Immunolocalization of estrogen‐producing and metabolizing enzymes in benign breast disease: Comparison with normal breast and breast carcinoma

Abstract

It is well known that estrogens play important roles in the cell proliferation of breast carcinoma. Benign breast disease (BBD) contains a wide spectrum of diseases, and some are considered an important risk factor for subsequent breast carcinoma development. However, the significance of estrogens in BBD has remained largely unknown. Therefore, in this study, we examined tissue concentrations of estrogens and immunolocalization of estrogen‐producing/metabolizing enzymes in BBD, and compared these findings with those in the normal breast and ductal carcinoma in situ (DCIS). Tissue concentration of estradiol in BBD (n = 9) was significantly (3.4‐fold) higher than normal breast (n = 9) and nearly the same (0.7‐fold) as in DCIS (n = 9). Immunoreactivity of estrogen sulfotransferase in BBD was significantly lower (n = 82) than normal breast (n = 28) but was not significantly different from DCIS (n = 28). Aromatase and steroid sulfatase immunoreactivities tended to be higher (P = 0.07) in BBD than in normal breast, and 17β‐hydroxysteroid dehydrogenase type 1 immunoreactivity was significantly higher in BBD than normal breast in the postmenopausal tissues. Immunoreactivity of estrogen and progesterone receptors was also significantly higher in BBD than normal breast. These results suggest that tissue concentration of estradiol is increased in BBD at a level similar to DCIS, which is considered mainly due to loss of estrogen sulfotransferase expression. Increased local estradiol concentration in BBD due to aberrant expression of estrogen‐producing/metabolizing enzymes may play important roles in the accumulation of estradiol‐mediated growth and/or subsequent development of breast carcinoma. (Cancer Sci 2010)

Benign breast disease (BBD) is the most common disorder of the breast in women, and encompasses a wide variety of histological entities.^{( 1} , ^{2 )} Some BBD are considered an important risk factor for subsequent development of breast carcinoma in the same patients. Among the histological subtypes of BBD, atypical ductal hyperplasia (ADH) has the highest (approximately 4‐fold) risk of developing breast carcinoma, but patients with common proliferating BBD, such as fibroadenoma (FA), papilloma, sclerosing adenosis (SA), or usual ductal hyperplasia (UDH), are also known to be at increased (approximately 2‐fold) risk of malignancy, although controversies have existed.^{( 2} , ³ , ^{4 )}

Human breast tissue is a well‐known estrogen target tissue. Biologically active estrogen, estradiol, contributes immensely to the growth of breast carcinoma, and endocrine therapy has been used in patients with breast carcinoma to inhibit intratumoral estrogen actions. Benign breast disease frequently expressed estrogen receptor (ER),^{( 5} , ⁶ , ⁷ , ⁸ , ⁹ , ^{10 )} and women who had used postmenopausal hormonal supplementation were reported to be associated with an increased risk of BBD.^{( 11} , ^{12 )} Treatment with an anti‐estrogen tamoxifen significantly reduced the volume and cell proliferation activity of FA^{( 13} , ^{14 )} and The National Surgical Adjuvant Breast Project Breast Cancer Prevention Trial showed that tamoxifen significantly reduced the risk of BBD, by approximately 28%.^{( 15 )} These findings all suggest the importance of estrogens in the development and/or progression of BBD.

In breast carcinoma, estradiol is locally produced from circulating inactive steroids by estrogen‐producing enzymes, such as aromatase (conversion from androstenedione to estrone, or testosterone to estradiol), steroid sulfatase (STS; hydrolysis of estrone sulfate to estrone) and 17β‐hydroxysteroid dehydrogenase type 1 (17βHSD1; reduction of estrone to estradiol).^{( 16 )} Estrogens are also inversely inactivated by estrogen sulfotransferase (EST; sulfonation of estrogen to estrogen sulfate)^{( 17 )} or 17βHSD2 (oxidation of estradiol to estrone).^{( 16 )} Pasqualini et al. ^{( 18 )} reported that estradiol concentration and STS activity was significantly higher in FA than the corresponding normal human breast tissue, and Ariga et al. ^{( 8 )} reported immunolocalization of 17βHSD1 and 17βHSD2 in UDH and ADH. However, other information regarding estrogen concentration or estrogen‐producing/metabolizing enzymes is unknown in BBD, and the significance of estrogens remains largely uncharacterized in BBD. Therefore, in our present study, we examined the tissue concentration of estrogens and immunolocalization of estrogen‐producing/metabolizing enzymes in BBD, and compared these findings with those in normal breast and ductal carcinoma in situ (DCIS), an early stage of breast carcinoma, in order to analyze the status of in situ estrogen production.

Materials and Methods

Patients and tissues.  Snap‐frozen specimens of nine BBD (FA [n = 7] and SA [n = 2]), and nine pure DCIS tissues were used in order to examine the tissue concentration of estrogens. These specimens were obtained from nine premenopausal female patients (aged 23–48 years for BBD and 33–49 years for DCIS) who underwent surgical treatment from 2001 to 2008 in the Departments of Surgery at Tohoku University Hospital and Tohoku Kosai Hospital (Sendai, Japan). Snap‐frozen specimens of nine breast tissues, which were distant from the invasive breast carcinoma associated with no significant histopathological abnormalities, were also used in this study as normal breast tissues. These specimens were all obtained from premenopausal patients (aged 41–50 years) who had invasive breast carcinoma and underwent mastectomy in the Department of Surgery at Tohoku Kosai Hospital. All of the specimens were stored at −80°C, and histology was confirmed before estrogen extraction using frozen tissue sections.

Eighty‐two BBD lesions, comprising FA (n = 29), papilloma (n = 20), SA (n = 16), and UDH (n = 17), were obtained from 68 pure BBD patients by surgical excision between 1998 and 2007 in the Department of Surgery, Tohoku University Hospital. Examination of atypical papilloma or ADH was not available in this study. The mean age of BBD patients was 43 years (range, 16–68 years). Among these patients, the breast tissue associated with no significant histopathological abnormalities was available for examination in 28 cases. They were used as normal breast tissue in this portion of the study. Twenty‐eight specimens of pure DCIS were obtained by surgical excision between 1998 and 2007 in the Department of Surgery, Tohoku University Hospital. The mean age of the patients was 52 years (range, 30–77 years). All the specimens were fixed with 10% formalin and embedded in paraffin wax. The menopausal status of patients was retrieved from patient charts. The number of tissues obtained from postmenopausal patients (at least 1 year after menopause, aged more than 50 years) was eight in normal breast, 16 in BBD, and 16 in DCIS.

Research protocols for this study were approved by the Ethics Committee at Tohoku University School of Medicine (No. 2008–382).

Liquid chromatography/electrospray tandem mass spectro‐metry (LC‐MS/MS).  Concentrations of estrone and estradiol were measured by LC‐MS/MS analysis in ASKA Pharma Medical (Kawasaki, Japan), as described previously.^{( 19 )} Briefly, breast tissue specimens (∼50 mg for each sample) were homogenized in 1 mL distilled water, and steroid fraction was extracted with diethyl ether. After application to a Bond Elut C18 column (Varian, Inc., Palo Alto, CA, USA), steroid derivatives were eluted with 80% acetonitrile solution.

An API‐5000 triple stage quadrupole mass spectrometer equipped with an electrospray ionization (ESI) ion source (Applied Biosystems, Foster City, CA, USA) and a Shimadzu HPLC system (Shimadzu, Kyoto, Japan) were used in our study. The chromatographic separation was carried out on Cadenza CD‐C18 column (150 mm × 3 mm i.d., 3 μm; Imtakt, Kyoto, Japan) at 40°C. The mobile phase, consisting of CH3CN–CH3OH (50:50, v/v) (Solvent A) and 0.1% HCOOH (Solvent B), was used with a gradient elution of A:B = 60:40–90:10 (0–5.5 min), 90:10–100:0 (5.5–7.5 min), 100:0 (7.5–8.5 min) and 40:60 (8.5–10 min) at a flow rate of 0.4 mL/min. ESI/MS conditions were as follows: spray voltage, 3300 V; collision gas, nitrogen, 1.5 psi (gas pressure); curtain gas nitrogen, 11 psi; ion source temperature, 600°C; and ion polarity positive.

The derived estrone and estradiol were determined using product ions (m/z 157 and 264, respectively) produced from their individual protonated molecular ions. The limit of quantification was 4 fmol/g for estrone, and 2 fmol/g for estradiol.

Immunohistochemistry.  The characteristics of primary antibodies for aromatase,^{( 20 )} STS,^{( 21 )} and EST^{( 22 )} were described previously. The primary antibodies for aromatase and STS were kindly provided by Dr. Dean B. Evans (Novartis Institutes for BioMedical Research Basel, Oncology Research, Basel, Switzerland) and Dr. Taisuke Nakata (Strategic Product Planning Department, Kyowa Hakko Kirin, Tokyo, Japan), respectively. Monoclonal antibody for 17βHSD1 (EP1682Y) and polyclonal antibody for 17βHSD2 (10978‐1‐AP) were purchased from Epitomics (Burlingame, CA, USA) and Proteintech Group (Chicago, IL, USA), respectively. Monoclonal antibodies for ER (ER1D5) and progesterone receptor (PR; MAB429) were purchased from Immunotech (Marseille, France) and Chemicon (Temecula, CA, USA). A Histofine kit (Nichirei, Tokyo, Japan), which uses the streptavidin–biotin amplification method, was used for immunohistochemistry in this study. Antigen retrieval was carried out by heating the slides in an autoclave at 120°C for 5 min for ER, PR, and EST immunostaining, and by heating the slides in a microwave at for 20 min for 17βHSD1 immunostaining. The dilution of primary antibodies used in this study was as follows: aromatase, 1/3000; STS, 1/200; 17βHSD1, 1/400; EST, 1/100; 17βHSD2, 1/200; ER, 1/50; and PR, 1/30. 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.

Immunoreactivity of the estrogen‐producing/metabolizing enzyme was detected in the cytoplasm of epithelial or carcinoma cells, and cases that had more than 10% positive cells were considered positive.^{( 23 )} Immunoreactivity of ER and PR was detected in the nucleus of epithelial or carcinoma cells. These immunoreactivities were evaluated in more than 1000 epithelial or carcinoma cells for each case and subsequently the percentage of immunoreactivity, the labeling index (LI), was determined.

Statistical analysis.  The statistical analyses between two groups were carried out using the Mann–Whitney U‐test and cross‐table using the χ²‐test. Analyses among groups were carried out using Kruskal–Wallis test and cross‐table using the χ²‐test. P‐values <0.05 were considered significant. The relative ratio between two groups was evaluated by their median values.

Results

Tissue concentration of estrogens in BBD.  We first examined tissue concentration of estrogens in normal breast, proliferating BBD, and DCIS tissues using LC‐MS/MS. The median (minimum–maximum) value of tissue concentration of estradiol was 29 (4–82) pg/g in normal breast, 98 (3–674) pg/g in BBD, and 139 (12–494) pg/g in DCIS (Fig. 1a). The median value of estradiol in BBD was significantly (P = 0.02 and 3.4‐fold) higher than that in normal breast, but it was similar (P = 0.89 and 0.7‐fold) to that in DCIS. However, the median (minimum–maximum) value of estrone concentration in normal breast, BBD, and DCIS was 86 (46–560), 114 (20–195), and 113 (40–159) pg/g, respectively (Fig. 1b).

Figure 1.

 Tissue concentrations of estradiol (a) and estrone (b) in benign breast disease (BBD). Data are represented as box and whisker plots. The median value is represented by a horizontal line in each box. The 75th (upper margin) and 25th (lower margin) percentiles of the values are shown. The upper and lower bars indicate the 90th and 10th percentiles, respectively. Statistical analysis was carried out using the Mann–Whitney U‐test. P‐values <0.05 were considered significant, and are indicated in bold. DCIS, ductal carcinoma in situ; normal, breast tissue with no significant pathological abnormalities.

Immunolocalization of estrogen‐producing/metabolizing enzy‐mes in BBD.  We then immunolocalized estrogen‐producing (aromatase, STS, and 17βHSD1) and metabolizing (EST and 17βHSD2) enzymes in normal breast, proliferating BBD, and DCIS cases. Aromatase immunoreactivity was negative in the normal breast (Fig. 2a), but was focally detected in the epithelial cells in 9 out of 82 (11%) BBD lesions (Fig. 2b,c). Aromatase immunoreactivity was positive in the carcinoma cells in 13 out of 28 (46%) DCIS cases, and was also detected in some intratumoral stromal cells as reported previously.^{( 23 )} Steroid sulfatase immunoreactivity was negative in normal breast (Fig. 2d) except for 1 (4%) case; the enzyme was focally detected in epithelial cells in 14 out of 82 (17%) BBD (Fig. 2e,f). Steroid sulfatase immunoreactivity was positive in carcinoma cells in 10 out of 28 (36%) DCIS cases. 17βHSD1 immunoreactivity was detected in epithelial cells or carcinoma cells in 8 out of 28 (29%) normal breasts, 32 out of 82 (39%) BBD, and 17 out of 28 (61%) DCIS cases. Both STS and 17βHSD1 immunoreactivity was not detected in stromal cells of any of the specimens examined in this study. Estrogen sulfotransferase immunoreactivity was detected in the epithelial cells in 25 out of 28 (89%) normal breasts (Fig. 2g), but was positive in only 40 out of 82 (49%) BBD (Fig. 2h,i) and 14 out of 28 (50%) DCIS cases. 17βHSD2 immunoreactivity was detected in epithelial cells or carcinoma cells in 5 out of 28 (18%) normal breasts, 13 out of 82 (16%) BBD, and 3 out of 28 (11%) DCIS cases, respectively.

Figure 2.

 Representative illustrations of immunohistochemistry for estrogen‐producing/metabolizing enzymes in benign breast disease. Aromatase immunoreactivity was negative in normal breast (a), but was detected in the epithelium of the papilloma (b,c). Aromatase immunolocalization was observed in a part of the hyperplastic (b) or simple epithelium in the same case (c). Steroid sulfatase immunoreactivity was not detected in the normal breast (d), but it was positive in the epithelium of the papilloma (e,f). Steroid sulfatase immunoreactivity was observed in a part of the hyperplastic (e) or simple epithelium in the same case (f). Estrogen sulfotransferase immunoreactivity was detected in normal epithelial cells (g), but was negative in fibroadenoma (h). Estrogen sulfotransferase immunoreactivity was positive in various degree of hyperplastic epithelium in usual ductal hyperplasia (i). M, mild hyperplasia component (three to four cell layers thick); S, severe (or florid) hyperplasia component (solid duct hyperplasia). Bar = 100 μm.

Association of immunoreactivity of these enzymes between BBD and normal breast or DCIS is summarized in Table 1. Immunoreactivity of aromatase and STS was more frequent in BBD than normal breast, although the association did not reach statistical significance (P = 0.07) (Table 1A). Immunoreactivity of aromatase, STS, and 17βHSD1 was significantly higher in DCIS than BBD (P < 0.0001 in aromatase; P = 0.04 in STS; P = 0.046 in 17βHSD1). However, EST immunoreactivity in BBD was significantly (P = 0.0004) lower than in normal breast, but was similar to that in DCIS (P = 0.91) (Table 1B). No significant association of 17βHSD2 immunoreactivity was detected between BBD and normal breast (P = 0.81) or DCIS (P = 0.51) in this study.

Table 1.

 Immunoreactivity for (A) estrogen‐producing enzymes, (B) estrogen‐metabolizing enzymes in normal breast, benign breast disease (BBD), and ductal carcinoma in situ (DCIS)

Enzymes	Immunoreactivity	Normal (n = 28)	BBD (n = 82)	DCIS (n = 28)	P‐value
					Normal vs BBD	BBD vs DCIS
A
Aromatase	+	0 (0%)	9 (11%)	13 (46%)
	−	28 (100%)	73 (89%)	15 (54%)	0.07	<0.0001
STS	+	1 (4%)	14 (17%)	10 (36%)
	−	27 (96%)	68 (83%)	18 (64%)	0.07	0.0400
17βHSD1	+	8 (29%)	32 (39%)	17 (61%)
	−	20 (71%)	50 (61%)	15 (39%)	0.32	0.0460
B
EST	+	25 (89%)	40 (49%)	14 (50%)
	−	3 (11%)	42 (51%)	14 (50%)	0.0004	0.9100
17βHSD2	+	5 (18%)	13 (16%)	3 (11%)
	−	23 (82%)	69 (84%)	25 (89%)	0.8100	0.5100

Data are presented as the number of cases and percentage in each histological group. P‐values were evaluated by a cross‐table using the χ²‐test between two histological groups. P‐values <0.05 were considered significant, and are shown in bold. EST, estrogen sulfotransferase; normal, breast tissue showing no significant pathological abnormalities; STS, steroid sulfatase; 17βHSD1, 17β‐hydroxysteroid dehydrogenase type 1.

Statistical association of the estrogen‐producing/metabolizing enzyme immunoreactivity between BBD and normal breast or DCIS according to the menopausal status is summarized in Table 2. Aromatase immunoreactivity was significantly (P = 0.003) higher in DCIS than BBD in premenopausal tissues, and a similar tendency (P = 0.06) was also detected in postmenopausal cases. 17βHSD1 immunoreactivity was significantly (P = 0.02) higher in BBD than normal breast in postmenopausal tissues. Statistical association of EST immunoreactivity was detected between BBD and normal breast regardless of the menopausal status of subjects examined (P = 0.01). 17βHSD2 immunoreactivity was significantly (P = 0.01) lower in DCIS than BBD in the postmenopausal tissues.

Table 2.

 Statistical association of immunoreactivity of estrogen‐producing/metabolizing enzymes between benign breast disease (BBD) and normal breast or ductal carcinoma in situ (DCIS) according to menopausal status

Enzymes	Premenopausal tissues		Postmenopausal tissues
	Normal (n = 20) vs BBD (n = 66)	BBD (n = 66) vs DCIS (n = 20)	Normal (n = 8) vs BBD (n = 16)	BBD (n = 16) vs DCIS (n = 16)
Estrogen‐producing enzymes
Aromatase	0.16	0.003	0.19	0.06
STS	0.19	0.180	0.19	0.24
17βHSD1	0.98	0.180	0.02	>0.99
Estrogen‐metabolizing enzymes
EST	0.01	0.73	0.01	>0.99
17βHSD2	0.45	0.43	0.37	0.01

Data are presented as P‐values, evaluated by a cross‐table using the χ²‐test between two histological groups. P‐values <0.05 were considered significant, and are shown in bold. EST, estrogen sulfotransferase; normal, breast tissue showing no significant pathological abnormalities; STS, steroid sulfatase; 17βHSD1, 17β‐hydroxysteroid dehydrogenase type 1.

Estrogen receptor and progesterone receptor immunoreactivity in BBD.  Nuclear ER immunoreactivity was detected in the epithelial cells or carcinoma cells of normal breast, BBD (Fig. 3a), and DCIS. Estrogen receptor LI was significantly (P < 0.0001 and 2.8‐fold) higher in BBD than that in the normal breast, and it was significantly (P < 0.0001 and 3.3‐fold) higher in DCIS than BBD (Fig. 3b). Progesterone receptor immunoreactivity was also detected in epithelial cells or carcinoma cells of normal breast, BBD (Fig. 3c), and DCIS. Progesterone receptor LI in BBD was significantly (P < 0.0001 and 6.5‐fold) higher than that in normal breast, but was not significantly different (P = 0.39 and 1.0‐fold) from that in DCIS (Fig. 3d). Progesterone receptor LI was positively associated (P = 0.002 and r = 0.34) with ER LI in 82 BBD tissues examined.

Figure 3.

 Estrogen receptor (ER) and progesterone receptor (PR) immunoreactivity in benign breast disease (BBD). (a,c) Estrogen receptor (a) and progesterone receptor (c) immunoreactivity was detected in the nuclei of epithelial cells in this case of fibroadenoma. Images show the same area. Bar = 100 μm. (b,d) Association of ER (b) and PR (d) labeling indexes (LI) between BBD and normal breast or ductal carcinoma in situ (DCIS). Data are represented as box and whisker plots. The statistical analysis was carried out using the Mann–Whitney U‐test. P‐values <0.05 were considered significant, and are shown in bold.

Immunohistochemical features of enzymes in BBD according to histological types.  As summarized in Table 3, no significant differences of immunoreactivity for aromatase, STS, 17βHSD1, EST, 17βHSD2, ER, or PR were detected among four different histological subtypes of proliferating BBD examined in this study. It is known that BBD includes variable histological appearances.^{( 24 )} However, by immunohistochemical observations, we could not find an association between such a histological appearance and immunoreactivity of estrogen‐producing/metabolizing enzymes. For instances, immunoreactivity of aromatase (Fig. 2b,c) and STS (Fig. 2e,f) was focally detected in the papilloma regardless of the degree of epithelial hyperplasia, and EST immunoreactivity was observed in various thicknesses of the epithelium in UDH (Fig. 2i).

Table 3.

 Immunoreactivity of estrogen‐producing/metabolizing enzymes and sex hormone receptors in benign breast disease (BBD) according to histological types

Enzymes	Immunoreactivity	FA	Papilloma	SA	UDH	P‐value
		(n = 29)	(n = 20)	(n = 16)	(n = 17)
Estrogen‐producing enzymes
Aromatase	+	2 (7%)	3 (15%)	3 (19%)	1 (6%)
	−	27 (93%)	17 (85%)	13 (81%)	16 (94%)	0.52
STS	+	6 (21%)	4 (20%)	3 (19%)	1 (6%)
	−	23 (79%)	16 (80%)	13 (81%)	16 (94%)	0.59
17βHSD1	+	12 (41%)	10 (50%)	5 (31%)	5 (29%)
	−	17 (59%)	10 (50%)	11 (69%)	12 (71%)	0.54
Estrogen‐metabolizing enzymes
EST	+	18 (62%)	9 (45%)	7 (44%)	8 (47%)
	−	11 (38%)	11 (55%)	9 (56%)	9 (53%)	0.54
17βHSD2	+	2 (7%)	4 (20%)	3 (19%)	4 (24%)
	−	27 (93%)	16 (80%)	13 (71%)	13 (76%)	0.41
Sex hormone receptors
ER LI†	NA	31 (0–74)	23 (4–88)	21 (4–89)	42 (0–82)	0.65
PR LI†	NA	45 (1–86)	24 (0–82)	28 (0–77)	44 (0–82)	0.11

†Data represent median (minimum–maximum), and P‐values were evaluated by the Kruskal–Wallis test. Other values are presented as the number of cases and percentage in each histological group, and P‐values were evaluated by a cross‐table using the χ²‐test. ER, estrogen receptor; EST, estrogen sulfotransferase; FA, fibroadenoma; NA, not applicable; normal, breast tissue showing no significant pathological abnormalities; PR, progesterone receptor; SA, sclerosing adenosis; STS, steroid sulfatase; UDH, usual ductal hyperplasia; 17βHSD1, 17β‐hydroxysteroid dehydrogenase type 1.

When we examined an association among immunoreactivity of estrogen‐producing/metabolizing enzymes in 82 BBD tissues, positive association was detected between aromatase and 17βHSD1 (P = 0.03), aromatase and 17βHSD2 (P = 0.003), and STS and 17βHSD1 (P = 0.02). However, these immunoreactivities were not necessarily colocalized in the same lesions or constitutive cells of the one lesion within a case by immunohistochemical observations.

Immunohistochemical features of enzymes in DCIS.  Associations between immunoreactivity of estrogen‐producing/metabolizing enzymes and clinicopathological parameters in the 28 DCIS cases are summarized in Table 4. Among these enzymes, immunoreactivity for EST and 17βHSD2 was inversely associated with ER LI (P = 0.01 and P = 0.04, respectively). Immunoreactivity for STS tended to be positively associated with Van Nuys classification (P = 0.051) in this study, and a significant positive association between these was previously reported in 83 DCIS cases.^{( 23 )} No significant association was detected among the immunoreactivity of these enzymes in DCIS cases examined.

Table 4.

 Statistical association between immunoreactivity of estrogen‐producing/metabolizing enzymes and clinicopathological parameters in 28 cases of ductal carcinoma in situ

Enzymes	Patient age	Menopausal status	Van Nuys classification	ER LI	PR LI
Estrogen‐producing enzymes
Aromatase	0.91	0.95	0.370	0.17	0.56
STS	0.58	0.99	0.051	0.25	0.98
17βHSD1	0.08	0.16	0.690	0.16	0.12
Estrogen‐metabolizing enzymes
EST	0.25	0.52	0.610	0.01	0.27
17βHSD2	0.37	0.71	0.250	0.04	0.37

Data are presented as P‐values, evaluated by a cross‐table using the χ²‐test or the Mann–Whitney U‐test. P‐values <0.05 were considered significant, and are shown in bold. 17βHSD1, 17β‐hydroxysteroid dehydrogenase type 1; ER, estrogen receptor; EST, estrogen sulfotransferase; LI, labeling index; PR, progesterone receptor; STS, steroid sulfatase.

Discussion

In our present study, tissue estradiol concentration in BBD was significantly (P = 0.02 and 3.4‐fold) higher than that in normal breast, and was similar to that in DCIS. Pasqualini et al. ^{( 18 )} reported that estradiol concentration of FA was three times higher than the corresponding normal breast in premenopausal women, which is in good agreement with results of our present study. Estradiol concentration in normal breast tissue was low regardless of the menopausal status, although it is very high in plasma before menopause.^{( 25 )} However, the estradiol concentration was fivefold higher in invasive breast carcinoma tissue than plasma in premenopausal women,^{( 26 )} and intratumoral estradiol level in DCIS was nearly comparable to that in invasive breast carcinoma.^{( 23 )} Therefore, these data suggest that estradiol is predominantly metabolized in normal breast tissue, but its concentration is significantly increased in the proliferating BBD at a level comparable with that of breast carcinoma.

Among the enzymes examined in this study, statistically significant association was detected between normal breast and BBD only in the cases of EST (Table 1), and EST immunoreactivity was significantly (P = 0.004) lower in BBD than normal breast. Estrogen sulfotransferase is the only sulfotransferase that displays affinity for estradiol in a physiological concentration range, and the sulfating activity for estradiol was stronger than that for estrone.^{( 17} , ²⁷ , ^{28 )} Estrogen sulfotransferase is expressed in a wide variety of human tissues including breast, and is considered to be involved in protecting peripheral tissues from circulating excessive estrogenic effects.^{( 22} , ^{29 )} Estrogen sulfotransferase immunoreactivity in breast carcinoma was frequently decreased compared to that in normal breast tissue, and was also shown to be inversely correlated with the tumor size or lymph node status.^{( 22 )} Falany et al. ^{( 30 )} reported that MCF‐7 breast carcinoma cells transfected with EST possessed EST at levels similar to normal human mammary epithelial cells, and these cells were actually associated with much lower estradiol‐stimulated cell proliferation than control MCF‐7 cells that do not express EST. Moreover, Fu et al. ^{( 31 )} very recently examined EST mRNA expression in an MCF‐10A‐derived lineage cell culture model, and reported that EST was abundantly expressed in MCF‐10A and preneoplastic MCF‐10AT1 cell lines, but was also markedly repressed in neoplastic MCF‐10A‐derived cell lines as well as in MCF‐7 cells. These results all suggest that the loss of EST expression results in the increased local estradiol level in BBD by reducing its metabolism. Thus, EST is a possible key regulator of the estradiol concentration in BBD.

A great majority of BBD is well known to occur in premenopausal women,^{( 1} , ³² , ³³ , ^{34 )} which may be partly explained by decreased EST expression in BBD, as indicated in our present study. However, it is also true that some cases of BBD arise after menopause when serum estradiol level is negligible. In our present study, the immunoreactivity of aromatase and STS tended to be higher in BBD than normal breast, although P values did not reach statistical significance (P = 0.07). 17βHSD1 immunoreactivity was also significantly higher in BBD than normal breast in postmenopausal tissues. Pasqualini et al. ^{( 18 )} reported that STS activity in BBD was significantly higher than in corresponding normal breast tissue. 17βHSD1 immunoreactivity was also shown to be progressively increased according to UDH, ADH, and DCIS,^{( 8 )} and Miyoshi et al. ^{( 35 )} reported that 17βHSD1 mRNA levels in invasive breast carcinoma were significantly higher in postmenopausal than in premenopausal tissue, suggesting the importance of 17βHSD1 associated with increased estradiol levels, especially in postmenopausal patients. Therefore, expression of estrogen‐producing enzymes, such as aromatase, STS, and 17βHSD1, may also contribute to an increment of estradiol concentration in BBD.

In our study, ER and PR immunoreactivity was significantly higher in proliferating BBD than normal breast. Previous reports also indicated that ER and PR expression was increased in BBD compared to normal breast at both mRNA and protein levels.^{( 5} , ⁶ , ⁷ , ⁹ , ¹⁰ , ³⁶ , ^{37 )} Results of our present studies are consistent with these results. In addition, our present study also showed significant (P = 0.002 and r = 0.34) positive association between ER and PR immunoreactivity in BBD. Progesterone receptor expression is mainly regulated by estradiol through ER, and PR immunoreactivity in general reflects functional estrogen actions in breast carcinoma.^{( 38} , ^{39 )} Cell proliferation activity of ER‐positive cells was higher in BBD than normal breast.^{( 9} , ^{36 )} As the immunoreactivity of estrogen‐producing/metabolizing enzymes and sex hormone receptors was not significantly different among the histological types of proliferative BBD, as shown in Table 3, increased estrogen actions are postulated to be mainly associated with the progression, rather than the morphogenesis or pathogenesis, of the BBD lesion. In addition, proliferating BBD represents a risk factor for breast carcinoma,^{( 2} , ³ , ^{4 )} and Shekhar et al. ^{( 40 )} showed that estradiol treatment frequently resulted in atypical hyperplasia, carcinoma in situ, and invasive carcinoma from proliferative BBD using an MCF‐10AT xenograft model. Therefore, an increment of estradiol level in BBD due to aberrant expression of estrogen‐producing/metabolizing enzymes may play an important role in accumulation of estradiol‐mediated growth and subsequent development of breast carcinoma in BBD. Further investigations are warranted to clarify this issue.