Intracrine Androgens Enhance Decidualization and Modulate Expression of Human Endometrial Receptivity Genes.

The expression of AKR1C3 was dynamically regulated across the decidualization time course. Concentrations of AKR1C3 mRNA were increased in ESC treated with decidualization media (DM) for 1, 2 and 4 days with a significant peak in expression detected after 2 and 4 days of treatment ( Fig. 1A , n = 8 p < 0.0001). Western blot analysis revealed a significant increase in total AKR1C3 protein in decidualized ESC compared to control after 4 ( Fig. 1B , n = 8 p < 0.05) and 8 days ( Fig. 1C , n = 8 p < 0.05). Immunofluorescence revealed positive staining for AKR1C3 (red) immunolocalised in the cytoplasm of both control and decidualized ESC treated for 4 days ( Fig. 1D ) although staining was more intense in decidualized cells consistent with the pattern of expression detected by Western blot. A significant increase in concentrations of T was detected in media recovered after 4 and 8 days of decidualization compared with controls ( Fig. 1E , n = 8 p < 0.01)). Metabolism of testosterone to A4 was assessed by TLC and was found to be significantly decreased in decidualized cells further augmenting net production of T ( Supplementary Figure 2 ; p < 0.05). The enzyme SRD5A1 catalyses the conversion of T to the potent endogenous androgen DHT. Interestingly, the expression of SRD5A1 mRNAs were significantly decreased after 1 ( Fig. 2A ; n = 8 p < 0.01) and 2 days ( Fig. 2A ; n = 8 p < 0.05) but unchanged compared to control at 8 days. Western blot analysis revealed expression of SRD5A1 in cell lysates from both control and decidualized cells that was unchanged after 4 days of treatment ( Fig. 2B ) but significantly decreased in decidualized ESC compared to control after 8 days of treatment ( Fig. 2C , n = 7 p < 0.0001). Immunofluorescence staining of ESC treated for 4 days confirmed positive staining for SRD5A1 (red) immunolocalised in the cytoplasm of both control and decidualized ESC ( Fig. 2D ). Although SRD5A1 protein was detected in both control and decidualized ESC, significant concentrations of DHT were only detected in supernatants from decidualized ESC ( Fig. 2E , n = 8 p < 0.0001); concentrations of DHT were significantly decreased in day 8 compared to day 4 decidualized ESC ( Fig. 2E , n = 8 p < 0.001). As decidualization of ESC resulted in biosynthesis of T and DHT we investigated whether these androgens could act as intracrine regulators of decidualization. Markers of decidual transformation were compared in ESC incubated with control media, DM or DM plus the specific AR antagonist flutamide. Cell morphology was assessed in ESC and control cells had an elongated fibroblast-like morphology and decidualized (DEC) ESC exhibited classic rounded, ‘epithelioid’ morphology ( Fig. 3A ). Consistent with inhibition of decidualization, flutamide treated cells (DEC Flut) exhibited limited morphological evidence of decidualization with mostly fibroblast-like morphology and few rounded cells ( Fig. 3A ). The presence of flutamide had a significant impact on expression of both IGFBP1 and prolactin: Concentrations of mRNAs encoded by IGFBP1 were significantly reduced by flutamide treatment after 1 ( Fig. 3B ; n = 6 p < 0.001), 2 (p < 0.0001) and 4 days (p < 0.05). PRL mRNA was significantly decreased at 2 ( Fig. 3C ; n = 6 p < 0.0001) and 4 days ( Fig. 3C ; n = 6 p < 0.05). Mean concentrations of secreted IGFBP1 protein detected in culture media from cells incubated with flutamide were significantly decreased after 4 days ( Fig. 3D ; n = 6 p < 0.01) and 8 days ( Fig. 3D ; n = 6 p < 0.0001) compared to DEC alone. Mean concentrations of secreted Prolactin protein detected in culture media from cells incubated with DM plus flutamide were significantly decreased after 1, 2 and 4 days ( Fig. 3E ; n = 6 p < 0.05, p < 0.01, p < 0.0001) compared to DM alone. Notably, mean concentrations of secreted IGFBP1 and Prolactin were decreased approximately 80% by flutamide treatment after 4 days. Candidate receptivity genes were selected by comparing androgen-regulated genes 4 with genes previously validated as playing an essential role in endometrial receptivity 22 23 . Incubation of ESC with DM resulted in significant time-dependent changes in concentrations of mRNAs and proteins encoded by SPP1 , MAOA and EDNRB : changes induced by addition of flutamide were consistent with a role for androgens in regulation of these 3 genes. Osteopontin/Secreted Phosphoprotein 1 (SPP1) is reported to be an essential mediator of implantation and receptivity in women 24 25 and is reported to be increased in decidualized ESC 26 . SPP1 mRNA was significantly increased in ESC treated to decidualize (DEC) for 4 ( Fig. 4A ; n = 8 p < 0.05) and 8 days ( Fig. 4A ; n = 8 p < 0.001). Concentrations of SPP1 mRNA were lower in cells treated with flutamide and this was significant at the 8-day time point ( Fig. 4A ; n = 8 p < 0.05). SPP1 is a phosphoprotein which is secreted in active form from cells to mediate effects in tissues. Concentrations of SPP1 protein in cell culture supernatants was therefore assessed by ELISA. Interestingly, the concentrations of SPP1 detected in culture media varied widely between patients but were generally lower after 8 days flutamide treatment compared to DEC ( Supplementary Figure 3 ) and undetectable in control supernatants. To account for inter-patient variability, protein concentrations were calculated as fold change relative to DEC which revealed SPP1 secretion was significantly decreased with flutamide treatment at 8 days ( Fig. 4B , n = 8 p < 0.01). Incubation with DM alone significantly increased monoamine oxidase ( MAOA) mRNA concentrations after 1 ( Fig. 4C ; n = 8 p < 0.001), 2 ( Fig. 4C ; n = 8 p < 0.001) and 4 days ( Fig. 4C ; n = 8 p < 0.05) but was not significantly different from controls at 8 days. Interestingly, the impact of flutamide treatment on MAOA mRNA expression was also time-dependent; on days 1–4 addition of flutamide had no effect on MAOA mRNAs compared to DM alone but on day 8 concentrations were significantly increased in ESC treated with flutamide compared to control ( Fig. 4D ; n = 8, p < 0.05). Western blot analysis of MAOA revealed that addition of flutamide had a striking effect on total protein concentrations in cells on day 8 compared to both control ( Fig. 4D , n = 8, p < 0.001) and DM alone ( Fig. 4D , n = 8, p < 0.01). Endothelin receptor B (EDNRB) binds members of the endothelin family of proteins that are reported to regulate endometrial blood flow 27 . The expression of EDNRB was also dynamically regulated during decidualization of ESC. Decidualization led to a significant increase in the concentrations of EDNRB mRNAs in ESC treated to decidualize for 4 ( Fig. 4E , n = 8 p < 0.05) and 8 days ( Fig. 4E , n = 8 p < 0.001). Incubation with flutamide resulted in an apparent (non-significant) decrease in EDNRB mRNA compared to DM alone although concentrations were increased relative to control at 4 ( Fig. 4E , n = 8 p < 0.01) and 8 days ( Fig. 4E , n = 8 p < 0.001). On Western blots EDNRB protein was readily detected in all treatment groups, was significantly increased after 8 days treatment with DM ( Fig. 4F , n = 8 p < 0.05) but this increase was completely abrogated by inclusion of flutamide ( Fig. 4F , n = 8 p < 0.001; DEC vs DEC Flut). Flutamide treatment also decreased EDNRB protein expression after 4 days of treatment compared to DM alone ( Supplementary Figure 4A , n = 4 p < 0.05) however MAOA was not detected by Western blot at this time point ( Supplementary Figure 4B ).

Human endometrial tissue (proliferative phase, n = 20) was obtained from women undergoing surgery for non-malignant gynaecological conditions. None of the women were receiving hormonal therapy or suffering from endometriosis. Cycle phase was determined as previously reported 37 . Written informed consent was obtained from all subjects prior to surgery, and ethical approval was granted by the Lothian Research Ethics Committee (LREC 10/S1402/59). Methods were carried out in accordance with NHS Lothian Tissue Governance guidelines. Primary endometrial stromal cells (ESC) were isolated from proliferative phase endometrium as described previously 37 . Decidualization was induced by addition of decidualization media (DM: RPMI 1640, 2% charcoal-stripped FCS, 0.1 mg/ml 8-Br-cAMP (Sigma B5386), 1 μM progesterone (Tocris, Cat no. 2835). Some cells were incubated with the antiandrogen flutamide for the duration of the culture period (10 μM; Sigma F9397). Control cultures were incubated with RPMI 1640, 2% charcoal-stripped FCS and equivalent volume of vehicle control (DMSO). To assess the time-dependent accumulation of secreted products treatments were maintained for the duration of each time point. ESC were treated for 1, 2, 4 and 8 days as indicated. Isolation of mRNAs, preparation of cDNAs, and analysis by qRTPCR was performed according to standard protocols (11); samples were quantified by standard curve method or by the comparative ΔΔCt method with CYC as internal control. Primers/probes are given in Supplemental Table 1. Western blotting was performed using 50 μg/lane total cell lysates; membranes were probed with rabbit or goat anti-SRD5A1 (Santa-Cruz Biotech; sc-20658 and sc-20396); rabbit anti-AKR1C3 (Abcam, Ab137546 ); rabbit anti-Monoamine Oxidase A (Abcam, ab126751) or rabbit anti-Endothelin B Receptor (Abcam ab129102). SRD5A1 was not detected at the predicted molecular weight (29 kDa) but consistently detected at around 100 kDa at all reducing conditions tested (50–250 mM dithiothreitol; DTT, β-mercaptoethanol and tributylphosphine; TBP). Identity of the protein was confirmed using two different specific antibodies; no bands were detected if the primary antibody was incubated with specific blocking peptide (SRD5A1 (A-18) peptide sc-20396 P) or in controls lacking the primary antibodies. In all cases loading control was goat anti-actin (Santa-Cruz biotech sc-1616; predicted molecular weight 43 kDa). Membranes were incubated with species-specific fluorescent-conjugated secondary antibodies and visualised using the Licor Odyssey system (Licor). Protein bands detected by Western blot were quantified by performing densitometry analysis using ImageJ (NIH.gov). Data were normalised to Actin and expressed as fold-change relative to control treatment. Immunofluorescence was carried out on ESC grown on chamber slides (BD Biosciences) incubated with control media or DM to identify the localisation of the proteins within the cell and as numbers of primary cells were limited only one time point was selected (4 days). In line with expectations, both AKR1C3 and SRD5A1 were localised to the cytoplasm of the primary ESC a finding consistent with expression within the endoplasmic reticulum. Primary antibodies; rabbit anti-SRD5A1 (Santa-Cruz Biotech; sc-20658) and rabbit anti-AKR1C3 (Abcam, Ab137546 ) were incubated overnight at 4 °C. Antigen detection was performed using a Tyramide signal amplification (Perkin Elmer) system followed by SYTOXGreen nuclear counterstain (Life technologies). Images were captured using a LSM 710 Confocal microscope (Zeiss). Cell morphology was assessed by phase-contract imaging using Axiovert 200 inverted microscope (Zeiss) with 5× objective. Testosterone (T), dihydrotestosterone (DHT) and Insulin-like Growth Factor Binding Protein 1 (IGFBP1) were measured in culture supernatants. Media were assayed in duplicate. Assay specificities are described in Supplemental Tables 2–4. Statistical analysis was performed using Graphpad prism. Student’s t test or One-way ANOVA were used to determine significance between treatments in data that were normally distributed. Non-parametric testing was utilised where sample sizes were insufficient to confirm normality of data distribution; Mann-Whitney test or Kruskal-Wallis test was used to assess differences between treatments. Where data were analysed as fold change significance was tested using one sample t test and a theoretical mean of 1. Criterion for significance was p < 0.05. All data are presented as mean ± SEM.

How to cite this article : Gibson, D. A. et al . Intracrine Androgens Enhance Decidualization and Modulate Expression of Human Endometrial Receptivity Genes. Sci. Rep . 6 , 19970; doi: 10.1038/srep19970 (2016).

Endometrial receptivity is a time-sensitive process underpinned by dynamic changes in the transformation of the stromal compartment and coordinated stromal-epithelial cross-talk. To our knowledge we present the first data that demonstrate endometrial stromal cells promote formation of a pro-androgenic microenvironment during decidualization. Critically, we found that changes in androgen metabolism and biosynthesis are time-dependent, suggesting androgens will be most abundant within areas of the tissue at early stages of decidualization. We found that incubation with the selective AR antagonist flutamide had a significant effect on gene expression within primary human ESC over the time course of decidualization. Notably in this study we did not add exogenous T or DHT therefore the impact of flutamide on gene transcription and protein biosynthesis provides evidence of intracrine androgen activity. We noted that blocking intracrine androgens both inhibited and delayed expression of proteins implicated as markers of the decidualization response and endometrial receptivity (Summarised in Fig. 5 ). The human endometrium has been extensively investigated using transcriptional profiling 22 23 . A customised array based on 238 genes (The Endometrial Receptivity Array (ERA) 22 ) has been validated and is in current use in the clinic to optimise management and timing of embryo transfer for women undergoing assisted reproduction, including those with repeated implantation failure. In studies using primary ESC our group has previously identified and validated a set of androgen responsive genes that were expressed in the human endometrium during the normal cycle 4 . In the present study we identified putative androgen-regulated receptivity genes by cross-referencing the ERA gene list 22 with our ‘androgen target gene’ set 4 . Three putative androgen-regulated genes were identified; MAOA , EDNRB and SPP1 . Blocking local androgen action resulted in time-dependent inhibition of SPP1 and EDNRB expression and altered expression of MAOA in decidualized ESC. Decidualization of ESC precedes implantation in the human endometrium and thus the expression of receptivity factors such as SPP1, MAOA and EDNRB may be affected by inadequate decidualization and alterations in androgen signalling. Expression of MAOA is reported to be enhanced in the endometrium during the window of implantation and deficient expression is associated with implantation failure 28 . SPP1 is the only factor identified that is common to all reported endometrial receptivity gene sets 22 23 and is reported to be an essential mediator of implantation and receptivity in humans 24 25 . The novel findings in the present study demonstrate that biosynthesis of androgens within endometrial tissue during the early secretory phase may play an essential role in maintaining SPP1 and EDNRB expression in ESC which could have important implications for the regulation of implantation. Timing of implantation and endometrial receptivity has a major impact on the establishment of a successful pregnancy and on the future health of offspring. It has been reported that the majority of conceptions can be detected between days 22 and 24 of the menstrual cycle 29 and that delayed implantation is associated with increased rates of early pregnancy loss 30 31 . The data in the current study suggest that a deficit in local bioavailable androgens may delay endometrial maturation and might be a previously unrecognised factor contributing to a transcriptional profile that is ‘out of phase’. Insufficient androgen signalling may be detrimental to the regulation of decidualization however exposure to excess androgens may also adversely affect endometrial function. Elevated AR expression has been reported in endometrium of women with polycystic ovarian syndrome (PCOS) 32 33 a complex metabolic disorder associated with ovarian dysfunction, irregular menstrual cycles and in some cases hyperandrogenism. In a recent study, Piltonen et al . reported that in ESC isolated from women with PCOS (fulfilling all Rotterdam criteria; oligo-amenorrhea, PCO and hyperandrogenism) response to decidualization induced by estradiol (E2) and progesterone was significantly reduced, however this was only in a small subset of the patients that were analysed 32 . These results suggest that altered androgen signalling may have a detrimental effect on endometrial function in women, however further studies are needed to elucidate the importance of these effects in complex disorders such as PCOS. Notably, complementary studies in rodents demonstrate that insufficient or excess concentrations of androgens can delay implantation or lead to aberrant gene expression in implantation sites respectively 33 . The widely reported age-related decline in fertility is often considered as an exclusively ovarian phenomenon but there is emerging evidence that reduced androgen bioavailability may result in endometrial ‘senescence’ a phenomenon that has received little attention to date. Concentrations of circulating androgens decline precipitously with age 34 , such that concentrations of adrenal androgens at age 40 are half those at age 21 6 . Changes in the availability of precursor androgens in circulation may impact on intracrine signalling in the endometrium and thus the regulation of implantation. We have previously reported that ESC express 3βHSD the enzyme responsible for converting adrenal androgens within target tissues 2 and we speculate that a reduction in the availability of circulating androgen precursors could limit the development of a full and robust decidualization response which may contribute to reduced fertility in older women. Studies evaluating the effect of DHEA supplementation on reproductive outcomes have produced conflicting results and have been limited by a lack of randomised controlled trials (RCT). A recent meta-analysis pooling data from 8 studies (including two RCTs) investigated the impact of DHEA supplementation on clinical pregnancy rate and number of oocytes retrieved in women with diminished ovarian reserve undergoing in vitro fertilisation or intracytoplasmic sperm injection 35 . The study reported that DHEA supplementation significantly increased clinical pregnancy rate but there was no significant effect on oocyte retrieval, implantation or abortion 35 . These findings are in agreement with the conclusions of the recent Cochrane review on androgen/DHEA supplementation for women undergoing assisted reproduction that reported that DHEA or testosterone supplementation may be associated with improved live birth rates but further evidence from well-designed studies is required before definitive conclusions can be made 36 . In the current study we provide further novel insights into the regulation of decidualization by profiling discrete responses of ESC across a time-course up to 8 days whereas many other studies only consider the response of cells after 6 or more days incubation 15 26 . We demonstrate rapid and sustained induction of AKR1C3 resulting in a significant increase in secretion of T which was maintained throughout the culture period. Notably, concentrations of DHT in media were significantly higher on day 4 than day 8 which may in part reflect reduced expression of SRD5A1. An alternative explanation for the reduced concentrations of DHT would be a switch towards conversion of T to E2 a hypothesis that would be consistent with our recent study in which we reported that E2 biosynthesis increases in a time-dependent manner during decidualization due to an increase in expression of aromatase (CYP19A1 ) by 8 days decidualization 2 . Furthermore, we have demonstrated that local E2 plays an important role in the endometrium by regulating immune and vascular function in early pregnancy 3 . These complementary data suggest perturbations in formation of androgens and the androgen-estrogen balance could have a major impact on the endometrial microenvironment as a result of intracrine androgen regulation of stromal decidualization and paracrine estrogen regulation of immune-vascular function. Taken together these data indicate that T may play a dual role in the endometrium as both a precursor to DHT in early decidualization and as a precursor to estrogens as decidualization progresses and pregnancy is established. The results of this study suggest a new paradigm for understanding control of endometrial remodelling in which local androgens modulate decidualization and promote coordinated development of the endometrial microenvironment. Critically, we describe a novel role for androgens that is both distinct from and complementary to the long-established role of progesterone in promoting post-ovulatory tissue remodelling in women. We demonstrate that androgens regulate early remodelling events by affecting endometrial transcription and altering the expression of decidualization and receptivity markers. These data suggest intra-uterine androgens may be critical for decidualization and thus impact on establishment and maintenance of pregnancy.