary epithelial ovarian carcinoma were selected from the pathology file of Shinshu University Hospital. The tissues were obtained from patients who underwent laparotomy at our department. None had received preoperative chemotherapy or radiotherapy. These specimens were fixed in 10% phosphate-buffered formalin and embedded in paraffin. Serial sections of 3-μm thickness were made for hematoxylin and eosin staining and for immunohistochemistry. Histologically, 27 of the 76 were serous, 7 were mucinous, 17 were endometrioid, and 25 were clear cell carcinomas. According to the classification of the International Federation of Gynecology and Obstetrics (FIGO), 37 were classified as stage I, 10 were stage II, 24 were stage III, and 5 were stage IV. Each tissue was used with the approval of the Ethical Committee of Shinshu University, after obtaining written informed consent from the patients. Immunohistochemical staining was performed on paraffin-embedded sections by the streptavidin-biotin-peroxidase complex method using a Histofine SAB-PO kit (Nichirei, Tokyo, Japan). The primary antibodies were anti-E-cadherin monoclonal antibodies (number 20820, clone 36) (Transduction Laboratories, Lexington, KY) and were used at a dilution of 1:250. After deparaffinization and rehydration, the sections were boiled in 0.01 mol/L citrate buffer (pH 6.0) for 15 minutes in a microwave oven. They were then treated with 0.3% hydrogen peroxide and incubated with 10% normal rabbit serum to block non-specific binding. The sections were incubated with a primary antibody at 4°C overnight. After washing in phosphate-buffered saline (PBS), they were incubated with biotinylated goat anti-mouse IgG, followed by treatment with peroxidase-conjugated streptavidin and stained with diaminobenzidine with 0.15% hydrogen peroxide. Counterstaining was performed with hematoxylin. Immunoreactivity was semiquantitatively estimated: negative (-) corresponded to 0% positive cells, weakly positive (+) to 1% to 50% positive cells, and strongly positive (++) to >50% positive cells. Serial, formalin-fixed and paraffin-embedded sections were used for HIF-1α immunostaining using the Catalyzed Signal Amplification System (DAKO, Carpinteria, CA), as described previously.10Horiuchi A Imai T Shimizu M Oka K Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of HIF-1α, VEGF and cell cycle-related molecules in ovarian cancer cells.Anticancer Res. 2002; 22: 2697-2702PubMed Google Scholar In brief, after deparaffinization and rehydration, the sections were treated with target retrieval solution (DAKO) at 95°C for 45 minutes. Primary antibody, mouse anti-HIF-1α monoclonal antibody (Novus Biologicals, Littleton, CO), was used at a dilution of 1:1000. They were counterstained with hematoxylin. Specific immunoreactivity for HIF-1α was observed in the cytoplasm and/or in the nuclei of the tumor cells. Cytoplasmic immunostaining was semiquantitatively estimated: negative (-) corresponded to 0% positive cells, weakly positive (+) to 1% to 50% positive cells, and strongly positive (++) to >50% positive cells. Since nuclear immunostaining was observed sporadically in the tumor cells, the cases were classified as positive (presence of tumor cells with nuclear staining) or negative (absence of tumor cells with nuclear staining). The ovarian cancer cell lines OVCAR3 and SKOV3 were purchased from the ATCC (Rockville, MD). SKOV3 was cultured in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA) with 5% fetal calf serum (FCS) (Biological Industries, Grand Island, NY) and 1% antibiotic-antimycotic solution (Invitrogen). OVCAR3 were cultured in RPMI-1640 (Sigma, St. Louis, MO) with 5% FCS and 1% antibiotic-antimycotic solution. Incubation was carried out at 37°C under 5% CO2 in air. Ovarian surface epithelium was obtained from five women who were treated surgically for benign gynecological disease, after obtaining written informed consent from each patient. At the operation, OSE cells were collected by scraping the surface of the ovaries with a surgical blade, and were immediately cultured on collagen type I-coated plastic dishes (Iwaki Glass, Chiba, Japan) in a 1:1 mixture of Medium 199 (Invitrogen) and MCDB 105 (Sigma) containing 15% FCS (Biological Industries).10Horiuchi A Imai T Shimizu M Oka K Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of HIF-1α, VEGF and cell cycle-related molecules in ovarian cancer cells.Anticancer Res. 2002; 22: 2697-2702PubMed Google Scholar The purity of cultured OSE cells was confirmed by positive immunostaining for cytokeratin and negative immunostaining for factor VIII.10Horiuchi A Imai T Shimizu M Oka K Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of HIF-1α, VEGF and cell cycle-related molecules in ovarian cancer cells.Anticancer Res. 2002; 22: 2697-2702PubMed Google Scholar For the reverse transcription-polymerase chain reaction (RT-PCR) and Western blotting, the cultured cells were transferred to a 75-cm2 flask (5 × 105Bukholm IK Nesland JM Karesen R Jacobsen U Borresen-Dale AL E-cadherin and α-, β-, and γ-catenin protein expression in relation to metastasis in human breast carcinoma.J Pathol. 1998; 185: 262-266Crossref PubMed Scopus (175) Google Scholar cells/flask) and incubated at 37°C under 20% O2 as control, or at 37°C under 5% O2 as hypoxia for 72 hours.10Horiuchi A Imai T Shimizu M Oka K Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of HIF-1α, VEGF and cell cycle-related molecules in ovarian cancer cells.Anticancer Res. 2002; 22: 2697-2702PubMed Google Scholar, 11Imai T Horiuchi A Wang C Nikaido T Konishi I Hypoxia-induced changes in the expression of cell adhesion molecules in ovarian cancer cells.in: Mok JE Quinn MA Namkoong SE Kim YT 9th Biennial Meeting of the International Gynecologic Cancer Society. Monduzzi Editore, Bologna2002: 99-101Google Scholar For in vitro invasion assay, it was difficult to perform the experiments using normal OSE cells. Instead, we used SV40 large T-immortalized OSE cells (IOSE 398), kindly provided by Dr. Nelly Auersperg at The University of British Columbia, Vancouver, Canada. Total RNA was extracted by the acid guanidinium-phenol-chloroform method as described previously.12Chomczynski P Sacchi N Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction.Anal Biochem. 1987; 162: 156-159Crossref PubMed Scopus (62909) Google Scholar, 13Horiuchi A Nikaido T Ito K Zhai Y Orii A Taniguchi S Toki T Fujii S Reduced expression of calponin h1 in leiomyosarcoma of the uterus.Lab Invest. 1998; 78: 839-846PubMed Google Scholar One microgram of total RNA was treated with 1 U/10 μl DNase I (Life Technologies, Gaithersburg, MD). RT was performed using an RNA PCR Kit (Takara Shuzo, Otsu, Japan), the 1 μg RNA sample being added to 20 μl of a reaction mixture consisting of 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 5 mmol/L MgCl2, 1 mmol/L dNTP mixture, 1 U/μl of RNase inhibitor, 0.25 U/μl of avian myeloblastosis virus-derived reverse transcriptase, and 0.125 μmol/L of oligo d(T)-adaptor primer. Using a thermal cycler (Gene Amp PCR System 2400-R; PerkinElmer, Norwalk, CT), the reaction mixture was incubated at 42°C for 30 minutes, heated at 99°C for 5 minutes, and then cooled down to 5°C for 5 minutes. One microliter of the RT products, containing 50 ng reverse-transcribed total RNA, was amplified by adding 20 μl of a PCR reaction mixture containing 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 2.5 U/100 μl of TaKaRa TaqDNA polymerase, with 0.2 μmol/L of a set of 20- to 21-mer oligonucleotide primers either for E-cadherin, HIF-1α, or SNAIL and for glyceraldehyde-3-phosphate dehydrogenase (G3PDH) cDNA. Primers were synthesized to encompass a specific segment of the cDNA sequence of E-cadherin14Mialhe A Levacher G Champelovier P Martel V Serres M Knudsen K Seigneurin D Expression of E-, P-, N-cadherins and catenins in human bladder carcinoma cell lines.J Urol. 2000; 164: 826-835Abstract Full Text Full Text PDF PubMed Google Scholar (sense, 5′-TCCATTTCTTGGTCTACGCC′ and antisense, 5′-CACCTTCAGCCAACCTGTTT-3′), HIF-1α15Wang GL Jiang BH Rue EA Semenza GL Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension.Proc Natl Acad Sci USA. 1995; 92: 5510-5514Crossref PubMed Scopus (4928) Google Scholar (sense, 5′-CCCTGCAGTAGGTTTCTGCT-3′ and antisense, 5′-CTCAAAGTCGGACAGCCTCA-3′),SNAIL16Cheng CW Wu PE Yu JC Huang CS Yue CT Wu CW Shen CY Mechanisms of inactivation of E-cadherin in breast carcinoma: modification of the two-hit hypothesis of tumor suppressor gene.Oncogene. 2001; 20: 3814-3823Crossref PubMed Scopus (194) Google Scholar (sense, 5′-AATCGGAAGCCTAACTACAAG-3′ and antisense, 5′-AGGAAGAGACTGAAGTAGAG-3′), or of G3PDH (sense, 5′-ACGACCACTTTGTCAAGCTC-3′ and antisense 5′-TCACA GTTGCCATGTAGACC-3′, spanning between exons 7 and 8). The corresponding cDNA fragments were denatured at 94°C for 30 seconds, annealed at 58°C for 1 minute, and extended at 72°C for 1 minute. After 35 cycles of amplification, the PCR products were analyzed on a 2% agarose gel, and the bands were visualized using ethidium bromide during exposure to an UV transilluminator. The density of the bands on the gel was quantified by densitometric analysis using a Quantity One Scan System (ATTO, Tokyo, Japan). Gene expression was presented by the relative yield of PCR product from the target sequence to that from the G3PDH gene. Mean values from three independent experiments were taken as results. To confirm the results of RT-PCR, mRNA expression was also analyzed using real-time PCR (Roche LightCycler; Roche Diagnostics, Mannheim, Germany). A master-mix of the following reaction components was prepared to the indicated end-concentration: 6.4 μl water, 1.2 μl MgCl2 (4 mmol/L), 0.2 μl forward primer (0.4 mol/L), 0.2 μl reverse primer (0.4 mol/L) and 1.0 μl LightCycler Fast Start DNA Master SYBR Green I (Roche Diagnostics). Nine microliters of master-mix was added to glass capillaries and 1 μl volume, containing 50 ng reverse-transcribed total RNA, and was added as a PCR template. The capillaries were closed, centrifuged, and placed into the rotor. To improve SYBR Green I quantification, the temperature of the fluorescence measurement point was set at 72°C. At the completion of cycling, melting curve analysis was carried out to establish the specificity of the amplicons produced. The level of expression of each mRNA and their estimated crossing points in each sample were determined relative to the standard preparation using the Light Cycler computer software. A ratio of specific mRNA/G3PDH amplification was then calculated to correct for any differences in efficiency. The relative abundance of the mRNAs, expressed as fold changes, was extrapolated from crossing point data. A difference of one PCR cycle in crossing-point number represents a twofold change in mRNA expression. Fifteeen micrograms of total RNA were electrophoresed on a denaturing 1% agarose gel containing formaldehyde and transferred to Hybond-N membranes (Amersham Biosciences Inc., Buckinghamshire, UK). The membranes were prehybridized at 42°C for 6 hours in 50% formamide, 5X saline-sodium phosphate-EDTA, 5X Denhardt's solution, 1% sodium dodecyl sulfate (SDS), and 100 μg/ml denatured herring sperm DNA. For probes, genomic DNA obtained from OSE was used as a template, [γ32P]deoxy-CTP (Amersham Biosciences Inc.) was labeled by PCR using templates and primers specific for E-cadherin, and G3PDH was used as an internal control. After prehybridization, the filters were hybridized with probe at 42°C for 12 hours. The hybridized filters were washed in 1X SSC with 0.1% SDS at 55°C for 20 minutes. The bands were analyzed using a MacBas system (Fuji Photo Film, Tokyo, Japan) and were then quantitated by densitometric analysis using a Quantity One Scan System. Cells were lysed in a lysis buffer: 50 mmol/L Tris-HCl (pH 8.0), 0.25 mol/L NaCl, 0.5% NP-40, 1 mmol/L PMSF (Sigma), 1 μg/ml aprotinin (Roche Diagnostics), 1 μg/ml leupeptin (Roche Diagnostics), and 20 μg/ml TPCK (Roche Diagnostics). The lysates were centrifuged at 13,000 × g for 20 minutes at 4°C and the supernatants were stored at −80°C. Extracts equivalent to 30 μg of total protein were separated by SDS-polyacrylamide gel electrophoresis (8% acrylamide) and transferred onto nitrocellulose membranes (Hybond TM-C super; Amersham Biosciences Inc.). The membranes were blocked in TBST (0.2 mol/L NaCl, 10 mmol/L Tris, pH 7.4, 0.2% Tween-20) containing 5% nonfat dry milk and 0.02% NaN3 for 1 hour, then incubated with mouse monoclonal antibodies against E-cadherin, HIF-1α and β-actin (Santa Cruz, St Louis, MO) in TBST containing 5% nonfat dry milk. The membranes were then incubated with sheep anti-mouse Ig (Amersham Biosciences Inc.) in TBST containing 2% nonfat dry milk. Bound antibodies were detected with an enhanced chemiluminescence system (Amersham Biosciences Inc.). Cell invasion through reconstituted basement membrane Matrigel was assayed by a method reported previously.17Albini A Iwamoto Y Kleinman HK Martin GR Aaronson SA Kozlowski JM McEwan RN A rapid in vitro assay for quantitating the invasive potential of tumor cells.Cancer Res. 1987; 47: 3239-3245PubMed Google Scholar Briefly, polycarbonate membranes (8.0-μm pore size) of the upper compartment of transwell culture chambers were coated with 5% Matrigel (BD Biosciences, Bedford, MA). Subconfluent cells were starved for 24 hours and harvested with 0.05% trypsin containing 0.02% EDTA, washed twice with PBS, and resuspended at 1 × 106Van Aken E De Wever O Correia da Rocha AS Mareel M Defective E-cadherin/catenin complexes in human cancer.Virchows Arch. 2001; 439: 725-751PubMed Google Scholar cells/ml in serum-free medium with 0.1% bovine serum albumin (BSA). The cell suspension (500 μl) was placed in the upper compartment, and the lower compartment was immediately filled with 500 μl of serum-free medium containing 0.1% BSA. After 20 hours of incubation, at 37°C under 20% O2 as control, or at 37°C under 5% O2 as hypoxia, the cells on the upper surface of the filter were removed carefully with a cotton swab, the membranes were stained with Diff-Quik solution (Kokusai-Shiyaku, Kobe, Japan) and the cells that had migrated through the membrane to the lower surface were counted in five different fields under a light microscope at ×100 magnification. Each experiment was performed in triplicate wells and repeated three times. The Kruskal-Wallis test, Scheffé test, and Mann-Whitney U test were used to compare the immunoreactivity of E-cadherin and HIF-1α according to the FIGO stage and histological type. The Mann-Whitney U-test was used to assess the differences in the invasion assay. Differences were considered to be significant at P < 0.05. These analyses were made using the StatView system (Abacus, Berkeley, CA). The immunohistochemical expression of E-cadherin was observed along the cytoplasmic membrane (Figure 1), and strongly positive in 56 (74%), weakly positive in 17 (22%), and negative in 3 (4%) of the 76 ovarian carcinomas (Table 1). According to the FIGO stage classification, strong expression of E-cadherin was found in 32 of the 37 stage I (86%), 7 of the 10 stage II (70%), 17 of the 24 stage III (71%), and none of the 5 stage IV (0%) tumors. Strong immunostaining of E-cadherin was significantly higher in stage I/II (83%) than in stage III/IV (59%) (P < 0.05). Regarding the histological type, the expression of E-cadherin was strongly positive in 17 of the 27 serous (63%), 7 of the 7 mucinous (100%), 12 of the 17 endometrioid (71%), and 20 of the 25 clear cell (80%) carcinomas. There was no significant difference according to the histological type.Table 1Immunohistochemical Expression of E-Cadherin in Ovarian Carcinomas According to FIGO Stage and Histological TypeTotalE-cadherin expression%−+++≧++762185656/7674FIGO stage I37053232/3786I+II II100377/107039/47 (83%)*P < 0.05. III24071717/2471III+IV IV52300/5017/29 (59%)*P < 0.05.Histology Serous27281717/2763 Mucinous70077/7100 Endometrioid17051212/1771 Clear cell25052020/2580−, 0% positive cells; +, 1–50% positive cells; ++, >50% positive cells.* P < 0.05. Open table in a new tab −, 0% positive cells; +, 1–50% positive cells; ++, >50% positive cells. The immunohistochemical expression of HIF-1α in the nuclei was observed in 32 of the 76 (42%) ovarian carcinomas (Table 2). According to the FIGO stage, the nuclear expression of HIF-1α was found in 17 of the 37 stage I (46%), 4 of the 10 stage II (40%), 9 of the 24 stage III (38%), and 2 of the 5 stage IV (40%) tumors. With regard to the histological type, it was observed in 11 of the 27 serous (41%), 2 of the 7 mucinous (29%), 6 of the 17 endometrioid (22%), and 13 of the 25 clear cell carcinomas (52%) (Table 2). There were no significant differences in the nuclear expression of HIF-1α according to the FIGO stage or histological type. Irrespective of the stage and histology, however, cancer cells with nuclear HIF-1α immunoreactivity were observed frequently in the tip of the papillary projection of the tumor (Figure 2a) or in the vicinity of the necrotic area (Figure 2b).Table 2Cytoplasmic and Nuclear Expression of HIF-1α in Ovarian Carcinomas According to FIGO Stage and Histological TypeTotalCytoplasmic expressionNuclear expression−+++≧+%−+%7614402262/7682443242FIGO stage I377161430/3781201746 II103527/10706440 III24317421/248815938 IV51224/5803240Histology Serous27514822/2782161141 Mucinous73314/7575229 Endometrioid1729615/178811635 Clear cell25414721/2584121352Cytoplasmic immunostaining was estimated as follows: −, 0% positive cells; +, 1–50% positive cells; ++, >50% positive cells.Nuclear immunostaining was classified as follows: −, absence of tumor cells with nuclear staining; +, presence of tumor cells with nuclear staining. Open table in a new tab Cytoplasmic immunostaining was estimated as follows: −, 0% positive cells; +, 1–50% positive cells; ++, >50% positive cells. Nuclear immunostaining was classified as follows: −, absence of tumor cells with nuclear staining; +, presence of tumor cells with nuclear staining. The immunohistochemical expression of HIF-1α in the cytoplasm was strongly positive in 22 (29%) (Figure 2, c and d), weakly positive in 40 (53%), and negative in the remaining 14 (18%) of the 76 ovarian carcinomas. There were no significant differences in the cytoplasmic HIF-1α expression according to the FIGO stage or histological type (Table 2). Closer observation of the immunoreactivity for E-cadherin and HIF-1α using the serial sections disclosed that the tumor cells with nuclear HIF-1α expression were associated with reduced expression of E-cadherin compared with the surrounding tumor cells that were negative for HIF-1α (Figure 3). Such reduced expression of E-cadherin along with nuclear HIF-1α expression was observed in 21 of the 32 cases (66%), especially in the cells at the tip of the papillary projection of the tumor (Figure 3, a and b) or in the vicinity of the necrotic area (Figure 3, c and d). In the remaining 11 cases, E-cadherin expression was preserved, although the tumor cells expressed HIF-1α in the nuclei. With regard to the expression of HIF-1α in the cytoplasm, there was no apparent topological correlation with reduced expression of E-cadherin. However, the 32 cases with nuclear HIF-1α expression were either strongly positive (14 cases) or weakly positive (18 cases) for cytoplasmic HIF-1α expression. Reduced expression of E-cadherin was observed in 9 of the 14 cases with strong expression, and in 12 of the 18 cases with weak expression for HIF-1α in the cytoplasm. To clarify whether hypoxic conditions decrease E-cadherin mRNA expression, we performed Northern blotting. Specific bands for E-cadherin and G3PDH mRNA were detected (Figure 4A). In the 2 ovarian cancer cell lines, the band intensities for E-cadherin were reduced under hypoxia compared with those under normoxia by 19% in SKOV3 and by 36% in OVCAR3. In normal OSE cells, however, the expression of E-cadherin was very low and showed no difference between the cells cultured in hypoxia and normoxia (Figure 4, A and B). Specific bands for E-cadherin and HIF-1α under normoxia were detected by Western blotting in both SKOV3 and OVCAR3, but not in OSE cells (Figure 5A). β-actin was expressed at 42 kd in all of the samples. In the ovarian cancer cell lines, the expression of E-cadherin was decreased under hypoxia by 21% in SKOV3 and by 56% in OVCAR3 (Figure 5, A and B). On the other hand, the expression of 120 kd HIF-1α was increased under hypoxia by 190% in SKOV3 and by 228% in OVCAR3 (Figure 5A). Apparently, normal OSE cells did not express E-cadherin or HIF-1α, and hypoxia did not affect the expression of either E-cadherin or HIF-1α. To address the mechanism involved in hypoxia-induced change in the E-cadherin expression, we examined the mRNA expression of SNAIL, a transcriptional
Referência(s)