0013.7227/92/1314-1657$03.00/0 Endocrinology Copyright 0 1992 by The Endocrine

Transforming of Estrogen

Vol. 131. No. 4 Printed in U.S.A.

Society

Growth Factor-a Is a Potential Action in the Mouse Uterus

KAREN G. NELSON, TSUNEO TAKAHASHI, DAVID C. LEE, NOREEN NANCY L. BOSSERT*, KIMBERLY ROSS, BENJAMIN E. EITZMAN, AND JOHN A. McLACHLAN

Mediator

C. LUETTEKE,

Laboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences (K.G.N., T.T., N.L.B., K.R., B.E.E., J.A.M.), Research Triangle Park, North Carolina27709; and the Lineberger Comprehensive Cancer Research Center and the Department of Microbiology and Immunology, University of North Carolina (D.C.L., N.C.L.), Chapel Hill, North Carolina 27599-7295 ABSTRACT

The induction of uterine TGFa mRNA is specific to estrogen; nonestrogenic steroids did not induce expression. Antibody specific to TGFcv significantly reduces estrogen-mediated uterine growth, which supports the concept that TGFo( is a mitogen for the reproductive tract. Analysis of TGFa/EGF receptors by binding, affinity labeling, and phosphorylation studies indicates that functional receptors are present in the mouse uterus after estrogen exposure. Thus, our data support a physiological role for TGFol and its receptor pathway in the female mouse reproductive tract. (Endocrinology 131: 1657-1664, 1992)

To better understand the role of peptide growth factors in sex steroid hormone-mediated growth of the female reproductive tract, the effect of estrogen on the expression of transforming growth factor-a (TGFol) in mouse uterus was investigated. Our results show that estrogen induces the expression of TGFa mRNA in the mouse uterus in a doseand time-dependent manner. The up-regulation of TGFol transcripts occurs predominantly in uterine epithelial cells. RIA and Western blot analysis demonstrate that immunoreactive TGFa protein is secreted at high levels into mouse uterine luminal fluid after estrogen treatment.

T

HE MECHANISM by which steroid hormones act on their target tissue(s)is not completely understood. Recent studies hypothesize that steroid hormones regulate cell function, growth, and differentiation by stimulating the local biosynthesis of peptide growth factors that act in an autocrine or paracrine fashion (l-7). Epidermal growth factor (EGF) is one such polypeptide growth factor that has been implicated in estrogen-induced uterine growth. This is supported, at least in the rodent uterus, by the presenceof immunoreactive EGF (4, 5), the enhancement of EGF and EGF receptor expression by estrogen (4-7), the potent mitogenicity of EGF for uterine epithelial cells both in culture (8) and in viva (9), and the inhibition of estrogen-induced mitogenesis in viva by EGF-specific antibodies (9). Furthermore, many of the cell types present in the uterus and other reproductive tissue(s) in different speciescontain high affinity binding sitesspecific for EGF (6-8,10-12). The fact that EGF receptors are present in the uterus at very early stages of development suggests that most uterine cell types may be potential targets of EGF action (6, 7, 10-12). In addition, recent in vitro studies demonstrate that the estrogen-induced peptide growth factors EGF (1, 2, 4, 5) and insulin-like growth factor-I (l-3) can replace estrogen not only in mediating mitogenesis,but also in inducing the progesterone receptor, which is thought to be a classical estrogen-regulated protein (13, 14). Collectively, these observations support the hypothesis that the Received March 18, 1992. Address requests for reprints to: Dr. Karen G. Nelson, National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, North Carolina 27709. * Current address: Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, North Carolina 27710.

EGF receptor pathway and its attendant regulation by steroid hormones are important in normal reproductive physiology. EGF sharesstructural and functional homology with transforming growth factor-a (TGFLu),which is a protein produced by both normal and malignant cells(15). TGFa is a 50-amino acid secretory polypeptide (mol wt, 5600) that is derived from a large glycosylated transmembrane precursor and is encoded by a 4.8-kilobase mRNA (15-17). The mature TGFa peptide sharesapproximately 30% sequencehomology with EGF and induces similar biological effects by binding to and activating the tyrosine kinase component of the EGF receptor. The presence of TGFa in a variety of cultured cells and tissues, including skin keratinocytes (1B), breast epithelial cells (19), activated macrophages (20), gastrointestinal mucosa(21), kidney (22), brain (23), pituitary (24), seminiferous tubule (25), ovary (26), liver (17,27), and decidua (28) implies that TGFa may play a role in normal cell physiology. Observations that TGFa is expressed in many normal tissues, including the decidualized uterus (28), led us to investigate whether TGFa is regulated by estrogen in the mouse uterus. Our results show that in the mouse uterus, TGFa is estrogen inducible, TGFa antibody inhibits uterine growth, and functional TGFLu/EGF receptors are present. Taken together, these data strongly support a role for TGFcll and its receptor in the regulation of uterine growth by estrogen. Materials AnimaLs,

treatment,

and Methods

and cell isolation

Immature female CD-l injection (0.1 ml) with

mice (17-21 days of age) were treated by SC various steroids (Sigma Chemical Co., St.

1657

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

1658

TGFa AND ESTROGEN

Louis, MO) dissolved in corn oil. At indicated times after treatment, the animals were killed by CO, asphyxiation and cervical dislocation. At all times, animals were treated as humanely as possible, using NIH guidelines. Uterine epithelial and stromal cells were isolated from CD-I mice (18 days old) 24 h after treatment with diethylstilbestrol (DES; 100 pg/ kg) or corn oil vehicle, using a modification of the procedures described previously (8, 11). The method for purification of the uterine epithelial cells involved enzymatic dissociation and Percoll density gradient centrifugation. After epithelial cell dissociation, uteri were treated further with proteolytic enzymes to obtain fractions enriched for uterine stromal cells. Isolated uterine cells were stored at -70 C until isolation of RNA.

Acid extraction fluid

and RIA of TGFa and EGF in mouse luminal

Uterine luminal fluid was collected from DES- or 17&estradiol (El)treated 21-day-old female mice (three daily SC doses of DES or Ez at 100 @g/kg starting on day 17 of age). The luminal fluid was acid extracted by the addition of an equal volume of 0.2 M acetic acid, followed by heating at 90 C for 10 min. The acidified luminal fluid was clarified by centrifugation (Beckman rotor 70.1 Ti, Palo Alto, CA; 100,000 X g; 60 min), and the supematant was lyophilized and resuspended in PBS. Aliquots were then analyzed using a RIA specific for TGFrv (Biotope, Redmond, WA) or EGF (Amersham, Arlington Heights, IL). It was not possible to collect uterine luminal fluid from control immature female mice because the lumen of the uterus in these animals was very small, and no fluid could be collected directly. We have tried injecting buffer into the uterine lumens of control mice and recollecting the fluid, but this was also difficult and unsuccessful. Analysis of tissue extracts from uteri from control and estrogen-treated (luminal fluid removed) mice revealed the presence of low levels of TGFcv (control, 19 ng TGFol/mg protein extracted; DES, 17 ng TGFo/mg protein extracted; El, 19 ng TGFn/mg protein extracted). However, the amount of TGFa extracted from uterine tissue did not differ among the experimental groups. The presence of immunoreactive TGFo in luminal fluid with little associated with the tissue suggested that estrogen treatment resulted in the induction and rapid secretion of TGFol, such that small amounts remained directly integrated into the tissue.

Immunoblotting Before immunoblot analysis, TGFa was partially purified from uterine luminal fluid collected from DES-treated female mice by acidification to a final concentration of 0.2 M acetic acid, addition to Sep-Pak Cl8 cartridges (Waters Associates, Milford, MA), and step elution with increasing percentages of acetonitrile. The purpose of the Sep-Pak procedure was to concentrate and purify TGFo from other components present in the acetic acid extract of the luminal fluid. Our procedure was standardized by using purified TGFa, which was found to elute from these cartridges upon washing with 20-40% acetonitrile. The eluted, partially purified uterine TGFol was then fractionated by electrophoresis on a sodium dodecyl sulfate (SDS)-polyacrylamide gel (15%), transferred to nitrocellulose, and immunoblotted using a specific monoclonal antibody against rat TGFa.

RNA isolation

and analysis

After treatment, the uteri were removed, immediately frozen in liquid nitrogen, and stored at -70 C until RNA extraction. Poly(A)+ mRNA was prepared from uteri using the method of Chirgwin et al. (29). A commercially available RNA isolation kit (5 Prime -+ 3 Prime, Inc., Paoli, PA) was used for the purification of poly(A)’ mRNA directly from isolated uterine stromal and epithelial cells. The poly(A)+ mRNA (2-7 pg) was fractionated by electrophoresis in a 6.2% formaldehyde-2% agarose gel and transferred to nylon membranes. The quality and quantity of the RNA were checked after transfer by staining with 0.04% methylene blue in 0.5 M sodium acetate (pH 5.2). The blots were then hybridized with a rat TGFa! cDNA (pr TGF& (16) or a mouse actin cDNA (nucleotides 34-315; Promega pGem4Z vector, Promega, Madison, WI) probe cloned in our laboratory (30).

ACTION

Endo. 1992 Vol 131. No 4

Uterine membrane preparation and EGF receptor binding, affinity labeling, and phosphorylation assays Uterine membranes were prepared from uteri of l&day-old female mice 24 h after treatment with DES (100 rg/kg) or corn oil vehicle, according to the method described by Mukku and Stance1 (7), in the presence of protease inhibitors (20 rg/ml phenylmethylsulfonylfluoride, 10 rg/ml leupeptin, and 100 kallikrein inhibitor units/ml aprotinin; Sigma Chemical Co.). Affinity labeling of uterine EGF receptors was performed using the procedure of Mukku and Stance1 (7), in which uterine membranes (100 pg) were incubated with 5 nM [?]EGF (Amersham Corp., Arlington Heights, IL) in the presence or absence of excess (500 nM) unlabeled EGF (Collaborative Research, Inc., Bedford, MA) and cross-linked using disuccinimidylsuberate (Calbiochem, La Jolla, CA). EGF-stimulated receptor autophosphorylation was analyzed by the method of Rubin and Earp (31), in which uterine membranes (100 rg) were preincubated for 10 min in the presence or absence of EGF (1 pg/ml), followed by the addition of [y-32P]ATP (Amersham) for 1 min before termination of the reaction. The affinity-labeling and phosphorylation reactions were terminated by the addition of SDSsample buffer and then boiled for 3 min. Proteins were separated by means of 8% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). The gels were then stained, dried, and exposed to x-ray film. EGF binding was carried out, as described by Carpenter (32), using varying concentrations (0.02-2.0 nM) of [‘251]EGF in the presence (nonspecific) or absence (total) of excess unlabeled EGF (3 PM). Specific binding was calculated as the difference between total and nonspecific binding. Data were analyzed using the Ligand procedure for best-fit.

Quantitation of uterine growth specific antibodies

after exposure to estrogen and

To determine whether antibodies specific for TGFol would inhibit estrogen-induced growth, slow release cholesterol-based pellets containing antibody against TGFL~ (otTGFa-IgG fraction; 0.67 mg/pellet; Biotope, Redmond. WA) were prepared by Innovative Research of America (Toledo, OH). This antiserum was produced using a synthetic immunogen against rat TGFa (residues 34-50) and recognizes both native and denatured forms of human and rat TGFa. No cross-reactivity with EGF was found, and the antibody blocked competition by synthetic rat TGFo( in a RRA (Biotope literature). The antibody-containing pellets were divided into fourths, and one quarter was implanted under the kidney capsule of S-week-old female CD-l mice that had been ovariectomized for at least 3 weeks. Adult mice were used for this study instead of immature mice because the larger size of the adult kidneys facilitated the insertion of the pellets, The adult females were ovariectomized at least 3 weeks before pellet implantation in order to remove endogenous sex steroids. Appropriate control pellets containing the IgG fraction of nonimmune goat serum (goat TGFo(; 0.57 mg/pellet; Biotope) were also implanted. Three days after pellet implantation, the animals were injected (SC) with a physiological dose of E2 (10 rg/kg) or the vehicle corn oil (CO), and DNA synthesis was measured by nuclear [3H]thymidine incorporation (2 &i 3H/g BW, injected 1 h before death) 16 h after estrogen administration (peak of estrogen-induced uterine DNA synthesis). The tissues for autoradiographic analysis were fixed in Bouin’s solution, paraffin embedded, sectioned (4-5 pm), and dipped in photographic emulsion (NTB2, Eastman Kodak, Rochester, NY). After 4-6 weeks of incubation, the slides were developed and stained with hematoxylin and eosin. For quantitation, tissue from at least 4 animals in each treatment group was evaluated. Counts of [3H]thymidine-labeled cells were made on at least 400 cells/uteri, and the data are presented as a percentage of the mean 2 SEM.

Results Estrogen induces the expression in mouse uterus

of TGFa protein

and mRNA

TGFcll expression after estrogen treatment was examined by RIA, immunoblot, and mRNA analyses. To determine

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

TGFa AND ESTROGEN ACTION whether estrogen treatment (E2or DES) induces the secretion of TGFa! and EGF into the lumen of the uterus, uterine luminal fluid was collected from immature female CD-l mice (21 days old) that had been treated for 3 consecutive days with the potent estrogen DES or EZ at 100 &kg, a dose shown to induce maximal secretion of luminal fluid by the uterus. The luminal fluid was then acid extracted and analyzed using RIAs specific for either EGF or TGFa. The results, as shown in Table 1, demonstrate that uterine luminal fluid collected from estrogen-treated animals contained high levels of TGFa. The level of TGFa after DES or EZ treatment was 30 times greater than that of EGF, which was also detected in luminal fluid (4). The amount of luminal fluid present in the uteri of mice treated with the vehicle (CO) or with nonestrogenic steroids was negligible and could not be collected for RIA analysis (see Materials and Methods). RIA analysis of whole tissue extracts demonstrated that TGFL~ was detectable at low levels in the uteri of these mice; however, the values did not differ between the estrogentreated and control mice (see Materials and Methods). This suggested that estrogen stimulated both the production and rapid secretion of these growth factors, such that low quantities stay associated with the tissue. Estrogen regulation of TGFa in mouse uterus was further investigated by Western blot analysis of TGFa in uterine luminal fluid and Northern analysis of RNA purified from intact uteri and isolated uterine stromal and epithelial cells. After partial purification and concentration of proteins in the uterine luminal fluid, immunoblot analysis confirmed the results obtained by the RIA and revealed the presence of TGFa! in the uterine luminal fluid of estrogen-treated mice that migrated similar to purified rat TGFa! with an approximate mol wt of 6000 (Fig. 1). To determine whether the increase in secreted TGFa! protein was related to a corresponding elevation of uterine mRNA encoding TGFa, dose-response and time-course RNA analyses were performed. Polyadenylated RNA [poly(A)+ RNA] isolated from whole uteri of estrogen-treated mice and analyzed by Northern hybridization revealed the presence of a 4% to 5.0-kb TGFa mRNA (Fig. 2). Estrogen induced uterine TGFa mRNA within 6 h after treatment and at doses as low as 1 pg/kg. TGFa mRNA was not detectable by Northern analysis in uteri from vehicle (CO)-treated 21-dayold mice. Actin mRNA was also evaluated to serve as an internal control for the quality and quantity of RNA from each experimental group, and as expected, estrogen upTABLE 1. RIA of TGFcv and EGF secreted into mouse uterine luminal fluid EGF TGFLY Treatment DES 2.00 -c 0.51 (7) 59.85 zk 8.56 (8) 2.68 f 0.25 (5) 60.00 + 11.09 (5) EO Values represent the mean + SE in nanograms per ml mouse uterine luminal fluid; the number of separate determinations is given in parentheses. RIAs were performed on acid-extracted uterine luminal fluid collected from immature CD-1 female mice after three daily treatments with DES or Ez at a dose of 100 rg/kg. There are no control values because uterine luminal fluid is negligible in control mice and cannot be collected for analysis.

1659

WESTERN BLOT MULF 20%ACN

Standard TGFa

40%ACN

long

MW Markers 46.0 -

30.0 -

21.5 14.3 6.5 -

1. Western blot analysis of TGFol in mouse uterine luminal fluid. Immunoblot analvsis demonstrates that mouse uterine luminal fluid (MULF) inducedby estrogen contains immunodetectable TGFol that migrates with a mobility similar to that of purified rat TGFa. The blot shows the stenwise elution of uterine TGFol using 20% and 40% acetonitrile (ACN) from Cl8 Sep-Pak cartridges. These Sep-Pak cartridges were used to concentrate and partially purify TGFol from other components present in acid-extracted uterine luminal fluid. MW, Mol wt.

FIG.

regulated actin expression (Fig. 2). In addition, equivalent RNA amounts and quality were found for each treatment group upon staining of the Northern blots with methylene blue (data not shown). To identify the cell type in the uterus that responded to estrogen with TGFol production, RNA was purified from enriched populations of stromal and epithelial cells isolated from uteri 24 h after estrogen treatment. Characterization of TGFa mRNA in stromal and epithelial cells from uteri of estrogen- or vehicle (CO)-treated mice demonstrated that estrogen induced the expression of TGFa mRNA predominantly in the uterine epithelium (Fig. 3). Little or no mRNA for TGFcxwas detectable in isolated uterine cells of the control animals. The presence of actin transcripts (Fig. 3) and staining of ribosomal RNA with methylene blue (data not shown) indicated that the RNA from each of the cell groups was of good quality (Fig. 3). To determine the steroid hormone specificity of TGFol RNA induction in the mouse uterus, uterine RNA was isolated from animals treated with dexamethasone (1 mg/kg), progesterone (2 mg/kg), 5oc-dihydrotestosterone (1 mg/kg),

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

1660

Estrogen

TGFa AND

ESTROGEN

Induction of Transforming Growth Factor cc Expression in Mouse Uterus Dose Response

ACTION

Endo. 1992 Vol131.No4

TGFa Expression in Isolated Mouse Uterine Epithelial and Stromal Cells

Time Course

Vehicle (Corn Oil) Treated

28s 28s

Estrogen

l

TGFol 18s.

TGFa

18s \

_I

28s - 28s

l

28s

l

.I_

Actin

I .,

.’

18S*

ACTIN

- 18s

FIG. 2. Northern analysis demonstrating the time- and dose-dependent induction of uterine TGFol mRNA by estrogen. A 4.8- to 5.0kilobase TGFa transcript is expressed in a dose-dependent manner in mouse uterus 24 h after estrogen treatment (DES; l-100 pg/kg). Timecourse analysis reveals that estrogen (DES; 5 rg/kg) increases the levels of TGFol mRNA within 6 h after treatment. Little or no transcripts for TGFol are detected in the uteri of vehicle (CO)-treated control animals. Analysis of actin mRNA serves as an internal control for RNA quality and quantity. Quantitation of TGFol mRNA by densitometric scanning show that doses of 5,20, and 100 pg DES/kg induce 1.22-, 1.71-, and 2.06-fold higher amounts, respectively, of TGFa! transcripts than does 1 Mg DES/kg. The ratio of TGFa to actin transcripts did not change greatly among the different estrogen doses (1.12-1.10) because of the simultaneous increase in actin expression. Densitometric analysis of TGFol expression in the time-course study indicates that 12, 16, and 24 h after estrogen treatment, TGFo transcript induction was 1.32-, 1.54-, and 1.24-fold greater, respectively, than that at 6 h. The ratio of TGFa to actin expression did not change a great deal from one time point to another (ranging from 0.63-0.77). DES (100 w.$kg), or EZ(100 &kg) and analyzed by Northern blotting for actin and TGFcx RNA expression (Fig. 4). The blot quite clearly demonstrated the presence of TGFa! transcripts in mouse uterus after estrogen treatment (DES or E2), whereas little or no TGFa RNA was detected after exposure to nonestrogenic steroids that did not promote uterine

FIG. 3. TGFa and actin mRNA expression by isolated uterine stromal and epithelial cells 24 h after estrogen (DES; 100 rg/kg) or vehicle (CO) treatment. Northern analysis demonstrates that estrogen stimulates TGFa expression predominantly in the uterine epithelium and that uterine cells from vehicle-treated mice do not express detectable levels. Analysis of actin mRNA indicates that the RNA obtained from the different groups is of good quality.

growth.Taken together, these data indicate that in addition to inducing EGF (4, 5), estrogen stimulated the mRNA and protein for TGFa in mouse uterus. Therefore, this suggests that both EGF and TGFa! may be important components of an autocrine mechanism by which estrogen controls uterine growth in vivo. Estrogen action is inhibited by antibody specificfor TGFol Further evidence to substantiate this hypothesis was obtained by investigating the effects of antibody specific for TGFa on estrogen-stimulated uterine growth in vivo. Slow releasepellets containing TGFa antibody (LYTGF~)or normal goat serum (NGS) were implanted under the kidney capsules of ovariectomized adult female mice 3 days before treatment with the natural estrogen E2 (10 pg/kg) or the vehicle CO. The effects of the TGFa antiserum on DNA synthesis in the uterus were measuredby quantitating nuclear [3H]thymidine incorporation 16 h after estrogen exposure, a time at which maximal uterine DNA synthesis is induced. Exposure to TGFa-specific antibody resulted in at least a 50% inhibition of estrogen-induced epithelial cell proliferation in mouse

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

TGFa AND ESTROGEN

Steroid

Specificity

DES

for the Induction in Mouse Uterus E2

P4

DHT

DEX

of TGFa

mRNA

CO

FIG. 4. Hormonal specificity of TGFa mRNA induction. TGFa and glyceraldehyde phosphate dehydrogenase (GAPDH) RNA expressions were analyzed by Northern blotting of poly(A) RNA (5 pg) isolated from the uteri of immature mice (20 days old) 24 h after treatment with EP (100 pg/kg), DES (100 pg/kg), dexamethasone (DEX; 1 mg/ kg), progesterone (Pl; 2 mg/kg), 5a-dihydrotestosterone (DHT, 1 mg/ kg), or the vehicle CO (control). The induction of uterine TGFa mRNA is specific to the estrogens (DES and E2); the nonestrogenic steroids do not induce expression. Expression of GAPDH mRNA was similar among the experimental groups, which indicated that equal qualities of RNA were evaluated for each sample. The Effects Estrogen-Induced

of Antibodies Against TGFa on Mouse Uterine DNA Synthesis

50 -I

FIG. 5. Inhibition of estrogen-induced epithelial proliferation in CD-1 female mouse uterus by antibody against TGFol. Slow release pellets containing antibody specific for TGFol or the normal goat Ig control were implanted under the kidney capsule. Three days after pellet implantation, animals were treated with E2 (10 pg/kg) or the vehicle CO, and DNA synthesis was measured by quantitation of nuclear [3H] thymidine incorporation. CO-NGS, CO injection, normal goat Ig pellet; CO+TGFa, CO injection, TGFa antiserum pellet; E,-NGS, E2 injection, normal goat Ig pellet; E*-aTGFa, E2 injection, TGFa antiserum pellet.

ACTION

1661

EGF receptors were present in the immature mouse uterus used in our studies, the receptor was analyzed by Scatchard analysis, affinity labeling, and phosphorylation studies 24 h after estrogen administration, a time point at which TGFa mRNA is detectable in the uterus. Scatchard analysis of TGFa/EGF receptor demonstrated that mouseuterine TGFcu/ EGF binding was similar to or slightly reduced [binding capacity, 0.690 nM (25 fmol EGF/mg protein); &, 0.0626 nM] compared to controls [binding capacity, 0.984 nM (36 fmol EGF/mg protein); Kd, 0.0637 nM; Fig. 61. Affinity labeling (Fig. 7A) and autophosphorylation (Fig. 7B) of the uterine TGFol/EGF receptor also indicated that functional receptor was present and that estrogen did not significantly affect receptor binding or autophosphorylation properties. In contrast to results reported for the rat uterus (6, 7), where estrogen transiently increased the TGFa/EGF receptor-binding capacity and mRNA levels, estrogen did not significantly affect TGFa/EGF receptor binding in the mouse uterus at several time points examined (6-48 h; data not shown). In addition, treatment with nonestrogenic hormones (1 mg/kg progesterone, 1 mg/kg 5ol-dihydrotestosterone, and 1 mg/kg dexamethasone) did not dramatically affect mouse uterine TGFa/EGF receptor binding 24 h after treatment (Fig. 8). Discussion Recent studies suggest that the mitogenic effects of sex steroids are partly mediated by the local production of peptide growth factors, which may play a physiological role in growth and differentiation of the male and female reproductive tract (l-13,25-27). Expression of TGFa/EGF ligand and receptor has been shown to be controlled by steroid hormones or gonadotropins in the ovary, uterus, seminiferous tubule, pituitary, and mammary cells (1, 2, 4-7, 14, 24-27). In light of the observations that EGF is expressed in the uterus, and the related peptide, TGFa, is expressedin normal tissues (17-28), including the decidualized uterus (28), we examined the role of TGFcx in estrogen-induced growth of the female mouse reproductive tract. Evidence necessary to support the hypothesis that a TGFa autocrine/paracrine MOUSE

UTERINE

EGF

RECEPTOR

.

Control

.

Estrogen

uterus (Fig. 5). The results of this study further suggest that TGFa is intimately involved in the mechanism by which estrogen stimulates growth in viva. Estrogen effectson mouseuterine TGFa/EGF receptor The presence of TGFa/EGF receptors in the uterus of several species, including the mouse, has been well documented (6-8, lo-12), but to insure that functional TGFa/

FIG. 6. Scatchard analysis of [‘251]EGF binding by uterine membranes isolated from estrogen-treated (DES; 100 pg/kg; 24 h) and vehicle (CO)-treated immature mice (20 days old).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

TGFa AND ESTROGEN ACTION

1662

Affinity labeling (A) and Phosphorylation of the Mouse Uterine EGF Receptor

STEROID

(B)

Endo. Voll31.

EFFECTS

ON UTERINE

1992 No 4

EGFR

150

A.

Affinity Labeling: (-) or (+) excess

Control -

-

+

Estrogen -

(24 hr) +

0 C

P

DHT

DEX

MS

TREATMENT

B.

Phosphorylation: (-) or (+) EGF

Estrogen

Control -

+ qy

* qym

(24 hr)

+

-

I

EGF Receptor MW=170,000

FIG. 8. Steroid effects on mouse uterine TGFo(/EGF receptor binding. The specific binding of 5 nM [?]EGF was determined for membranes prepared from uteri of immature mice (20 days old) 24 h after treatment with vehicle CO (C), progesterone (P; 2 mg/kg), 50c-dihydrotestosterone (DHT; 1 mg/kg), dexamethasone (DEX, 1 mg/kg), or DES (100 pg/ kg). The data are presented as a percentage of the control value with the SE. Pairwise comparison was performed on each treatment group vs. the CO control using Dunnett’s t test. Only the DES treatment group showed a statistically significant difference from the control group (P < 0.01).

----c

FIG. 7. Affinity labeling (A) and EGF-stimulated autophosphorylation (B) of the TGFa/EGF receptor of uterine membranes isolated from estrogen-treated (DES; 100 rg/kg; 24 h) and control vehicle (CO)treated animals. Affinity labeling (A) was performed by incubation of the membranes with [iz51]EGF in the presence (+) or absence (-) of excess unlabeled EGF, followed by cross-linking of EGF to the receptor with disuccinimidylsuberate and resolution by SDS-PAGE. EGF-stimulated receptor autophosphorylation (B) was performed by incubation of membranes with (+) or without (-) EGF to activate the receptor kinase, followed by the addition of [32P]ATP to label the receptor and then resolution of the phosphorylated receptor by SDS-PAGE.

pathway is operating in the uterus would include the demonstration that 1) TGFa is expressed in the uterus during estrogen-induced proliferation; 2) uterine growth is shown to be at least partially dependent on this growth factor; and 3) functional receptors for TGFa are present to transmit the mitogenic signal. The results of our study demonstrate that estrogen induces the expression of uterine TGFa! protein and mRNA. Estrogen induces uterine TGFa as early as 6 h after treatment and at physiological doses. The induction of TGFa occurs before the peak of DNA synthesis in uterine luminal epithelial cells (16 h), which suggeststhat this growth factor may be one of several components (e.g. insulin-like growth factor-I) that triggers cell cycle progression. Analysis of isolated uterine cells indicates that estrogen induces much higher levels of TGFa mRNA in the epithelial than in the stromal cells.

Uterine cells other than those in the epithelium (e.g. muscle, stroma, and endothelium) express TGFa/EGF receptors and respond with some DNA synthesis at various times after estrogen treatment. Thus, it is possiblethat TGFcvproduction and secretion by the epithelium may function asan autocrine as well as a paracrine factor, resulting in the stimulation of DNA synthesis in the multiple cell types of the mouseuterus in response to estrogen. The induction of uterine TGFa is specific to estrogen; nonestrogenic steroids did not affect the expression. Antibody specific for TGFa reduces estrogenmediated growth of the uterus in vivo. Binding, affinitylabeling, and phosphorylation analysesof the uterine TGFa/ EGF receptor demonstrate that functional receptors are present and that estrogen treatment slightly reduces the number of receptors. Collectively, our data provide supportive evidence that both EGF and TGFa and their receptors are essentialcomponents of the normal pathway of cell growth that occurs in the female mouse uterus in responseto estrogen. There is not a clear explanation for the existence of both EGF and TGFa in estrogen-stimulated uteri, since the two factors share sequencehomology, bind to the samereceptor, and have similar biological effects. However, several studies suggest that these factors have different potencies, with TGFa being more effective and potent than EGF (15, 18,3335). For example, TGFa is more potent than EGF in the induction of angiogenesis(15), in the stimulation of osteoclast cells and bone resorption (15), in the enhancement of keratinocyte growth in vitro (18,33), in epidermal wound healing in vivo (34), and in the regulation of vascular blood flow (35). In fact, it has been proposed that compared to EGF, TGFa! is a superagonist for the TGFa/EGF receptor (34).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

TGFa AND ESTROGEN Thus, the resulting biological effects mediated by the TGFa/ EGF receptor upon binding to TGFa would be greater and more persistent than those induced by EGF. Alternatively, the mitogenic stimulus of low levels of EGF may be amplified by the expression of higher levels of TGFa. Recent studies have demonstrated that the expression of TGFcll is up-regulated by EGF or TGFa (15, 18, 33, 36), suggesting that TGFa is under positive autoregulatory control. An autoregulatory autocrine/paracrine pathway for TGFa would allow a cell to respond to low initial levels of EGF by inducing the synthesis of more TGFa and, thus, resulting in an amplification of the initial signal. A continuous exposure to TGFa may be required to initiate complete cell cycle progression. Such an amplification mechanism for TGFcx has been proposed for skin, pituitary, cultured breast epithelial cells, and regenerating liver (18, 19, 24, 27, 33, 36). Similarly, our data suggest that a positive amplification system for TGFLu/EGF probably also exists in vim for the uterus, and that estrogen is the trigger that regulates the expression of these polypeptide mediators of growth. The response of a tissue to a growth factor is dependent on the expression of functional receptors that transduce the mitogenic signal into the cell. Therefore, we examined the effects of estrogen on mouse uterine TGFa/EGF receptor by binding, affinity labeling, and phosphorylation analyses. The mouse uterine TGFcx/EGF receptors were either not affected or slightly reduced after estrogen treatment. A reduction in the number of uterine receptors may be explained by downregulation of the receptor by the binding of the endogenous ligands, TGFa and EGF, that are induced by estrogen. In contrast to what is seen in the mouse, estrogen treatment of rats has been shown to significantly stimulate TGFa/EGF receptor binding and receptor mRNA (6, 7). This discrepancy may simply reflect differences in the responses of the two different species or in the experimental design and technique. However, we have not yet investigated the effects of estrogen on the expression of mRNA for the TGFa/EGF receptor in the mouse uterus. Considering the high concentration of TGFa that is present in uterine luminal fluid after estrogen treatment, it is quite likely that estrogen does stimulate TGFa/EGF receptor synthesis in the mouse uterus to prevent depletion of the receptor due to down-regulation. Presumably, a critical number of receptors must be maintained to transmit the mitogenic signal required for growth stimulation. Nevertheless, our results indicate that functional TGFLu/ EGF receptors, as measured by binding and phosphorylation studies, are present in mouse uterus, thereby suggesting a potential physiological role for this receptor pathway in the induction of uterine growth by estrogen. That an autocrine/ paracrine growth factor regulates cell growth requires not only demonstration of the expression of both ligand and receptor in a cell system, but also evidence that DNA synthesis is dependent on that growth factor. Our study demonstrates growth factor dependence by showing that an antibody specific for TGFa significantly reduces estrogenmediated growth. These data support the notion that TGFa binding to its receptor acts as a mitogenic stimulus for the female reproductive tract. The lack of total inhibition of DNA

ACTION

1663

synthesis that is obtained in these experiments may be due to the existence of privileged sites, where the growth factor is protected from the antibody, or possibly due to the existence of active forms of TGFcv produced during synthesis and processing that are not recognized by the antibody. In addition, TGFa! is probably only one of multiple growth factors involved in uterine growth, each of which may act independently of TGFa; therefore, blocking TGFcv action alone would not completely suppress estrogen-induced mitogenesis. In conclusion, our data strongly suggest that TGFa is an autocrine/paracrine mediator of mouse uterine growth by estrogen. The TGFa/EGF receptor pathway seems to be an important regulator of uterine growth involving estrogeninduced secretion of these growth factors, which then interact with uterine receptors to initiate a mitogenic response.

References 1. Dickson RB, Lippman ME 1987 Estrogenic regulation of growth and polypeptide growth factor secretion in human breast carcinoma. Endocr Rev 8:29-43 2. Pollard JW 1990 Regulation of polypeptide growth factor synthesis and growth factor-related gene exJ%ssion-in the rat and mouse uterus before and after implantation. I Reurod Fertil 88:721-731 3. Murphy LJ, Murphy LC,& Freisen HG 1987 Estrogen induces insulin-like growth factor-l expression in the rat uterus, Mol Endocrino1 1:445 4. DiAugustine RP, Petruez P, Bell GI, Brown CF, Korach KS, McLachlan JA, Teng CT 1988 Influences of estrogens on mouse uterine epidermal growth factor precursor protein and messenger ribonucleic acid. Endocrinology 122~2355-2363 5. Huet-Hudson YM, Chakraborty C, De SK, Suzaki Y, Andrews GK, Dey SK 1990 Estrogen regulates synthesis of epidermal growth factor in mouse uterine epithelial cells. Mol Endocrinol 4:510 6. Gardner RM, Verner G, Kirkland JL, Stance1 GM 1989 Regulation of uterine epidermal growth factor (EGF) receptors by estrogen in the mature rat and during the estrous cycle. J Steroid Biochem 32:339 7. Mukku VR, Stance1 GM 1985 Regulation of epidermal growth factor receptor by estrogens. J Biol Chem 260:9820-9824 8. Tomoaka Y, DiAugustine RP, McLachlan JA 1985 Proliferation of mouse uterine epithelial cells in vitro. Endocrinology 118:lOl l-1017 9. Nelson KG, Takahashi T, Bossert NL, Walmer DK, McLachlan JA 1991 Epidermal growth factor replaces estrogen in the stimulation of female genital tract growth and differentiation. Proc Nat1 Acad Sci USA, 88:21 10. Lin T-H, Muuku VR, Verner G, Kirkland JL, Stance1 GM 1988 Autoradiographic localization of epidermal growth factor receptors to all maior uterine cell tvoes. Biol Reurod 38:403-411 11. Bossert NL, Nelson KG;‘Ross KA, Takahashi T, McLachlan JA 1990 Epidermal growth factor binding and receptor distribution in the mouse reproductive tract during development. Dev Biol 142:7585 12. Hofmann GE, Rao CV, Barrows G, Schultz GS, Sanfilippo JS 1984 Binding sites for epidermal growth factor in human uterine tissues and leiomyomas. J Clin Endocrinol Metab 58:880-884 13. Sumida C, Pasqualini JR 1989 Antiestrogens antagonize the stimulatory effect of epidermal growth factor on the induction of progesterone receptor in fetal uterine cells in culture. Endocrinology 124:591-597 14. Katzenellenbogen BS, Norman MJ 1990 Multihormonal regulation of the progesterone receptor in MCF-7 human breast cancer cells: interrelationships among insulin/insulin-like growth factor-l, serum, and estrogen. Endocrinology 126:891-898 15. Derynck R 1988 Transforming growth factor-a. Cell 54:593-595 16. Lee DC, Rose TM, Webb NR, Todaro GJ 1985 Cloning and

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.

TGFa

AND

ESTROGEN

sequence N&e

analysis of a cDNA for rat transforming “Vmowth factor-a. 313:489-491 17. Luetteke NC, Michalopoulos GK, Teixido J, Gilmore R, Massaque I, Lee DC 1988 Characterization of high molecular weight transforming growth factor-a produced cells. Biochemistry 27:6487-6494

by rat hepatocellular

transforming 328:817

growth

factor-a

in

human

keratinocytes.

growth factor-a may be a physiological regulator of liver regeneration by means of an autocrine mechanism. Proc Nat1 Acad Sci USA 86:1558

28. Han VKM, Hunter III ES, Prati RM, Zendegui JG, Lee DC 1987 Expression of rat transforming growth factor alpha development occurs predominantly in the maternal Cell Biol 7:2335

29. Chirgwin

Nature

19. Bates SE, Valverius EM, Ennis BW, Bronzert DA, Sheridan JP, Stampfer MR, Mendelsohn J, Lippman ME, Dickson RB 1990 Expression of the transforming growth factor-alepidermal growth factor receptor pathway in normal human breast epithelial cells. Endocrinology 126:596-607

activated

human

epiderr& mucosa.

alveolar

macrophages.

Cell 5%285 erowth factor-a and in normal hu”man gastrointestinal

SA, Elder IB 1989 Transformine growth facior levels Br J Cancer 60:657

22. Gamella LG, Sargent ER, Wade TP 1989 Expression ing growth expression carcinoma.

23. Kudlow

factor-a in normal human adult of transforming growth factor-a Cancer Res 49:6972

JE, Teung AWC, Kobrin

Transforming

growth

factor-a

264:3880-3883 24. Samsoonder J, Kobrin

MS, Paterson AJ, Asa SL 1989

in the mammalian

MS, Kudlow

growth factor secreted by untransformed cells in culture. J Biol Chem 261:14408

25. Skinner 26.

of transformkidney and enhanced and p1 in renal cell

JE 1986 bovine

brain.

J Biol Chem

a-Transforming anterior pituitary

MK, Takacs K, Coffey RJ 1989 Transforming growth factor-a gene expression and secretion in the seminiferous tubule: peritubular cell-seritoli cell interactions. Endocrinology 124:845 Skinner MK, Coffey Jr RJ 1988 Regulation of ovarian cell growth through the local production of transforming growth factor-a by theta cells. Endocrinology 123:2632

JM, Przylbyla

AE, MacDonald

Isolation of biologically active ribonucleic in ribonuclease. Biochemistry 18:5294

mRNA during decidua. Mol

RJ, Rutter WJ 1979

acid from sources

enriched

30. Alonso S, Mi,ntz A, Bourlet Y, Buckingham

31.

20. Madtes DK, Raines EW, Sakariassen KS, Assoian RK, Sporn MB, Bell GI, Ross R 1988 Induction of transforming growth factor-a in 21. Cartildee

Endo - 1992 Voll31. No 4

27. Mead JE, Faust0 N 1989 Transforming

carcinoma

18. Coffey Jr RJ, Derynck R, Wilcox JN, Bringman TS, Goustin AS, Moses HL, Pittelkow MR 1987 Production and auto-induction of

ACTION

32.

M 1986 Comparison of three actin-coding sequences in the mouse; evolutionary relationships between the actin genes of warn-blooded vertebrates. J Mol Evol 23:ll Rubin RA, Earp HS 1983 Solubilization of EGF receutors with Triton X-100 alters stimulation of tyrosine phosphorylatibn by EGF and dimethvl sulfoxide. 1 Biol Chem 258:5177 Carpenter ‘G 1985 Binding assays for epidermal growth factor. Methods Enzymol 109:lOl

33. Pittelkow MR, Lindquist PB, Abrahams RT, Graves-Deal R, Derynck R, Coffey RJ 1989 Induction of transforming growth factor-a expression Chem 264:5164

in human

keratinocytes

by phorbol

esters. J Biol

34. Schultz GS, White M, Mitchell R, Brown G, Lynch J, Twardzik DR, Todaro GJ 1987 Epithelial wound healing enhanced by transforming

growth

factor-a

and

vaccinier

growth

factor.

Science

235:350 35. Hollenberg MD, Muramatsu I, Itoh H, Pate1 P, Yang S-G, Lederis K 1989 Contractile actions of epidermal growth factor-urogastrone in isolated smooth muscle preparations from guinea pig stomach: structure-activity relationships and comparison with the effects of human transforming growth factor-alpha. J Pharmacol Exp Ther

248:384 36. Bjorge JD, Paterson AJ, Kudlow JE 1989 Phorbol growth mRNA factor-a

ester or epidermal factor (EGF) stimulates the concurrent accumulation of for the EGF receptor and its ligand transforming growth in a breast cancer cell line. J Biol Chem 264:4021

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 07 March 2015. at 13:26 For personal use only. No other uses without permission. . All rights reserved.