Estrogen Inhibits Cell Proliferation Through in Situ Production in Human

Human Cancer Biology

Estrogen Inhibits Cell Proliferation through In situ Production in Human Thymoma Hironori Ishibashi,1,2,4 Ta k a s hi S u z u k i, 1Satoshi Suzuki,2 Takuya Moriya,1Chika Kaneko,1 Taisuke Nakata,5 Makoto Sunamori,4 Masashi Handa,3 Ta k a s hi Ko n d o, 2 and Hironobu Sasano1

Abstract Purpose: We showed previously estrogen receptor (ER) a as an independent prognostic marker in human thymoma. Estrogen sulfotransferase (EST), steroid sulfatase (STS), 17h- hydroxysteroid dehydrogenase (17h-HSD), and aromatase are considered to play important roles in hormone metabolism of estrogen-dependent tumors. Experimental Design: We examined estrogen production using primary cultures of human thymoma epithelial cells (TEC), intratumoral estradiol (E2) concentrations, and status of these enzymes above using immunohistochemistry or semiquantitative reverse transcription-PCR. We then correlated these findings with clinicopathologic variables and/or clinical outcome in 132 patients. Results: E2 inhibited cell proliferation via ERa inTEC, which synthesized estrone and E2.Intratu- moral E2 concentrations were inversely correlated with EST, positively correlated with STS or 17 h-HSD type 1, and significantly higher in lower-grade or early-stage thymoma. ESTstatus was positively correlated with tumor size, clinical stage, histologic differentiation, and Ki-67 labeling index and significantly associated with adverse clinical outcome and turned out to be a potent independent prognostic factor. STS and/or 17h-HSD type 1status was inversely correlated with Ki-67 labelingindexandassociated withlowerhistologic grade orearlyclinical stages. Conclusions: E2 inhibits proliferation of TEC through ERa, which suggests that E2 may be effective in treatment of thymoma, especially inoperable tumor, possibly through suppressing its cell proliferation activity. EST status is a potent prognostic factor in thymoma through inac- tivating estrogens. In situ estrogen synthesis through intracrine mechanism therefore may play important roles in tumorigenesis and/or development of thymoma through regulation of cell proliferation in an intracrine manner.

Estrogen regulates cell proliferation and/or other biological relatively long half-life in the peripheral blood (4), in which functions in various neoplasms derived from hormone- serum levels of E1-S are 10-fold higher than those of dependent tissues, such as breast and endometrial cancer (1). unconjugated E1 or estradiol (E2; ref. 5). E1-S is transformed In situ estrogen levels are also considered to play very important into a biologically active form, E1, by steroid sulfatase (STS; roles in the pathogenesis of these neoplasms (2). We reported refs. 6, 7), whereas aromatase catalyzes circulating androgen previously that estrogens play an important role via binding androstenedione into E1 (8). E1 is subsequently converted to to estrogen receptor (ER) a in the development of human E2 by 17h-hydroxysteroid dehydrogenase (17h-HSD) type 1 in thymomas (3). several estrogen target tissues, including breast cancer cells (9). A major circulating form of plasma estrogen is estrone sulfate E2 acts on breast cancer cells primarily via its binding to ERa (E1-S), a biologically inactive form of estrogen. E1-S has a and/or ERh. STS, aromatase, and 17h-HSD type 1 have been all considered to be involved in in situ production of estrogen in neoplastic tissues, such as breast cancer. Estrogen sulfotransfer- Authors’ Affiliations: 1Department of Pathology, Tohoku University School of ase (EST), SULT1E1 or STE gene, is a member of the 2 Medicine; Department of Thoracic Surgery, Institute of Development, Aging superfamily of cytosolic steroid sulfotransferases (10) and and Cancer, Tohoku University, Sendai, Japan; 3Department of Thoracic Surgery, Iwate General Hospital, Morioka, Japan; 4Thoracic Cardiovascular Surgery, catalyzes E1 into biologically inactive E1-S. Graduate School,Tokyo Medical and Dental University; and 5Kyowa Hakko Kogyo In this study, we examined the expression and biological Co., Ltd.,Tokyo, Japan significance of EST, STS, 17h-HSD type 1, and aromatase in Received 12/6/04; revised 4/26/05; accepted 5/31/05. human thymoma to further elucidate the status of in situ The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance estrogen production in thymoma. We first studied effects of with 18 U.S.C. Section 1734 solely to indicate this fact. estrogens on cell proliferation and in situ E1 or E2 production Requests for reprints: Hironori Ishibashi, Department of Thoracic Surgery, from E1-S using primary cultures of thymoma epithelial cells Institute of Development, Aging and Cancer,Tohoku University, 4-1Seiryo-machi (TEC). We then determined E2 concentrations and immunolo- Aoba-ku, Sendai, Miyagi Prefecture 980-0875, Japan. Phone: 81-22-717-8521; h Fax: 81-22-717-8526; E-mail: [email protected]. calized EST, STS, and 17 -HSD type 1 in human thymoma. In F 2005 American Association for Cancer Research. addition, mRNA levels of EST, STS, 17h-HSD type 1, and doi:10.1158/1078-0432.CCR-04-2495 aromatase were examined using semiquantitative reverse

www.aacrjournals.org 6495 Clin Cancer Res 2005;11(18) September 15, 2005 Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Human Cancer Biology transcription-PCR in human thymoma. We then correlated EST, STS, or 17h-HSD type 1 immunoreactivity with clinicopathologic factors and clinical outcome in 132 patients with thymomas to establish clinical significance of these findings.

Materials and Methods

Cell culture analysis Establishment of thymoma-derived neoplastic epithelial cellsin primary culture. Fresh thymoma tissues (five thymoma cases: two cases for type A, two cases for type B2, and one case for type B3), which expressed ERa, STS, or 17h-HSD type 1 but not ERh were repeatedly washed in PBS and carefully minced and digested in a collagenase- containing mixture composed of collagenase type I (1 mg/mL; Wako, Osaka, Japan), DMEM, 10% fetal bovine serum (FBS), and antibiotics (penicillin and streptomycin) for 8 hours at 37jC (11). These minced and digested thymoma tissues were subsequently filtrated through three layers of sterile gauze and cultured by a 10-day culture in DMEM supplemented with 10% FBS and antibiotics but without any growth factors. On confluence, cells were mechanically dislodged by 0.25% trypsin and 0.02% EDTA and subjected to secondary cultures. Cells were subsequently characterized by their morphology and immunoreactivity of cytokeratin using immunocytostaining. Coexisting fibroblasts were eliminated by anti-fibroblast microbeads (MACS, Miltenyi Biotec, Gladbach, Germany) according to the manufacturer’s instructions. The purity of the TEC, which expressed ERa, STS, or 17h-HSD type 1 but not ERh, was examined by flow cytometry and the proportion of cytokeratin-positive cells was nearly always >97% (data not shown). TEC was grown to confluence in DMEM in the presence of 10% FCS and antibiotics under standard concentrations. Cell proliferation assay. TEC was seeded at a density of 5 Â 103 per mL into 24-well plates (Falcon; Becton Dickinson Labware, Lincoln Park, NJ) in a final volume of 1 mL phenol red–free DMEM

Ta b l e 1. Summary of clinical data in132 patients and in 3 20 patients with thymoma examined in this study Fig. 1. Effects of E2 and anti-estrogen on cell growth ofTEC. Cells (5 Â 10 /mL) were plated in triplicate well onto a 24-well plates. After 24 hours for cell attachment, cells were treated for 73 hours with E2 at various concentrations with or without n =132 n =20 anti-estrogen, ICI182,780. Cells treated with vehicle (absolute ethanol) were used as control. Columns, mean (n = 5); bars, SD. *, P < 0.001, compared with control. Age (y)* 54.0 F 15.1 5 4.4 F 13. 2 Sex Male 58 (43) 10 (50) with 5% charcoal-stripped FBS. Medium was then replaced with Female 74 (57) 10 (50) phenol red–free DMEM without FBS to arrest the growth in 24 Premenopausal 38 (29) 6 (30) hours after seeding. The medium was replaced again with phenol Postmenopausal 36 (28) 4 (20) red–free DMEM with 5% charcoal-stripped FBS together with either Myasthenia gravis a vehicle (0.1%), increasing concentration of E2 (0.1, 1, 10, 100, and 1,000 nmol/L), and/or 100 nmol/L ICI 182,780 in 24 hours later. +24(18)3(15) Following 72-hour incubation with these agents, the cells were tryp- À 108(82) 17(85) sinized and resuspended. E was purchased from Sigma (St. Louis, F F 2 Tumor size (mm)* 60.0 25.5 61.3 17.0 MO) and ICI 182,780 from Tocris Cookson Ltd. (Ellisville, MO). Clinical stage Estrogens and anti-estrogens were dissolved in absolute ethanol I 67(51) 10(50) (Sigma) and added to the medium daily. Cell cultures that were not II 28 (21) 4 (20) treated with estrogenic compounds received absolute ethanol as a III 21 (16) 3 (15) vehicle control. Total additive ethanol concentrations never exceeded IV 16 (12) 3 (15) 0.2% throughout the culture period. The cells were refed with freshly WHO classification prepared medium every other day. We then used a Cell Counting A28(21)4(20)Kit-8 system (Dojindo Technologies, Kumamoto, Japan) to determine the cell number. Then, 100 AL from each sample was aliquoted into AB 23 (17) 4 (20) a 96-well microtiter plate with 10 AL of the working solution B1 29 (22) 4 (20) containing WST-8 and then incubated for an additional 2 hours in a B2 39 (30) 5 (25) CO2 incubator at 37jC. The absorbance of each well was measured B3 13(10) 3 (15) at 450 nm with a reference wavelength at 650 nm with a M-UVmax microscope reader (Molecular Devices Corp., Menlo Park, CA). An *Mean F SD. All other values represent n (%). aliquot was taken from the medium to count the number of cells with a Burker-Turk counter (Nitirin, Tokyo, Japan).

Clin Cancer Res 2005;11(18) September 15, 2005 6496 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Estrogen Inhibits Cell Proliferation in Human Thymoma

Terminal deoxynucleotidyl transferase–mediated dUTP nick-end label- Tohoku University School of Medicine, Sendai Kousei Hospital and ing analysis. Visualization of apoptotic cells in chamber slides was Tokyo Medical and Dental University School of Medicine. done using the terminal deoxynucleotidyl transferase–mediated dUTP Antibodies. Rabbit polyclonal antibody for EST was raised against nick end labeling method (Apoptosis In situ Detection kit, Wako; ref. the synthetic NH2-terminal peptide of human EST corresponding to 12). As a negative control, fixed and permeabilized cells incubated amino acids 1 to 13 (PV-P2237, Medical Biological Laboratory, Nagoya, without terminal deoxynucleotidyl transferase, but with secondary Japan; 1:750 dilution). STS antibody was raised against the enzyme antibodies and 3,3V-diaminobenzidine, were used. As a positive control, purified from human placenta, which recognizes the STS peptide permeabilized cells were incubated with bovine pancreatic DNase I corresponding to amino acids 420 to 428 (KM1049, kindly provided by before terminal deoxynucleotidyl transferase treatment to induce DNA Kyowa Hakko Kogyo Co., Ltd., Tokyo, Japan; 1:3,000 dilution) and has fragmentation. The presence of terminal deoxynucleotidyl transferase– been reported previously in the evaluation of human breast cancer (15). mediated dUTP nick end labeling positivity was determined by visual 17h-HSD type 1 antibody was a rabbit polyclonal antibody against the inspection of cultures using light microscopy. enzyme purified from human placenta (kindly provided by Dr. Van

Estrone and estradiol production from thymoma epithelial cells. E1-S Luu-The, Laboratory of Molecular Endocrinology, CHUL Research was added after a 48-hour culture in phenol red and FCS–free DMEM Center, Quebec, Quebec, Canada; ref. 16). (final concentration, 0.5, 2.5, and 12.5 nmol/L). Medium (2 mL) was The antibodies used were mouse monoclonal ERa (NCLER-6F11, extracted twice with diethyl ether. Extracts were evaporated to dryness Novocastra Lab, Newcastle, United Kingdom; 1:50 dilution), proges- under nitrogen and reconstituted in 100 AL dextran-coated, charcoal- terone receptor-B (PR-B; hPRa2, NeoMarkers Co. Ltd., Fremont, CA; treated FBS. E1 and E2 levels of the extracts were then determined by RIA 1:200 dilution), and Ki-67 (MIB-1, Immunotech, Marseilles, France; using an assay kit (Diagnostic Products, Tokyo, Japan). The sensitivity 1:50 dilution) described in our previous study (3). of the assay was 3 pmol/L. Immunohistochemistry. Immunohistochemical procedures employ-

Estradiol concentrations in human thymoma and thymus. E2 ed in this study have been described previously in detail (2, 17, 18). concentrations were examined in 20 cases of fresh frozen thymoma. Briefly, immunohistochemical staining was done by the streptavidin- These specimens were carefully dissected to eliminate all adjacent biotin method with a Histofine kit (Nichirei Co. Ltd., Tokyo, Japan). The adipose and connective tissues. Cytosol and nuclear fractions were antigen-antibody complex was then visualized with 3,3V-diaminobenzi- prepared by centrifugation of the homogenates (1.0 g specimens in dine solution and counterstained with hematoxylin. Antigen retrieval for 10 mL cold physiologic saline) at 4jC for 60 minutes at 15,000 Â g in ERa, PR-B, and Ki-67 immunostaining was done by heating the slides in an ultracentrifuge. Tissue E2 concentrations were determined according an autoclave at 120jC for 5 minutes in citric acid buffer. Similarly, to the methods described above. antigen retrieval for EST immunostaining was done by heating the slides Patients and tissue specimens. Thymoma cases (n = 132) were in a microwave (500 W) for 15 minutes in citric acid buffer. No antigen retrieved from surgical pathology files at Sendai Kousei Hospital retrieval was done for STS and 17h-HSD type 1 immunostaining. As a (Sendai, Japan) and Tokyo Medical and Dental University (Tokyo, positive control, normal liver was used for EST (18), normal full-term Japan). All specimens were fixed for 24 hours in 10% formalin at room placenta for STS (18) and 17h-HSD type 1 (19), breast cancer for ERa, temperature and embedded in paraffin wax. Relevant clinical findings, and endometrium for PR-B (3). Normal rabbit and mouse IgG was used including patient age, sex, menopausal status, presence or absence of instead of the primary antibody as a negative control. No specific myasthenia gravis, tumor size, clinical stage (13), and WHO histologic immunoreactivity was detected in these tissue sections. classification (14) are summarized in Table 1. Scoring of immunoreactivity. Immunoreactivity of EST, STS, and Freshly frozen specimens were also available for reverse transcrip- 17h-HSD type 1 was analyzed according to a previously described tion-PCR analysis, real-time PCR, and E2 concentration analysis in 20 method (20). After completely reviewing all the slides of immunos- cases (Table 1) among these 132 cases of thymoma examined for tained sections for each thymoma, two of the authors (H.I. and T.S.) immunohistochemistry. Specimens for RNA isolation were immediate- independently and blindly divided the thymomas into the following ly frozen in liquid nitrogen and stored at À80jC until used for RNA three groups: ++, >50% immunopositive cells; +, 1% to 50% isolation. RNA was extracted within 2 weeks following surgery. immunopositive cells; and À, no immunoreactivity. Cases with Informed consent was obtained from all the patients. The research discordant results among the observers were simultaneously reeval- protocols used in this study were approved by the Ethics Committee at uated using a multiheaded microscope. Semiquantitative analysis of

Fig. 2. Time course of E1 (A)andE2 (B) concentration produced from E1-S in primary culture of TEC. E1 and E2 concentrations were significantly increased in proportion to the duration and concentration of E1-S. Points, mean of three determinations; bars, SE. *, P < 0.05.

www.aacrjournals.org 6497 Clin Cancer Res 2005;11(18) September 15, 2005 Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Human Cancer Biology

Ta b l e 2 . Correlation between E2 tissue concentrations and other variables in 20 human thymomas

c E2 tissue concentration (pg/g)* P r EST immunoreactivity ++ (n =4) 27.0F 3.6 0.013 +(n = 8) 48.3 F 7.5 À (n =8) 95.1F15.8 STS immunoreactivity ++ (n =3) 126.7F 32.9 0.008 +(n =9) 65.7F 9.4 À (n =8) 35.5F 5.0 17 h-HSD type1immunoreactivity ++ (n = 4) 125.5 F 21.3 0.007 +(n =7) 65.3F 7.7 À (n = 9) 32.9 F 3.3 Aromatase mRNA levelb 0.601 (P =0.002) ERa H-scoreb 0.774 (P < 0.001) PR-B H-scoreb 0.749 (P =0.001) Clinical stage I(n =9) 93.9F 14.3 0.006 II (n =5) 47.4F 5.7 III (n = 3) 32.3 F 4.9 IV (n =3) 25.3F 4.1

*Mean F SEM. cP was listed only if significant. P < 0.05 was considered significant. bMean F 95% confidence interval.

immunoreactivity of ERa and PR-B H-score and Ki-67 labeling index 1, and aromatase (17). Negative control experiments lacked cDNA (LI) were done according to a previous report (3, 21, 22). Cases substrate to check for the presence of exogenous contaminant DNA. No associated with a H-score of >50 were regarded as steroid receptor– amplified products were observed under these conditions and PCR positive thymomas (23). products were purified and subjected to direct sequencing to verify amplification of the correct sequences as described previously. The Reverse transcription-PCR mRNA levels for STS, EST, and aromatase in each case are summarized RNA extraction and cDNA synthesis. Total RNA was extracted by as a ratio of glyceraldehyde-3-phosphate dehydrogenase and evaluated homogenizing frozen tissue samples in 1 mL TRIzol reagent (Life as a ratio (%) compared with that of each positive control. Technologies, Inc., Grand Island, NY) followed by a phenol/chloroform Statistical analysis. Values for patient age, tumor size, Ki-67 LI, phase extraction and isopropanol precipitation. The SuperScript H-scores of ERa or PR-B, and mRNA levels for EST, STS, 17h-HSD type Preamplification System RT kit (Life Technologies) was employed in 1, or aromatase were summarized as a mean F 95% confidence interval. the synthesis and amplification of cDNA according to the manufac- Statistical analyses between E2 tissue concentration and aromatase turer’s instructions. mRNA level, ERa H-score, or PR-B H-score were done using a Real-time reverse transcription-PCR. The Light Cycler System (Roche correlation coefficient (r) and regression equation. An association Diagnostics GmbH, Mannheim, Germany) was used to semiquantify between immunoreactivity for EST, STS, or 17h-HSD type 1 and these the level of STS, EST, 17h-HSD type 1, and aromatase mRNA expression variables were evaluated using Kruskal-Wallis tests. Statistical differences in 20 cases of thymoma using real-time PCR (24). Settings for the PCR between immunoreactivity of EST, STS, or 17h-HSD type 1 and sex, thermal profile were as follows: initial denaturing step of 95jC for 1 status of myasthenia gravis, clinical stage, or WHO histologic minute followed by 40 cycles, respectively, of 95jC for 0 second, 15- classification were evaluated in a cross-table using the m2 test. Overall second annealing at 58jC (EST) and 60jC (glyceraldehyde-3-phos- survival curves were generated according to the Kaplan-Meier method, phate dehydrogenase, STS, and aromatase), and extension for 15 and the statistical significance was calculated using the log-rank test. seconds at 72jC. The primer sequences used in this study are as follows: Univariate and multivariate analyses were evaluated by a proportional EST (NM005420; forward 5-AGAGGAGCTTGTGGACAGGA-3 and hazard model (Cox) using PROC PHREG in SAS software. P < 0.05 was reverse 5-GGCGACAATTTCTGGTTCAT-3; ref. 25), STS (M16505; considered significant. forward 5-AGGGTCTGGGTGTGTCTGTC-3 and reverse 5-ACTG- CAACGCCTACTTAAATG-3; ref. 26), 17h-HSD type 1 (XM 012644; forward 5V-AGGGCCGCGTGGACGTGCTGGTGTGTAAC-3V and reverse Results 5V-CCATCAATCCTCCCACGCTCCCGG-3V; ref. 17), aromatase (X13589; forward 5V-GTGAAAAAGGGGACAAACAT-3V and reverse 5V-TGGAA- Cell proliferation assay. E2 inhibits cell proliferation of TEC TCGTCTCAGAAGTGT-3V; ref. 17), and glyceraldehyde-3-phosphate in proportion to the concentration of E2 (10, 100, and 1,000 dehydrogenase (M33197; forward 5V-TGAACGGGAAGCTCACTGG-3V nmol/L; P < 0.001; Fig. 1A). ICI 182,780, which is a steroidal and reverse 5V-TCCACCACCCTGTTGCTGTA-3V; ref. 3). Liver (25) cells anti-estrogen with no agonist activities, blocks inhibition of cell were used as a positive control for EST, whereas frozen tissues of proliferation exerted by E2 (Fig. 1B; ref. 28). Terminal placenta were used as a positive control for STS (26, 27), 17h-HSD type deoxynucleotidyl transferase–mediated dUTP nick end labeling

Clin Cancer Res 2005;11(18) September 15, 2005 6498 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Estrogen Inhibits Cell Proliferation in Human Thymoma

analysis showed few apoptotic cells after treatment of E2 (data phenol red and FBS–free DMEM were below the limits of not shown). detection. Estrone and estradiol synthesis in primary culture of thymoma- Correlation among estradiol concentrations and estrogen derived neoplastic epithelial cells. E1 and E2 production in sulfotransferase, steroid sulfatase, or 17b-hydroxysteroid dehy- primary cultures of TEC are summarized in Fig. 2. E1 and E2 drogenase type 1 immunoreactivity, aromatase mRNA level, concentrations are detected in proportion to the duration and estrogen receptor a and progesterone receptor-B H-score, and concentration of E1-S. Concentrations of E1 and E2 in the clinical stage in 20 human thymomas. Results are summarized medium of TEC treated with E1-S (2.5 and 12.5 nmol/L) were in Table 2. E2 concentrations in thymoma were significantly significantly higher than those with E1-S of 0.5 nmol/L after higher than those of normal thymus (59.3 F 9.5 versus 19.2 F 8 and 24 hours (P < 0.05). E1 and E2 concentrations in 3.6; P < 0.05), inversely correlated with that of EST

Fig. 3. Immunohistochemistry for EST (A and B), STS (C and D), and 17h-HSD type 1 (E and F) in tissue specimens of human thymoma. Immunoreactivity for EST, STS, and 17h-HSD type 1was detected in the cytoplasm ofTEC but not in lymphocytes. A, C,andE, type A byWHO classification (same field); B, D,andF, type B3 (same field). Original magnification, Â400. Bar, 25 Am.

www.aacrjournals.org 6499 Clin Cancer Res 2005;11(18) September 15, 2005 Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Human Cancer Biology

Ta b l e 3 . Correlation between EST, STS, and17h-HSD type1immunoreactivity and pathologic variables in132 human thymomas

EST P STS immunoreactivity immunoreactivity

++ + À ++ + À (n =22) (n =55) (n =55) (n =20) (n = 57) (n =55) Age (y)* 50.0 F 3.7 57.2 F 1. 8 5 1. 4 F 2.0 0.290 53.6 F 3.6 54.0 F 1. 9 51. 6 F 2.1 Sex Male 7 (31.8) 32 (58.2) 19 (34.5) 0.405 12 (60.0) 24 (42.1) 22 (40.0) Female 15(68.2) 23 (41.8) 36 (65.5) 8 (40.0) 33 (57.9) 33 (60.0) Pre 5 (22.7) 11 (20.0) 22 (40.0) 0.138 5 (25.0) 16 (28.1) 17 (30.9) Post 10 (45.5) 12 (21.8) 14 (25.5) 3 (15.0) 17(29.8) 16 (29.1) Myasthenia gravis + 2 (9.1) 13 (23.6) 10 (18.2) 0.333 8 (40.0) 9 (15.8) 8 (14.5) À 20 (90.9) 42 (76.4) 45 (81.8) 12 (60.0) 48 (84.2) 47 (85.5) Tu m o r s i z e 78.0 F 6.5 58.7 F 3.3 54.8 F 3.0 0.013 50.6 F 3.5 62.2 F 3.6 61.8 F 3.6 (mm)* Clinical stage I 2(9.1) 17(30.9) 44(80.0) <0.001 12 (60.0) 33 (57.9) 18(32.7) II 2 (9.1) 21 (38.2) 9 (16.4) 5 (25.0) 14 (24.6) 13 (23.6) III 4 (18.2) 16 (29.1) 1 (1.8) 2 (10.0) 4 (7.0) 15 (27.3) IV 14 (63.6) 1 (1.8) 1 (1.8) 1 (5.0) 6 (10.5) 9 (16.4) Ki-67 LI* 3.2 F 0.2 1.6 F 0.1 1.3 F 0.1 0.045 1.42 F 0.13 1.58 F 0.11 2.11 F0.13 Aromatase mRNA 5.12 F 0.21 3.56 F 0.11 1.24 F 0.24 0.029 1.99 F 0.51 3.32 F 1.42 2.43 F 0.96 level (%), n =20 ERa H-score* 46.3 F 10.4 7 1.5 F 7.9 93.3 F 4.2 0.009 105.3 F 1. 8 8 3 . 6 F 4.9 54.4 F 7.7 PR-B H-score* 22.3 F 11.0 46.3 F 9.6 68.4 F 6.2 0.015 83.0 F 1. 2 5 6 . 4 F 7.6 3 1.1 F8.4

NOTE: P < 0.05 was considered significant. Abbreviations: Pre, premenopausal; Post, postmenopausal. *Mean F 95% confidence interval. All other values represent mean F SE.

immunoreactivity (P = 0.013), and positively correlated with Real-time PCR analysis. mRNA expression of EST, STS, 17h- that of STS (P = 0.008), 17h-HSD type 1 (P = 0.007), aromatase HSD type 1, aromatase, and glyceraldehyde-3-phosphate dehy- mRNA level (r = 0.601; P = 0.002), ERa H-score (r = 0.774; P < drogenase was identified as a specific single band at 114, 290, 0.001), and PR-B H-score (r = 0.749; P =0.001).E2 201, 215, and 307 bp, respectively (data not shown). There were concentrations were significantly higher in cases with earlier significant positive correlations between EST immunoreactivity clinical stage (P = 0.006). and its mRNA level (P = 0.001), STS immunoreactivity and its Immunolocalization of estrogen sulfotransferase, steroid sulfa- mRNA level (P = 0.010), and 17h-HSD type 1 immunoreactivity tase, and 17b-hydroxysteroid dehydrogenase type 1 in human and its mRNA level (P = 0.007; data not shown). thymoma. EST immunoreactivity was detected predominant- Correlation between estrogen sulfotransferase, steroid sulfatase, ly in the cytoplasm of epithelial cells of thymoma (Fig. 3A) and 17 b-hydroxysteroid dehydrogenase type 1 immunoreactivity but not in lymphocytes (Fig. 3B). The number of the cases and clinicopathologic variables. An association between EST, positive for EST in 132 human thymomas was summarized STS, or 17h-HSD type 1 immunoreactivity and clinicopath- as follows: ++, n = 22 (16.6%); +, n = 55 (41.7%); and À, ologic factors in thymoma patients were summarized in n = 55 (41.7%). Table 3. EST immunoreactivity was positively correlated with STS immunoreactivity was also detected predominantly in the tumor size (P = 0.013), clinical stage of the patients (P < cytoplasm of epithelial cells of thymoma (Fig. 3C) but not in 0.001), Ki-67 LI (P = 0.045), and mRNA level of aromatase lymphocytes (Fig. 3D). The number of cases positive for STS (P = 0.029). EST immunoreactivity was inversely correlated immunoreactivity was summarized as follows: ++, n =20 with ERa H-score (P = 0.009) and PR-B H-score (P = 0.015). (15.2%); +, n = 57 (43.2%); and À, n = 55 (41.6%). There were significant inverse correlations between STS 17h-HSD type 1 immunoreactivity was detected predomi- immunoreactivity and clinical stage of the patients (P = nantly in the cytoplasm of epithelial cells of thymoma (Fig. 3E) 0.036) or Ki-67 LI (P = 0.011). There were significant positive but not in lymphocytes (Fig. 3F). The number of cases positive correlations between STS immunoreactivity and ERa H-score for 17h-HSD type 1 immunoreactivity in these cases was (P = 0.005) or PR-B H-score (P = 0.006). summarized as follows: ++, n = 27 (20.5%); +, n =46 There were significant inverse correlations between 17h-HSD (34.8%); and À, n = 59 (44.7%). type 1 immunoreactivity and clinical stage of the patients

Clin Cancer Res 2005;11(18) September 15, 2005 6500 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Estrogen Inhibits Cell Proliferation in Human Thymoma

Ta b l e 3 . Correlation between EST,STS, and17h-HSD type1immunoreactivity and pathologic variables in132human thymomas (Cont’d)

P 17 B-HSD type 1 P immunoreactivity

++ + À (n =27) (n =46) (n =59) 0.697 56.8 F 2.8 53.7 F 2.0 50.6 F 2.1 0.191

0.286 10 (37.0) 21 (45.7) 27 (45.8) 0.720 17 (63.0) 25 (54.3) 32 (54.2) 0.572 7 (25.9) 13 (28.3) 18 (30.5) 0.138 10 (37.1) 12 (26.0) 14 (23.7)

0.053 9 (33.3) 11 (23.9) 15 (25.4) 0.141 18 (67.7) 35 (76.1) 44 (74.6) 0.185 56.9 F 4.6 53.3 F 3.0 60.3 F 3.7 0.076

0.036 14 (51.9) 28 (60.9) 21 (35.6) 0.008 9 (33.3) 12(26.1) 11 (18.6) 2 (7.4) 3 (6.5) 16 (27.1) 2 (7.4) 3 (6.5) 11 (18.6) 0.011 1.50 F 0.13 1.47 F 0.11 2.07 F 0.14 0.012 0.539 3.11 F0.82 2.58 F 1. 21 1. 8 8 F 1.02 0.053

0.005 103.3 F 2.2 86.4 F 3.6 53.9 F 6.6 0.001 0.006 78.5 F 3.8 64.9 F 6.4 26.4 F 6.4 0.001

(P = 0.008) or Ki-67 LI (P = 0.005) and positive correlations Discussion between 17h-HSD type 1 immunoreactivity and ERa H-score (P = 0.001) or PR-B H-score (P = 0.001). Estrogens have been considered to play important roles in Correlation between WHO classification and other variables. biological features of thymoma. However, this is the first study An association between WHO classification and clinicopathologic to show an inhibition of thymic epithelial cell proliferation by factors in thymoma patients was summarized in Table 4. E2 estrogens, the presence of in situ production of estrogens in concentrations were significantly higher in type A thymoma human thymoma using primary cell culture, and a correlation than type B thymoma (P = 0.005). There was a significant between the status of estrogen metabolizing or producing positive correlation between WHO classification and EST enzymes and clinicopathologic variables or endogenous estro- immunoreactivity (P < 0.001) and inverse correlations between gen concentrations in human thymus and thymoma. WHO classification and STS immunoreactivity (P < 0.001), ERa In our present study, TEC proliferation was inhibited by E2 H-score (P = 0.015), or PR-B H-score (P = 0.028). via ERa. This finding is consistent with our previous study, Correlation between estrogen sulfotransferase, steroid sulfatase, which suggests that E2 or E2-like agents may be effective in and 17b-hydroxysteroid dehydrogenase type 1 immunoreactivity treatment of thymoma, especially in cases of inoperable or and overall survival of patients with thymoma. There was a disseminated tumor possibly through suppressing its cell significant positive correlation between EST immunoreactivity proliferation activity (3). In this study, E1 and E2 were both and clinical outcome of the patients (P = 0.0001; Fig. 4A). No produced from E1-S in ERa, STS, and 17h-HSD type 1 positive h significant correlations were detected between STS or 17 -HSD TEC in proportion to the duration and concentration of E1-S type 1 immunoreactivity and clinical outcome in patients with administered. In addition, E2 tissue concentrations were thymoma (Fig. 4B and C). inversely correlated with the status of EST immunoreactivity, Following a univariate analysis (Table 5), clinical stage positively correlated with the status of STS or 17h-HSD type 1 (P = 0.0017), ERa immunoreactivity (P = 0.0021), EST immunoreactivity, and significantly higher in earlier clinical immunoreactivity (P = 0.0023), and tumor size (P = 0.0024) stages and lower histologic grades. These findings showed that all turned out to be significant prognostic factors of overall lower histologic grade tumors or well-differentiated thymoma survival in 132 thymoma patients. A subsequent multivariate cells were associated with in situ estrogen biosynthesis and analysis revealed that only clinical stage (P = 0.0239), ERa estrogenic actions via binding to ERa, which result in E2- immunoreactivity (P = 0.0355), and EST immunoreactivity dependent inhibition of tumor growth. (P = 0.0436) were independent prognostic factors with relative EST immunoreactivity was also detected in TEC and positively risks of >1.0 in our series of 132 thymoma patients. correlated with Ki-67 LI or tumor size and inversely correlated

www.aacrjournals.org 6501 Clin Cancer Res 2005;11(18) September 15, 2005 Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Human Cancer Biology

Ta b l e 4 . Correlation betweenWHO classification and other variables in human thymomas

AABB1B2B3P* c E2 tissue concentration (pg/g) 111. 5 F 27.8 89.5 F 9.0 31.5 F 5.2 34.0 F 5.2 28.7 F 2.6 0.005 EST immunoreactivity ++ (n =22) 2(9.1) 1(4.5) 1(4.5) 7(31.8) 11(50.0) <0.001 +(n = 55) 14 (25.5) 12(21.8) 8 (14.5) 17(30.9) 4 (7.3) À (n =55) 10(18.7) 12(21.8) 20(36.4) 12(21.8) 1(1.8) STS immunoreactivity ++ (n = 22) 11 (55.0) 3 (15.0) 2 (10.0) 3 (15.0) 1 (5.0) <0.001 +(n = 55) 14 (24.6) 14 (24.6) 15 (26.3) 10 (17.5) 4 (7.0) À (n = 55) 1 (1.8) 8 (14.5) 12(21.8) 23 (41.8) 11(20.0) 17 h-HSD type1immunoreactivity ++ (n = 27) 11(40.8) 3 (11.1) 4 (14.8) 6 (22.2) 3 (11.1) 0.057 +(n = 46) 8 (17.4) 11(23.9) 11(23.9) 14 (30.4) 2 (4.4) À (n = 59) 7 (11.9) 11(18.6) 14 (23.7) 16 (27.2) 11(18.6) c Aromatase mRNA level 3.11 F0.12 3.42 F 0.22 3.24 F 0.18 2.98 F 0.32 3.98 F 0.22 0.283 c ERa H-score 111.1 F9.8 104.9 F 16.2 96.3 F 14.4 64.9 F 12.6 36.5 F 7.4 0.015 c PR-B H-score 85.7 F 15.8 64.6 F 14.0 51.3 F 11.9 45.1 F 9.8 18.0 F 9.6 0.028

*P was listed only if significant. P < 0.05 was considered significant. cMean F 95% confidence interval. All other values represent mean F SE. with H-score of ERa or PR-B. In addition, EST-positive thymoma for overall survival as well as ERa and clinical stage following patients tended to be associated with advanced clinical stages multivariate analysis. Tissue E2 concentrations were also with higher histologic grades and significantly adverse clinical inversely correlated with EST immunoreactivity. Therefore, outcome than EST-negative thymoma. In addition, EST immu- EST-negative thymoma is considered to result in an increased noreactivity was shown to be an independent prognostic factor in situ concentration of biologically active estrogens, which may

Fig. 4. Overall survival of132patients with thymoma with respect to EST (A), STS (B), and 17h-HSD type 1 (C) immunoreactivity (Kaplan-Meier method). EST immunoreactivity was significantly associated with an improved overall survival (P = 0.0001) but STS and 17h-HSD type1immunopositive thymoma was relatively associated with better clinical outcome, although this association did not reach statistical significance.

Clin Cancer Res 2005;11(18) September 15, 2005 6502 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Estrogen Inhibits Cell Proliferation in Human Thymoma

Ta b l e 5 . Univariate and multivariate analyses of overall survival in132thymoma patients

Variable Univariate Multivariate PPRelative risk (95% confidence interval) Clinical stage (II, III, IV/I) 0.0017 0.0239 5.783 (1.709-47.196) ERa (À/+) 0.0021 0.0355 4.786 (1.143-21.959) EST (++, +/À) 0.0023 0.0436 2.509 (1.126-5.263) Tu m o r s i z e ( >60/Q60 mm) 0.0024 Female (premenopausal/postmenopausal) 0.0620 PR-B (À/+) 0.4694 Sex (male/female) 0.7033

contribute to inhibition of cell proliferation in thymoma to be increased by progesterone. These results suggest that through ERa. Qian et al. showed that MCF-7 breast cancer cells progesterone is capable of specifically inducing EST and transfected with EST expressed EST at levels similar to those estrogen sulfation in the human Ishikawa endometrial adeno- observed in normal human mammary epithelial cells and are carcinoma cell line (37). Furthermore, progestin has also been associated with much lower estrogen-stimulated DNA synthesis reported to induce 17h-HSD enzyme protein in the T-47D or cell proliferation than MCF-7 cells not associated with EST expression (29). EST-negative breast cancer is therefore considered to be associated with an increased in situ estrogen concentration, which results in an increased incidence of tumor recurrence and subsequent poor clinical outcome for the patients diagnosed with breast cancer (20, 30). EST is more frequently detected in TEC in higher histologic grades or advanced clinical stages, which are consistent with decreased expression of STS or 17h-HSD type 1. All of these are considered to result in decreased levels of E2 in these thymoma cases. STS and 17h-HSD type 1 immunoreactivity was also positively correlated with the H-score of ERa and PR-B and inversely correlated with Ki-67 LI. STS- positive and 17h-HSD type 1–positive thymoma cases were significantly correlated with earlier clinical stages and lower histologic grades. In addition, 17h-HSD type 1–positive thymoma tends to be associated with favorable clinical outcome. STS catalyzes E1-S to E1 and 17h-HSD type 1 catalyzes E1 to E2 in human thymoma, which contributes to increase in situ estrogen concentration (31, 32). Results of our present study suggest that STS and 17h-HSD type 1 also contribute to estrogenic actions in human thymoma through in situ estrogen production. In contrast, a significant inverse correlation was reported between 17h-HSD type 1 immunoreactivity and Ki-67 LI or histologic grade in patients with breast cancer (33, 34). 17h-HSD type 1 immunoreactivity is considered to reflect its enzymatic activity (16), and ER immunoreactivity has been shown to be correlated with estrogen-dependent biological phenomena (35). Therefore, results of our present study showed that in situ produced E2 exerts its effects through ERa in human thymoma, which may be consistent with the relatively better clinical outcome of 17h-HSD type 1–positive human thymoma patients. Further investiga- tions are required for clarification. PR-B H-score was positively correlated with STS and 17h- HSD type 1 immunoreactivity and inversely correlated with EST immunoreactivity in human thymoma. These findings are also Fig. 5. Summary of local production of estrogens in high-grade (A)andlow-grade consistent with the fact that expression of PR is estrogen related, (B) thymoma. High concentrations of circulating inactive steroids, androstenedione because PR has been regarded as one of the markers of a and E1-S, are major precursor substrates of local estrogen production in these tissues. Aromatase catalyzes androstenedione into E1, and STS hydrolyzes E1-S to E1. functional estrogen pathway (36) and ERa-positive thymomas E1is subsequently converted to potent E2 by 17h-HSD type 1and acts onTEC are generally positive for PR-B (3). The levels of immunoreac- via ERa. ESTsulfonates E1to biologically inactive E1-S.HighintratumoralE2 concentration as a result of in situ production in thymoma with early-stage or tive EST and EST mRNA in Ishikawa cells, a cell line established well-differentiated histologic features is therefore considered to result in inhibition from human endometrial adenocarcinoma, were both shown of cell proliferation ofTEC.

www.aacrjournals.org 6503 Clin Cancer Res 2005;11(18) September 15, 2005 Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Human Cancer Biology human breast cancer cell line (38). Therefore, coexpression of We show the findings of in situ estrogen production in EST, STS, and/or 17h-HSD type 1 and PR detected thymoma in human thymoma (Fig. 5). Advanced stage or high-grade this study suggests that progesterone may play an important thymoma produces low level of E2, which cannot inhibit the role in regulating the expression of these enzymes involved in proliferation of TEC (Fig. 5A). On the other hand, high in situ estrogen metabolism in thymoma. However, it awaits concentration of E2 is produced in early-stage or low histologic further investigating to identify the specific roles of these grade thymoma and inhibits the proliferation of TEC (Fig. 5B). enzymes in metabolizing estrogens in thymoma. In summary, high intratumoral E2 concentration as a result Aromatase cytochrome P450 (CYP19 gene) is an enzyme of in situ production in thymoma with early-stage or well- located in the endoplasmic reticulum of estrogen-producing differentiated histologic features is therefore considered to cells and a key enzyme mainly involved in the aromatization of result in inhibition of cell proliferation of TEC. androstenedione to E1 (29). In the present study, there was a statistically significant positive correlation between aromatase Acknowledgments mRNA level and immunoreactivity for EST but not for STS, 17h-HSD type 1, and aromatase mRNA levels tended to We thank Drs. Touichirou Takizawa and Takumi Akashi (Department of Patholo- gy, Graduate School, Tokyo Medical and Dental University) and Dr. Hideki correlate with E2 tissue concentrations. These results also Akamatsu (Department of Thoracic Cardiovascular Surgery, Graduate School, suggest that estrogen can still be produced via aromatase Tokyo Medical and Dental University) for their kind efforts in retrieving the speci- pathway in EST-positive thymoma case even if the activity of mens of thymoma and AndrewD. Darnel (Department of Pathology,Tohoku Uni- aromatase was 0.5% to 2% of STS (3). versity School of Medicine) for careful editing of this article.

References 1. Alexieva-FiguschJ,Van Putten WL, Blankenstein MA, 14. Rosai J, Sobin LH. Histological typing of tumours of 27. Salido EC, Yen PH, Barajas L, Shapiro LJ. Steroid Blonk-Van Der Wijst J, Klijn JG. The prognostic value the thymus. International histological classification of sulfatase expression in human placenta: immunocy- and relationships of patient characteristics, estrogen tumours. 2nd ed. NewYork: Springer; 1999. p. 9 ^ 13. tochemistry and in situ hybridization study. J Clin and progesterone receptors, and site of relapse in 15. SaekiT, Takashima S, Sasaki H, Salomon DS. Local- Endocrinol Metab 19 9 0;70 :15 6 4 ^ 7. primary breast cancer. Cancer 1988;61:758 ^ 68. ization of estrone sulfatase in human breast carcinoma. 28. Wakeling AE, Dukes M, Bowler J. A potent specific 2. Pasqualini JR, Chetrite G, Blacker C. Concentrations Breast Cancer 1999;6:331 ^ 7. pure antiestrogen with clinical potential. Cancer Res of estrone, estradiol, and estrone sulfate and evalua- 16. Poutanen M, Isomaa V, Lehto VP,Vihko R. Immuno- 1991;51:3867 ^ 73. tion of sulfatase and aromatase activities in pre- and logical analysis of 17h-hydroxysteroid dehydrogenase 29. Qian Y, Deng C, Song WC. Expression of estrogen postmenopausal breast cancer patients. J Clin Endo- in benign and malignant human breast tissue. Int J sulfotransferase in MCF-7 cells by cDNA transfection crinol Metab1996;81:1460^4. Cancer1992;50:386^90. suppresses the estrogen response: potential role of 3. Ishibashi H, Suzuki T, Suzuki S, et al. Sex steroid 17. Sasano H, Frost AR, Saitoh R, et al. Aromatase and the enzyme in regulating estrogen-dependent growth hormone receptors inhuman thymoma. JClinEndocri- 17 h-hydroxysteroid dehydrogenase type 1 in human of breast epithelial cells. J Pharmacol Exp Ther nol Metab 2003;88:2309 ^ 17. breast carcinoma. J Clin Endocrinol Metab 1996;81: 1998;286:555 ^60. 4. Ruder HJ, Loriau L, Lipsett MB. Estrone sulfate: pro- 4042 ^ 6. 30. Suzuki T, Moriya T, Ishida T, Kimura M, Ohuchi N, duction rate and metabolism in man. J Clin Invest 18. MikiY,NakataT, SuzukiT, et al. Systemic distribution Sasano H. In situ production of estrogens in hu- 1972;51:1020^ 33. of steroid sulfatase and estrogen sulfotransferase in man breast carcinoma. Breast Cancer 2002;9: 5. Samojlik E, Santen RJ, Worgul TJ. Plasma estrone human adult and fetal tissues. JClin Endocrinol Metab 296 ^302. sulfate: assessment of reduced estrogen production 2002;87:5760 ^ 8. 31. Selcer KW, Hegde PV, Li PK. Inhibition of estrone during treatment of metastatic breast carcinoma. Ste- 19. TakeyamaJ, Sasano H, Suzuki T,Iinuma K, Nagura H, sulfatase and proliferation of human breast cancer cells roids 1982;39:497 ^ 507. Andersson S. 17h-Hydroxysteroid dehydrogenase by nonsteroidal (p-O-sulfamoyl)-N-alkanoyl tyr- 6. Naitoh K, Honjo H,YamamotoT, et al. Estrone sulfa- types1and2inhumanplacenta:animmunohistochem- amines. Cancer Res 1997;57:702^ 7. tase and sulfotransferase activity in human breast ical study with correlation to placental development. 32. Pasqualini JR, Chetrite GS. Estrone sulfatase versus cancer and endometrial cancer. J Steroid Biochem JClin Endocrinol Metab1998;83:3710 ^ 5. estrone sulfotransferase in human breast cancer: 1989;33:1049^54. 20. Suzuki T, Nakata T, Miki Y,et al. Estrogen sulfotrans- potential clinical applications. J Steroid Biochem Mol 7. PasqualiniJR, Gelly C. Biological response of the anti- ferase and steroid sulfataseinhumanbreast carcinoma. Biol1999;69:287^92. estrogen ICI 164,384 in human hormone-dependent Cancer Res 2003;63:2762 ^70. 33. SuzukiT, MoriyaT, Ariga N, Kaneko C, Kanazawa M, and hormone-independent mammary cancer cell lines. 21. Goulding H, Pinder S, Cannon P, et al. A newim- Sasano H. 17h-hydroxysteroid dehydrogenase type 1 Cancer Lett 1990;50:133 ^ 9. munohistochemical antibody for the assessment of and type 2 in human breast carcinoma: a correlation 8. Miller WR, Hawkins RA, Forrest AM. Significance of estrogen receptor status on routine formalin-fixed to clinicopathological parameters. Br J Cancer 2000; aromatase activity inhuman breast cancer. Cancer Res tissue samples. Hum Pathol 1995;26:291 ^ 4. 82:518^ 23. 1982;42:3365^8. 22. PanCC,HoDM,ChenWY,HuangCW,ChiangH. 34. Sasano H, SuzukiT, Takeyama J, et al.17-h-Hydrox- 9. Luu-The V, Labrie C, Zhao HF, et al. Characterization Ki-67 labeling index correlates with stage and histolo- ysteroid dehydrogenase inhuman breast and endome- of cDNAs for human estradiol17h-dehydrogenase and gy but not significantly with prognosis in thymoma. trial carcinoma. A newdevelopment in intracrinology. assignment of the gene to chromosome 17: evidence Histopathology1998;33:453 ^ 8. Oncology 2000;59 Suppl 1:5 ^ 12. for two mRNA species with distinct 59-termini in 23. Thike AA, Chng MJ, Fook-Chong S, Tan PH. 35. Marrazzo A, La Bara G, Taormina P,Bazan P.Deter- human placenta. Mol Endocrinol 1989;3:1301 ^ 9. Immunohistochemical expression of hormone recep- mination ofoestrogen receptors with monoclonal anti- 10. Falany CN. Enzymology of human cytosolic sulfo- tors in invasive breast carcinoma: correlation of results bodies in fine needle aspirates of breast carcinoma. transferase. FASEB J1997;11:206 ^ 16. of H-score with pathological parameters. Pathology BrJCancer1989;59:426 ^ 8. 11. Kadota Y, Okumura M, Miyoshi S, et al. Altered 2001;33:21 ^ 5. 36. Horwitz KB, McGuire WL. Estrogen control of pro- T-cell development in human thymoma is related to 24. Dumoulin FL, Nischalke HD, Leifeld L, et al. Semi- gesterone receptor in human breast cancer. Correla- impairment of MHC class II transactivator expression quantification of human C-C chemokine mRNAs with tion with nuclear processing of estrogen receptor. induced by interferon-g (IFN-g). Clin Exp Immunol reversetranscription/real-timePCRusingmulti-specific J Biol Chem1978;253:2223 ^ 8. 2000;121:59^68. standards. JImmunol Methods 2000;241:109 ^19. 37. Falany JL, Falany CN. Regulation of estrogen sulfo- 12 . Salakou S, Tsamandas AC, Bonikos DS, et al. The 25. Aksoy IA,Wood TC,Weinshilboum R. Human liver transferase in human endometrial adenocarcinoma potential role of bcl-2, bax, and Ki67 expression in estrogensulfotransferase: identificationbycDNAclon- cells by progesterone. Endocrinology 1996;137: thymus of patients with myasthenia gravis, and their ing and expression. Biochem Biophys Res Commun 1395^401. correlation with clinicopathologic parameters. Eur J 1994;200:1621 ^ 9. 38. Poutanen M, Isomaa V, Kainulainen K, Vihko R. Cardiothorac Surg 2001;20:712^ 21. 26. Utsumi T, Yoshimura N, Takeuchi S, et al. Steroid Progestin induction of 17 h-hydroxysteroid dehy- 13. Yamakawa Y, Masaoka A, Hashimoto T, et al. A sulfatase expression is an independent predictor of drogenase enzyme protein in the T-47D human tentative tumor-node-metastasis classification of recurrence in human breast cancer. Cancer Res breast-cancer cell line. Int J Cancer 1990;46: thymoma. Cancer 1991;68:1984 ^ 7. 1999;59:377 ^ 81. 897^901.

Clin Cancer Res 2005;11(18) September 15, 2005 6504 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 2, 2021. © 2005 American Association for Cancer Research. Estrogen Inhibits Cell Proliferation through In situ Production in Human Thymoma

Hironori Ishibashi, Takashi Suzuki, Satoshi Suzuki, et al.

Clin Cancer Res 2005;11:6495-6504.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/11/18/6495

Cited articles This article cites 35 articles, 6 of which you can access for free at: http://clincancerres.aacrjournals.org/content/11/18/6495.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/11/18/6495.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://clincancerres.aacrjournals.org/content/11/18/6495. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.