The progesterone receptor (PR) has three isoforms, PR-A, PR-B, and PR-C, which have different physiological effects. PR-A may inhibit PR-B-mediated transcription. Parturition requires withdrawal of progesterone (P4). This could occur through decreased P4 concentrations and/or a change in PR isoforms to diminish the effect of P4. We measured mRNA for PR isoforms in rat uterine tissues through late gestation and investigated the effects of antagonists to estrogen (tamoxifen) and P4 (RU-486). Two specific probes were used for ribonuclease protection assays; one (PR-total) measured PR-A, PR-B, and PR-C, and the other recognized only PR-B. PR-total mRNA increased significantly through late gestation, whereas PR-B was unchanged. The ratio of PR-total to PR-B peaked on the day before parturition. Tamoxifen delayed parturition and inhibited the increase in PR-total without affecting PR-B mRNA. RU-486 caused early parturition associated with increased PR-total mRNA, with no change in PR-B. We conclude that there are significant changes in PR isoforms in late-gestation rat uterus. These changes may be regulated by estrogen and P4and may influence the timing of parturition.
in the rat and many other species, progesterone (P4) is responsible for uterine quiescence throughout pregnancy, and P4withdrawal is an important step to initiate parturition (5). Withdrawal of P4 could be accomplished by decreasing the concentrations of P4 or by diminishing the effective concentrations of the specific receptors that mediate the actions of the steroid.
The genomic effects of P4 are mediated by specific receptors that are members of the nuclear receptor superfamily of transcription factors. The progesterone receptor (PR) has at least three isoforms, all originating from the same gene (30). The rat uterus contains all three isoforms: PR-A, PR-B, and PR-C (18). PR-A is a truncated form of PR-B, arising from a different transcription start site and lacking a 164-amino acid sequence of the NH2-terminal-transactivating region of PR-B (12). In general, PR-B is the stronger transactivator, and PR-A is a dominant inhibitor of PR-B and other nuclear receptors (26, 27). PR-C is also an NH2-terminal truncated transcriptional product, but it is much smaller than PR-A (32). It has an intact hormone-binding domain but only the second zinc finger of the DNA-binding domain, and therefore it does not directly interact with DNA. Thus PR-C may act as a selective suppressor of P4 action by binding P4 in the cytosolic fraction to curtail the binding of progesterone to active receptor forms (30, 32). In human breast cancer cells, PR-C was a minor isoform (32). Conversely, Ogle et al. (18) reported that it was the predominant form in decidua basalis in early gestation. In summary, the PR-A and PR-C isoforms may have an inhibitory influence on transcriptional activity of PR-B. Thus a significant change in the ratio of PR isoform expression could regulate the biological activity of P4 and result in functional P4 withdrawal in the absence of changes in maternal serum concentrations or total progesterone-binding activity of the uterus.
PR isoform expression in uteri from late rat gestation has not yet been reported. The PR-A and PR-B isoforms each have a distinct promoter region (12, 14). Uterine PR may be regulated by estrogen and P4. Estrogen increases PR mRNA levels (12, 14,15) and P4 binding (16, 17), whereas P4 decreases the number of its own receptors (22,28). There is no information regarding the effect of estrogen or P4 on the relative quantities of the PR isoforms in the late-pregnant uterus.
Our previous studies detected no significant change in the binding of P4 to rat uterine tissues around the time of parturition (6). We hypothesized that the changing concentrations of estrogen and P4 in maternal rat serum through late rat gestation and around the time of parturition could affect the absolute as well as the relative expression of PR-A, PR-B, or PR-C isoforms in uterine tissues. In this study, our objectives are to measure the concentrations of mRNAs encoding the PR isoforms in rat uterus through late gestation and to determine the effects of the estrogen antagonist tamoxifen (TAM) and the PR antagonist RU-486. To accomplish these objectives, we have developed ribonuclease protection assays with two specific cRNA probes, one that will measure all three PR isoforms (PR-total), and one that measures only PR-B.
MATERIALS AND METHODS
The Institutional Animal Care Committee approved all experiments. Time-mated primigravid Sprague-Dawley rats (∼250 g each) were transferred from Charles River Canada (St. Constant, QC, Canada) to our animal facility on day 12 of pregnancy. Food and water were available ad libitum. A 12:12-h light-dark cycle was used. In this strain of rats, parturition usually occurs early in the afternoon ofday 22. The rats (n = 3–9 at each time point) were killed between 0900 and 1000 on day 13, 15, 16, 18, 19, 21, or 22 of pregnancy, after delivery of the first pup (delivery), or on postpartum day 1. Rats were killed using an intraperitoneal injection of Euthanyl (100 mg/kg body wt). The uteri were removed immediately after euthanasia. Uterine tissues were frozen in liquid nitrogen and stored at −80°C until RNA extraction.
The animals (n = 3–5 at each time point) in the tamoxifen (TAM) group were treated daily with TAM (200 μg per rat, sc; Sigma Chemical, St. Louis, MO) in 0.4 ml of oil from day 15 of pregnancy until they were killed. Control animals received only oil. Animals were killed on day 19, 21, 21.5 (the evening of day 21 of pregnancy), or 22 of pregnancy or during labor (after delivery of the first pup). The animals in the RU-486 group (n = 4–6 at each time point) were injected once with RU-486 (2.5 mg per rat sc) on day 15 of pregnancy and killed at time 0, or 6, 12, or 24 h after treatment, or during labor.
Rat PR-isoform mRNA measurement.
Rat PR isoform probes were generated using RT-PCR. The primer sets are shown in Fig. 1. The PB1/PB2 primer set flanks part of the 5′-untranslated region and the NH2-terminal region of the B form (390 bp). The PT1/PT2 set flanks the hormone-binding domain for all A, B, and C isoforms (320 bp). The reverse transcription products were generated using random primers, and PCR was performed for 30 cycles of 1 min at 95°C, 2 min at 56°C, and 3 min at 72°C. The RT-PCR products with primer sets of PB1/PB2 and PT1/PT2were generated using total RNA (1 μg) from a day 21pregnant rat uterus. The mRNA measured using the probe from the 5′-untranslated region is referred to as PR-B, and the mRNA measured using the probe from the 3′ region is referred to as PR-total.
The amplified DNA (PR-B or PR-total) was cloned using pPCR-Script Amp Cloning kit (Stratagene, La Jolla, CA). Restriction and sequence analyses were performed to confirm the sequence of the inserted PCR product. The PCR-generated rat PR-B and PR-total cDNA clones in pPCR-Script plasmid were linearized using the restriction enzymeNotI to generate an antisense DNA template transcribed by T7 polymerase.
Immediately after euthanasia, the fetal, placental, and fetal membrane tissues were scraped from the uteri with the blade of a scissors. Whole uterine tissues, including the myometrium and endometrium, were homogenized, and total RNA was extracted using TRIzol (GIBCO-BRL, Grand Island, NY). The concentration of RNA was determined by spectrophotometry at A260. PR-total and PR-B mRNA levels in the uterine tissues were measured using ribonuclease protection assays (RPA). The optimal concentrations of total RNA (μg) were determined by increasing concentrations of total RNA (from 5 to 80 μg). The minimal radioactivity of 32P-CTP-labeled probes was determined by increasing the amount of radioactivity (from 0.125 to 2.0 cpm × 106) titrated with 20 μg of total RNA (the optimal concentration) in the RPA. The assay was performed as previously described (6). Briefly, 20 μg of total RNA were hybridized with gel-purified antisense 32P-labeled cRNA PR-total or PR-B probes, and a rat β-actin probe was used as an internal control. Hybridization was performed in 80% formamide and 5-fold concentrated salts containing 200 mM 1,4-piperazinediethanesulfonic acid, 2 M NaCl, and 5 mM EDTA for 18 h at 55°C. After incubation with 0.75 μg ribonuclease A and 300 U ribonuclease T1 (both from Boehringer Mannheim Canada, Laval, QC, Canada) for 30 min at 30°C, protected fragments were analyzed on 6% denaturing polyacrylamide gels. Samples from 2–5 animals at each gestational age were used in this analysis. The gel was exposed to XAR X-ray film (Eastman Kodak, Rochester, NY) for 30 min for β-actin mRNA and for 14 h for PR-total or PR-B mRNA. Autoradiograms were scanned using a high-performance scanner and quantitated by analyzing the density of each blot by means of NIH-imaging analysis software. In this assay, negative controls were performed by hybridization ofday 21 pregnant uterine total RNA (20 μg) with a sense probe and also by hybridization of tRNA (10 μg) with the antisense probe.
As noted in Fig. 2 A, double bands were protected for PR-total and β-actin. This is commonly seen in RPAs that use a combination of ribonucleases that cut at different sites. Depending on the nucleotides adjacent to the protected double-strand fragments, there may be “tails” on the protected fragments from one enzyme that are not present with the other. Hence there may be two protected fragments differing in size by the length of the tail. We have found that the relative amount of RNA in each of the two bands is constant within a given assay. Therefore, we have used only the measurement of the denser band to make our calculations.
Data are presented in graphs as means ± SE. Results were first analyzed by one-way ANOVA (Instat; GraphPad Software, San Diego, CA) to examine changes with advancing gestational age. Post hoc comparisons of the means were performed using the Tukey-Kramer test. Differences between the experimental and control groups over time were sought by use of two-way ANOVA (Prism; GraphPad Software) to detect changes. If any significance occurred, the two-tailed unpaired Student'st-test was performed between two groups at the same time point. Differences were considered to be significant when aP value <0.05 was obtained. If Bartlett's test revealed nonhomogeneity of variance, the corresponding nonparametric test was used.
The PR-total mRNA in the normal pregnant uterus changed significantly through late gestation, with the most significant increase occurring over the 24–36 h before parturition (P < 0.01; Fig. 2). There was no significant change in PR-B mRNA. The ratio of PR-total to PR-B also demonstrated a significant increase on the day of parturition.
Treatment with TAM significantly prolonged the duration of pregnancy by 24 ± 1.2 h (P < 0.01) (6). In the control animals, there was a significant increase in the ratio of PR-total to β-actin mRNA on day 22 of pregnancy (Fig.3 A) similar to that in the normal pregnancy group just noted. Treatment with TAM completely prevented the increase in PR-total mRNA and, on day 22, the uterine tissue concentration of mRNA for PR-total in the TAM group was significantly less than that in the controls. PR-B mRNA levels did not change through late gestation in the control animals and were not influenced by treatment with TAM (Fig. 3 B).
In the animals treated with RU-486, the mean time to delivery of the first fetus was 27.0 ± 1.2 h after treatment (7). None of the control animals underwent spontaneous labor and delivery. There was a significant increase in PR-total mRNA in the animals treated with RU-486 (P < 0.05), with peak levels occurring at the time of delivery (Fig.4 A). The level of PR-B mRNA did not change significantly after RU-486 administration (Fig.4 B).
There was no effect of gestational age or of steroid antagonists on expression of the β-actin gene that was used as an internal control.
These data report the mRNA concentrations of rat uterine PR isoforms through late pregnancy. They demonstrate that the predominance of the PR-B isoform decreases in late rat gestation. Although we cannot distinguish the mRNA for PR-A or PR-C separately, it is clear that their combined mass increases significantly relative to PR-B and that they become the dominant isoforms just before parturition. PR-B and PR-A have similar ligand-binding affinities and similar affinities for DNA binding (2, 4, 32). However, it is clear that the different isoforms have different functions. Using human breast cancer cell lines, Richer et al. (23) determined that, of 94 progesterone-responsive genes, 65 were uniquely regulated by PR-B, 4 uniquely by PR-A, and 25 by both isoforms. This differential effect may be not only gene specific but also tissue specific, because PR-A inhibits transcription of mRNA for gonadotrophin-releasing hormone in both human placenta and pituitary cells, whereas PR-B stimulates transcription in placenta but represses it in pituitary cells (3).
PR-A has been demonstrated to have an inhibitory influence on transcription, mediated by PR-B and other members of the steroid receptor family, including the glucocorticoid androgen and mineralocorticoid receptors (27). Furthermore, Bennett and colleagues [Pieber et al., (20) and (21)] have overexpressed PR-A in amnion or myometrial cells from pregnant women and demonstrated strong inhibition of PR-B-mediated transcription. The precise mechanism of this inhibitory influence is uncertain. However, this effect does not appear to depend on DNA binding. It has been proposed that PR-A may compete with PR-B (or other nuclear receptors) for binding to a limiting co-factor required for transcriptional activation (27). An alternative explanation has been provided by Hovland et al. (10), who described an inhibitory function domain in the NH2-terminal region of PR-A and PR-B. However, in PR-B, the inhibitory function of this domain is constrained by the activator function in the upstream NH2-terminal region. Only PR-A can independently and strongly repress the transcription of the target genes, because it lacks this NH2-terminal upstream segment. PR-C lacks one of the zinc fingers necessary for association with DNA response elements and may influence the biological effects of P4 actions by binding the hormone or by dimerizing with one of the two other PR isoforms (33). In human breast cell lines, PR-C was a minor isoform (32). However, it is known from measuring protein levels in early pregnant rat decidua that PR-C is the dominant form (19). Our data suggest that PR-A and/or PR-C becomes increasingly dominant before delivery. The increasing dominance of these isoforms that are inhibitory to PR-B may be an important aspect of preparation of the gravid uterus for parturition, in that it may effectively diminish the ability of P4 to maintain uterine quiescence.
In most animal models, including the rat, parturition proceeds normally only when there is a withdrawal from the influence of P4(5). This usually is accomplished by a decrease in the production rate of P4, with a corresponding decline in maternal serum P4 concentrations. Maternal serum P4 concentrations in the rat decline in late gestation and reach their lowest levels by day 22 (6). A switch in PR isoform expression would augment the P4withdrawal. This mechanism could be particularly important in primate species in which there appears to be no significant decline in maternal P4 concentrations around the time of parturition (25).
As with our previous studies, we have left the uterine wall intact. There are important paracrine interactions between the endometrium and the myometrium, and these interactions may influence the timing of parturition. By not separating the tissues, we have avoided potential artifacts due to tissue trauma and also have avoided possible misinterpretation of the data resulting from incomplete separation of the two tissue types. It will be of interest to determine the tissue and cellular localization of the changes in PR isoform expression by use of in situ hybridization or other means.
The probe for PR-total measures the mRNA for all PR isoforms. Because the radiolabeled probe may hybridize with different efficiency to the mRNA for the different isoforms, we have avoided attempting to calculate PR-A plus PR-C by simply subtracting PR-B from the PR-total.
Administration of the estrogen antagonist TAM resulted in significant delay in parturition. Although TAM may have mixed agonistic and antagonistic properties, it acts as a strong antagonist in the high estrogen milieu of late pregnancy (11). We have demonstrated that TAM administration does not significantly change maternal serum P4 concentrations (6). It is possible that the effect of TAM in prolonging gestation is mediated at least in part by a change in the relative expression of PR isoforms. In the presence of TAM, the increase in PR-A/PR-C noted before parturition in control animals failed to occur, suggesting that it is dependent on estrogen. The 5′-untranslated and 5′-flanking regions of the rat PR-A gene contain five imperfect palindromic estrogen response elements (ERE) and four widely spaced ERE half-sites (15). Estrogen upregulates PR in breast cancer cells (8, 31), rat uterus (13), and sheep endometrial and myometrial cells (29). The estrogen-induced isoform ratio appears to be tissue specific. Estrogen stimulated mRNA for both PR-A (10-fold) and PR-B (5-fold) in human endometrial stromal cells (24), but it only increased PR-B in T47D cells (8). Estrogen increased PR-A protein levels fivefold greater than PR-B in rhesus monkey endometrium (1). These findings support our data and suggest that the PR-A isoform may account for most of the estrogen-dependent increase in PR mRNA that we demonstrated before parturition. It is possible that the increase in estrogen concentrations in late rat gestation causes an increase in PR-A expression that adds to the P4 withdrawal leading to parturition. Inhibition of this increased expression of PR-A may be one mechanism of action of TAM in delaying the time of parturition.
The administration of the PR antagonist RU-486 on day 15 of gestation caused markedly premature parturition approximately a day after administration. The present data demonstrate that, in this model of preterm delivery, changes in PR isoform expression appear to be similar to those observed at normal parturition. Treatment with RU-486 significantly increased the concentration of PR-total mRNA but not of PR-B, suggesting that P4 has a tonic inhibitory effect on PR-A and/or PR-C expression and that this inhibition is removed by the P4 antagonist. Interestingly, the transdominant inhibitory effects of PR-A on PR-B are also induced by antiprogestins (27), and this may contribute to the biological sequelae of RU-486 treatment.
The effects of P4 appear also to be tissue specific and isoform specific. For example, within human endometrium, P4increased the PR-A and PR-B mRNA 2- to 10-fold in the stromal cells but reduced them in glandular epithelial cells (24). In contrast, Ogle et al. (19) reported that RU-486 treatment reduced decidual PR-A and PR-B but not PR-C protein levels in day 9 pregnant ovariectomized rats. They also reported that P4 replacement after ovariectomy had no effect on PR mRNA on day 10 but caused a significant increase on day 14. The differences between their findings and ours may relate to the difference in tissue types as well as the difference in gestational age. It also is possible that translation and posttranslational processing are subject to different control mechanisms from those regulating transcription.
Our experiments in late rat gestation suggest that antagonists to estrogen and progesterone receptors regulate transcription of PR-A and PR-C but not PR-B. Little is known regarding the regulation of expression of PR isoforms. Although the three isoforms transcribe from the same gene, at least two functionally distinct promoters have been found on this gene (14). Consensus progesterone response elements have not been identified. Whereas estrogen may influence gene transcription directly through the weak ERE sites in the promoters, it is likely that there are more important indirect or nongenomic mechanisms of steroid hormone action. In addition to transcriptional regulation, these mechanisms may influence mRNA stability, translational activity, or posttranslational modifications of the PR gene products.
These experiments were designed to study the influence of sex steroid antagonists on relative expression of mRNA for PR isoforms. We acknowledge that it is highly likely that estrogen and P4influence transcription of many other genes that play an important role in the mechanisms of parturition. Similarly, the steroids, and potentially their antagonists, may have important nongenomic effects, including interaction with membrane receptors such as oxytocin receptors, which may influence uterine contractility (9). It is not clear how important the changes in PR isoform expression are in relation to the complex mechanisms that regulate the timing of parturition. However, this physiological phenomenon does offer an explanation that could underlie a mechanism for significant “P4 withdrawal” in the absence of changes in P4 concentrations or total P4 binding to uterine tissues. It will be important to demonstrate corresponding changes in protein expression using Western blot or immunohistochemical techniques if isoform-specific antibodies become available.
In summary, these data have provided information regarding the uterine concentrations of mRNAs for PR isoforms through late rat gestation and around the time of parturition. The results strongly support a role for estrogen and P4 in the regulation of PR isoform expression. Our findings provide another mechanism that may contribute to the withdrawal of P4, which is essential in the mechanism of parturition in many species. This mechanism may be particularly important in primates, where decreases in maternal serum P4concentrations do not occur at the time of parturition. Further investigation of the changes in PR isoforms at translational and posttranslational levels may provide more information regarding the roles of PR isoforms in regulation of the timing of parturition.
This work was supported by a grant from the Canadian Institutes of Health Research (to B. F. Mitchell).
Address for reprint requests and other correspondence: B. F. Mitchell, Perinatal Research Centre, Dept. of Obstetrics and Gynecology, Rm. 227 HMRC, Univ. of Alberta, Edmonton, Alberta, Canada T6G 2S2 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
August 6, 2002;10.1152/ajpendo.00116.2002
- Copyright © 2002 the American Physiological Society