Regulation of Epidermal Growth Factor Receptor by Estrogen*

Administration of 17B-estradiol (E,) to immature female rats produces a 3-fold increase in ‘261-epidermal growth factor (EGF) binding to uterine membranes with no change in the affinity of membrane receptors for EGF. E, treatment also increases the EGF receptor visualized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis after affinity labeling of uterine membranes and the EGF-stimulated receptor auto- phosphorylation activity. In addition, E, administration stimulates EGF-dependent tyrosine kinase activ- ity in an assay system using exogenous angiotensin I1 as substrate. Following hormone treatment, EGF re- ceptor levels increase between 6 and 12 h, remain elevated at 18 h, and decline between 24 and 36 h. This stimulation of EGF receptor levels by Ez is spe- cific, since the non-estrogenic hormones progesterone, dexamethasone, and dihydrotestosterone fail to elevate receptor levels. Ez-stimulated increases in EGF receptor levels are also blocked by cycloheximide and acti- nomycin D, suggesting that the observed effect represents de novo synthesis of the EGF receptor and may be mediated by a transcriptional mechanism. These results demonstrate that estrogen can regulate acutely the levels of EGF receptor in vivo and raise the possi-bility that events coupled to this receptor may play a role in estrogen-stimulated growth.

These effects are thought to be mediated by the interaction of the peptide with a specific plasma membrane receptor (2)(3)(4)(5) . A very rapid response to EGF binding is autophosphorylation of the receptor; it appears that EGF binding, kinase activity, and the phosphorylation site(s) reside in the same polqpeptide chain of the receptor (6-10). Most studies of the growth stimulatory properties of EGF have utilized cultured cell svstems, and less information is available about the possible role of EGF or related peptides in physiological growth processes i n uiuo.
One in Eiuo growth system which has been widely studied is stimulation of uterine growth by estrogens. We have shown that uterine membranes contain EGF receptors with properties similar to those in other systems (ll), and Gonzalez et al. using a radioimmunoassay procedure. This study was therefore undertaken to determine if the uterine content of EGF receptor is altered by acute estrogen treatment in uiuo. Our results illustrate that the measurable levels of uterine EGF receptor increase approximately 3-fold within 12-18 h after hormone administration.

Materials
Unless specifically noted below, all chemicals and reagents were as previously described (11, 13) or the highest grade commercially available. The lz5I-EGF used for the binding studies was prepared as previously described (11, 13).

Methods
Animals-Immature Sprague-Dawley female rats (20-22 days of age, weight of 35-40 g) were obtained from Harlan-Sprague-Dawley (Houston) and ovariectomized 4-7 days prior to use. Animals received standard rat chow and water ad libitum. At the indicated times prior to death, animals received a subcutaneous injection of 17g-estradiol (40 pg/kg) administered in 0.5 ml of 95% saline, 5% ethanol; controls received the vehicle alone. Animals were killed by decapitation at the indicated times. For the experiments illustrated in Fig. 5 only, the steroids were administered in 0.1 ml of propylene glycol containing 2% acetone. This vehicle was used because of the decreased solubility of the non-estrogenic steroids used in aqueous solutions. For the indicated experiments, cycloheximide (4 mg/kg) and actinomycin D (8 mg/kg) were administered by intraperitoneal injection in a volume of 0.5 ml of saline. Actinomycin D was administered once at the same time as E2, and the animals were killed 12 h later. Cycloheximide (4 mg/kg) was initially administered at the same time as E2, and the same dose was then administered 6 h later; animals were killed 12 h after the E2 injection.
Membrane Preparation and Assays-All steps in the preparation of membranes were carried out a t 0-4 "C. Uteri were pooled, minced, and homogenized with two 15-s bursts in a Polytron P-10 homogenizer in 15 volumes of homogenization buffer (20 mM PIPES, 0.25 M sucrose, 1 mM EGTA, 1 mM EDTA, and 0.5 mM phenylmethylsulfonyl fluoride, pH 7.4). The homogenates were centrifuged at 800 X g for 15 min, and the supernatants were transferred to separate centrifuge tubes. The pellets were rehomogenized as above in the same volume of buffer and recentrifuged at 800 X g. The 800 X g supernatants were combined and centrifuged at 105,000 x g for 60 min. The resulting pellets were suspended in homogenization buffer and recentrifuged at 105,000 X g. The final pellets were resuspended in 20 mM PIPES, 0.15 M NaCI, pH 7.4 (2 ml/g of starting material), quickly frozen, and stored at -70 "C until use. For the estimation of protein content, an aliquot of the suspended membranes was diluted with an equal volume of 2 N NaOH, heated at 90 "C for 15 min, and then diluted with water prior to measurement of protein by the method of Lowry et al. (15). 5'-Nucleotidase activity in diluted membrane preparations was measured as described by Aronson and Touster (15), except that the substrate concentration in the assay was doubled.
Assay of EGF-stimulated Kinase Activity with Exogenous Angiotensin I I as Substrate-Frozen membranes prepared from 6-10 animals/ group (see above) were thawed and extracted for 30 min at 0 "C with 1% Triton X-100 and 10% glycerol and centrifuged a t 105,000 X g for 60 min. The supernatant (0.5 ml containing 1-1.5 mg of protein) was mixed with an equal volume of packed wheat germ agglutinin-Sepharose (Pharmacia) and rotated end-over-end for 2 h at 0-4 "C. This mixture was poured into a small column, washed with 5 ml of 40 mM imidazole, pH 7.2, containing 0.5 M NaCl, 10% glycerol, and 0.05% Triton X-100. The column was eluted with 2 ml of the same buffer containing 0.3 M N-acetylglucosamine (16). The eluent was then concentrated by filtration using Centricon-30 filters (Amicon).
EGF-stimulated kinase activity was performed in a volume of 25 yl containing the final concentrations of the following: 20 mM imidazole, pH 7.2, 85 mM NaCl, 5% glycerol, 0.05% Triton X-100, 20 mM MgC12, 10 mM p-nitrophenyl phosphate, 100 p M sodium vanadate, 2 mM dithiothreitol, and 100 PM [T-~'P]ATP. Assays contained 2 mM angiotensin I1 (Sigma) as the substrate and were linear for 10 min at 25 "C. The source of kinase activity was 2-5 pg of the wheat germ agglutinin-purified material (see above) mixed with 15-20 r g of bovine serum albumin. This kinase preparation was preincubated with or without EGF (2 yg/ml) for 15 min at 25 "C and then added to initiate the reaction. Reactions were carried out for 10 min at 25 "C and stopped by the addition of trichloroacetic acid (to 3% final concentration). After cooling on ice (10 rnin), the samples were centrifuged at 12,000 X g for 20 min, and incorporation was determined by spotting 20-pl aliquots of supernatant on phosphocellulose paper (17). Blank values (determined by adding the angiotensin I1 after the trichloroacetic acid) were subtracted for each assay, and EGF-stimulated activity was then determined for each preparation by subtracting the activity in the absence of the preincubation with EGF.

RESULTS
The results in Fig. 1 illustrate the effect of a single injection of estrogen on EGF binding in uterine membranes. results are obtained if EGF binding is expressed on the basis of 5"nucleotidase activity in the membrane preparations. The results illustrate that there is a sharp rise in EGF binding between 6 and 12 h after estrogen treatment; EGF binding remains elevated at 18 h and subsequently begins declining toward control levels. These results were obtained using 5 nM T -E G F which, on the basis of previous work ( l l ) , is a saturating concentration of growth factor.
To further ensure that these observed changes represent an increase in the number of binding sites per se, Scatchard analysis (18) was performed with membrane preparations from control animals and animals receiving hormone treatment 18 h prior to death. These results (Fig. 2) illustrate that estradiol administration leads to an increase in the number of uterine EGF-binding sites without a major change in affinity. K d values are 0.14 and 0.21 nM for control and estrogentreated samples, respectively; B,,, values are 22 (control) and 66 (estrogen-treated) fmol of EGF bound per mg of membrane protein.
In order to further substantiate this increase in EGFbinding sites, we performed affinity labeling studies and measurements of EGF-stimulated kinase activity. These results are shown in Figs. 3 and 4, respectively. For affinity labeling studies, membranes were incubated with lZ5I-EGF, briefly washed by centrifugation to remove unbound lZ51-EGF, and cross-linked with disuccinimidyl suberate. The cross-linked membranes were then extracted and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis; labeled material was observed by autoradiography.
As seen in Fig. 3, major (Mr = 170,000) and minor (Mr = 150,000) species of cross-linked Y -E G F are present, and hormone treatment produces a substantial increase in both species. It is important to note that binding in both samples is abolished by the presence of excess unlabeled EGF in the initial incubation.
EGF also stimulates the phosphorylation of a species of similar M, in extracts from control and hormone-treated membranes (Fig. 4); it is clearly seen that hormone treatment also leads to a substantial increase in the level of ;his species, which represents receptor autophosphorylation We have previously shown (11) that this autophosphorylation occurs largely a t tyrosine residues. Since the EGF receptor may function by phosphorylation of cellular substrates, tyrosine kinase activity was also measured with angiotensin I1 as an exogenous substrate (19). For these studies, uterine membranes were prepared from animals receiving estrogen treatment 18 h prior to death and from vehicle-treated control animals. The EGF receptor was solubilized by detergent treatment and partially purified by passage over a wheat germ agglutinin column. The EGF-stimulated kinase activity was then determined by adding angiotensin I1 as an exogenous substrate. The results of these studies are illustrated in Table I.
As seen in Table I, estrogen treatment causes a 2-3-fold increase in activity. This is in general agreement with results from the previous binding ( Figs. 1 and 2), affinity labeling (Fig. 3), and autophosphorylation data (Fig. 4).
In order to examine the hormonal specificity of the estrogen-induced increase in EGF receptor level, we tested the TIME AFTER ESTRADIOL  effects of other non-estrogenic steroids. As seen in Fig. 5, neither dexamethasone (600 pg/kg), nor dihydrotestosterone (400 pg/kg), nor progesterone (4 mg/kg) treatment produced the increase in EGF receptor content caused by estradiol (40 pg/kg) administration. As an initial approach to elucidating the mechanism hv which estrogen increases uterine EGF receptor levels, we tested the effects of cycloheximide and actinomycin D on the induction process. For these studies, animals were treated with estradiol and inhibitors 12 h prior to death, the earliest time we observed a consistent increase in EGF receptor levels ( Fig. 1). This time was chosen (rather than the 18-h treatment used in most previous experiments) to inimize the in vivo exposure of the test animals to these IY oxic inhibitors. The results of these studies are illustrated in Fig. 6. Neither inhibitor had any effect on the levels of EGF receptor in vehicle-treated control animals. Cycloheximide completely abolished the increase in uterine EGF levels following estrogen administration, and actinomycin D substantially decreased the effect of the hormone.

DISCUSSION
In a previous study, we demonstrated that uterine membranes contain EGF receptors with properties similar to EGF receptors from other sources (11). These include a high affinity for EGF (Kd = 0.4 nM), specificity, saturability, molecular weight of 170,000-150,000, and an EGF-stimulated kinase activity which phosphorylates both tyrosine and serine residues on the receptor itself. Based on studies in other systems, it is likely that the minor species of M , = 150,000 i s a proteolytic fragment of the major M, = 170,000 species (7, 9, 20). At present, we do not know the distribution of EGF receptors in the various uterine cell types in the rat.
In this study, we have shown that 17B-estradiol elevates uterine EGF receptor levels approximately 3-fold within 12 h. This increase in EGF receptor levels assessed by Y -E G F binding (Figs. 1 and 2) is also accompanied by increases in EGF receptor levels visualized by affinity labeling (Fig. 3), EGF-stimulated receptor autophosphorylation (Fig. 4) EGF-stimulated tyrosine kinase activity measured with angiotensin I1 as substrate (Table I). In the absence of information about levels of EGF (or other endogenous ligands) and receptor dynamics, this &fold increase could be caused by a variety of factors, and further studies will be required to determine the mechanistic basis of this effect.
The studies with different steroids indicate that the increases in uterine EGF receptor levels are specific for estrogens. Strictly speaking, these results only indicate that this effect is specific under the conditions of our study, i.e. acute administration in a defined dose range. This suggests that the observed effect is likely mediated by the estrogen receptor system. However, the uterus clearly contains receptors for androgens (21), glucocorticoids (22), and progesterone (23), and it would not thus be surprising if these hormones altered EGF receptor levels under other conditions or with different dosage regimens.
The studies with actinomycin D and cycloheximide provide some insight into the mechanism by which estrogen elevates uterine EGF receptor levels. The finding that cycloheximide abolishes the increase in EGF receptor suggests that the effect of E, represents de m u 0 synthesis of the receptor, and the effect of actinomycin D is consistent with a transcriptional mechanism of estrogen action. Sarff and Gorski (24) demonstrated that the doses of inhibitors used in our studies effectively block protein and RNA synthesis in the uterus. Conclusions based upon such inhibitor studies are only suggestive, however, especially under in vivo conditions. Studies in progress are thus aimed at quantitating EGF receptor levels by the use of antibodies and at measuring mRNA levels for the receptor. These should provide more definitive information about the mechanism by which estrogens regulate EGF receptor levels.