Molecular Feminization of Mouse Seminal Vesicle by Prenatal Exposure to Diethylstilbestrol: Altered Expression of Messenger RNA

0022-5347/94/1515-1370$03.00/0 Vol. 151, 1370-1378, May 1994

THE JOURNAL OF UROLOGY Copyright © 1994 by AMERICAN UROLOGICAL ASSOCIATION, INC.

Printed in U.S.A.

MOLECULAR FEMINIZATION OF MOUSE SEMINAL VESICLE BY PRENATAL EXPOSURE TO DIETHYLSTILBESTROL: ALTERED EXPRESSION OF MESSENGER RNA W. C. BECKMAN, JR., RETHA R. NEWBOLD,* CHRISTINA T. TENG AND JOHN A. McLACHLAN From the Developmental Endocrinology and Pharmacology Section, Laboratory of Reproductive and Developmental Toxicology, Division of Intramural Research, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina

ABSTRACT

Exposure to estrogens during critical stages of development has been reported to cause irreversible changes in estrogen target tissues such as the reproductive tract. In fact, recent studies using mice describe prenatal estrogen exposure resulting in the expression of the major estrogen-inducible uterine secretory protein, lactoferrin (LF), by the seminal vesicles of the male offspring. Thus, we have studied the role of estrogens in abnormal and normal gene expression in the developing male reproductive tract using LF and seminal vesicle secretory protein IV (SVS IV), an androgenregulated murine seminal vesicle secretory protein, as markers. Lactoferrin and SVS IV protein and mRNA expression were studied in histological samples by using the techniques of in situ hybridization (ISH) and immunohistochemistry (IHC). Seminal vesicle secretory protein IV was expressed in all (100%) epithelial cells of the control seminal vesicle, but this protein was decreased by castration. However, LF expression was undetectable by ISH or IHC in control seminal vesicle epithelium. Lactoferrin was inducible in 2% of the seminal vesicle epithelial cells from adult castrated mice treated with estradiol 17{J (E 2 ; 20 ~g./kg./day for 3 days), indicating that a small percentage of the seminal vesicle cells could be induced to secrete LF after modification of the endocrine environment. Prenatal DES treatment (100 ~g./kg. maternal body weight on days 9 through 16 of gestation) resulted in the male offspring exhibiting constitutive expression of LF in 5% of the seminal vesicle epithelial cells, while expression of the androgen-regulated protein SVS IV was slightly decreased. The maximal contrast between LF and SVS IV expression was observed in prenatally DES-treated mice that were subsequently castrated as adults and further treated with E 2 ; LF was detected in 40% of the epithelial cells in these mice. Double immunostaining techniques revealed that epithelial cells which were making LF had ceased production of SVS IV. Since a large percentage of the epithelial cells in the intact prenatal DES exposed male was capable of expressing the normal gene product, SVS IV, it was concluded that DES treatment during prenatal development appears to imprint or induce estrogenic sensitivity in the adult seminal vesicle, causing increased production of LF. The results suggest that this altered protein response may be an example of atypical gene expression in male reproductive tract tissues following hormonal manipulation early in development. KEY WORDS:

diethylstilbestrol, lactoferrin, seminal vesicles, in situ hybridization, immunohistochemistry

Exposure of the developing fetus to potent estrogens, such as diethylstilbestrol (DES), at critical stages in gestation has been associated with abnormalities. These include benign and malignant lesions in the reproductive tract of both rodents and humans. This finding is of particular concern since many humans were exposed to DES when the drug was administered to pregnant women. 1 Furthermore, since many chemicals, insecticides, drugs and natural compounds found in our environment have estrogenic activity,2 exposure to these substances could have significant lasting effects on the susceptible developing fetus. Thus, the murine model is valuable and predictable in evaluating possible insults to humans exposed to DES and other environmental drugs and compounds possessing estrogenic activity. Recent studies from our laboratory describing the lasting effects of estrogens on the developing male mouse reproductive tract suggest that embryonic mesonephric (Wolffian) derived-structures such as the seminal vesicle are affected by prenatal DES exposure, as are tissues derived from paramesonephric (Mullerian) duct structures. 3• 4 We continue to Accepted for publication April 23, 1993. * Requests for reprints: Laboratory of Reproductive and Developmental Toxicology, MD E4-02, National Institute of Environmental Health Science, P.O. Box 12233, Research Triangle Park, North Carolina 27709.

investigate the role of estrogens in the developing male reproductive tract by studying the expression and localization of two proteins, lactotransferrin (lactoferrin, LF) and seminal vesicle secretory protein IV (SVS IV), in seminal vesicle tissues as markers for abnormal and normal gene expression, respectively. Lactoferrin, an estrogen-inducible mouse uterine protein, belongs to the class of proteins involved in the transport of Fe+++ cations in mammalian biological fluids. 5 Lactoferrin was expressed as a 65-70 kD component in other female reproductive organs such as the oviduct, vagina and mammary glands. 5 However, LF is not uniquely expressed in female tissues; it has also been demonstrated in mouse dorsolateral prostate. 3 Indeed, its distribution is quite varied among mammals. It is found in such diverse tissues as the reticuloendothelial system,6 salivary glands 7 and bronchi8 as well as in reproductive tract organs. 9 Even within the reproductive tract, the tissue sources of LF vary significantly among species. In the rabbit, LF is found in the ductus deferens and seminal vesicle, but not in the prostate,lO whereas in the mouse it occurs in the dorsolateral prostate but not in the seminal vesicle under normal endocrine conditions. 3 A polyclonal antibody to mouse uterine LFll and a mouse LF cDNN 2 have been prepared and characterized. It has been

1370

1371

DES FKMINIZATION OF SEMINAL VESICLE TABLE

1. Experimental design

CONTROL

(C+E) CONfROL INTACf PLUS

INTACf

ESTROGEN

(C)

CORN OIL - - - CORN OIL-TREATED OFFSPRING (C ()')

CONfROL, ORCHIECfOMIZED

DES-------

DES-TREATED

OFFSPRING

<

(DES+E) (DES) _ _ _ _ _ _ DES-TREATED DES-TREATED INTACT INTACf PLUS ESTROGEN (DES
ORCHlECTOMIZED

DAILY MATERNAL INJECTIONS (DES OR CORN OIL)

t

4,

VAGINAL PLUG

(DES&+E) DES-TREATED ORCHlECTOMIZED PLUS ESTROGEN

DAILY INJECTIONS

(1713-E,)

m

I Day 0 2

(C0+E) CONTROL, ORCHIEcroMIZED PLUS ESTROGEN

6 8 10 12 14 16 18 19 ~e...---;Wk;;;--;2'----'4;--;;6:--;;8----'~---"D:--ay--;:CO--;-1--"-2"""3;:-"4-:--5"..-,6;:-;7::;---,8,---;9,--------

T

DAYO!BIRTH PERIOD OF GENITAL TRACT DIFFERENTATION

shown that uterine expression of both LF mRNA 12 and protein 5 are increased by 3 days of estrogen administration. Lactoferrin secretion is not usually demonstrable in the male mouse seminal vesicle. Previous studies have shown that estrogen treatment alone was ineffective in inducing LF secretion in this tissueY However, it was found that castration of adult male mice or castration followed by estrogen treatment was able to cause a small increase in LF expression in the seminal vesicleY In contrast, prenatal exposure of male mice to DES was found to induce the constitutive expression of LF mRNA in the seminal vesicle. 13 Furthermore, an 8-fold increase in LF mRNA expression could be obtained by castrating the mice, The administration of estrogen to the castrated DES-mice led to a further increase in LF expressionY It was concluded from these studies that prenatal DES treatment had either altered the molecular mechanisms which androgens regulate the function of the seminal vesicle or selectively altered a subpopulation of seminal vesicle epithelial cells, or both. 13 Pentecost et al. 13 had measured LF expression gel electrophoresis of either protein or mRNA from whole tissue extracts. The techniques employed did not permit an assessment of the cell types involved in the production and control of the LF secretion. Subsequent studies from this laboratory 3 using immunohistochemical techniques showed that LF protein was localized in the epithelial cells of the seminal vesicle from mice that had been prenatally exposed to DES. The location of the message for this protein remained unknown, and it was not clear if these LF -containing epithelial cells were also producing the characteristic seminal vesicle secretory protein, SVS IV. Seminal vesicle secretory protein IV is one of several androgen-dependent seminal vesicle proteins. 14 It has biological significance in formation of the copulatory plug l5 • 16 and hence is important for male fertility.17 The eDNA coding for SVS IV mRNA in both rae 8 • 19 and mouse 20 has been cloned and sequenced. Earlier studies suggested that male mice prenatally exposed to DES secreted SVS IV into the seminal vesicle fluid,

DAYOFCLTION (8--12 wks of age)

DAJOF SACRIFICE

but that the amounts could be decreased by castration and treatment with estradioL 13 The present investigation was designed to study the patterns of expression of lactoferrin and seminal vesicle secretory protein IV at the cellular level in mice treated prenatally with DES, The objective was to determine whether the same cells in the seminal vesicle are capable of both LF and SVS IV production or if DES exposure produces a new subpopulation of cells that carry the message for LF alone. MATERIALS AND METHODS

Female CD-l mice [CrL:CD-l nCR) BR; Charles River Laboratories, Raleigh, North Carolina] were housed with experienced breeder males and checked daily for vaginal plugs. Vaginal plug detection was considered day 0 of pregnancy. Mice were maintained in the NIEHS animal facility under approved laboratory animal care and feeding standards (21-22°C; 12/12 light cycles; NIH 31, with free access to chow and fresh water). Pregnant females were separated into control (corn oil vehicle treated) and DES-treated groups. Mice in the DES-treated group received daily injections of chromatographically purified DES (Sigma Chemical Co., St. Louis, Missouri; >99% pure). Mice received diethlystilbestrol (DES; 100 ~g.jkg. maternal body weight in 0.01 ml./gm. corn oil) subcutaneously in the loose skin on the back of the neck from days 9 through 16 of gestation. Male offspring born to this group were designated as DES-IOO males (DES group). Pregnant control females received corn oil alone, and their male offspring were designated control males (C group). Day 19 of gestation was standardized for delivery, either naturally or by Caesarean section, and litter sizes were adjusted to eight. At 22 days of age, pups were weaned, segregated by sex and housed in groups of five per cage. Between 8 and 12 weeks of age, one-half of the males from C and DES groups were castrated (C~ and DES~). Under pentobarbital sodium anesthetic supplemented with metafane anesthesia, castration

FIG. 1. Immunohistochemical localization of lactoferrin (LF) in accessory reproductive organs of mice from four treatment groups: intact (A, B), intact plus E2 (3 daily injections, 20 Ilg./kg.) (C, D), orchiectomized (6 days) plus vehicle (3 days) (E, F) and orchiectomized (6 days) plus E2 (3 days). Left-hand panel (A, C, E, G) represents animals from controls. Right-hand panel (B, D, F, H) represents tissues from animals treated with DES prenatally. Expression of LF was observed in only 2% of cells of orchiectomized, ~ treated control animals (G), but not observed in other control groups. G illustrates maximal amount of LF immunoreactivity observed in C~+E samples. Light brown reaction product was present over intracellular granules in all prenatal DES animals (B, D, arrows). Immunostaining was also observed in secretory fluid of these treatment groups. Castration was more effective than E2 alone in increasing intensity of both extra- and intracellular staining (F vs. D). A few cells in this group showed uniform cytoplasmic staining which obscured detail of individual granules (arrow). Combination of castration and E2 treatment was most effective in inducing LF expression, both in control (G) and DES-treated (H) groups. (hematoxylin counterstain, 430x).

DES FEMINIZATION OF SE]VHNAL VESICLE

1373

FIG. 2. In situ hybridization (ISH) autoradiograms of LF mRNA expression in accessory reproductive organs of mice from two treatment groups: intact (A, B), and orchiectomized plus 3 days of E2 (C, D). Left- hand panel represents animals from controls. Right-hand panel represents animals from mothers prenatally treated with DES. Tissues illustrated were from frozen tissue samples. Increased silver grain intensity was observed in animals treated with both castration and E 2 • Findings were similar to those observed with immunohistochemistry (See fig. 1, A, B, G, H) (methyl green-pyronin counterstain, 430x).

was performed through a scrotal incision; mice were allowed to recover for 6 days. One-half of the males in the above four groups received daily subcutaneous injections of 17(3-estradiol (E 2; 20 ,ug./kg./day) for 3 days (C + E, DES + E, C~ + E, DES ($ + E). The other one-half received vehicle alone (C, DES, 6~, DES~). Mice were sacrificed by cervical dislocation on the day following the last injection. The experimental protocol is summarized in Table 1. A minimum of three mice were included in each of the eight experimental groups. As a positive control for the SVS IV probe, intact adult male mice were used. Seminal vesicle tissues were collected and processed as described below. As a positive control for the LF probe, immature (17 days old) female CD-1 mice (Charles River Laboratories) were used. Female mice received subcutaneous injections of DES in corn oil (100 ,ug./kg. body weight) or corn oil alone for 3 consecutive days. At 20 days of age, mice were sacrificed and uterine tissues were collected, fixed and processed as described for male tissues. Tissue collection. After three days of E2 treatment, male mice were sacrificed by cervical dislocation. Seminal vesicle tissues were removed and either immediately frozen on cryostat stubs in liquid nitrogen or prepared for paraffin histology as previ-

described. 3 Briefly, tissues were immersed in Bouin's fixative overnight and rinsed in three changes of cold deionized water, followed by serial washes in 50% and 70% ethanol prior to embedding in paraffin. Either frozen or paraffin sections (6 ,urn.) from each experimental group were cut and mounted on microscope slides coated with polylysine (Sigma Chemical Co.). Paraffin was cleared through a series of xylene and alcohol washes, and tissues were hydrated in RNase-free 2x saline sodium citrate (SSC; 0.89% NaCl, 0.1 M. sodium citrate) buffer for 1 hour before processing for immunohistochemistry (IHC) or in situ hybridization (ISH). Immunohistochemistry. Single immunostaining. Tissues were immunostained with rabbit primary antisera to LF and SVS IV diluted from whole serum to working dilutions of 1:1000 for overnight incubations at room temperature. The ABC (Vectastain Elite Kit, Vector Laboratories, Burlingame, California) method was used as the bridging system, with diaminobenzidine/hydrogen peroxide as the chromogen. Specificity of the SVS IV and LF antisera was determined in previous studies using Western Blot analysis. 3 ,13 Uterine tissue sections were run with each batch of seminal vesicle tissue as a positive control. Other controls employed in these studies included

1374

DES FEMINIZATION OF SEMINAL VESICLE

omission of the primary antiserum to control for background staining and the comparison of immunolocalization of protein with distribution of mRNAs in the seminal vesicle and coagulating gland. Before immunostaining, sections were treated for 3 minutes with a solution of 1.2 mg./ml. pronase (Sigma, type XIV) to ensure detectability of antigenic sites. Slides were rinsed in phosphate buffered saline and incubated in normal sheep serum 10 minutes before application of the primary antiserum. Double immunostaining. Tissues were prepared as for the single immunostaining protocol. The slides were first incubated overnight at room temperature with LF antiserum (1:1000). This antiserum was detected using the peroxidase antiperoxidase (PAP) bridge kit (DAKO Corp., Carpinteria, California). The chromogen was diaminobenzidine, which yields a brown reaction product. The immunostained sections were photographed and then soaked for 2 hours at room temperature in a solution of 0.1 M. glycine buffer (pH 2.2)21 to remove the first antibody/enzyme bridge. The sections were again blocked with normal serum and incubated with SVS IV antiserum (1:1000, overnight at room temperature). The second antibody was detected with the same PAP method, but using 4-chloronaphthol (blue reaction product) as the chromogen. Sections were again photographed. Controls consisted of omitting the primary antisera, and use of several positive and nonpositive reproductive accessory organ tissues for each of the antigens. Quantitation of immunostaining. Slides immunohistochemically stained for lactoferrin were examined with the 20x lens of a photomicroscope. A hand-held cell counter was used to evaluate the percentage of LF -containing cells in three separate areas of seminal vesicle tissue on each slide. A minimum of 200 cells was counted in each area. A cell was considered positive if it had a brown reaction product localized over secretory granules or uniformly distributed within the cytoplasm. The number of LF-containing cells was divided by the total number of cells counted. Groups were compared by an unpaired twotailed T test. In situ hybridization. Complementary DNA probes for LF (2.2 Kb) and SVS IV (556 bp) were random primed with alphaP2P]-dCTP (Dupont-NEN) to a specific activity of 1 X 106 cpm/ng. probe. Hybridizations were performed on both frozen and paraffin embedded tissues. Tissues for both procedures were fixed overnight at 4C in 10% neutral buffered formalin or 4% paraformaldehyde. The next day, tissues were rinsed (4 X 15 minutes) in phosphate buffer and either snap frozen onto cryostat stubs for frozen sectioning or dehydrated through alcohols and xylenes for paraffin embedding. Frozen sections (6 ~m.) were thawed onto slides coated with polylysine, postfixed in 4% paraformaldehyde and stored in saline/sodium citrate buffer (2X SSC)22 before prehybridization. Paraffin sections were deparaffinized through graded xylenes and alcohols and hydrated in 2X SSC. Hydrated tissue sections were prehybridized for 1 hour with hybridization buffer, blotted and covered with 100 ~l. of hybridization buffer containing 2 ng. of probe. Sections were hybridized at 37C overnight. Following hybridization, sections were gently rinsed with a stream of a 1:1 mixture of formamide and 4X SSC (37C), then washed 2 X 30 minutes in the same solution (37C). Subsequent washes were in 2X SSC, 1X SSC and O.lx SSC (each 2 X 30 minutes at room temperature). Sections were then rinsed in distilled water, dehydrated through a graded series of ethanols (3 minutes each), air dried and coated with a 1:1 dilution of Kodak NTB-3 emulsion (Eastman Kodak Co., Rochester, New York) in water. Sections were exposed in black opaque autoradiography boxes for 1 to 4 weeks, developed in D-19 developer (Eastman Kodak Co.) for 45 seconds, fixed in Kodak Fixer (Eastman Kodak Co.) for 5 minutes, rinsed, stained with hematoxylin and eosin or methyl green-pyronin and photographed.

RESULTS

In intact male mice, the seminal vesicle is a large, commashaped gland containing white secretory fluid. Seminal vesicles of mice prenatally treated with DES varied in size and lacked the curved shape seen in the glands from control mice. In DESmice which had been castrated and treated with E2, the seminal vesicles appeared involuted. Histologically, this involution was characterized by an increase in the proportion of stroma to epithelium, a decrease in cytoplasmic volume in the epithelial cells and a loss of secretory fluid in the glandular lumen of the seminal vesicle. The pseudo stratified columnar epithelium of both the intact control and DES mouse seminal vesicles was progressively reduced to low cuboidal tissue by estrogen treatment, orchiectomy and combined orchiectomy and estrogen treatment. Immunostaining for LF revealed that this protein was not localized in the SV of control (C), estrogenized control (C + E) or castrated control (C~) mice (fig. 1, A, C and E). However, subsequent treatment of control castrated mice with E2 (~ + E) did induce detectable LF expression in 2% of the epithelial cells of the seminal vesicle (fig. 1, G). In these few epithelial cells, LF appeared to completely fill the cytoplasm, both basally and luminally, and was only occasionally seen as a diffuse granular deposit. Lactoferrin was not detectable in the luminal secretory fluid by immunohistochemical techniques. In contrast, prenatal exposure to DES caused a slight increase in LF protein as determined by immunolocalization of the protein in DES treated intact mice (group DES) (fig. 1, B). Small granular intracellular deposits of LF were evident in 5% of seminal vesicle epithelial cells (arrow), and a diffuse reaction product appeared in the luminal secretions. Additional treatment of intact DES mice with estradiol (DES + E) increased the amount of LF expression to 8% of the epithelial cells; granules staining positive for LF were seen in the apical cytoplasm (fig. 1, D). Granular concentrations of LF were evident in a larger number of cells (arrows), and the concentration of LF in the luminal fluid was increased. Castration alone (DE~) induced detectable amounts of LF protein in the low cuboidal epithelium in approximately 25% of cells (fig. 1, F). A few cells increased their intracellular storage of LF to the point that granules were no longer evident (arrow). Additional treatment of DES~O mice with estradiol (DES~ + E) resulted in a substantial increase in LF secretion (fig. 1, H); approximately 40% of the epithelial cells in the seminal vesicle tissue of this group contained either intense uniform intracellular deposits of LF or diffuse intracellular granularity observed in the other DES treatment groups. Lactoferrin immunostaining was observed in the luminal fluid of all DES groups which had immunopositive cells. Increases observed in the percentage of epithelial cells showing LF expression were statistically significant (p <.001). Table 2 summarizes the relative amounts of LF and SVS IV expression found in each of the experimental groups. Analysis of in situ hybridization auto radiograms revealed changes in mRNA expression that were similar to those observed with immunohistochemistry (fig. 2, table 2). Lactoferrin message was not detectable in control intact mice [groups C (fig. 2, A), C~ or C + E (data not shown)]. However, LF mRNA was marginaily detectable in group C~ + E (fig. 2, C). Lactoferrin mRNA was readily detectable in DES intact mice (fig. 2, B). Its expression in intact DES + E and DES~ was similar to that in the intact DES group (data not shown). Only group DES~ + E was characterized by increased expression of LF mRNA (fig. 2, D). Expression of SVS IV protein and mRNA was also studied. The estrogen-treated and castrated mice from both the control and DES-treated groups exhibited only slight decreases in the amount of SVS IV message and protein expression. In comparison to the intact control mice, a slight decrease in both SVS

1375

DES FEIVIINIZATIOrJ OF SEMII\JAL VESICLE TABLE 2.

Effects of treatments on LF and SVS TV expression Analysis LF

Treatment Groups Control Intact Control Control + E, Control Orchiectomy Control Orchiectomy + E, Prenatally DES-Treated Intact DES Intact DES + E, DES Orchiectomy DES Orchiectomy + E2 ND

(C)

(C+E) (C~) (CO +E) (DES) (DES + E) (DE~) (DE + E)

SVSIV

Protein

mRNA

% Total Cells

Protein

mRNA

ND ND ND +

ND ND ND +

ND ND ND

++++ ++++ ++++ +++

++++

+ ++ +++ ++++

+ + ++ +++

5% 8%

+++ +++

+++ +++

~% 40%

-+

-+

2%

++++ ++++ +++

not detectable by immunohistochemistry. Differences in percentage of LF -containing cells among experimental groups were significant at p <.001.

IV protein and mRNA was observed in the control castrated plus estradiol group (C~ + E). Intact DES mice and castrated DES mice plus estradiol (DES ~ + E), showed further decreases in protein and mRNA. Initial inspection of the single antigen immunostaining and ISH results for both SVS IV and LF suggested that most of the epithelial cells of the seminal vesicle were producing SVS IV protein while only certain cells produced LF. However, due to the ubiquitous distribution of SVS IV throughout the SV secretions and the more selective distribution of LF, it was difficult to ascertain whether all the cells were making SVS IV. The possibility remained that LF might be secreted by a different cell type or at least in a cell which was not making SVS IV. Therefore, a double staining technique was used to stain for LF and SVS IV in the same tissue section. Diaminobenzidine (brown reaction product) was used to detect LF in the first immunodetection protocol, followed by removal of the LF antiserum and restaining for SVS IV with 4-chloronaphthol (blue reaction product) as the second chromogen. If the primary antiserum were not completely removed by this method, all the LF -containing cells as well as the SVS IV -containing cells would have been stained by the second immunostaining sequence since the primary antisera were both made in rabbit. However, when LF was used as the first antiserum it was found to be completely removable by the acid treatment, allowing a clear distinction to be made between LF - (brown) and SVS IVcontaining (blue) cells (fig. 3, A and B). Lactoferrin was found to occur in a few epithelial cells of the seminal vesicle from control castrated mice treated with estradiol (C~ + E). In contrast, it was readily apparent in all four groups of prenatally DES-treated mice. Seminal vesicle secretory protein IV was found in the seminal vesicle epithelium (fig. 3, B), but not in other accessory organs, such as dorsal prostate (fig. D) or coagulating gland (fig. 4). Moreover, LF, but not SVS detected in the dorsal prostate of all groups (fig. 3, C and Comparison of the distribution of LF and SVS IV in the double-stained slides indicated that cells which were stained for LF had very low SVS IV immmunoreactivity (fig. 3, A and B). Seminal vesicle epithelial cells located toward the base of epithelial folds appeared to contain more SVS IV protein than those located closer to the lumen. Occasional cells could be found which appeared to be stained for both LF and SVS IV; however, these represented less than 1% of the total number of cells. The overall pattern suggested that the greater number of cells, which were stimulated to produce LF, do not simultaneously synthesize SVS IV. Seminal vesicle secretory protein IV immunohistochemical staining of slides containing reproductive accessory organs from the intact control mice revealed SVS IV protein only in the seminal vesicle epithelium. The coagulating gland did not stain for SVS IV (fig. 4, A and B). Hybridization of similar sections with SVS IV cDNA likewise detected mRNA only in the seminal vesicle tissue (fig. 4, C and D).

DISCUSSION

The role of estrogens in the development of the reproductive tract of both sexes has been the subject of much study23-26 but remains to be determined. However, the exposure of fetuses to exogenous estrogen during this critical period of reproductive tra~t development has profound consequences. In the female, pennatal treatment with estrogen induces reproductive tract abnormalities. 27 -32 In the male, estrogen induces genital tract abnormalities, including metaplasia, nodular masses in the ampullary region of the reproductive tract33 and inhibition of epithelial branch morphogenesis,34 as well as the transcription of estrogen-regulated genes in male reproductive accessory organs. 3. 13. 35 Many of these estrogen-induced changes are not reversible by testosterone in adulthood. 36 In the present study, we have compared the expression of t~o genes which are expressed in the male or female reproductIve systems. Lactoferrin is a highly basic (pI> 10) single chain glycoprotein with an approximate molecular weight of 70 kDY Its function in the uteru.s, in which it is estrogen regulated, is presently unknown, but Its levels as a constituent of milk vary inver.sely with tran~fe:rin in a. variety of species,37 suggesting that It may playa sImIlar role m promoting the bioavailability of iron. Seminal vesicle secretory protein IV is an androgen-regulated, heat-stable, basic protein that plays an important role in creation of the copulatory plug as well as in suppression of the female immune response to the ejaculated sperm. 15 Under ordinary developmental conditions, the expression of the SVS IV gene in the male genital tract is restricted to the seminal vesicle, a reproductive accessory organ of mesonephric origin, and SVS IV has not been reported in the female reproductive tract. 13. 15 Under the conditions of our study, LF secretion could be identified by immunohistochemistry in a few cells of the seminal vesicle of control mice after castration and subsequent stimulation with E 2; the intact control mice did not express LF even when treated with E2 for 3 days. This suggests that ~ subpopulation of cells exists in control male mice which is capable of secreting LF under the appropriate endocrine conditions, that is, removal of endogenous androgens (castration) plus treatment with exogenous E 2 • Thus, the increased expression of LF observed in the animals treated prenatally with DES may represent an increase in the number of this subpopulation of cells. Intact male mice from the DES-treatment group produced amounts of SVS IV protein which were comparable to those seen in the control mice. Prenatal DES treatment alone did induce an increase in constitutive expression of the LF mRNA, but the amount was low and was found in only 5% of the seminal vesicle epithelial cells. Full expression of the LF mRNA required 3 days of estradiol treatment to castrated DEStreated mice. Substantial LF expression could be obtained by castration of mice prenatally treated with DES, suggesting that the LF gene may be negatively regulated by testosterone in

1376

DES FEMINIZATION OF SEMINAL VESICLE

FIG. 3. Sequential immunostaining of seminal vesicle (A and B) and dorsal prostate (C and D) for LF and SVS IV. Tissue came from animal in prenatally DES-treated group which had been orchiectomized and treated for 3 days with E 2 • Lactoferrin was localized by brown diaminobenzidine reaction product, while SVS IV was detected by blue 4-chloronaphthol reaction product. Non-coincident pattern of localization in SV suggested that, in general, the two proteins are not translated simultaneously by individual cells. Occasional cells (arrow) appeared to make both proteins, suggesting that capacity to make both proteins may be present, but that type of protein secreted may be regulated by endocrine condition of animal. Dorsal prostate was found to make LF in all experimental groups. 430x .

these animals. Thus, although the seminal vesicle secretions were clearly "feminized" by the administration of DES prenatally, the nature of the feminization was one of increasing the percentage of cells capable of responding to castration and! or estrogen treatment with LF secretion, rather than de novo regulation of transcription of a gene that is not normally expressed in the male seminal vesicle. The observation that the synthesis of LF and SVS IV correlates inversely in the animals treated prenatally with DES suggests that the cells that make SVS IV are also capable of making LF. In view of the high content of SVS IV in the intact prenatally DES-treated groups, it is possible that most of the affected cells are capable of making both SVS IV and LF, but that the SVS IV synthetic pathway is down-regulated by the combination of castration and administration of exogenous estradiol, whereas LF synthesis is stimulated. Thus, it may be that prenatal DES exposure selects for the clonal growth of a subpopulation of cells that exists naturally and is especially sensitive to estrogens. Alternatively, the early estrogenization may actually alter the development of some cells, creating a new subpopulation that responds to E2 stimulation by slowing

production of SVS IV and increasing the synthesis of LF. The regulation . of the constitutive expression of SVS IV by the seminal vesicle tissue of mice prenatally treated with DES was also studied. In a previous study, Pentecost et al,13 found that prenatal DES treatments did not significantly reduce SVS IV gene expression over that observed in intact male seminal vesicle tissue, although the subsequent treatment of these mice with E2 for 3 days was found to reduce the amount of SVS IV mRNA. In the present study, prenatal DES treatment also failed to significantly reduce SVS IV expression. However, IHe demonstrated that the decrease in SVS IV expression observed in DES-treated animals following castration and estradiol treatment is preferentially distributed within the gland and is much more evident in the luminal folds of epithelium than in the basal cells. In summary, the alterations in LF mRNA and protein secretion probably represent a change in the program of differentiation of the seminal vesicle itself rather than a primary alteration in the hormonal status of the mice, as circulating levels of estrogen and testosterone were not significantly different between the control and prenatally DES-treated groupS.28 DES-

DES FEMINIZATION OF SEMINAL VESICLE

1377

FIG. 4. Immunohistochemical (A, B) and in situ hybridization (C, D) specificity controls. Coagulating gland (CG) is shown on left, while seminal vesicle epithelium (SV) is on right in all illustrations. Both SVS IV protein (A, B) and mRNA and (C, D) were found to be preferentially localized in cells of SV rather than CG. In situ hybridization revealed mRNA to be only in cells that produced SVS IV protein (see silver grains over SV). Hematoxylin counterstain, A and C, 160x; Band D, 430x.

castrated mice that were subsequently treated with E2 showed an increase in expression of both LF mRNA and protein in comparison to those receiving castration alone. This suggests that LF gene expression in these mice is regulated by the estrogen receptor which was detected in our earlier study.3 These findings suggest that the mechanism of action of DES on the male reproductive tract is not primarily to block its masculinization. On the contrary, DES appears to feminize the reproductive tract at the molecular level by (1) inducing estrogen receptor expression 3 and (2) increasing the number of cells capable of expressing the estrogen-regulated protein LF in response to estradiol. This is accomplished without significantly decreasing the constitutive expression of the native secretory product of the seminal vesicle, SVS IV, although, as would be expected, total SVS IV expression is decreased under the combined effects of castration plus estradiol. It is anticipated that further studies with this model system will provide a clearer understanding of the mechanisms by which DES accomplishes the molecular feminization of seminal vesicle cells and the processes by which other environmental pathogens disrupt the normal development of their target cells. Finally, this model may provide general information on the role of estrogens in the normal differentiation of the male reproductive tract. REFERENCES 1. Herbst, A. L. and Bern, H. A.: Developmental Effects of Diethylstilbestrol (DES) in Pregnancy. New York: Thieme-Stratton, Inc., 1981. 2. McLachlan, J. A., Korach, K. S., Newbold, R. R. and Degen, G.

3.

4.

5.

6.

7.

8. 9. 10. 11.

H.: Diethylstilbestrol and other estrogens in the environment. Fundam. App!. Toxico!., 4: 686, 1984. Newbold, R. R., Pentecost, B. T., Yamashita, S., Lum, K., Miller, J. V., Nelson, P., Blair, J., Kong, H., Teng, C. and McLachlan, J. A.: Female gene expression in the seminal vesicle of mice after prenatal exposure to diethylstilbestrol. Endocrinology, 124: 2568,1989. Newbold, R. R. and McLachlan, J. A.: Neoplastic and non-neoplastic lesions in male reproductive organs following perinatal exposure to hormones and related substances. In: Toxicity of Hormones in Prenatal Life. Edited by N. Mori and T. Nagasawa. Boca Raton, Florida: CRC Press, pp. 89-109, 1988. Teng, C. T., Pentecost, B. T., Chen, Y. E., Newbold, R. R., Eddy, E. M. and McLachlan, J. A.: Lactotransferrin gene expression in the mouse uterus and mammary gland. Endocrinology, 124: 992,1989. Bagby, G. C., Jr., Rigas, V. D., Bennett, R. M., Vandenbark, A. A. and Garewal, H. S.: Interaction of lactoferrin, monocytes, and T lymphocyte subsets in the regulation of steady-state granulopoiesis in vitro. J. Clin. Invest., 68: 56, 1981. Reitamo, S., Konttinen, Y. T. and Segerberg-Konttinen, M.: Distribution of lactoferrin in human salivary glands. Histochemistry, 66: 285, 1980. Masson, P. 1., Heremans, J. F. and Schonne, E.: Lactoferrin, an iron binding protein in neutrophilic leukocytes. J. Exper. Med., 130: 643, 1969. Wichmann, 1., Vaalasti, A., Vaalasti, T. and Tuohimaa, P.: Localization of lactoferrin in the male reproductive tract. Int. J. Andro!., 12: 179, 1989. Hekman, A. and Riimke, P.: The antigens of human seminal plasma with special reference to lactoferrin as a spermatozoa coating antigen. Ferti!. Steri!., 20: 312, 1969. Teng, C. T., Walker, M. P., Bhattacharyya, S. N., Klapper, D. G.,

1378

12. 13.

14.

15.

16. 17. 18.

19.

20.

21.

22. 23.

DES FEMINIZATION OF SEMINAL VESICLE

DiAugustine, R P. and McLachlan, J. A.: Purification and properties of an estrogen stimulated mouse uterine glycoprotein (MW 70,000). Biochem. J., 240: 413, 1986. Pentecost, B. T. and Teng, C. T.: Lactotransferrin is the major estrogen inducible protein of mouse uterine secretions. J. BioI. Chern., 262: 10134, 1987. Pentecost, B. T., Newbold, R R, Teng, C. T. and McLachlan, J. A.: Prenatal exposure of male mice to diethylstilbestrol alters the expression of the lactotransferrin gene in seminal vesicles. Mol. Endocrinol., 2: 1243, 1988. Ostrowski, M. C., Kistler, M. K. and Kistler, W. S.: Purification and cell-free synthesis of a major protein from rat seminal vesicle secretion: a potential marker for androgen action. J. BioI. Chern., 254: 383, 1979. Fawell, S. E., Pappin, D. J., McDonald, C. J. and Higgins, S. J.: Androgen -regulated proteins of rat seminal vesicle secretion constitute a structurally related family present in the copulatory plug. Mol. Cell. Endocrinol., 45: 205, 1986. Wagner, C. L. and Kistler, W. S.: Analysis of the major large polypeptides of rat seminal vesicle secretion: SVS I, II, and III. BioI. Reprod., 36: 501, 1987. Pang, S. F., Chow, P. H. and Wong, T. M.: The role of the seminal vesicles, coagulating glands and prostate glands on the fertility and fecundity of mice. J. Reprod. Fertil., 56: 129, 1979. Mansson, P. E., Sugino, A. and Harris, S. E.: Use of a cloned double stranded cDNA coding for a major androgen dependent protein in rat seminal vesicle secretion: the effect of testosterone in gene expression. Nucleic Acids Res., 9: 935, 1981. Kistler, M. K., Taylor, R. E., Jr., Kandala, J. C. and Kistler, W. S.: Isolation of recombinant plasmids containing structural gene sequences for rat seminal vesicle secretory proteins IV and V. Biochem. Biophys. Res. Commun., 99: 1161, 1981. Chen, Y. H., Pentecost, B. T., McLachlan, J. A. and Teng, C. T.: The androgen-dependent mouse seminal vesicle protein IV: characterization and complementary deoxyribonucleic acid cloning. Mol. Endocrinol., 1: 707, 1987. Nakane, P. K.: Simultaneous localization of multiple tissue antigens using the peroxidase-labeled antibody method: a study on the pituitary gland of the rat. J. Histochem. Cytochem., 16: 557, 1968. Maniatis, T., Fritsch, E. F. and Sambrook, J.: Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory, p. 447, 1982. Jost, A.: Problems of fetal endocrinology: the gonadal and hypo-

physeal hormones. Rec. Prog. Horm. Res., 8: 379, 1953. 24. Burns, R K.: Role of hormones in the differentiation of sex. In: Sex and Internal Secretions. Edited by W. C. Young. Baltimore: Williams and Wilkins, pp. 76-158, 1961. 25. Jost, A.: Hormonal factors in the sex differentiation of the mammalian foetus. Philos. Trans. Royal Soc. Lond. BioI., 259: 119, 1970. 26. George, F. W. and Wilson, J. D.: Sex determination and differentiation. In: The Physiology of Reproduction. Edited by E. Knobil and J. Neill. New York Raven Press, pp. 3-26, 1988. 27. Plapinger, L. and Bern, H. A.: Adenosis-like lesions and other cervicovaginal abnormalities in mice treated perinatally with estrogen. J. Natl. Cancer Inst., 63: 507, 1979. 28. McLachlan, J. A., Newbold, R R, Pentecost, B. T. and Teng, C. T.: Adverse hormonal regulation of reproductive organ development. In: Serono Symposium on Reproductive Organ Development. Edited by M. Parvinen. New York: Raven Press, 1988. 29. Forsberg, J. G. and Kalland, T.: Neonatal estrogen treatment and epithelial abnormalities in the cervicovaginal epithelium of adult mice. Cancer Res., 41: 721, 1981. 30. Newbold, R. R and McLachlan, J. A.: Vaginal adenosis and adenocarcinoma in mice exposed prenatally or neonatally to diethylstilbestrol. Cancer Res., 42: 2003, 1982. 31. Newbold, R. R and McLachlan, J. A.: Diethylstilbestrol associated defects in murine genital tract development. In: Estrogens in the Environment. II: Influences in Development. New York: Elsevier, pp. 288-318, 1985. 32. Newbold, R. R., Bullock, B. C. and McLachlan, J. A.: Uterine adenocarcinoma in mice following developmental treatment with estrogens: a model for hormonal carcinogenesis. Cancer Res., 50: 7677, 1990. 33. McLachlan, J. A., Newbold, R R and Bullock, B.: Reproductive tract lesions in male mice exposed prenatally to diethylstilbestrol. Science, 190: 991, 1975. 34. Donjacour, A. A. and Cunha, G. R: The effect of androgen deprivation on branching morphogenesis in the mouse prostate. Devel. Biology, 128: 1, 1988. 35. Pylkkanen, L., Santti, R, Newbold, R R and McLachlan, J. A.: Regional differences in the prostate of the neonatally estrogenized mouse. Prostate, 18: 117, 1991. 36. Rajfer, J. and Coffey, D. S.: Effects of neonatal steroids on male sex tissues. Invest. Urol., 17: 3, 1979. 37. Jennes, R: Composition and characteristics of goat milk: review 1968-1979. J. Dairy Sci., 63: 1605, 1980.