Alk2 cKO females are sterile and show delayed blastocyst implantation and defective decidualization

We next investigated how ablation of ALK2 in the uterus affects female fertility by performing a 6-month breeding trial in which sexually mature control and cKO females (n=10 for each group) were mated with wild type males of proven fertility. Alk2 cKO female mice are sterile, while control littermates display normal fertility and fecundity (9.1±0.2 pups/litter and 1.0±0.2 litter/month), comparable to wild type (8.1±0.3 pups/litter and 1.2±0.1 litter/month) and Pgr^cre/+ females (7.8±0.1 pups/litter and 1.1±0.4 litter/month) [17]. The presence of vaginal plugs in Alk2 cKO females showed that the infertility was not due to abnormal mating behavior. We then performed timed mating experiments in which control and cKO females were mated with wild type males and sacrificed at sequential time points during pregnancy. The attachment of the embryo to the uterine epithelium causes a local increase in vasculature permeability, and implantation sites can be visualized at an early stage by injecting Chicago blue dye intra venous (i.v.). On the day of attachment (i.e., day 4 of pregnancy), we did not observe any difference in the number of sites of attachment between control and cKO females (data not shown). The presence of a normal number of attached embryos shows that the hypothalamic–pituitary–gonadal axis is intact, and ovulation is normal. This suggests that the infertility in Alk2 cKO females is likely caused by uterine defects.

Upon embryo attachment, the embryo starts invading into the uterine stroma (day 5) and stimulates the differentiation of the stromal cells into decidual tissue, around each implanted embryo. When we inspected the implantation sites at day 5.5 post coitum (5.5 dpc) (Figure 2A–B), we observed that embryos in Alk2 cKO mice were attached to the luminal epithelium, but the epithelial layer was still intact because the invasion into the stroma had not occurred (Figure 2E–F). Moreover, the uteri of Alk2 cKO mice still showed similar numbers of implantation sites compared to controls (Figure 2D), but their size and weight were significantly reduced compared to controls (Figure 2C). To further characterize this defect, we collected uteri from pregnant females at 6.5 and 7.5 dpc and examined the histology of the implantation sites (Figure 3). While the embryos eventually invade into the stroma by 6.5 dpc in Alk2 cKO mice, the process is delayed compared to controls (Figure 3G,L). In addition, embryos that implant into the Alk2 cKO uteri are consistently smaller compared to controls (Figure 3G–I, L–N) and start to be reabsorbed at 7.5 dpc (Figure 3N). The implantation sites are also smaller in cKO mice compared to controls at all time points observed beginning at 5.5 dpc.

Open in a separate window

Figure 2

Delayed blastocyst invasion in cKO mice at 5.5 dpc.

A–B) Gross morphology of the uterus of Alk2 control (A) and cKO (B) mice. Black arrows indicate the implantation sites in the Alk2 cKO uterus. C) Wet weight of the implantation sites in milligrams (mg.) recovered from pregnant control (black column) and cKO (red column) mice collected at 5.5 dpc. The size of the implantation sites in Alk2 cKO mice is significantly smaller (*p<0.05). N=5. D) Number of implantation sites in control (black column) and cKO mice (red column). E–F) CDH1 (epithelial marker) staining of implantation sites collected from control (E) and cKO (F) mice. Brown color indicates positive staining. Tissue is counterstained with hematoxylin. G–H) PTGS2 staining of implantation sites collected from control (G) and cKO (H) mice. Brown color indicates positive PTGS2 signal. M=mesometrial side, AM=antimesometrial side. Arrowhead indicates the embryos. Data are mean ± SEM.

Open in a separate window

Figure 3

Defective implantation in Alk2 cKO mice.

Implantation sites collected from control (A–C) and cKO mice (D–F) at sequential time points during pregnancy and stained with H&E.

To further study implantation in Alk2 cKO mice, we analyzed prostaglandin-endoperoxide synthase 2 (PTGS2) protein levels and localization by immunohistochemistry. PTGS2 shows a dynamic localization that precisely correlates with different phases of the implantation process [18]. In particular, PTGS2 localizes to luminal epithelial and subepithelial stromal cells at the antimesometrial pole of the implantation site at the moment of embryo attachment (~4.0 dpc), and at 5.5 dpc, PTGS2 is produced in decidualizing cells at the mesometrial pole. In control mice, PTGS2 correctly localized at the mesometrial pole at 5.5 dpc. In contrast, in Alk2 cKO mice, at 5.5 dpc PTGS2 displayed a pattern resembling the moment of attachment (Figure 2G–H), confirming that Alk2 cKO mice have a delay in the implantation process.

These results show that ALK2 is dispensable for the attachment of the embryo to the uterine epithelium, but is required at a later step during implantation. Ablation of BMP2 has also been shown to disrupt the implantation process soon after the embryo attachment, although the differences in the phenotypes observed between Alk2 and Bmp2 cKO mice suggest that BMP2-ALK2 is not the only ligand-receptor pair involved in the regulation of this stage of pregnancy. In fact, the presence of other BMP signaling components in the pregnant mouse uterus suggests that other BMP ligand(s) and receptor(s) can compensate with various degrees of efficiency for the loss of one component of the pathway.

Ablation of ALK2 compromises uterine decidualization

Invasion of an embryo into the uterine wall stimulates stromal cells to undergo a functional and morphological transformation called decidualization. During this process, stromal cells of the uterus extensively proliferate and differentiate to create a permissive environment for the implanting embryo. Because Alk2 cKO mice show smaller implantation sites, we decided to further investigate how ablation of ALK2 affects the proliferation and differentiation of the uterus when a decidualizing stimulus is given. The uterus is receptive only for a limited period of time during which the sex hormones prepare it to adequately respond to the attachment of the embryo. Because sex hormones are produced by the ovaries in cyclic waves, we ovariectomized and injected mice with sex hormones to artificially induce receptivity and exogenously control and synchronize the cycle of the mice in our experiments. We mimicked the implantation of the embryo by injecting one horn with oil. Consistent with the reduced size of implantation sites during natural pregnancy, we observed that Alk2 cKO females undergo a significantly reduced increase in uterine horn weight one day after oil injection compared to control littermates (Figure 4A–C). At five days, the difference in size of the stimulated to untreated horn of Alk2 cKO mice to control mice was even greater (Figure S2A–C). These results confirm that the uteri of cKO mice are not able to decidualize in response to the attachment of the embryo.

Open in a separate window

Figure 4

Alk2 is required during uterine stromal decidualization.

Gross morphology of the uteri of Alk2 control (A) and cKO (B) mice collected one day after one horn (right) was injected with oil to stimulate the stromal cells to undergo decidualization. The other horn (left) was untreated and used as a control. C) Ratio between the wet weight of oil-injected and control horn collected from Alk2 control (black column) and cKO (red column) mice. The ratio of Alk2 cKO mice is significantly smaller than in controls (*p<0.05) indicating that a decidualization defect is occurring in the former. N=5 per genotype. D–E) Proliferation of stromal cells during decidualization visualized through Mki67 staining on sections of uteri collected from Alk2 control (D) and cKO (E) mice one day after the decidualization was induced. Brown color indicates positive Mki67 staining, thus, proliferating cells. Cell nuclei are stained in blue (hematoxylin). F) Expression of some of the main markers of the S phase of the cell cycle (Ccne1, Ccne2, Cdk2 and Cd25) in the uteri of Alk2 control (black columns) and cKO (red columns) mice, one day after the decidualization was induced, quantified by qPCR analysis. Alk2 cKO mice show a significant lower expression of all of these genes compared to controls (** p<0.01). N=5. G,K) Epithelial proliferation in Alk2 cKO mice. Sections of uterine tissues obtained from implantation sites of control (G) and cKO mice (K) were stained for the proliferation marker Mki67. A striking increase in the amount of proliferating cells was observed in cKO mice compared to controls. E=epithelium, S=stroma. H–I) Differentiation of stromal cells during decidualization visualized through the evaluation of endogenous alkaline phosphatase (ALP) activity. Black color indicates positive ALP staining. J) Expression of differentiation markers (Prl8a2 and Prl3c1) in decidualizing uteri of Alk2 control (black) and cKO (red) mice, evaluated by qPCR. Alk2 cKO mice show a significant lower expression of both genes (** p<0.01). N=5. Data are means ± SEM.

We next investigated the ability of stromal cells to proliferate and differentiate in response to the decidual stimulus. Both control and cKO mice show extensive stromal proliferation in the uterus (detected by MKi67 staining), but the amount of cells in the stromal compartment of cKO mice was reduced compared to controls, suggesting an impairment of the process (Figure 4D–E). Moreover, cKO mice showed a significantly reduced expression of several molecular markers of the S phase of the cell cycle (Ccne1, Ccne2, Cdk2, and Cdc25) (Figure 4F). To investigate whether the differentiation of stromal cells is also impaired by the ablation of ALK2, we evaluated alkaline phosphatase activity (ALP) in stromal cells one day after decidualization was induced. While control mice showed a strong purple signal in the uterine stromal compartment (indicating the presence of alkaline phosphatase activity, and thus differentiation), we observed a lack of differentiating cells in the uteri of cKO females (Figure 4I–J). In addition, qPCR analysis of the two differentiation markers Prl8a2 and Prl3c1 confirmed a significant impairment of differentiation at a molecular level (Figure 4K). These data demonstrate that signaling through ALK2 is required during decidualization to achieve stromal cell proliferation and differentiation.

Although stromal cell proliferation and differentiation was reduced, Ki67 staining also revealed a pronounced proliferative activity in the epithelium of cKO mice that is not present in control mice (Figure 4G–H). E2 stimulates the proliferation of uterine epithelial cells in preparation for implantation, and this proliferative action is inhibited by P4 acting through its receptor, PGR [16]. These observations led us to hypothesize that ALK2 acts to regulate stromal cell proliferation and repress epithelial cell proliferation by regulating the expression and/or activity of PGR. However, when we assayed the levels of PGR in the uterus one day after the induction of decidualization, we did not observe any difference in the total amount nor cellular localization of PGR protein in control and cKO mice, with PGR mainly localized in the nuclei of stromal cells (Figure S4A–D).

Microarray analysis of Alk2 cKO uteri indicates that Cebpb is regulated by BMPs during decidualization

To investigate the downstream pathways controlled by BMPs during the decidualization process, we performed microarray analysis on control and cKO uteri collected one day after artificial induction of decidualization. We found that the expression of 909 unique genes is significantly affected by the ablation of ALK2 (p<0.01, fold change <0.7, >1.3) (Figure 5A). We used the Database for Annotation, Visualization and Integrated Discovery (DAVID) to identify the main pathways controlled by BMPs during decidualization. During decidualization, stromal cells undergo a wave of proliferation and differentiation that lead to the formation of a specialized tissue (the decidua) able to sustain the growth of the implanted embryo. Our data show that without ALK2, stromal cells cannot respond to a decidual stimulus by normally proliferating and differentiating. Thus, the significant alteration in the nucleotide synthesis pathways and DNA metabolic process that we observe in Alk2 cKO mice is likely a consequence of the defect in proliferation; moreover, because there is a profound remodeling of the vasculature during decidualization that is tightly regulated by immune system factors, it was not surprising to see an effect of ALK2 ablation on pathways involved in vasculature development and immune response (Table 1).

Open in a separate window

Figure 5

Microarray analysis of Alk2 control and cKO uteri at day one of decidualization.

A) Heat map of genes upregulated (yellow) and downregulated (blue) in Alk2 cKO (right) mice compared to controls (left). Each column represents the expression data obtained from a pool of RNA collected from 3 mice with the same genotype. Three samples of pooled RNA per each genotype were submitted for microarray analysis. B) Expression of Cebpb in the uteri of Alk2 control (black column) and cKO (redcolumn) mice collected one day after the induction of decidualization, quantified by qPCR analysis. Alk2 cKO mice show a significantly lower expression of Cebpb compared to controls (* p<0.05). N=5. C–D) CEBPB staining on sections of Alk2 control (C) and cKO (D) decidualized uteri collected one day after oil injection. Brown color indicates positive CEBPB staining. Cell nuclei are stained in blue (hematoxylin). Data are means ± SEM.

Another Gene Ontology (GO) term that we found to be significantly affected in the array analysis comprises genes involved in the regulation of in utero development (Table 2). Among these genes is CCAAT/enhancer-binding protein beta (Cebpb), which has been shown to regulate the proliferation and differentiation of stromal cells during decidualization. Similar to Alk2 cKO females, mice with a null mutation in Cebpb are not able to decidualize [19]. We used qPCR and immunohistochemistry to confirm that the expression of Cebpb mRNA and CEBPB protein levels are significantly reduced in Alk2 cKO mice compared to controls during decidualization (Figure 5B–D). Although both ALK2 and CEBPB are required for stromal cell proliferation during decidualization, the two factors seem to regulate different phases of the cell cycle; while Cebpb cKO mice show a disruption of expression of factors controlling the G2 to M phase transition of the cell cycle, Alk2 cKO mice display an alteration at the S phase. These data suggest that BMP signaling acts upstream to regulate the expression of the transcription factor Cebpb, and this regulatory mechanism is required for the progression of the cell cycle in uterine stromal cells during decidualization.

ALK2 is required for decidualization in human endometrial stromal cells (hESC) and controls the expression of CEBPB and PGR

In many species, endometrial stromal cell decidualization is induced upon implantation of the embryo. In humans, this process takes place during the mid-secretory phase of each menstrual cycle, independently of pregnancy. Similar to the mouse, decidualization in humans is a process of dramatic transformation of endometrial stromal fibroblasts into secretory, epithelioid-like decidual cells. To study decidualization in humans, this process can be recapitulated in primary cultures of human endometrial stromal cells (hESC) [20]. Treatment of hESC with a cyclic AMP analog triggers the expression of decidual marker genes like prolactin (PRL) and insulin-like growth factor binding protein 1 (IGFBP1) within hours and, via activation of protein kinase A (PKA), sensitizes hESC to progesterone action. When hESC are treated with both cAMP and P4, the expression of decidual markers is enhanced and the decidual phenotype is maintained for a longer period. We utilized hESC to investigate the role of ALK2 during decidualization in human. hESC were first treated with either non-targeting (NT) small interfering RNA (siRNA) or siRNA targeted to ALK2 (ALK2 knock down, KD), then stimulated to decidualize with a hormonal cocktail containing estradiol, medroxyprogesterone acetate (MPA) and cAMP (EPC). After 3 days of treatment, we used qPCR to measure the expression of IGFBP1 and PRL. Both decidualization markers show a significant reduction in ALK2 KD cells compared to NT-treated cells (Figure 6A–B), indicating that ALK2 is required during decidualization both in mouse and human. Upon differentiation of the stromal compartment, many genes first expressed in the epithelium, are induced in decidual cells. CEBPB has been shown to be required for human and murine decidualization and the physical interaction between CEBPB and PGR has been demonstrated using an in vitro system [21]–[23]. Because we found Cebpb to be downregulated during decidualization in the mouse, we were interested in testing whether BMPs control the expression of CEBPB in human endometrial stromal cells as well. Both CEBPB mRNA and protein level were significantly lower in ALK2 KD cells compared to NT-treated cells, indicating that the involvement of the BMP-CEBPB pathway in the regulation of decidualization is conserved in mice and humans (Figure 6C,E).

Open in a separate window

Figure 6

Alk2 is required during the decidualization of human endometrial stromal cells.

Expression of decidual markers, IGFBP1 (A) and PRL (B) in human endometrial stromal cells (hESC) cultured in growth medium (blue columns), treated with a hormonal cocktail containing estradiol, medroxyprogesterone acetate (MPA) and cAMP (EPC, green columns), treated with EPC after transfection of non targeting siRNA (red columns) or siRNA targeted to ALK2 (purple columns). Day 0 and day 3 refer to EPC treatment; the siRNA transfections were performed at day −1 and −2. (* p<0.05, ** p<0.01). N=3. C–D) Expression of CEBPB (C) and PGR (D) at day 0 and after 3 days of EPC treatment in cells treated with non targeting (NT) siRNA (black column) and siRNA targeted to ALK2 (ALK2 siRNA, red column) (* p<0.05). N=3. E) CEBPB and PGR (isoform A and B) protein levels in NT siRNA-treated and ALK2 siRNA-treated hESC, quantified by western blotting at day 0 and day 3 of EPC treatment. β-ACTIN was used as loading control. F–H) Expression of HAND2 (F), WNT4 (G) and FOXO1 (H) at day 0 and day 3 of EPC treatment in NT siRNA-treated (black column) and ALK2 siRNA-treated (red column) hESC (* p<0.05). N=3. Experiments were performed on cells from 3 patients. Data are means ± SEM.

Progesterone acts through its cognate receptor, the nuclear progesterone receptor (PGR), to regulate implantation and decidualization via the activation of several downstream pathways [24]. Although Alk2 cKO mice show a desensitization to progesterone, neither the PGR mRNA nor protein levels show a significant alteration in the cKO mice. It has been suggested that BMPs are required for the expression of FK-506-binding proteins (FKBPs) that, in turn, control the activity of PGR by binding and maintaining the receptor in a functional state. This hypothesis is confirmed by the observation that Fkbp4 and Fkbp5 are downregulated in Alk2 cKO mice during decidualization (Figure S3A–B). To investigate whether the same mechanism was involved in the control of decidualization in hESC, we first measured the expression and translation of PGR in NT-treated and ALK2 KD cells. In contrast to the mouse, we found that ablation of ALK2 in human cells causes a downregulation of progesterone receptor at both the mRNA and protein levels. Because the levels of PGR are reduced in ALK2 KD cells even before decidualization was induced (day 0), the downregulation of PGR is not a consequence of the lack of decidualization but is a direct effect of the ablation of ALK2. To confirm the hypothesis that ALK2 is required for functional PGR activity, we measured the expression of known PGR targets that are normally activated by PGR during decidualization. The expression of all the PGR targets that we quantified by qPCR [i.e., wingless-type MMTV integration site family, member 4 (WNT4), forkhead box O1 (FOXO1), and heart and neural crest derivatives expressed transcript 2 (HAND2)] is significantly reduced in ALK2 KD cells (Figure 6F–H). These findings show that ALK2 is required during the decidualization of hESC, and the mechanism through which it acts likely involves the regulation of CEBPB and PGR. The regulation of CEBPB, but not PGR levels, by BMPs during decidualization, is conserved between mice and humans.

Hormone treatments

Physiological concentrations of E2 and P4 were given exogenously to mice to recapitulate the hormonal levels occurring during early pregnancy as previously described [16]. Briefly, 6–8 week old mice were ovariectomized via a dorsal incision under 2,2-tribromoethanol (2.5% Avertin) anesthesia and left untreated for 2 weeks to achieve complete withdrawal of endogenous ovarian hormones; then, mice were “primed” with daily subcutaneous (s.c.) injections of 100 ng of 17β-estradiol (E2) for 2 days. After 2 days of rest, mice were injected for 3 days with 1 mg of progesterone (P4) daily and on the fourth day with 1 mg of P4 and 50 ng of E2. Mice were sacrificed 6 hours after the last injection.

The artificial induction of decidualization has been previously described [26]. Briefly, control and cKO mice were ovariectomized and, after 2 weeks, treated with s.c. injections of 100 ng of E2 once a day for 3 days. After 2 days rest, mice were injected daily with P4 (1 mg, s.c.) and E2 (6.7 ng, s.c.) once a day for 3 days. One uterine horn was then treated with a decidual stimulus by injecting 200 µl of sesame oil into the lumen 6 hour after the last hormone injection. The contralateral horn was not injected and served as a control. The next day, mice were given a final injection of P4 (1 mg, s.c.) and E2 (6.7 ng, s.c.) and euthanized 6 hours later; wet weight of the traumatized and control horns was recorded. Another group of mice was injected for 5 days after the oil injection and sacrificed 6 hours after the last injection. Upon euthanasia, the uterine tissue was collected and either fixed in 4% paraformaldehyde (PFA) or frozen for RNA extraction.

Epithelial-stromal separation

To separate the epithelial and stromal components of the uterine tissue, uteri were collected, cut in 1–2 mm pieces and incubated in Hank's balanced salt solution (HBSS, Invitrogen) +1% trypsin (Sigma) for 3–5 hour at room temperature. Afterwards, tissues were washed in PBS and luminal epithelium was mechanically separated from the stromal compartment under a dissecting microscope using dissecting forceps and a mouth pipette. To ensure adequate purity of the isolated epithelia and stromal cell populations, the expression of epithelial (cadherin1, Cdh1) and stromal (vimentin, Vim) markers was evaluated by qPCR (data not shown).

Fertility analysis

To examine the fertility of female mice, wild type (WT), Pgr^cre/+, control (Alk2^flox/flox), and Alk2 cKO females were mated independently with WT fertile male mice for a 6-month period (n=10 per genotype). Cages were monitored daily, and the numbers of litters and pups were recorded.

To investigate the different phases of early pregnancy we performed timed mating experiments: adult female mice were mated to fertile wild type males and the presence of vaginal plugs was monitored every morning. The morning of the plug was designated as 0.5 day post coitus (dpc); to investigate embryo attachment, 4.5 dpc pregnant animals were injected intravenously (i.v.) in the tail vein [49] with Chicago blue dye and sacrificed 10 minutes after the injection.

Histological and immunohistochemical analysis

Uteri were dissected and fixed in 4% PFA for histology. Tissue processing and embedding in paraffin were performed in the Baylor College of Medicine Department of Pathology Core Laboratory. Paraffin sections (5 µm) were deparaffinized, rehydrated and boiled in citrate buffer for antigen retrieval. Blocking of non-specific signal was performed by incubating the sections with 10% normal goat serum (NGS) or 3% BSA in TBS-T for 1–3 hours at room temperature. Tissue sections were incubated overnight at 4°C with primary antibodies specific to ALK2 (1200, LifeSpan Biosciences, Inc., LS-B6835), CDH1 (1400, Cell Signaling, 24E10), PTGS2 (11000, Thermo Scientific, RB-9072), MKI67 (11000, BD BioScience, 550609), CEBPB (1200, Santa Cruz Biotechnology, Inc. sc-150), PGR (1500, DAKO, IR068) or HAND2 (1100, R&D Systems, AF3876). After washing with TBS-T, sections were either incubated with biotin-labeled (CDH1, PTGS2, MKI67, CEBPB, PGR, HAND2) or fluorescent-labeled (ALK2, PGR) secondary antibody for 1 hour at room temperature. Sections treated for immunohistochemistry were then incubated with VECTASTAIN ABC solution (Vector Labs) for 30 minutes followed by peroxidase substrate (Vector Labs) for signal development. Finally, sections were briefly counterstained with CAT hematoxylin (Biocare Medical), dehydrated and mounted.

Alkaline phosphatase staining was performed as previously described [50] Frozen sections were incubated with a 100 mM Tris buffer (pH 9.5) containing 5-bromo-4-chloro-3-indolyl phosphate and Nitro blue tetrazolium chloride (Roche Applied Science). Nuclear Fast Red (Vector) was used for counterstaining.

RNA isolation, quantitative real-time PCR and microarray hybridization

Uterine tissues were collected from control and Alk2 cKO female mice and stored immediately at −20°C until RNA extraction. Total RNA for each sample was extracted using the RNeasy Mini kit (Qiagen). Gene expression was analyzed by real-time quantitative polymerase chain reaction (qPCR). Either SYBR Green detection system (Life technologies) or TaqMan Assays-On-Demand PCR primer and probe sets (Life technologies) were used for individual gene analysis. Melt curve analysis was performed when using Sybr Green to verify a single amplification peak. Relative mRNA levels of transcript were calculated by the 2^−ΔΔC method after normalizing to the endogenous reference (Rn18S), and plotted as mean ± SEM.

For microarray analysis, we pooled RNA from three mice per genotype and performed the experiment in triplicate. Samples were tested for quality assurance on the Agilent 2100 Bioanalyzer by the Genomic and RNA Profiling Core at Baylor College of Medicine. The labeled cRNA was hybridized to Illumina Mouse WG-6 v.2.0 by the Genomics and Proteomics Core at Texas Children Hospital (Houston, TX). Array data were quantile normalized. Differentially expressed genes were defined, using two-sided t-test and fold change (on log-transformed data). Array data have been deposited on the Gene expression omnibus (GEO, accession number GSE46689).

Human endometrial stromal cell decidualization and rhBMP2 treatment

Samples were collected from patients undergoing gynecological surgery at Ben Taub General Hospital or the Obstetrics and Gynecology Clinic of Baylor College of Medicine (Houston, TX). The study was approved by the Institutional Review Board of Baylor College of Medicine. Undifferentiated stromal cells were obtained from endometrial biopsies from normally cycling fertile women with no history of uterine disease in the proliferative phase of the cell cycle. Written informed consent was obtained before sample collection. All samples were collected under Baylor College of Medicine Institutional Review Board (IRB) approval and IRB approval from each individual hospital. After collection, samples were transported to the laboratory and processed to obtain hESC. Cells were cultured at 37°C, 5% CO₂, 90% RH in growth medium (DMEM/F12 +10% csFBS +1% Antibiotic-Antimycotic +1% HEPES).

To investigate the role of ALK2 in controlling the decidualization of stromal cells in human we utilized siRNA specifically targeting ALK2 to knock down the receptor (Figure S5). The day of decidualization induction was designed as day 0. During the two days prior to induction (day −2 and −1), cells in 6-well plates were treated for 6–8 hours with DMEM/F12 medium +2% csFBS (Invitrogen) containing a mix of Lipofectamine 2000 (Invitrogen) and 100 pmol of either non-targeting (NT) siRNA (ON-TARGETplus Non-targeting Control Pool, Thermo Scientific) or siRNA targeting to ALK2 (ON-TARGETplus Mouse Acvr1 siRNA, Thermo Scientific) dissolved in Opti-MEM (Invitrogen). On day 0, an aliquot of cells for each treatment (NT or ALK2 siRNA) was collected and the remaining cells were treated with decidual medium containing 10⁻⁸ M 17β-estradiol, 10⁻⁶ M 17α-hydroxy-6α-methylprogesterone acetate (MPA) and 50 µM cyclic AMP (cAMP) in Opti-MEM +2% csFBS +1% Antibiotic-Antimycotic (Invitrogen). The medium was changed (to freshly prepared decidual medium) on day 2, and on day 3 cells were harvested and stored at −80°C.

To investigate the effects of rhBMP2 on SMAD1,5,8 phosphorylation in human stromal cells, we dissolved rhBMP2 (R&D Systems, 355-BM) in DMEM/F12 +10% csFBS at a concentration of 100 ng/ml and we treated the cells with this solution for 30, 45, 60, 180 or 360 minutes. To test the effect of ALK2 ablation on SMAD1,5,8 phosphorylation, we performed the same experiment but transfected the cells with either NT or ALK2 siRNA during the two days before rhBMP2 treatment as described above.

Western blotting

After the treatment, cells were trypsinized, washed in PBS, pelleted and stored at −80°C. Protein extraction was performed by incubating the pelleted cells with RIPA buffer (Thermo Scientific) in agitation for 30–45 minutes. Protein concentration was determined by Bradford's method using BSA as the standard. Samples containing 20 µg of total protein were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel (NuPAGE 4–12% Bis-Tris gel, Invitrogen) electrophoresis. The separated proteins were transferred onto a nitrocellulose membrane and the membrane was blocked for 1 hour with 5% skim milk in Tris-buffered saline with 10 mM Tris-HCl +150 mM NaCl +0.1% Tween 20 (TBS-T). Then, the membrane was probed overnight with antibodies against C/EBPβ (1500, Santa Cruz, sc-150), PGR (1500, Santa Cruz, sc-7208), p-SMAD1,5,8 (1500, Cell Signaling, 9511) or β-actin. After 1 hour-incubation with horseradish peroxidase-linked secondary antibodies, the membrane was treated with enhanced chemiluminescence reagents (SuperSignal West Pico Chemiluminescent Substrate, Thermo Scientific) to visualize the immunoreactivity.

We thank the Matzuk and DeMayo laboratory members and Yiqun Zhang for technical assistance.