Ovary-Independent Estrogen Receptor Expression in Neonatal ...

BIOLOGY OF REPRODUCTION 58, 1009-1019 (1998)

Ovary-Independent Estrogen Receptor Expression in Neonatal Porcine Endometrium' Becky J. Tarleton,4 Anne A. Wiley,4 Thomas E. Spencer, 3 '4 Anthony G. Moss, 5 and Frank F. Bartol 2 4,

Departments of Animal and Dairy Sciences 4 and Zoology and Wildlife Science,5 Auburn University, Auburn, Alabama 36849-5415 Between birth (Day 0) and postnatal day (PND) 120, the porcine uterine wall is transformed from histoarchitectural infancy to maturity, and uterine tissues acquire the capacity to respond to conceptus signals (reviewed in [4]). In the pig, endometrial glands are absent at birth but develop rapidly thereafter, forming coiled, branched epithelial tubes that extend radially from the luminal surface, through endometrial stroma, to the adluminal border of the myometrium by approximately PND 30. Intensely glandular folds of endometrium are apparent by PND 60. Events that accompany the appearance and proliferation of endometrial glands during this period include an abrupt, spatially focused increase in glandular, but not luminal, epithelial DNA synthesis, and alterations in the distribution and/or orientation of -D-galactose and/or -N-acetyl-D-galactosamine residues associated with glandular, but not luminal, epithelial cell surfaces [6, 7]. These developmental patterns suggest that uterine gland genesis in the neonatal pig involves local regulation of both cell proliferation and differentiation. Estrogen is uterotropic in the neonatal pig [8]. Administration of estradiol-1713 valerate (EV) to crossbred gilts for 7 days prior to hysterectomy on PND 7, 14, or 49 increased uterine wet weight and endometrial thickness, and altered patterns of uterine protein synthesis on all days. Effects were related to age at first exposure. Uterotropic effects of EV were confined largely to the endometrium on PND 7 and 14 and, overall, were greater in older animals [8]. Age-specific responses of neonatal porcine uterine tissues were interpreted to suggest that acquisition of estrogen sensitivity occurs progressively in the developing uterine wall and that estrogen-sensitive uterine cell populations develop first in the endometrium. Ontogeny of ER expression in the neonatal porcine endometrium has not been characterized, nor is it known whether ovarian factors affect patterns of endometrial ER expression during the period between birth and PND 120 as the porcine endometrium is becoming structurally and functionally mature. Moreover, delineation of mechanisms through which estrogens or related xenobiotics may affect normal uterine growth and development, or induce developmental lesions in these tissues, requires that ER-positive cell populations be identified and that patterns of development of ER-positive cell populations be characterized in situ. Therefore, objectives of the present study were to determine effects of age and of ovariectomy at birth on patterns of uterine growth, endometrial development, distribution of ER-positive endometrial cells, and endometrial ER gene expression in situ between PND 0 and 120 in the porcine uterus.

ABSTRACT Effects of age and ovariectomy (OVX) at birth on uterine growth, endometrial development, and estrogen receptor (ER) expression were determined for intact and OVX gilts (n = 5 per day) hysterectomized on postnatal days (PND) 0, 15, 30, 60, 90, or 120. Uteri were evaluated histologically, and ER protein and mRNA expression were characterized immunohistochemically and by in situ hybridization. OVX did not affect uterine weight or endometrial thickness until after PND 60, when both increased more rapidly in intact gilts. Neither did it affect genesis of uterine glands, which were present and which proliferated after PND 0, or endometrial ER expression patterns in glandular epithelium (GE), luminal epithelium (LE), or stroma (S)between PND 0 and 120. Endometrium was ER negative at birth. On PND 15, the ER signal was strong in GE, weak in S, and effectively absent in LE. Thereafter, although the ER signal remained strong in GE and increased through PND 60 in S, it was not evident consistently until after PND 30 in LE. The data indicate that 1) porcine uterine growth and endometrial morphogenesis are ovary-independent processes before PND 60; 2) uterine gland genesis is associated temporally with development of ER-positive endometrial GE and S; and 3) regulation of endometrial ER expression isovary independent between PND 0 and 120. The results establish the ER as a marker of GE differentiation and implicate this receptor in mechanisms regulating endometrial morphogenesis in the neonatal pig. INTRODUCTION Organizational mechanisms regulating growth, morphogenesis, and cytodifferentiation of uterine tissues are poorly defined. However, it is clear that adult uterine structure and function can be compromised if developmental events critical to the success of uterine wall organization are disrupted by exposure of tissues to estrogen or related compounds during specific periods of fetal or neonatal life [1-5]. Generally, such observations can be interpreted to indicate that at least some organizational events critical to the success of uterine development are estrogen sensitive and, therefore, are likely to depend upon the presence, distribution, and relative state of activation of the estrogen receptor (ER) system. Consequently, it is important that such relationships be defined and evaluated with respect to their potential to affect adult uterine function, reproductive efficiency, and health. Accepted November 24, 1997. Received October 16, 1997. 'This work was supported by USDA Grants 91-37203-6605 and 9537203-1995 to F.F.B. Data were presented, in part, at the Fourth International Conference on Pig Reproduction at the University of Missouri, Columbia, in May, 1993. This is publication number 4-975851 of the Alabama Agricultural Experiment Station. 2 Correspondence. FAX: (334) 844-1519; e-mail: [email protected] 3 Current address: Center for Animal Biotechnology, Texas A&M University, Institute of Biosciences and Technology, 444 Kleberg Center, College Station, TX 77843-2471. 1009

MATERIALS AND METHODS Animals and Tissue Collection At birth (Day 0), crossbred gilts were assigned to remain intact (n = 30) or to be ovariectomized (OVX, n = 25) on

1010

TARLETON ET AL.

Day 0. Animals in each treatment group were then assigned to be hysterectomized (n = 5 gilts per day) under general anesthesia on either Day 0 (intact only), 15, 30, 60, 90, or 120. Procedures were approved by the Auburn University Institutional Animal Care and Use Committee and were in accordance with the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching [9]. Each uterus was trimmed free of supportive connective tissues; oviducts and ovaries (intact only) were removed, and the uterus was separated from the cervix at the level of the internal cervical os. Uterine wet weights were obtained using an analytical balance. A cross section (approximately 1 cm in length) from the middle of one uterine horn was fixed immediately in 4% paraformaldehyde, embedded in Paraplast Plus (Sherwood Medical, St. Louis, MO), and sectioned at 5-7 pm. The remainder of the uterus was snap-frozen in liquid nitrogen and stored at -80 0 C. A segment of small intestine from one pig was obtained at the time of hysterectomy, fixed in 4% paraformaldehyde, embedded, and processed for use as a negative control tissue for ER immunohistochemistry. After embedding, tissues for histological analyses were processed as described previously [7], then stained with hematoxylin and counterstained with eosin. Procedures for histological measurements were as described by Spencer et al. [7]. Tissue sections were mounted on poly-L-lysine (0.1% v:v; Sigma Chemical Co., St. Louis, MO)-coated glass slides. Western Blotting ER protein was detected on Western blots and by immunohistochemistry using the H222 [10] rat anti-human ER monoclonal antibody (H222 mAb; gift of Dr. G.L. Greene, University of Chicago). The precise epitope recognized by the H222 mAb is not known. Though preliminary, current evidence indicates that this antibody is specific for ERa (Dr. G.L. Greene, personal communication). Specificity of the H222 mAb was demonstrated by Western blot analysis of total protein extracted from porcine uteri (0, 15, and 120 days of age) according to the procedure of Glatstein and Yeh [11], exactly as described by Goyal et al. [12] with the following modifications. Extracted protein (50 jIg per lane) was subjected to SDS-PAGE [13], using a 10% separating gel and the Protean II Electrophoresis Cell system (Bio-Rad, Richmond, CA). The separated proteins were transferred to nitrocellulose membrane according to procedures recommended for the Bio-Rad Trans-Blot Electrophoretic Transfer Cell. To visualize transferred proteins, the membrane was stained with 0.1% (w:v) Ponceau S in 1% acetic acid so that individual lanes could be identified and cut into strips. Strips were destained in Tween 20 (0.5% w:v)/Tris-buffered saline and probed with H222 [12]. Controls included substitution of an irrelevant rat IgG (Jackson Immunological Research Laboratories, West Grove, PA) for H222 mAb and elimination of the primary antibody. Protein bands were visualized using the ECL chemiluminescence system (Amersham Life Science, Cleveland, OH) and Kodak X-OMAT LS film (Eastman Kodak, Rochester, NY). Immunohistochemistry Nuclear ER protein was localized in neonatal uterine tissue sections using the mAb H222 [10] as described previously [12]. Procedures for localization of ER in paraffinembedded tissue sections were modified from those described by Cartun and Pederson [14] and Papadimitriou et

al. [15]. Sections were deparaffinized in xylene, rehydrated through a series of ethanol washes, and rinsed in water. Endogenous peroxidase activity was blocked by incubating sections in 3% H2 0 2 for 5 min at room temperature. Next, slides were rinsed in water and placed in PBS for 5 min at room temperature. Tissue sections were immersed in 37°C Pronase E solution (0.5 mg/ml; protease type XIV; Sigma) for 8 min. The H222 mAb was applied to tissue sections at a concentration of 5 ixg/ml in PBS, with 1% (w:v) BSA. Visualization of nuclear ER protein was accomplished using the Vecta-Stain Elite ABC kit (Vector Laboratories, Burlingame, CA), which employs the peroxidase enzyme detection system. Negative controls included substitution of irrelevant rat IgG for H222 in uterine sections, elimination of the pronase digestion step, and evaluation of porcine small intestine sections processed identically to uterine tissues stained with the H222 mAb. Nuclear staining intensity (absent, weak, moderate, or strong) for luminal epithelium (LE), glandular epithelium (GE), and stroma (S) was scored independently by two observers. Multiple nonsequential sections (3-5) from each uterus were examined. Northern Blotting Northern hybridization was performed to confirm specificity of the ovine oER8 [16] probe for porcine RNA. The sequence of this partial cDNA for the ovine ER [17] shares significant homology with ERua of other species, but not with the recently cloned ER3 [18, 19]. Total cellular RNA was extracted from uteri using TRIzol Reagent (Life Technologies, Gaithersburg, MD) according to manufacturer's instructions. RNA (15 g) was solubilized in sample buffer (single-strength MOPS [3-[N-morpholino]propanesulfonic acid], 50% formamide, 2.2 M formaldehyde, 25% glycerol, and 0.025% bromophenol blue), heat denatured, and electrophoresed through a 1.25% agarose gel (1.1 M formaldehyde, single-strength MOPS). RNA was transferred to Nytran Plus (Schleicher and Schuell, Keene, NH) by downward capillary blotting [20] and UV cross-linked. Antisense cRNA probe was produced by in vitro transcription with T7 RNA polymerase from the EcoRI-linearized oER8 template, incorporating [ot- 32 P]UTP (specific activity: 800 Ci/mmol; ICN Pharmaceuticals, Costa Mesa, CA). Membrane was prehybridized for 4 h; probe was denatured in hybridization buffer for 5 min at 65°C and then added to the hybridization tube. The membrane was hybridized overnight at 55°C, rinsed three times in 0.1-strength SSC (single-strength SSC is 0.15 M sodium chloride and 0.015 M sodium citrate) with 0.1% SDS for 20 min at 68°C, wrapped in plastic film, and exposed to a BI Imaging Screen (BioRad) for 12 h. Radioactivity was localized using a GS-525 Molecular Imager System (Bio-Rad). In Situ Hybridization In situ hybridization (ISH) [21] was performed, using [ 35S]UTP-labeled (specific activity: 800 Ci/mmol; Amersham, Arlington Heights, IL) antisense and sense cRNA probes produced by in vitro transcription from the EcoRIand BamHI-linearized oER8 cDNA template [16], with a MaxiScript kit (Ambion, Austin, TX). T7 and SP6 RNA polymerases were used to transcribe antisense and sense probes, respectively. Procedures were essentially identical to those previously described [16, 22]. Each slide included tissue sections that received either radiolabeled sense (negative control) or antisense cRNA probe. Hybridization was carried out overnight at 55°C. After posthybridization pro-

1011

ESTROGEN RECEPTOR IN THE NEONATAL PORCINE UTERUS cessing, slides were dipped in NTB-2 liquid photographic emulsion (Eastman Kodak) and exposed at 4°C for 12 wk. Slides were developed at 15°C using D19 developer (Eastman Kodak) and counterstained with mercury-free Harris modified hematoxylin with acetic acid (Fisher Scientific, Pittsburgh, PA).

9 U C

3:

c S

Light Microscopy and Digital Imaging Brightfield images of porcine endometrial tissues stained with hematoxylin and eosin, and for nuclear ER protein, were obtained using x10, x20, or x40 planachromat objectives on a conventional compound light microscope (model BHS; Olympus, New York, NY). Brightfield images were captured directly to a digital image capture board (model Flashpoint; Integral Technologies, Indianapolis, IN) using an instrumentation-grade Newvicon camera designed for high-quality video-enhanced microscopy (VE1000; Dage/MTI, Michigan City, IN). Slides of porcine endometrium developed to display in situ-hybridized ER antisense and sense cRNAs were examined at x10, x20, and x40 by darkfield microscopy on the same microscope equipped with an Olympus darkfield condenser. Darkfield images were captured from a silicon intensified target (SIT) camera (model 68; Dage/MTI). Both brightfield and darkfield images were background subtracted and were processed digitally using constant parameters with a commercial image-processing package (Image-Pro Plus; Media Cybernetics, Bethesda, MD) on a Pentium-based microcomputer. Images were printed on a high-resolution laser printer (Tektronix, Wilsonville, OR). Great care was taken to keep image capture and processing conditions identical for controls and corresponding experimental conditions, and between iterations of the image capture and processing effort for related experiments. All images were corrected for the same background level by manual adjustment of the camera black level to the same image bit-value by comparison against a pseudocolor standard, and the gain and the SIT camera-intensified target voltage were maintained at the same value for all related images to ensure an accurate representation of the relative local image intensities. Thus, all images of a given type (i.e., all darkfield images and, separately, all brightfield images) were directly comparable to each other, and all images shown herein represent the relative local image signal realistically. Statistical Analyses Data for uterine weight and endometrial thickness were subjected to least-squares analysis of variance using General Linear Models procedures [23]. The statistical model included effects of age, treatment (ovariectomy), and their interaction. Tests of significance were based on expectations of the error mean squares. Least-squares regression analyses and tests for heterogeneity of regression were performed to determine whether patterns of change in uterine wet weight or endometrial thickness differed between intact and OVX gilts in association with advancing postnatal age. Data are presented as least-squares means with standard errors. RESULTS Uterine Growth Effects of age and OVX at birth on uterine wet weight and endometrial thickness are summarized in Figure 1. Both uterine weight and endometrial thickness were af-

=r 0

30

60

90

120

Postnatal Age (days)

E

1400

-

1200

*

1000

°

800

-

600

E o w

400 200 0

0

30

60

90

120

Postnatal Age (days) FIG 1. Effects of neonatal age and OVX at birth on porcine uterine wet weight (A) and endometrial thickness (B). Both uterine wet weight and endometrial thickness were affected by postnatal age (p < 0.01) and OVX (p < 0.01), and an age x OVX interaction was detected (p < 0.05). For each response, least-squares means ( SE) for intact (solid circles) and OVX (open circles) gilts are presented, together with best-fit least-squares regression curves. Patterns of change in uterine wet weight (A) followed a cubic trend for intact (Y = -0.003 + 0.203X - 0.006X 2 + 0.00006X 3 ; R2 = 0.97, p < 0.01; solid line) and a linear trend for OVX gilts (Y = -0.136 + 0.097X; R2 = 0.74, p < 0.01; broken line). Patterns of change in endometrial thickness (B) followed different quadratic trends for both intact (Y =259.5 + 4.73X + 0.034X 2 ; R2 = 0.87, p < 0.01; solid line) 2 and OVX gilts (Y = 288.4 + 10.3X - 0.05X2; R = 0.62, p < 0.01; broken line).

fected by age and OVX (p < 0.01), and an interaction (p < 0.01) between age and OVX was detected for both of these dependent variables. Overall, OVX at birth did not affect patterns of change in uterine weight or endometrial thickness between birth and PND 60. Both parameters increased (p < 0.01) in a similar way in intact and OVX gilts before PND 60. After PND 60, these values were greater and increased (p < 0.01) more rapidly in intact gilts (Fig. 1). However, both uterine wet weight and endometrial thickness continued to increase in OVX gilts to PND 120. Histology Photomicrographs depicting age-related changes in endometrial histology between birth and PND 120 in intact and OVX gilts are shown in Figure 2. On Day 0, the porcine endometrium had no glands but consisted of a simple columnar epithelium overlying developing stroma. By PND 15, GE was present and tubular glands had begun to coil and penetrate the endometrial stroma. Between PND 15 and PND 120, endometrial folds developed and uterine glands continued to proliferate such that the deep stratum spongiosum became intensely glandular. OVX at birth did not inhibit uterine gland genesis. Western Blotting Results of Western blot analyses of uterine tissue protein from PND 0, 15, and 120 are shown in Figure 3. In tissue

1012 FIG. 2. Representative photomicrographs depicting histogenesis of the porcine endometrium from birth (DO; top) through PND 120 (D120; bottom). Uteri were obtained at hysterectomy on Days 0, 15, 30, 60, 90, and 120 from both intact gilts and gilts OVX at birth. Tissue sections were stained with Harris hematoxylin and eosin. Note the aglandular endometrium on DO. GE is present by PND 15, and glands proliferate throughout the S as animals age. M = myometrium. Bar = 100 ptm.

TARLETON ET AL.

ESTROGEN RECEPTOR IN THE NEONATAL PORCINE UTERUS

1013

of endometrial glands, strong positive nuclear staining was observed in GE. In striking contrast, nuclear ER staining was effectively absent in LE on PND 15, although a few individual LE nuclei did stain weakly for ER protein. Strong nuclear ER staining was observed for GE in all tissues obtained on and after PND 15. Staining of LE nuclei was weak but more regular in tissues from PND 30 and 60 and was moderate but regular in tissues from PND 90 and 120. Stromal cell nuclear staining for ER was weak to moderate on PND 15, moderate on PND 30, and strong thereafter. Positive nuclear immunostaining, indicating presence of ER protein, was observed only when the H222 primary antibody was applied to Pronase-treated tissue sections. No nuclear staining was observed when irrelevant rat IgG was substituted for the H222 primary antibody or when primary antibody was omitted, nor was nuclear staining observed in sections of porcine small intestine when H222 was applied as described above (data not shown). FIG. 3. Identification of uterine ER by Western blot analysis. Total protein (50 pIg) from homogenates of uterine tissue obtained at birth (Day 0), PND 15 (Day 15), and PND 120 (Day 120) was applied to each lane. For each sample, the first (left) lane was incubated with primary rat antihuman ER H222 antibody and secondary antibody (H222+/2nd Ab+), while the second lane received an equivalent amount of irrelevant rat IgG instead of H222 (H222-/2nd Ab+). Relative migration positions for molecular weight markers are indicated at 66, 45, and 31 kDa (left). The approximate migration position for ER protein is indicated by the arrow (right). No ER-specific signal was detected in tissue homogenate from PND 0. A faint, ER-specific signal (H222+/2nd Ab+) was detected at the expected position in tissue homogenate from PND 15, and a strong signal was seen in tissue homogenate from PND 120. This signal was not detected when irrelevant rat IgG was substituted for H222 (H222-/2nd Ab+) or when primary antibody was omitted (Day 120; (H222 = 0/2nd Ab+). Nonspecific binding of AbI and Ab2 was observed for all samples (bands below 66 kDa; see text for details).

homogenate from PND 0, no specific signal of the expected size for ER protein (approximately 66 kDa) was detected. However, a specific positive ER signal of the expected size was observed as a faint band in tissue homogenate from PND 15, and a distinct specific signal of appropriate size was observed in tissue homogenate from PND 120. The ER-specific signal observed in tissue homogenates from PND 15 and 120 was not seen when irrelevant rat IgG was substituted for the H222 antibody or when buffer alone was substituted for primary antibody. In addition to the ERspecific signal, bands indicative of nonspecific binding of secondary reagents were also observed. In tissue homogenate from PND 0, these included major bands at approximately 52.4 kDa and 27.5 kDa. The nonspecific 52.4-kDa band was also observed in homogenates from PND 15 and 120. Immunohistochemistry Figure 4 shows representative photomicrographs depicting age-related changes in nuclear immunostaining patterns for ER protein between birth and PND 120 in endometrium from intact and OVX gilts. Extranuclear background staining was observed in tissue sections from all days and was most pronounced in tissue from PND 0. Patterns of change in nuclear immunostaining for ER protein were not affected by OVX at birth. Age-related changes in ER immunostaining intensity observed for GE, LE, and S are summarized in Figure 5. Nuclear staining was absent in endometrium from PND 0. However, by PND 15, with the appearance

Northern Blotting and ISH Analyses Porcine uterine ER mRNA was detected by Northern blot and ISH analyses using [ot- 32 P]UTP- and [35S]UTPlabeled antisense cRNA probes generated from oER8 cDNA template. Using the [a- 32 P]UTP-labeled antisense ER cRNA probe, a single major transcript of approximately 6.2 kilobases (kb) was detected by Northern blot analysis of total uterine RNA (Fig. 6). Darkfield photomicrographs depicting age-related changes in the presence and distribution of ER mRNA, detected by ISH in uterine tissues obtained from intact and OVX gilts between birth and PND 120, are presented in Figure 7. In addition, nuclear ER immunostaining and ER ISH signal patterns for epithelial and stromal compartments of the endometrium on PND 0 and 15, the interval associated with appearance of endometrial GE, are illustrated in detail in Figure 8. Results of ISH for ER mRNA were complementary to immunohistochemical data for detection of nuclear ER protein in the developing porcine endometrium (Figs. 4 and 8). Consistently, ER mRNA signal intensity was not affected by OVX at birth. Signal above background, observed when labeled sense cRNA probe was substituted for antisense probe in ISH procedures (Fig. 7, Ctl), was not significant in tissue sections from PND 0. Thus, there was no evidence of ER gene expression in uterine tissues obtained at birth (Figs. 7 and 8B). On PND 15, with the appearance of endometrial glands, a strong ISH signal was observed in association with GE, while a moderate signal was observed in endometrial S (Figs. 7 and 8D). Signal above background was negligible in LE for tissues obtained on PND 15 (Fig. 8D). In tissues obtained on and after PND 15, ER mRNA signal intensity remained strong for GE, while signal for S and LE increased such that maximum signal was observed in all cell types by PND 90 to 120 (Fig. 7). DISCUSSION The results indicate clearly that OVX at birth did not affect porcine uterine growth or endometrial morphogenesis between birth and PND 60. Thus, uterine growth and endometrial development during this period are ovary-independent phenomena. In contrast, uterine growth from PND 60 to PND 120 is ovary dependent. Results are consistent with those reported by Wu and Dziuk [24], indicating that OVX of gilts on PND 20 did not affect uterine growth until after PND 60, when uteri of intact controls were larger and

1014 FIG. 4. Immunolocalization of nuclear ER protein in developing porcine endometrium. Photomicrographs depict effects of postnatal age, from birth (PND 0 = DO; top) to PND 120 (D120; bottom), on presence and distribution of nuclear ER protein in endometrium of intact gilts (i; left column) and gilts OVX at birth (o; right column). Nuclear immunostaining intensity, designated as absent, weak, moderate, or strong (see Fig. 5), was evaluated in endometrial S, GE, and LE. No effects of OVX were observed. Nonspecific, extranuclear background staining was evident in all tissues. Nuclear staining indicative of ER protein was not observed in tissue from PND 0 (top left), or in control sections (Ctl; top right; D120 shown) in which irrelevant rat IgG was substituted for the primary H222 antibody. On PND 15 (D1 5i and D1 5o), specific nuclear ER staining was strong in GE and weak to moderate in S, but was absent in LE, with the exception of a few individual LE cells that appeared weakly ER positive (arrowhead). In GE, nuclear ER staining remained strong through PND 120. Consistent, strong nuclear ER staining was observed in endometrial S by PND 60. Regular nuclear ER staining was not observed in LE until PND 60 and later. Bar = 150 pLm.

TARLETON ET AL.

ESTROGEN RECEPTOR IN THE NEONATAL PORCINE UTERUS -

GE

.9 4i 3

--

LE

-0__

1015

S

rrony

v

C

a

C c C

2

IX /O

1

-

J-A

e

Moderate

(n 'iV

01 M Z

0/-/

4-Weak

a Co

z

0

I'

0

' ' ' '