The high concentration of progesterone is harmful for endometrial receptivity and decidualization

Progesterone is required for the establishment and maintenance of mammalian pregnancy and widely used for conservative treatment of luteal phase deficiency in clinics. However, there are limited solid evidences available for the optimal timing and dose of progesterone therapy, especially for the possible adverse effects on implantation and decidualization when progesterone is administrated empirically. In our study, mouse models were used to examine effects of excess progesterone on embryo implantation and decidualization. Our data indicate that excess progesterone is not only harmful for mouse implantation, but also impairs mouse decidualization. In excess progesterone-treated mice, the impaired LIF/STAT3 pathway and dysregulated endoplasmic reticulum stress may lead to the inhibition of embryo implantation and decidualization. It is possible that the decrease in birth weight of excess progesterone-treated mice is due to a compromised embryo implantation and decidualization. Furthermore, excess progesterone compromises in vitro decidualization of human endometrial stromal cells.

unstimulated cycles 13 . More studies should be done to determine the optimal dosage and possible adverse effects on implantation and decidualization when P is used empirically in clinical trial for luteal phase support 18 .
On the other hand, ovarian stimulation program is routinely used to induce multiple ovulation in human in vitro fertilization (IVF), which inevitably leads to ultra-physiological level of P on the day of hCG administration, defined as 'premature luteinization' . Premature luteinization may be caused by multiple follicles, the overdose of gonadotropins and poor ovarian response. The frequency of elevated serum P level varies between 5% and 38% due to the discrepancy on stimulation regimen, method of P assessment and P cut-off level 19 . It is still controversial whether the high P serum level at the end of follicular phase has any adverse impacts on ongoing pregnancy outcome 20 . Several studies show that there is no significant difference on IVF pregnancy outcome between normal and high P serum level (≥0.9 ng/ml) on hCG day 21,22 . However, premature P elevation (≥1.5 ng/ml) in stimulated IVF cycles seems to have a detrimental influence on the pregnancy outcome [23][24][25] . Therefore, further evidences are badly needed to clarify these controversial issues.
In this study, the effects of P at different concentrations on embryo implantation and decidualization were evaluated in mouse models. Effects of excess P on human in vitro decidualization were also examined. Our data suggested that endometrial receptivity and decidualization are compromised by a high level of P in mice, and human in vitro decidualization is also impaired by supplementation of excess P.

Results
Effects of excess P on mouse endometrium receptivity. Leukemia inhibitory factor (LIF) is strongly expressed in the glandular epithelium and required for mouse implantation 26 . The phosphorylation of Stat3, as a receptivity marker on day 4 of pregnancy in mice, is at the downstream of LIF 27 . Therefore, pregnant mice were treated with 1,4 and 8 mg P/mouse on day 3 9:00, compared to control, the level of LIF mRNA expression on day 4 9:00 was inhibited by 4 or 8 mg/mouse P (Fig. 1A). Accordingly, the level of phosphorylated Stat3 in the luminal epithelium was sharply decreased on days 4 of pregnancy after day 3 pregnant mice were treated with 4 mg P /mouse, ( Fig. 1B and C). When pregnant mice are treated with 4 mg P/mouse on days 3 and 4, the number of implantation sites are significantly reduced compared to control at midnight on day 4 of pregnancy ( Fig. 1D and E). To evaluate whether excess P has any negative effect on embryo development, then we examined the blastocyst development at the 14:00 of day 4. The morphology of blastocysts from different dose of P treated mice is normal and similar to vehicle control (Fig. 1F).
Effects of excess P on the expression of PR, estrogen receptor (ER) and P target genes. Because P executes its function through PR, effects of excess P on PR expression were examined. The levels of total PRB and PRAB expression were reduced by 4 or 8 mg/mouse P, not by 1 mg/mouse P ( Fig. 2A and B). Compared to control, the level of PR immunostaining was also decreased by 4 mg P/mouse (Fig. 2C). In mouse uterus, P inhibits estrogen-induced cell proliferation 28 . Therefore, effects of excess P on ER were also examined. Compared to control, ER immunostaining was slightly inhibited in 4 mg P-treated mouse uterus (Fig. 2C).
Ihh and Areg are P target genes and essential for mouse embryo implantation [29][30][31] . When day 3 pregnant mice were treated with different concentrations of P, the levels of both Ihh and Areg were obviously downregulated by 4 or 8 mg/mouse P. Ihh expression was also reduced by 1 mg/mouse P ( Fig. 2D and E).
Effects of excess P on mouse decidualization and birth weight. In order to analyze effects of excess P on mouse decidualization, day 3 pregnant mice were treated with different doses of P daily (from days 3 to7). Compared to control, the weight of implantation sites on day 8 was significantly declined by 1, 4 and 8 mg/mouse P ( Fig. 3A and B). In order to exclude effects of excess P on embryonic development, pseudo-pregnant mice under artificial decidualization were treated with 4 mg/mouse P. Treatment of 4 mg/mouse P caused a significant decrease on the weight of deciduoma ( Fig. 3C and D). To further verify effects of excess P on mouse decidualization, mouse stromal cells under in vitro decidualization were treated with P. Under in vitro decidualization, Dtprp, a marker for mouse decidualization 32 , was significantly induced, while Dtprp expression was significantly suppressed by 4 and 20 μM, but not by 0.8 μM P (Fig. 3E).
Because treatment of excess P during early pregnancy had significant effects on embryo implantation and decidualization, we would like to explore whether these effects during early pregnancy affect the whole pregnant outcome. After day 3 pregnant mice were treated with 1, 4 and 8 mg/mouse P daily for 5 days from days 3 to 7, respectively, the birth weight of P treated mice was significantly reduced by 1 and 8 mg/mouse P, not by 4 mg/mouse P (Fig. 3F). Effects of excess P on endoplasmic reticulum stress. ER stress is shown to be required for mouse decidualization 33 . Ovariectomized mice were treated with different concentrations of P to examine its effects on ER stress. In P-treated uterus, endoplasmic reticulum stress was activated, especially for GRP78/p-eIF2a/ ATF4 pathway (Fig. 5A). P treatment had little effects on IRE1a/XBP1 pathway. Spliced XBP1 (sXBP1) mRNA level didn't show obvious changes following P treatments (Fig. 5B). Previous investigation indicated that GRP78/ IREα/XBP1 pathway is physiologically activated in mouse decidualization 33 . Then in order to analyze if excess P has any effect on ER stress of day 8 uteri, day 3 pregnant mice were treated with 4 mg/mouse P daily from days 3 to 7 for 5 days, we found that GRP78/eIF2a/ATF4 pathway was aberrantly upregulated by excess P in decidua of day 8 (Fig. 5C). Similarly, excess P didn't show any obvious effects on IRE1a/XBP1 pathway because spliced Xbp1 remained unchanged following P treatments (Fig. 5D).

Discussion
The endometrium is a highly hormone responsive tissue. Under the influence of steroid sex hormones, the endometrium undergoes dynamic changes prepared for embryo implantation and decidualization 37 . Estrogen and P are the major mediators for embryo implantation and decidualization 38 . Estrogen is critical for determining (F) Effects of excess P on the birth weight after pregnant mice were treated with oil or P (1,4 and 8 mg/ mouse) daily from days 3 to 7, the weights of individual newborn fetus were counted. The real-time values are normalized to the RPL7 expression level and indicated as the mean ± SEM. n = 3. *P < 0.05. the duration of implantation window. A high level of estrogen will lead to the close of implantation window 39 . Estrogen administration during early gestation can disrupt implantation 40 . The increased estrogenic responses caused from uterine deletion of gp130 or Stat3 also result in implantation failure 41 . Although effects of excess estrogen on receptivity and decidualization have been extensively explored, whether excess P affects receptivity and decidualization remains to be clarified. P is widely used to treat women with threatened abortion for maintaining pregnancy. However, up to now, there is no real consensus on the timing, dose and routes of P administration 42,43 .
LPD ubiquitously exists in IVF with controlled ovarian hyper-stimulation and other pathological conditions. The aetiology and diagnosis criterion of LPD are still not well established in current clinics 15,44 . It is empirical for clinicians to treat all the possible LPD with P for that they may ignore the adverse effects of excessive supplementation of P on pregnancy. Almost all of current evidence suggests that luteal phase support using P can play positive effects on main pregnancy outcome 43 . While studies conducted by Kyrou et al. don't show any improvement of pregnancy rates in women received routine P supplementation 45,46 . In spite of the potential risks for excess P supplementation, little attention has been paid to the possible detrimental impacts of excess P supplementation on pregnancy outcome. Our data showed that excessive P dramatically destroys decidualization process in mice and humans in a dosage-dependent manner, which may give a reference on the clinical use dosage of P. The side-effects of excessive administration of P on the reproduction outcomes should be carefully taken into consideration.
There are accumulating evidences indicating that premature P over 1.5 ng/ml in stimulated IVF cycles seems to have detrimental influences on the pregnancy outcome [23][24][25] . High P level (1.7 ng/ml) before oocyte retrieval is associated with an obvious reduction of endometrial receptivity 47 . The gene expression profile of the endometrium is indeed affected when P level is above 1.5 ng/ml at the end of the follicular phase 48 . Elevated P levels on the day of hCG during the initial fresh cycle are correlated with poor pregnancy in the fresh transfer cycles but not in subsequent frozen-thawed embryo transfer cycles 49 . A previous retrospective study including 4,106 IVF/intracytoplasmic sperm injection cycles reported that patients with P ≥ 2 ng/mL exhibit more high-quality embryos than patients with P < 1 ng/mL 50 , while our data show that excess P has no detrimental effect on embryo development, which suggests that it is endometrium, not the oocyte, that is compromised by P elevation. In a recent study, women in GnRH down-regulation cycles were treated with different dose of P (2.5, 5, 10, 40 mg/ day) and compared to control group female with normal ovulation. The high dose of P (≥5 mg/day) is harmful for endometrium receptivity due to aberrant gene expression in spite of a normal histology 51 . Our data also showed a harmful impact on mouse receptivity and decidualization in excess P group, in addition to impaired decidualization of human endometrial stromal cells in vitro.
In our study, excess P-treated mice exhibit a lower birth weights than control, which is in line with previous studies that neonatal birth weights are lower in fresh blastocyst transfer cycles after controlled ovarian stimulation than in frozen-thawed embryos transfer cycles without ovarian stimulation 52,53 . Another latest retrospective analysis indicates that in fresh embryo transfers cycle, patients with elevated P levels (>2.0 ng/mL) suffer from lower birth weight compared to P levels ≤ 2.0 ng/mL counterparts 54 . The cumulating evidences indicate that it is not advisable to perform embryo transfer for patients with high levels of P in fresh cycle.
Successful implantation requires a synchronous cross-talk between a competent blastocyst and a receptive endometrium 55 . The primary masters that coordinate the endometrium receptivity are estrogen and P 56 . In natural mouse reproduction cycle, uterine epithelial cell proliferation is stimulated by a pre-ovulatory estrogen. P secreted from newly formed corpus luteum enhances uterine stromal cell proliferation. The combined actions of P4 and estrogen is required to establish receptivity for implantation 39,57 . In our study, Lif expression and p-Stat3 immunostaining were inhibited by excess P. LIF expression is also downregulated following ablation of epithelial PR or PRA overexpression in in whole uterus or uterine epithelium 58,59 . However, ER immunostaining was reduced by excess P. These evidences suggest that the downregulation of PR and ER may contribute to the decrease of Lif expression.
As an estrogen-responsive gene, LIF is crucial for uterine receptivity and implantation, for that its deletion leads to implantation failure in mice 26,60 . There is a decline of LIF in endometrial glandular epithelium of women with recurrent implantation failure after IVF 61 . As a direct downstream target of LIF, signal transducer and activator of transcription 3 (STAT3) is phosphorylated during the establishment of uterine receptivity 27 . Implantation is impaired when STAT3 phosphorylation is inhibited or uterine conditional deletion of STAT3 is performed 41,62 . It is reported that endometrial p-STAT3 is reduced in some women with unexplained infertility 63 . In our study, a significant reduced expression of LIF and p-STAT3, resulted from excess P treatment, may underline the deficiency in embryo implantation.
Indian hedgehog (Ihh) has been identified as a target gene of P and is expressed in epithelium, mediating epithelial-mesenchymal interactions in the mouse uterus. Conditional knockout of Ihh in the murine uterus results in infertility for defective embryo implantation and decidualization 64 . Another P regulated gene, amphiregulin (Areg), has also been identified as a receptivity marker for implantation 29 . In our study, Ihh and Areg are conspicuously down-regulated by excess P in mouse uterus on day 4 pregnancy, indicating a compromised endometrium receptivity.
Perturbation of endoplasmic reticulum (ER) protein homeostasis leads to the accumulation of misfolded proteins in its lumen and subsequently causes a stress that is called ER stress, consisting of three pathways: glucose regulated protein 78 (GRP78)/inositol requiring enzyme 1 a (IRE1a)/X-box protein 1(XBP1) signaling pathway, activating transcription factor 6(ATF6) pathway and pancreatic ER kinase (PERK)/eukaryotic translation initiation factor 2a (eIF2a)/activating transcription factor 4 (ATF4) pathway 65 . GRP78/IRE1a/ XBP1 pathway is activated and essential during mouse decidualization. excessive or chronic ER stress is harmful to mouse decidualization 33 . Sustained endoplasmic reticulum stress-induced apoptosis in decidualization may play an ignominious role in early pregnancy loss 66 . The markers of ER stress, GRP78, IRE1a, and spliced XBP1 (sXBP1), are significantly increased in fetal membranes and myometrium after term and preterm labor 67 . Excessive potentiation of uterine ER stress fails to maintain uterine caspase-3 and 7 levels, leading to preterm birth 68 . Our results showed that excess P can promote ER stress by predominantly up-regulating GRP78/eIF2a/ ATF4 pathway in ovariectomized mice. Similarly, GRP78/eIF2a/ATF4 pathway is activated by excess P4 in day 8 pregnant mouse uterus, resulting in a repression of decidualization. In humans, developmentally impaired embryos elicit an anomalous endoplasmic stress response in human decidual cells 69 . PERK/eIF2a and ATF6 signaling pathway is activated in fetal growth restriction 70 . Actually, in present study, mice received excess P deliver a lower birth weight, with an aberrant elevation of GRP78/eIF2a/ATF4 signaling at implantation sites on day 8 pregnancy.
In conclusion, our study indicates that excess P has a detrimental effect on endometrium receptivity and decidualization. Excess P treatment may cause fetal growth restriction through compromising embryo implantation and decidualization in mouse models.

Materials and Methods
Animal Treatments. All animal experiments were approved by Animal Care and Use Committee of South China Agricultural University. All of the experiments were carried out in accordance with the approved guidelines by South China Agricultural University. Adult CD1 mice were housed in a temperature-and light-controlled environment with 14 h light: 10 h dark cycle. Pregnant or pseudo-pregnant female mice (8-10 weeks) were obtained by mating with fertile or vasectomized males of the same strain (day 1 is the day of vaginal plug), respectively. From days 1-4, pregnancy was verified by flushing the embryos from the fallopian tube and uterus, respectively. The implantation sites on day 5 were confirmed by tail intravenous injection of Chicago blue dye (Sigma). Artificial decidualization was performed as previously described 71 .
In order to examine the effects of P on the expression of endometrial receptivity-related genes, pregnant mice were subcutaneously injected with 1, 4, and 8 mg /mouse P (Sigma, dissolved in 100 μl sesame oil) at 9:00 on day 3 of pregnancy. The control mice received 100 μl of sesame oil. Mice were sacrificed at 9:00 on day 4 of pregnancy to collect uteri for further analysis.
For examining the effects of excess P on implantation sites, pregnant mice were subcutaneously injected with 4 mg/mouse P (Sigma, dissolved in 100 μl sesame oil) twice (9:00 on day 3 and 9:00 on day 4). The control mice received 100 μl sesame oil/mouse. Mice were sacrificed at midnight on day 4 to collect uteri for counting implantation sites. Blastocysts were flushed from uterine horns of mice at 14:00 on day 4 of those mice.
To analyze the effects of excess P on the weight of mouse implantation sites on day 8, pregnant mice were subcutaneously injected with 1, 4, and 8 mg P/mouse (Sigma, dissolved in 100 μl sesame oil) daily from 9:00 on day 3 to 9:00 on day 7 for 5 days. The control mice received 100 μl sesame oil/mouse. Mice were sacrificed at 9:00 on day 8 to collect uteri for weighing implantation sites.
To exclude the effects of P on embryonic development, pseudo-pregnant mice induced for artificial decidualization were treated with a daily injection of 4 mg /mouse P (dissolved in 100 μl sesame oil) on days 5, 6 and 7 of pseudo-pregnancy. Uteri were collected and weighed on day 8 of pseudo-pregnancy.
Ovariectomized mice were subcutaneously injected with 1, 4, and 8 mg /mouse P (Sigma, dissolved in 100 μl sesame oil) daily for 3 days. The control mice received 100 μl sesame oil. Mice were sacrificed 24 h after last injections to collect uteri for further analysis.
Immunohistochemistry. Immunohistochemistry was performed as described previously 72  Isolation and treatment of mouse endometrial stromal cells. Primary endometrial stromal cells were enzymatically isolated from day 4 pregnant mice and cultured as described previously 71 . Briefly, mouse uteri were digested with Hanks' balanced salt solution (Sigma) containing 1% trypsin (AMRESCO) and 6 mg/ml dispase (Roche). Luminal epithelial cells were removed after HBSS washing. After the remaining uteri were treated with 0.15 mg/ml collagenase I (Invitrogen), endometrial stromal cells were collected and cultured in DMEM/F12 (Sigma) containing 10% charcoal-treated FBS (Biological Industries, Israel). For inducing in vitro decidualization, primary endometrial stromal cells were treated with 10 nM estradiol-17β and 1 μM P. Under in vitro decidualization, cultured stromal cells were treated with 0.8, 4 and 20 μM P (Sigma), respectively. The highest treatment dose of P has no toxic effect on cell viability. Real-time PCR and detection of spliced XBP1. Real-time PCR was performed as previously described 74 .
Briefly, total RNAs from each sample were isolated using TRIzol reagent kit (Invitrogen), digested with RQ1 deoxyribonuclease I (Promega, Fitchburg, WI) and reverse-transcribed into cDNA with PrimeScript reverse transcriptase reagent kit (TaKaRa). For real time PCR, cDNA was amplified using a SYBR Premix Ex Taq kit (TaKaRa) on the CFX96 Touch ™ Real-Time System (Bio-Rad). Data from real-time PCR were analyzed using the 2 △△Ct method and normalized to Rpl7 or GAPDH expression. XBP1 primers were designed to contain the 26 base pairs which were used to detect the spliced and un-spliced XBP1 mRNA. PCR products were separated on 2.5% agarose gel electrophoresis as described previously 75 . The corresponding primer sequences were provided in Table 1.