Peri- and Postpubertal Estrogen Exposures of Female Mice Optimize Uterine Responses Later in Life

Abstract

At birth, all female mice, including those that either lack estrogen receptor α (ERα-knockout) or that express mutated forms of ERα (AF2ERKI), have a hypoplastic uterus. However, uterine growth and development that normally accompany pubertal maturation does not occur in ERα-knockout or AF2ERKI mice, indicating ERα-mediated estrogen (E2) signaling is essential for this process. Mice that lack Cyp19 (aromatase knockout, ArKO mice), an enzyme critical for E2 synthesis, are unable to make E2 and lack pubertal uterine development. A single injection of E2 into ovariectomized adult (10 weeks old) females normally results in uterine epithelial cell proliferation; however, we observe that although ERα is present in the ArKO uterine cells, no proliferative response is seen. We assessed the impact of exposing ArKO mice to E2 during pubertal and postpubertal windows and observed that E2-exposed ArKO mice acquired growth responsiveness. Analysis of differential gene expression between unexposed ArKO samples and samples from animals exhibiting the ability to mount an E2-induced uterine growth response (wild-type [WT] or E2-exposed ArKO) revealed activation of enhancer of zeste homolog 2 (EZH2) and heart- and neural crest derivatives-expressed protein 2 (HAND2) signaling and inhibition of GLI Family Zinc Finger 1 (GLI1) responses. EZH2 and HAND2 are known to inhibit uterine growth, and GLI1 is involved in Indian hedgehog signaling, which is a positive mediator of uterine response. Finally, we show that exposure of ArKO females to dietary phytoestrogens results in their acquisition of uterine growth competence. Altogether, our findings suggest that pubertal levels of endogenous and exogenous estrogens impact biological function of uterine cells later in life via ERα-dependent mechanisms.

Estrogen (E2) is the primary female sex hormone that plays a critical role in reproductive function (1). In the rodent uterus, E2 activity is mediated by its receptor, estrogen receptor α (ERα), which binds directly to target sequences within the deoxyribonucleic acid, referred to as E2-responsive elements (2). Studies indicate that stromal cell nuclear ERα/E2 signaling leads to secretion of paracrine growth factors that can induce mitosis in adjacent luminal epithelial cells, ultimately leading to the transition of a single layer of cuboidal cell epithelium into a hyperplastic secretory layer (3, 4). Interestingly, administering the insulin-like growth factor 1 (IGF1) directly to mice partially mimics the uterine epithelial growth response that is induced by E2 (5). Studies by this lab utilizing a global ERα knockout mouse line demonstrated that ERα is necessary for the wave of uterine growth subsequent to E2 or IGF1 exposure, as ERα knockout uterine epithelial cells lacked proliferative response to either substance (5, 6). A similar experiment conducted with a global ERβ knockout mouse line showed luminal epithelial cells proliferate after E2 injection, indicating that ERβ is not necessary for luminal cell proliferation and may repress it to some extent (7). It is therefore accepted that in the mouse, E2/ERα action is responsible for the inducing of uterine proliferation. In order for ERα to act as a ligand-specific transcription factor, both of its activation function domains must be functional. Female mice that have a mutated activation function (AF) domain 2 (AF2ERKI) have a phenotype similar to that of ERα knockout mice. No uterine proliferation is observed, even in mice treated with E2 or IGF1 (8). The AF2-mutated ERα retains its affinity for E2 and its ability to bind to E2-responsive elements; however, transcriptional activity in the presence of E2 is significantly reduced due to the inability to recruit coactivators such as the p160 transcriptional coactivators (8, 9). Both the ERα knockout and the AF2ERKI share the characteristic that their uterine cells are insensitive to the pubertal E2, and thus the uterus remains in its hypoplastic prepubertal state. Interestingly, the mutated AF-2 domain also causes ICI182780 and tamoxifen, normally ER antagonists, to act as agonists thereby activating the receptor (8). Once the receptor is activated by ICI182780 or tamoxifen, proliferation occurs in the luminal epithelium, mimicking E2 action (8) and indicating that the hypoplastic uterine tissue of the AF2ERKI is competent to mount a growth response via its mutated ERα but still appear hypoplastic. These findings led us to explore the role of fully functioning E2 via ERα in mediating the maturation process that prepubertal uterine cells undergo following the initiation of estrous cyclicity. Here, we utilized the aromatase knockout (ArKO) mouse model because it lacks the Cyp19 gene and is unable to synthesize E2 (10) but expresses ER, therefore allowing manipulation of periods of E2 exposure. Earlier work with ArKO mice indicated the necessity to ensure that no estrogenic compounds were present in the chow they were fed in order to observe the full impact of E2 deficiency (11), indicating its tissues were extremely sensitive to both exogenous and endogenous substances with estrogenic activity.

Methods

Mice

All mice were handled in accordance with protocols approved by the National Institute of Environmental Health Sciences Animal Care and Use Committee in compliance with the National Research Council’s Guide for the Use and Care of Laboratory Animals (1). For experiments using ArKO, male and female mice heterozygous for the knocked-out Cyp19 gene were bred together (10). Earlier work with ArKO mice indicated the necessity to ensure that no estrogenic compounds were present in the chow they were fed (11), therefore the ArKO colony was maintained on a phytoestrogen-reduced diet (Zeigler, Garners, PA) except when specified otherwise. Offspring were genotyped by Transnetyx Inc. (Cordova, TN). ArKO and wild-type (WT) control littermates mice were weaned and separated by genotype at 3 weeks of age.

24-hour E2 treatment

ArKO mice and WT littermates were ovariectomized (OVX) at 10 weeks of age and then rested for 14 days. The OVX mice were divided into 2 treatment groups: V or E2. E2 (10 µg E2 per kg body weight) or saline vehicle (V) were administered by intraperitoneal injection. Twenty-four hours after initial treatment, uterine tissue was collected. Half of each uterus was snap frozen in liquid nitrogen, and the other half was fixed in formalin for immunohistochemistry.

3-day pubertal treatment

Both WT and ArKO mice were treated once a day with V or E2 (10 µg E2 per kg body weight), starting on postnatal day 28 and ending on day 30, for a total of 3 treatments. Some uteri were collected on day 31, weighed, and fixed in formalin. The rest of the mice were rested until 10 weeks of age and then OVX. Two weeks postovariectomy, mice were treated with V or E2 as above. Half of each uterus was fixed in formalin while the other half was snap frozen and kept at –80°C.

3-week to 10-week estradiol benzoate treatment

All animals were treated weekly with injections of estradiol benzoate (EB; 10 µg EB per kg body weight) starting at 3 weeks of age and continuing until 10 weeks of age, for a total of 7 treatments. At 10 weeks of age, the mice were OVX and rested for 10 to 14 days to clear any endogenous hormones. At the conclusion of the rest period, the OVX mice were divided into 2 treatment groups: V or E2. Twenty-four hours after initial treatment, uterine tissue was collected and fixed in formalin.

9-week to 10-week EB treatment

All animals were treated with EB once a week starting at 9 to 10 weeks of age for a total of 2 treatments. At 10 weeks of age, the mice were OVX and rested for 10 to 14 days to clear any endogenous hormones. At the conclusion of the rest period, the OVX mice were divided into 2 treatment groups: V or E2. Twenty-four hours after initial treatment, uterine tissue was collected, and half of each uterus was fixed in formalin while the other half was snap frozen.

Dietary phytoestrogen treatment

At 3 weeks of age, mice in this group were weaned onto 5K20 chow (LabDiet, St Louis, MO) that contained higher levels of phytoestrogen in the form of soybean meal. Mice were fed this chow until 10 weeks of age and then switched to a phytoestrogen-reduced diet and OVX and then allowed to rest for 14 days. At the conclusion of the rest period, the OVX mice were divided into 2 treatment groups: V or E2. Twenty-four hours after initial treatment, uterine tissue was collected, and half of each uterus was fixed in formalin while the other half was snap frozen.

IGF1 treatment

IGF1 was administered using an osmotic pump as previously described (12) and then tissue was collected 24 hours after pump placement and fixed in formalin as above.

Immunohistochemistry

Uterine tissue was embedded in paraffin and cut into 4-µm sections onto charged slides. Sections were deparaffinized, decloaked in a Biocare Decloaking Chamber with Biocare Antigen Decloaking buffer (BiocareMedical, Pacheco, CA), and blocked with 5% hydrogen peroxide. Sections were blocked with 10% normal horse serum (Jackson Immunoresearch, West Grove, PA) then incubated with mouse antihuman Ki-67 (BD Pharmingen/BD Biosciences, San Jose, CA; catalog no. 550609; research resource identifier (RRID): AB_393778 (13)) diluted 1:100 in 10% normal horse serum or with mouse anti-human ERα (Biocare Medical ER 1D5 catalog no. ACA 054C RRID AB_2651037 (14)) diluted 1:200 in 10% normal horse serum. Sections were incubated with 1:500 biotinyl-anti-mouse immunoglobulin G (Vector Laboratories, Burlingame, CA). Ki67 sections were then incubated with 1:50 extravidin peroxidase (Sigma Chemical, St Louis, MO), whereas ERα sections were incubated with RTU-Elite (Vector) and both Ki67 and ERα sections were then developed with Dako Products DAB substrate (Agilent Technologies, Santa Clara, CA).

Microarrays

Uterine ribonucleic acid (RNA) was made from samples from treatment groups, as previously described (12). Samples were submitted to the National Institute of Environmental Health Sciences Microarray Group for analysis using the Affimetryx MTA chip, as previously described (12). Datasets are deposited in GEO GSE147900.

Results

To study E2-mediated requirements for uterine pubertal maturation, we utilized mice that have the Cyp19 (aromatase) gene deleted (ArKO), which were therefore unable to synthesize E2 (10). Like the ERα knockout and AF2ERKI mice (2), ArKO uteri remained hypoplastic as they were not exposed to endogenous E2. The ArKO uterus did, however, have ERα in its cells (Fig. S1 All supplementary material and figures are located in a digital research materials repository. (15)), and thus the impact of administering exogenous E2 on uterine maturation and response could be studied. Acute (24-hour) uterine growth responses of ArKO females to exogenous E2 had not previously been examined. OVX adult (10 weeks old) ArKO and WT littermates were given a 24-hour treatment of either V or E2. The ArKO mice did not respond to E2 as little to no Ki-67 staining was observed, and uteri exposed to E2 looked similar to the vehicle-treated controls (Fig. 1). We hypothesized that lack of exposure to E2 during puberty might impact competence for mounting a growth response to E2 in adulthood.

Figure 1.

Aromatase knockout (ArKO) mouse uterine response to E2 is impaired. (a) OVX wild-type (WT) or ArKO mice were treated with saline vehicle (V) or estrogen (E2) for 24 hours. Sections were evaluated by immunohistochemistry for the Ki67 proliferative marker. Extensive presence of positive (brown) cells was observed in WT E2-treated samples. Representative sections are shown. Note that because the ArKO uterine tissue is smaller than WT, the images were taken at higher magnification, as evidenced by the different scale bars. (b) Percentage of luminal epithelial cells that stained positive for Ki67 in samples from OVX adult V- or E2-treated WT (blue; n = 6 V, 8 E2) and ArKO (red; n = 6 V, 7 E2) mice. *P < 0.01 versus V by 2-way analysis of variance with the Fisher least significant difference posttest.

To evaluate whether exposure to E2 at the age of puberty was sufficient to confer later E2 responsiveness, ArKO mice were treated with either V or E2 once a day for 3 days at the time WT mice were beginning to have estrous cycles (days 28, 29, and 30 after birth). The E2 treatment was meant to mimic the rising levels of E2 normally observed at the onset of puberty. To evaluate the immediate effect of the treatments, some samples were collected on day 31. Both WT and ArKO uteri treated for 3 days with E2 showed an increase in uterine weights (Fig. 2a), indicating the ArKO uterus could and did respond to E2 treatment. Examination of uterine histology indicated a similar response of WT and ArKO to E2, however after 3 days of E2 dosing, the Ki67 proliferative marker was sporadically seen in WT samples, whereas the ArKO luminal epithelial cells maintained continued proliferation in most epithelial cells (Fig. 2b and 2c).

Figure 2.

Aberrant growth of peripubertal aromatase knockout (ArKO) uterine epithelial cells induced by estrogen (E2). (a) Uterine wet weights of wild-type (WT) and ArKO females on pnd31 after daily saline vehicle (V) or E2 injections on post natal day (pnd)28 to 30. (b) Histology of uterine samples of WT and ArKO females on pnd31 after daily V or E2 injections on pnd28 to 30. The sections were stained for Ki67. The right panels show a high magnification image to illustrate the increased epithelial proliferation in AKO versus WT. (c) Percentage of luminal epithelial cells that stained positive for Ki67 in samples from V- or E2-treated WT (blue; n = 5 V, 6 E2) and ArKO (red; n = 6 V, 6 E2) mice. *P < 0.01 versus V; +P < 0.01 versus WT by 2-way analysis of variance with the Fisher least significant difference posttest. h, hours.

The remainder of the pubertal window-treated mice were aged to adulthood (≥10 weeks old), OVX, and treated with V or E2, for 24 hours. The ArKO mice treated on post natal day (pnd)28 to 30 with E2 exhibited an increase in the epithelial proliferative response to E2, (Fig. 3). The E2 treatment on pnd 28 to 30 led to increased Ki67-positive epithelial cells (Fig. 3b and 3c), but the growth response of 4 of 6 of the mice was less robust than observed in WT mice that were not treated during the pubertal window (Fig. 3c).

Figure 3.

Peripubertal estrogen (E2) treatment partially restores aromatase knockout (ArKO) uterus E2-induced growth competence later in life. (a) Treatment scheme used to evaluate the impact of replacing peripubertal E2 using daily injections on post natal day (pnd) 28 to 30. Mice were OVX at 10 weeks of age and then rested for 2 weeks before treatment with saline vehicle (V ) or E2 for 24 hours. (b) Histological sections from uteri of wild-type (WT) or ArKO mice that were injected with V or E2 on pnd 28 to 30, then OVX at 10 weeks of age and treated with V or E2 for 24 hours. Sections were stained for the Ki67 proliferative marker. Note that because the ArKO uterine tissue is smaller than WT, some of the images were taken at higher magnification. (c) Percentage of luminal epithelial cells that stained positive for Ki67 in samples from pnd 28 to 30 V-treated WT (blue; n = 2 V, 2 E2) and ArKO (green; n = 3 V, 4 E2) mice or pnd 28 to 30 E2-treated WT (red; n = 2 V, 2 E2) and ArKO (purple; n = 3 V, 6 E2) mice. *P < 0.01 versus V by 2-way analysis of variance with the Fisher least significant difference posttest. h, hours; wk, week.

To evaluate whether E2 replacement throughout the postpubertal period might further facilitate later E2 responsiveness, WT and ArKO mice were treated with EB once a week for 7 weeks, starting at week 3 of life and ending at week 10. EB is a prodrug of estradiol, and the treatment administered was meant to mimic episodic endogenous E2 availability that occurs during estrous cycles. Following the EB treatment period, mice were OVX, and after a 2-week rest period, mice were given a 24-hour treatment of either V or E2. As expected, the luminal epithelium of WT uteri treated with V did not proliferate (Fig. 4), and EB-treated WT uteri dosed with E2 for 24 hours showed a visible increase in the number of luminal cells proliferating, as evidenced by the increase in Ki-67 staining observed (Fig. 4b and 4c). In the EB-treated ArKO uteri, proliferation was induced by E2 at levels comparable with WT uteri (Fig. 4b and 4c), indicating the E2 replacement resulted in acquisition of growth responsiveness.

Figure 4.

Weekly estradiol benzoate (EB) treatment (weeks 3-10) further increases estrogen (E2)-induced growth of aromatase knockout (ArKO) uterus. (a) Treatment scheme used to evaluate the impact of repeated E2 exposure using weekly EB injections from ages 3 to 10 weeks. (b) Histological sections from uteri of wild-type (WT) or ArKO mice treated as described in a and then treated with saline vehicle (V) or E2 for 24 hours. Sections were stained for Ki67. (c) Graph of percentage of luminal epithelial cells per section with Ki67 marker counted in samples. *P < 0.01 versus V by 2-way analysis of variance with the Fisher least significant difference posttest (n = 6 WT V, 6 WT E2, 5 ArKO V, 9 ArKO E2). h, hours; wk, week.

Since we observed increased responsiveness as a result of the lengthier EB treatment (7 weeks) as compared with the pubertal treatment (3 days), we wondered whether the improved response was the result of the longer treatment period or whether it reflected the shorter time between the last E2 treatment and the experimental assessment of response (8 weeks vs 2 weeks). Therefore, we administered EB for a shorter time (Fig. 5), only on weeks 9 and 10, corresponding to the last 2 weeks of the 7-week treatment and then OVX and rested the mice for 2 weeks as before. When mice were treated with E2 for 24 hours, we observed a response like that of the lengthier treatments (Fig. 5). This suggests that the treatment period can be shorter and still result in a growth-competent tissue as long as the rest period is not too lengthy.

Figure 5.

Estradiol benzoate (EB) treatment during weeks 9 to 10 is sufficient for estrogen (E2)-induced aromatase knockout (ArKO) uterine growth. (a) Treatment scheme used to evaluate the impact of recent E2 exposure using EB injections from during weeks 9 and 10. (b) Histological sections form uteri of wild-type (WT) or ArKO mice treated as described in a and then treated with saline vehicle (V) or E2 for 24 hours. Sections were stained for Ki67. Note that because the ArKO uterine tissue is smaller than WT, the images were taken at higher magnification. (c) Graph of percentage of luminal epithelial cells per section with Ki67 marker counted in samples. *P < 0.01 versus V by 2-way analysis of variance with the Fisher least significant difference posttest (n = 6 WT V, 4 WT E2, 3 ArKO V, 3 ArKO E2). h, hours; wk, week.

To evaluate the impact of the EB treatment that underlies acquisition of E2-induced growth response, we assessed the transcriptome of RNA samples isolated from V-treated OVX adult WT or ArKO females (“V” samples from Fig. 1 experiment), and compared the impact of the 9- to 10-week EB treatment (“V” samples from Fig. 5 experiment) on gene expression using microarray. Analysis comparing RNA samples that lacked growth response competence (ArKO OVX V) with RNA samples from animals exhibiting the ability to mount an E2-induced uterine growth response (OVX V-treated WT or EB-treated ArKO) revealed differential expression of 411 Refseq genes between the 2 groups (Table S1(15), Fig. 6). Analysis of the differentially expressed genes using Ingenuity Pathway Analysis tools revealed activation of the upstream regulators heart- and neural crest derivatives-expressed protein 2 (HAND2) and enhancer of zeste homolog 2 (EZH2) (Table S2A, Fig. S2 (15)) in the nonresponsive ovariectomized ArKO in comparison with the fully responsive groups (OVX V-treated WT or EB-treated ArKO). Both EZH2 and HAND2 are key uterine regulators that facilitate proper uterine development and function by inhibiting the rate of epithelial proliferative responses (16-19). We have previously classified ERα–binding sites in mouse uterine tissue as super-enhancers or typical enhancers, based on H3K27Ac chromatin immunoprecipitation sequencing signal (20, 21). Examination of the Hand2 gene revealed that there was an ERα-binding super-enhancer that overlapped the transcript (Fig. S3A (15)). Our additional studies have indicated that ERα-binding super-enhancers were present prior to pubertal uterine maturation (21) and that the genes they regulate acquired E2 responsiveness during puberty (21). We next examined the role of ERα in potentially mediating the level of Hand2 expression during pubertal maturation. We evaluated uterine transcripts in microarray data from both prepubertal and OVX adult uteri 2 hours after treatment with V or E2 (Fig S2B). Hand2 was significantly lower in adult samples than in prepubertal samples (Fig S3B) and was decreased by E2 treatment. These observations are consistent with a developmental role for E2 in establishing proper expression of Hand2.

Figure 6.

Estradiol benzoate (EB) treatment of aromatase knockout (ArKO) (weeks 9-10) alters uterine transcriptome. (a) Heatmap cluster of log2 signal of differentially expressed genes (DEG) in saline vehicle (V)-treated OVX ArKO (responds: “No”) versus V-treated growth-competent samples (ArKO EB ovx, wild-type [WT] EB ovx and WT ovx; responds: “Yes”). Scale is log2 signal intensity (3.5-20). (b) Comparison of DEG (Bioset 1) to Bioset 2 from mouse embryonic fibroblasts (MEF) overexpressing HAND2, NKX2-5 GATA5, and MEF2C versus empty vector indicated positive correlation of differential expression of 171 genes.

Conversely, the upstream regulator, GLI Family Zinc Finger 1 (GLI1), a mediator of Indian hedgehog (IHH) signaling, was expressed at a significantly lower level in the OVX V-treated ArKO uterus than in the growth-competent samples (OVX V-treated WT or EB-treated ArKO; Table S2A and Fig. S2 (15)), and the analysis of its downstream targets indicated that GLI1 signaling was inhibited (Table S2A (15)). Deletion of uterine Ihh inhibited cell cycle progression (22) and overexpression of Smoothened, which is a transducer of IHH initiated signals, in the mouse uterus led to hypertrophy. Therefore, one role of E2 in pubertal maturation may be optimization of HAND2-, EZH2-, and IHH-mediated signaling to facilitate uterine responses. Using the NextBio tool, we observed a positive correlation of this list of genes (bioset) (OVX V-treated ArKO uterus vs OVX V-treated WT or EB-treated ArKO) to a bioset from mouse embryonic fibroblasts that were overexpressing Hand2 (Fig 6B), which aligned with the observation of increased HAND2 activity in our analysis. We also observed a positive correlation with a bioset from OVX V-treated ERα knockout uteri compared with OVX V-treated WT uteri (Fig. S3C (15)), consistent with a requirement for ERα for development of uterine growth competence during pubertal development. Analysis of the impact of the differential gene expression of OVX V-treated ArKO versus growth-competent samples included predicted inhibition of cell migration and development together with predicted increase in morbidity and genitourinary system malformation (Table S2B (15)).

Since we observed an impact of E2 exposure either at puberty or later in life on uterine growth response, we wondered whether exposure of ArKO females to dietary estrogens might have an impact on later uterine responsiveness. Therefore, we fed weanlings a diet that contained soy proteins from weeks 3 to 10 of life, OVX the mice, and replaced their food with the phytoestrogen-reduced chow and rested the mice for 2 weeks. When these mice were treated with E2 for 24 hours, we observed a response similar to that seen after the pubertal E2 treatment (Fig. 7; compare with Fig. 3). The ArKO exhibited a recovery of growth response to E2 that was less robust than the normal WT growth response (Fig. 7b and 7c).

Figure 7.

Exposure to phytoestrogen-containing diet partially restores aromatase knockout (ArKO) uterus estrogen (E2)-induced growth. (a) Treatment scheme used to evaluate the impact exposure to diet containing phytoestrogens from weeks 3 to 10. (b) Histological sections from uteri of wild-type (WT) or ArKO mice treated as described in a and then treated with saline vehicle (V) or E2 for 24 hours. Sections were stained for Ki67. Note that because the ArKO uterine tissue is smaller than WT, the images were taken at higher magnification. (c) Graph of percentage of luminal epithelial cells per section with Ki67 marker counted in samples. *P < 0.01 versus V by 2-way analysis of variance with the Fisher least significant difference posttest (n = 4 WT V, 4 WT E2, 4 ArKO V, 5 ArKO E2). h, hours; wk, week.

Finally, we have previously noted that E2 induces growth factors, including IGF1, and that administering IGF1 can mimic the uterine growth response that E2 induces (5). We were interested in evaluating whether the acquisition of uterine growth competence subsequent to puberty also impacted the ability of IGF1 to induce a growth response. To test this, we evaluated the impact of IGF1 administration on the ArKO uterus. Interestingly, the IGF1-induced growth response was minimal unless the ArKO females first received the EB treatment (Fig. S4A and B (15)), and a 3- to 10-week EB exposure was required to attain a growth response comparable with WT (Fig. S4A and B (15)). ArKO females exposed to dietary phytoestrogens did not acquire a uterine growth response to IGF1 (Fig. S4 A and B (15)).

Discussion

Appropriate uterine epithelial cell function is critical for reproductive health, as the luminal epithelium is the initial interface for implanting embryos while the glandular epithelium secretes molecules needed for establishing pregnancy, such as leukemia inhibitory factor (23). For pregnancy to be successfully established, uterine cells must respond appropriately to cyclical levels of ovarian hormones and form an optimal receptive endometrial environment (24, 25). Thus, uterine endometrial cells must be acutely responsive in anticipation of hormonal and potential embryonic signals but must also sense and adapt following ovulatory cycles that do not produce embryos in order to return to a nonreceptive state. Perturbations in endometrial cell responses contribute to diseases of the endometrium, including endometrial hyperplasia and adenocarcinoma as well as endometriosis (24, 26-29), and consequently greatly impact a woman’s fertility and general health. Therefore, it is important that we understand the mechanisms involved in establishing optimal hormonal responsiveness. Our findings indicate that the E2 exposure normally provided after onset of estrous cycles (Fig. 8) or experimentally in ArKO mice, by EB dosing, is necessary for later luminal epithelial cell proliferation. The mechanisms affected by EB priming that confer later uterine competence in response to E2 have not been fully elaborated (Fig. 8). We did observe indications of decreased GLI1 signaling, suggesting an impact on IHH signaling, which is known to play essential roles in uterine function (22). We also observed enrichment of transcripts correlated with increased HAND2 and EZH2 signaling in ArKO versus WT uterine RNA. Both HAND2 and EZH2 are known inhibitors of uterine epithelial cell growth responses, HAND2 via inhibition of fibroblast growth factor signaling (17), and EZH2 via control of ERα activity (18, 19). In fact, uterine deletion of Hand2 results in a hyperproliferative epithelial response to E2 (17). Although we did not observe a significant difference in Hand2 expression between untreated ArKO and WT, the differentially expressed genes profile was correlated with HAND2 signaling (Table S2 (15)). Additionally, in our analysis of the uterine chromatin landscape, we identified an ERα-binding super-enhancer that overlaps the Hand2 gene (21), and we also observed that the level of Hand2 decreases significantly after puberty (Fig S3B). Although most of the ERα-binding super-enhancers identified in our study were observed prior to pubertal development, the prepubertal Hand2 super-enhancer exhibits less ERα binding than the adult. Altogether, this indicates that ERα-mediated signaling might impact developmental expression of factors important to normal uterine responses.

Figure 8.

Mechanism by which the endometrium acquires competence for estrogen (E2)-induced growth. The uterus undergoes postpubertal maturation as a result of repeated exposure to E2 with each estrous cycle. Following maturation, epithelial cell proliferation following acute E2 treatment demonstrates competency. E2 working through estrogen receptor α (ERα) decreases HAND2 and EZH2 signaling and increases GLI1 signaling. Without pubertal E2 response, as in the aromatase knockout (ArKO) or ERα knockout (ERαKO), epithelial competency is not observed. Providing exogenous pubertal E2 to ArKO mice, via estradiol benzoate (EB) dosing or through dietary E2, can induce endometrial competency. Exposure to endocrine-disrupting chemicals (EDCs) during the window of pubertal maturation might impact the process. Abbreviations: EZH2, enhancer of zeste 2 polycomb repressive complex 2; GLI1, GLI family zinc finger 1; HAND2, heart- and neural crest derivatives expressed 2.

We showed that dietary E2 exposure, which might compensate for the lack of postpubertal E2 exposure (Fig. 8), enables similar effects in later responsiveness of uterine epithelial cells. Our observation, although preliminary and made using an animal model, suggests a period of heightened susceptibility of young individuals, which has potential implications regarding exposures to environmental substances. Here, we found that pubertal E2 can impact later response of the mouse uterus because in mice engineered to lack the aromatase enzyme, and thus unable to produce E2, uterine response is impaired unless exogenous E2 is provided. Conversely, it is well known that when prepubertal mice are exposed to exogenous estrogens, such as E2, diethylstilbestrol, or the soy protein genistein, their endometrial development is permanently altered, resulting in genomic, epigenetic, and pathologic alterations (30-32). Thus, it is important to pay attention to this peripubertal period during which growth competence is established as it may be susceptible to perturbations as a result of exposure to endocrine-disrupting chemicals (Fig. 8). Observations such as these, made in mouse models, highlight the sensitivity of female reproductive tissues during the prepubertal time period that can have potentially lifelong effects. Additionally, they underline the importance of preserving the uterine hormone balance early in life. Hormone disruptions that occur during childhood could have important implications for reproductive health as an adult. Future work will continue to reveal the mechanistic details that contribute to these processes and better enable interventions of prevention and treatment.

Glossary

Abbreviations

AF

activation function

AF2ERKI

mice that have a mutated activation function (AF) domain 2

ArKO

aromatase knockout

E2

estrogen

EB

estradiol benzoate

ER

estrogen receptor

GLI1

GLI Family Zinc Finger 1

HAND2

heart- and neural crest derivatives-expressed protein 2

IGF1

insulin-like growth factor 1

OVX

ovariectomized

EZH2

Enhancer of zeste homolog 2

RNA

ribonucleic acid

V

saline vehicle

WT

wild-type

Elson S. Floyd College of Medicine, Washington State University, Spokane, WA (M. Carmona); Northwestern University, Chicago, IL (K. G. Foley); University of North Carolina, Chapel Hill, NC (L. J. Donoghue); North Carolina State University, Raleigh, NC (S. L. Lierz); and Washington State University, Pullman, WA (W. Winuthayanon).

Additional Information

Disclosure Summary: The authors declare that there is no conflict of interest regarding the publication of this article.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.