Estrogen Sulfotransferase and Steroid Sulfatase in Human Breast Carcinoma

Breast carcinoma is one of the most common malignancies in women worldwide. Estrogens play an important role in the development of hormone-dependent breast carcinomas (1, 2), and ER² status is well known to affect the prognosis of patients with breast neoplasms (3). Previous studies have demonstrated that tissue concentrations of estradiol (E2) in breast carcinomas are 10 times higher than the levels found in plasma (4, 5), suggesting that estrogens in human breast cancer tissues may be produced in situ from circulating biologically inactive precursors. Aromatase catalyzes circulating androgen androstenedione into estrone (E1; 4, 6), whereas STS hydrolyzes circulating estrone sulfate (E1-S) to E1 (7, 8). E1 is subsequently converted to the potent estrogen E2 by 17β-HSD type 1 (9, 10) and acts on breast cancer cells through ERα and/or ERβ.

EST (SULT 1E1 or STE gene) is a member of the superfamily of steroid-sulfotransferases and sulfonates estrogens to biologically inactive estrogen sulfates (11, 12, 13). Therefore, it is suggested that EST may play an important role in the regulation of in situ estrogen levels in human breast carcinoma, in a manner similar to STS, aromatase, and 17β-HSD type 1. EST enzymatic activity has been reported in various breast cancer cell lines (14, 15) and breast carcinoma tissues (16, 17). However, to date, the expression of EST mRNA and protein has not been examined in human breast carcinoma tissues, and thus, the biological significance of EST remains unclear. Although STS expression has recently been reported using RT-PCR (18) and immunohistochemistry (19), information regarding the expression of STS in human breast carcinoma tissues is still very limited, compared with that of aromatase or 17β-HSD type 1. The purpose of this study was to examine the expression and biological significance of EST and STS in human breast carcinomas. To accomplish this, we first studied the expression of EST and STS using immunohistochemistry, RT-PCR, and enzymatic assay in 35 specimens of human breast carcinoma and analyzed the correlation among these variables. We then examined the immunolocalization of EST and STS in 113 cases of human breast carcinoma and correlated EST and STS immunoreactivity with various clinicopathological indices, including clinical outcome in these 113 patients.

Thirty-five specimens of invasive ductal carcinoma were obtained from patients who underwent mastectomy in 2000 in the Departments of Surgery at Tohoku University Hospital and Tohoku Kosai Hospital, Sendai, Japan. Breast tissue specimens were obtained from patients with a mean age of 51.9 years (range 31–79). Clinical findings in these patients were obtained from the history charts retrieved by two of the authors (T. I. and H. H.). Specimens of adipose tissue adjacent to the carcinoma were available for examination in 10 of these 35 cases, but normal breast tissues were not available for study in any of the cases examined in this study. Specimens for RNA isolation were snap frozen and stored at −80°C, and those for immunohistochemistry were fixed with 10% formalin and embedded in paraffin wax. Informed consent was obtained from all patients before their surgery and examination of specimens used in this study.

Specimens (113) of invasive ductal carcinoma of the breast were obtained from female patients who underwent mastectomy from 1984 to 1989 in the Department of Surgery, Tohoku University Hospital. Breast tissue specimens were obtained from patients with a mean age of 53.1 years (range 27–82). None of the patients examined in this study used oral contraceptives. The patients did not receive irradiation or chemotherapy before surgery or hormonal therapy, including tamoxifen, after surgery. Clinical data were recorded in the standardized form, and missing data were obtained from the charts by two of the authors (T. S. and T. I.). The mean follow-up time was 106 months (range 5–154 months). Disease-free survival data were available for all patients. The histological grade of each specimen was evaluated by three pathologists (T. S., T. M., and H. S.), based on the method of Elston and Ellis (20). All specimens were fixed with 10% formalin and embedded in paraffin wax. Snap-frozen tissues were not available for examination in these cases.

Research protocols for this study were approved by the Ethics Committee at both Tohoku University School of Medicine and Tohoku Kosai Hospital.

Rabbit polyclonal antibody for EST (PV-P2237) was purchased from the Medical Biological Laboratory (Nagoya, Japan). This antibody was raised against the synthetic NH₂-terminal peptide of human EST, corresponding to amino acids 1–13. Results of immunoblotting analyses demonstrated that the sensitivity limit of EST antibody (1/500 dilution) was <10 μg of human liver (13) high-speed supernatant/lane.³ In addition, this antibody did not recognize other members of the human steroid-sulfotransferase superfamily, including phenol-sulfating phenol sulfotransferase (SULT 1A1 or STP1 gene), M-PST (SULT 1A3 or STM gene), and dehydroepiandrosterone-sulfotransferase (SULT 2A1 or STD gene).⁴ The affinity purified monoclonal STS (KM1049) antibody was raised against the STS enzyme purified from human placenta and recognized the peptide corresponding to amino acids 414–434. STS antibody was used for immunohistochemistry analysis, as described previously (19). 17β-HSD type 1 antibody was a rabbit polyclonal antibody (21) kindly provided by Dr. Poutanen at the University of Oulu (Oulu, Finland). Monoclonal antibodies for ERα (ER1D5), PR (MAB429), and Ki-67 (MIB1) were purchased from Immunotech (Marseille, France), Chemicon (Temecula, CA), and DAKO (Carpinteria, CA), respectively. Rabbit polyclonal antibodies for ERβ (06–629) and HER-2/neu (A0485) were obtained from Upstate Biotechnology (Lake Placid, NY) and DAKO, respectively.

A Histofine Kit (Nichirei, Tokyo, Japan), which uses the streptavidin-biotin amplification method, was used for the identification of EST, STS, ERα, PR, 17β-HSD type 1, Ki-67, and HER-2/neu immunoreactive staining, whereas EnVision⁺ (DAKO) was used for ERβ immunohistochemical analysis. Antigen retrieval for ERα, ERβ, PR, Ki-67, and HER-2/neu immunostaining was performed by heating the slides in an autoclave at 120°C for 5 min in citric acid buffer [2 mm citric acid and 9 mm trisodium citrate dehydrate (pH 6.0)]. Similarly, antigen retrieval for EST immunostaining was done by heating the slides in a microwave oven for 15 min in a citric acid buffer. No antigen retrieval was performed for STS and 17β-HSD type 1 immunostaining. The dilutions of the primary antibodies used in this study were as follows: 1/1500 EST, 1/9000 STS (0.37 μg/ml), 1/50 ERα, 1/50 ERβ, 1/30 PR, 1/800 17β-HSD type 1, 1/50 Ki-67, and 1/200 HER-2/neu. The antigen-antibody complex was visualized with 3,3′-diaminobenzidine solution [1 mm 3,3′-diaminobenzidine, 50 mm Tris-HCl buffer (pH 7.6), and 0.006% H₂O₂] and counterstained with hematoxylin. Human tissues of liver (13) and placenta (22) were used as positive controls for EST and STS antibodies, respectively. As a negative control, normal rabbit or mouse IgG was used instead of the primary antibodies. No specific immunoreactivity was detected in these sections.

Total RNA was carefully extracted from 35 specimens of invasive ductal carcinoma and 10 specimens of adipose tissues adjacent to the carcinoma with guanidinium thiocyanate followed by ultracentrifugation in cesium chloride. A reverse transcription kit (SUPERSCRIPT II Preamplification system; Life Technologies, Inc., Grand Island, NY) was used in the synthesis of cDNA.

The Light Cycler System (Roche Diagnostics GmbH, Mannheim, Germany) was used to semiquantify the level of EST and STS mRNA expression in 35 specimens of invasive ductal carcinoma and 10 specimens of adipose tissue by real-time PCR (23). Settings for the PCR thermal profile were: initial denaturation at 95°C for 1 min followed by 40 amplification cycles of 95°C for 1 s, annealing at 60°C (STS, aromatase, and GAPDH) or 58°C (EST) for 15 s, and elongation at 72°C for 15 s. The primer sequences used in this study are as follows: EST [NM005420; FWD 5′-AGAGGAGCTTGTGGACAGGA-3′ (cDNA position; 751–771) and REV 5′-GGCGACAATTTCTGGTTCAT-3′ (cDNA position; 844–864)], STS (M16505; FWD 5′-AGGGTCTGGGTGTGTCTGTC-3′ and REV 5′-ACTGCAACGCCTACTTAAATG-3′; Ref. 18), aromatase [X13589; FWD 5′-GTGAAAAAGGGGACAAACAT-3′ (cDNA position; 1286–1305) and REV 5′-TGGAATCGTCTCAGAAGTGT-3′ (cDNA position; 1481–1500)], and GAPDH [M33197; FWD 5′-TGAACGGGAAGCTCACTGG-3′ (cDNA position; 731–750) and REV 5′-TCCACCACCCTGTTGCTGTA-3′ (cDNA position; 1018–1038)]. Oligonucleotide primers for EST (11), aromatase (24), and GAPDH (25) were designed using cDNA sequences published previously. To verify amplification of the correct sequences, PCR products were purified and subjected to direct sequencing. HuH7, a well-differentiated human hepatocellular carcinoma cell line, was used as a positive control for EST, whereas human placental tissue was used as a positive control for STS and aromatase. Negative control experiments lacked cDNA substrate to check for the possibility of exogenous contaminant DNA. No amplified products were detected under these conditions. The mRNA level for EST and STS in each case has been summarized as a ratio of GAPDH and subsequently evaluated as a ratio (%) compared with that of the positive controls.

To examine the localization of EST and STS mRNA, microdissection was conducted using the Laser Scissors CRI-337 (Cell Robotics, Inc., Albuquerque, NM). A detailed procedure has been described previously (26, 27). Briefly, ∼500 carcinoma or intratumoral stromal cells were collected separately under the microscope from breast carcinoma frozen tissue sections embedded in Tissue-Tek optimal cutting temperature (OCT) compound (Sakura Finetechnical Co., Ltd., Tokyo, Japan). Total RNA was extracted according to an RNA micro-isolation protocol described by Niino et al. (27). Protocols for RT-PCR are described in the above section. The synthesized cDNAs were amplified by PCR for 40 cycles. The products were resolved on a 2% agarose ethidium bromide gel, and the images were captured with Polaroid film under UV transillumination.

EST was assayed as described previously (12). Briefly, frozen specimens were homogenized in a reaction buffer at 4°C and centrifuged for 15 min at 1000 × g. The upper layer was subsequently used as the enzyme source. Approximately 0.2 mg of protein was added in each assay, and the reaction mixture included 50 mm Tris-HCl (pH 7.4) and 7 mm MgCl₂; E1 contained [³H] E1 at 20 nm. Reactions were initiated with the addition of adenosine 3′-phosphate 5′-phosphosulfate (PAPS) to a final concentration of 20 μm, in a final volume of 0.125 ml. The reaction mixtures were then incubated at 37°C for 30 min. The reactions were terminated with the addition of 4 ml of chloroform, followed by the addition of 0.375 ml of 0.25 m Tris-HCl (pH 8.7) to alkalinize the solution. The reaction mixtures were subsequently centrifuged at 600 × g for 5 min to separate the aqueous and organic phases. Synthesis of the tritiated E1-S was determined with a liquid scintillation counter (LC-6500; Beckman).

The STS activity was assayed according to Utaaker and Støa (28) with slight modifications. Briefly, enzyme solution (∼0.2 mg of protein) was mixed with E1-S containing [6,7-³H] E1-S (1.6 × 10⁵ dpm, 0.5 pmol/liter) at 20 μm and added to a reaction volume ≤0.15 ml with PBS (−) containing 25 mm sucrose and 4 mm Nicotinamide. The reaction mixture was incubated at 37°C for 60 min in a shaking water bath. The enzyme reaction was terminated with the addition of toluene and mixed vigorously with a vortex mixer for 1 min. The reaction mixtures were centrifuged at 600 × g for 5 min to separate the aqueous and organic phases. The toluene layer was collected, and [³H] radioactivity was measured via a liquid scintillation counter (LC-6500; Beckman), which is equivalent to the amount of E1 formed.

Incubation conditions for these assays were designed so that the formation of product was linear.

Immunoreactivity for EST and STS was analyzed according to a method described previously (29). After completely reviewing the entire slides of immunostained sections for each carcinoma, three pathologists [T. S., T. M., and H. S.] independently and blindly divided the carcinomas into the after three groups: ++, >50% positive cells; +, 1–50% positive cells; and −, no immunoreactivity. Interobserver differences were <5% [EST immunostaining: 1.8% (2 cases, between + and −), 2.7% (3 cases, between + and −), and 4.4% (5 cases, between + and −); STS immunostaining: 1.8% (2 cases, between + and −), 2.7% (1 case, between + and −, and 2 cases, between ++ and +), and 2.7% (1 case, between + and −, and 2 cases, between ++ and +)]. Discordant results were mainly attributable to differences in the evaluation of weak immunopositive or negative staining (background). Cases with discordant results among the observers were simultaneously reevaluated using a multiheaded microscope. ERα, ERβ, PR, and Ki-67 immunoreactivity was scored in >500 carcinoma cells for each case and counted independently by the same three authors, and the percentage of immunoreactivity, i.e., LI, was determined. In the present study, interobserver differences were <5%, and the mean of the three values was obtained. Cases in this study that were found to have ERα- or PR-labeling indices <10% were noted as ERα- or PR-negative breast carcinomas, according to a report by Allred et al. (30).

Values for mRNA levels and enzymatic activity for EST and STS, patient age, tumor size, and LIs for ERα, ERβ, PR, and Ki-67 were summarized as a mean ± 95% CI. The association between immunoreactivity for EST or STS and these parameters was evaluated using a one-way ANOVA and Bonferroni test. Statistical analyses between EST or STS enzymatic activity and ERα LI was performed using a correlation coefficient (r) and regression equation. Statistical differences between immunoreactivity for EST or STS and menopausal status, stage, lymph node status, histological grade, ERα status, PR status, 17β-HSD type 1, or HER-2/neu were evaluated in a cross-table using the χ² test. Overall and disease-free survival curves were generated according to the Kaplan-Meier method. Univariate and multivariate analyses were evaluated by a proportional hazard model (Cox) using PROC PHREG in our SAS software. Significant variables evaluated by univariate analyses were only examined in the multivariate analyses. Differences with Ps < 0.05 were considered significant.

Immunoreactivity for EST was detected in the cytoplasm of invasive ductal carcinoma cells (Fig. 1,A). The number of cases expressing immunoreactive EST in each group of 35 breast carcinoma tissues are summarized as follows: ++, n = 6 (17.1%); +, n = 12 (34.3%); and −, n = 17 (48.6%). EST immunoreactivity was also detected in epithelial cells of morphologically normal glands (Fig. 1 B), whereas the mammary myoepithelium, stroma, or adipose tissue adjacent to the carcinoma was found to be immunohistochemically negative for EST.

STS immunoreactivity was detected in the cytoplasm of carcinoma cells (Fig. 1 C). The number of cases in each group in the 35 breast carcinomas is summarized as follows: ++, n = 11 (31.4%); +, n = 15 (42.9%); and −, n = 9 (25.7%). In morphologically normal glands, immunoreactivity for STS was weakly detected in epithelial cells. STS immunoreactivity was not detected in stroma or adipose tissues adjacent to the carcinoma.

To examine the localization of mRNA for EST and STS in breast cancer tissues, we performed laser capture microdissection followed by RT-PCR analyses. mRNA expression for EST, STS, and GAPDH was detected as a specific single band (114 bp for EST, 290 bp for STS, and 307 bp for GAPDH) using RT-PCR. mRNA expression for EST was detected in both carcinoma and intratumoral stromal cells adjacent to the carcinoma cells (Fig. 2,A), whereas that of STS was detected only in microdissected carcinoma cells (Fig. 2 B).

Correlation between EST immunoreactivity and EST mRNA level or EST enzymatic activity was summarized in Table 1,A. Significant correlations were detected between EST immunoreactivity and EST mRNA level that was semiquantified by real-time PCR (P = 0.0027, ++ versus − and P = 0.0468, ++ versus +; Bonferroni test) or EST enzymatic activity (P = 0.0005, ++ versus − and P = 0.0019, ++ versus +; Bonferroni test) in 35 cases of breast cancer tissues. In adipose tissues adjacent to the carcinoma (n = 10), a low level of EST mRNA or EST enzymatic activity was detected. No association was detected between EST immunoreactivity and aromatase mRNA level, ERα LI, or PR LI, although EST enzymatic activity was significantly correlated to ERα LI (Fig. 2 C) in the 35 cases of breast carcinoma [P = 0.005, r = 0.464, R² = 0.215, correlation coefficient (r), and regression equation], as was reported previously (16, 17).

Correlation between STS immunoreactivity and its mRNA level or enzymatic activity is summarized in Table 1 B. Significant correlations were detected between STS immunoreactivity and STS mRNA level (P = 0.0158, ++ versus −; Bonferroni test) or STS enzymatic activity (P = 0.0089, ++ versus − and P = 0.0119, ++ versus +; Bonferroni test) in 35 cases of breast cancer. In adipose tissues adjacent to the carcinoma (n = 10), mRNA levels for STS or STS enzymatic activity were negligible. No statistically significant association was detected between STS immunoreactivity and aromatase mRNA level, ERα LI, or PR LI.

Results of an association between EST immunoreactivity and clinicopathological parameters in 113 breast carcinomas are summarized in Table 2, A and B. The number of cases expressing immunoreactive EST in each group is summarized as follows: ++, n = 18 (15.9%); +, n = 32 (28.3%); and −, n = 63 (55.8%). A significant inverse association was detected between EST immunoreactivity and tumor size, which was histologically measured (P = 0.003, ++ versus −; Bonferroni test), or lymph node status (P = 0.0027, χ² test). There was, however, no significant relationship between EST immunoreactivity and patient age, menopausal status, history of pregnancy, history of exogenous hormone use, stage, histological grade, ERα status, ERα LI, ERβ LI, PR status, PR LI, 17β-HSD type 1, Ki-67 LI, or HER-2/neu. A significant correlation between EST immunoreactivity and lymph node status was also confirmed when classification of lymph node status was increased to three groups (negative, one to three nodes positive and more than four nodes positive; P = 0.0154, χ² test).

Correlation between STS immunoreactivity and clinicopathological parameters in 113 breast carcinomas is summarized in Table 2, C and D. The number of cases in each group for STS immunoreactivity is summarized as follows: ++, n = 20 (17.7%); +, n = 64 (56.6%); and −, n = 29 (25.7%). A significant positive correlation was detected between STS immunoreactivity and tumor size (P = 0.0047, ++ versus −; Bonferroni test). There were, however, no significant associations between STS immunoreactivity and other clinicopathological parameters examined, including EST immunoreactivity.

Significant correlations described above were confirmed in increased rankings of positivity for EST and STS immunoreactivity to five groups (0, 1–25, 26–50, 51–75, and 75–100% positive cells), EST immunoreactivity and tumor size (P = 0.0089; 75–100 versus 0%), Bonferroni test, EST immunoreactivity and lymph node status (P = 0.0161, χ² test), and STS immunoreactivity and tumor size (P = 0.0101; 75–100 versus 0%; Bonferroni test).

Disease-free survival curves are illustrated in Fig. 3, A and B. EST immunoreactivity was associated with a decreased risk of recurrence (Fig. 3,A), whereas STS immunoreactivity was associated with an increased risk of recurrence (Fig. 3,B). A similar tendency was detected when EST and STS immunoreactivity was further categorized into five groups (0, 1–25, 26–50, 51–75, and 75–100% positive cells; Fig. 3,C). After univariate analysis by Cox (Table 3,A), lymph node status (P < 0.0001), tumor size (P = 0.0040), EST immunoreactivity (P = 0.0044), and STS immunoreactivity (P = 0.0118) were demonstrated to be significant prognostic factors for disease-free survival in 113 breast carcinoma patients. A multivariate analysis (Table 3 A), however, revealed that only lymph node status (P = 0.0011) and EST immunoreactivity (P = 0.0429) were independent prognostic factors with relative risks >1, whereas the other factors described above were not significant.

Overall survival curves are shown in Fig. 3, D and E. There was a significant correlation between clinical outcome and the expression of immunoreactive EST (Fig. 3,D) or STS (Fig. 3,E) in this study. A similar tendency was detected when EST and STS immunoreactivity was further subclassified into five groups (data not shown). Using a univariate analysis (Table 3,B), lymph node status (P = 0.0008), EST immunoreactivity (P = 0.0026), tumor size (P = 0.0124), STS immunoreactivity (P = 0.0325), and HER-2/neu immunoreactivity (P = 0.0333) turned out to be significant prognostic factors for overall survival in this study. Multivariate analysis revealed that only lymph node status (P = 0.0087) and EST immunoreactivity (P = 0.0149) were independent prognostic factors with a relative risk >1; however, other factors were not significant (Table 3 B).

In this study, one patient markedly immunopositive (++ group) for EST (n = 18) had recurrent disease and died after surgery. This patient was also evaluated as ++ for STS immunoreactivity.

In this study, EST immunoreactivity was detected in carcinoma cells and significantly associated with its mRNA level and enzymatic activity. EST immunoreactivity was also positive in the epithelial cells of morphologically normal glands. Enzymatic activity for EST has been demonstrated previously in various human breast carcinoma cell lines (14, 15) and normal human mammary epithelial cells (31). EST enzymatic activity has also been reported in human breast carcinoma and normal breast tissues (16, 17). In addition, the concentration of E1-S in breast cancer tissues has been reported to be significantly higher than plasma levels (32). Results from our present study are in good agreement with previous findings and suggest that EST is involved in the inactivation of local estrogens in human breast cancer tissues. In this study, EST mRNA expression was also detected in intratumoral stromal cells and adipocytes adjacent to the carcinoma. EST immunoreactivity was, however, not detected, suggestive of low levels of EST expression in these cells.

We found that EST immunoreactivity was inversely correlated with tumor size or lymph node status and associated with a decreased risk of recurrence or improved prognosis. Results of multivariate analyses demonstrated that EST immunoreactivity is an independent prognostic factor for both recurrence and overall survival, as well as lymph node status, a well-established diagnostic modality (33), although it may not be as robust of a prognostic factor for recurrence. Previous studies have demonstrated that MCF-7 breast cancer cells transfected with EST possess EST at levels similar to normal human mammary epithelial cells and are associated with much lower estrogen-stimulated DNA synthesis or cell proliferation than control MCF-7 cells that do not possess EST (31, 34). These findings suggest that the loss of EST may result in altered estrogen metabolism in breast cancer cells (31). Results from our present study are also consistent with these reports, and thus, it is possible to speculate that EST-negative breast carcinomas may be associated with an increment of in situ estrogen concentrations, thereby resulting in an increased recurrence and/or poor prognosis in these patients.

In our present study, STS expression was detected in breast carcinoma cells at both protein and mRNA levels. STS immunoreactivity was significantly associated with levels of mRNA and enzymatic activity and positively correlated with tumor size, risk of recurrence, and worse prognosis. E1-S is the most abundant estrogen in peripheral blood (35). Enzymatic activity for STS has been reported to be higher in breast cancer tissues than in normal breast tissues (36, 37). In addition, Saeki et al. (19) reported the immunolocalization of STS in carcinoma cells in 22 of 25 cases (88%). STS inhibitors have been demonstrated to be effective for depressing the proliferation of estrogen-dependent MCF-7 cells when E1-S was the source of estrogen (38), and Utsumi et al. (18) reported that patients with high mRNA levels for STS were associated with an increased risk of recurrence. Results from our present study are consistent with these reports published previously and may suggest that STS is involved in the in situ activation of E1 from E1-S, thereby contributing to the increment of estrogenic actions in human breast cancer tissues.

In our study, a statistically significant correlation was not detected between EST immunoreactivity and ER status, although EST enzymatic activity was correlated with both EST immunoreactivity and ERα LI in 35 breast carcinomas. This may be possibly because other members of the steroid-sulfotransferase superfamily, such as M-PST and phenol-sulfating phenol sulfotransferase, may also contribute to sulfonating estrogens to estrogen sulfates (14, 15, 39, 40). However, the role of M-PST in estrogen biotransformation has been in dispute (41), and expression of these enzymes remains largely unknown in breast carcinoma tissues. Therefore, additional investigations are required to fully characterize the biological significance of estrogen-sulfonation in breast carcinoma. Chetrite et al. (15) have reported that progesterone derivatives stimulate EST in breast cancer cell lines, but the correlation between EST and progesterone could not be demonstrated in this study, possibly because patients who had used exogenous hormones were markedly limited (only two cases). Moreover, no correlation was detected between EST immunoreactivity and PR status in patients examined in this study. This may be attributable, in part, to the fact that transcription of the PR gene is enhanced and maintained by estrogens. Furthermore, a positive PR status has been generally regarded as a marker of functional estrogenic pathways in breast carcinoma (42, 43). However, additional investigations are required to verify and/or clarify these hypotheses.

In addition, no association was detected between the expression for EST, STS, 17β-HSD type 1, and aromatase in our study. These findings may suggest that the expression of these enzymes is regulated by different mechanisms. Antiestrogens, such as tamoxifen, which block ER, have been used as an endocrine therapy in breast carcinoma for many years. Recently, the inhibition of intratumor estrogen production has also been considered as an endocrine therapy. Use of an aromatase inhibitor, e.g., is clinically useful in reducing the progression of breast tumors in postmenopausal women (44). Therefore, results from our present study suggest that induction of EST and/or inhibition of STS may also have possible therapeutic potential as an endocrine therapy for breast cancer, similar to the aromatase inhibitor.

In summary, we have examined the expression of EST and STS, both of which may regulate in situ production of estrogens, in human breast carcinomas using immunohistochemistry, RT-PCR, and enzymatic assay. In our study, EST immunoreactivity was found to be inversely correlated with tumor size or lymph node status and significantly associated with a decreased risk of recurrence and/or improved prognosis. STS immunoreactivity, in contrast, was found to be significantly associated with tumor size and an increased risk of recurrence and worsened prognosis. Results from our present study appear to suggest that both EST and STS immunoreactivity is associated with their levels of mRNA and enzymatic activity and that EST immunoreactivity is a potent prognostic factor in human breast carcinoma.

We thank Kumiko Hidaka, Department of Pathology, Tohoku University School of Medicine, for her skillful technical assistance. We also thank Andrew D. Darnel, Department of Pathology, Tohoku University School of Medicine, Sendai, Japan, for careful editing of this manuscript.