Estrogen receptor β regulates endometriotic cell survival through serum and glucocorticoid-regulated kinase activation

Endometriosis is a debilitating disease that affects 5–10% of women of reproductive age ( 1 , 2 ). Clinical manifestations of endometriosis include chronic pelvic pain and infertility; a conclusive diagnosis requires laparoscopy and pathological examination of the lesions ( 3 , 4 ). Although the etiology of the disease is unknown, several theories have been proposed to explain the development of endometriosis. These theories include retrograde menstruation, coelomic metaplasia, and vascular/lymphatic dissemination ( 5 – 8 ). Though the theory of retrograde menstruation is the most accepted, the lack of one defined theory highlights the complex nature of the disease. A genetic basis for endometriosis has not been clearly established; however, among sisters, there is a five-fold increased probability of developing endometriosis ( 9 , 10 ). Genome-wide association studies have identified chromosomal regions near the coding regions of HOXA10, HOXA11, WNT4, IL33 , and IL1A that are associated with the disease ( 11 – 13 ). Epigenetic defects are also present in endometriosis ( 14 , 15 ), with hypomethylation present in the promoters of several key genes, including ESR2 ( 16 ), NR5A1 ( 17 ), GATA6 ( 15 ), HOXA10 ( 18 ), and CYP19A1 ( 19 ). Hypomethylation of the ESR2 promoter is associated with increased gene and protein levels in diseased tissues ( 16 ). The inflammatory state of endometriosis is also correlated with the levels of ERβ in diseased tissues ( 20 ). Because of the strong estrogen-mediated effects on the disease and the high expression levels of ERβ in affected tissues, studies are needed to determine ERβ transcriptional targets in endometriosis. We recently found that ERβ drives the transcription of several genes with altered expression in endometriotic stromal cells ( 21 ). One of the genes identified in that screen was the serum and glucocorticoid regulated kinase ( SGK1 ), which is overexpressed in endometriosis and has an ERβ binding site in its promoter region. The anti-apoptotic roles of SGK1 have been described in several studies, which indicate that SGK1 mediates the cell’s response to environmental stress ( 22 , 23 ). SGK1 has also been shown to regulate the pro-apoptotic factor FOXO3a ( 24 ). In this study, we tested the hypothesis that SGK1 is a transcriptional target of ERβ and is upregulated in endometrial tissues to promote endometriotic stromal cell survival. We examined SKG1 expression in tissue retrieved from women with endometriosis and from healthy women undergoing hysterectomy for benign conditions. We also used siRNA knockdown of ERβ, as well as estradiol (E2) and prostaglandin E 2 (PGE2) treatments to examine the regulation of SGK1 and its downstream targets in endometriosis.

We analyzed SGK1 protein levels in stromal cells derived from the endometrium of four normal subjects (NoEM) and from the endometriotic lesions of five subjects with endometriosis (E-Osis; Figure 1A ). Based on densitometric analysis, SGK1 protein levels were 3.59-fold higher ( p<0.05 ) in E-Osis compared NoEM cells ( Figure 1B ). Quantitative real-time PCR (qRT-PCR) revealed that, compared to NoEM, SGK1 mRNA levels were 2.1-fold (n=4, p<0.0005) higher in E-Osis cells ( Figure 1C ). We also performed immunohistochemistry to detect SGK1 in tissue samples of NoEM and E-Osis ( Figure 1D–E ). As shown in Figure 1D–E , SGK1 protein levels were more abundantly expressed in the diseased tissues compared to the controls. Quantification of SGK1 immunofluorescence levels confirmed that SGK1 expression was significantly elevated in E-OSIS compared to the NoEM tissues (mean florescence intensity, NoEM 2444.5 ± 429.4 versus E-Osis 5639.9 ± 696.1) ( Figure 1F ). To test whether E2 induces SGK1 expression, we treated E-Osis cells with 10 −7 M E2 for 2 and 6 hours. We then quantified SGK1 protein expression before and after E2 treatment by immunoblot ( Figure 2A ). We observed a strong induction of SGK1 protein expression in E-Osis cells following E2 treatment. Although the expression of SGK1 strongly increased after E2 treatment, the levels of SGK1 were not detected under basal conditions, possibly due to the patient-to-patient variation in SGK1 expression. Densitometric analyses demonstrated that E2 increased SGK1 expression 7.49-fold at 2h (p<0.05) and 6.5-fold at 6h (n.s.; Figure 2B ). To verify whether the E2-mediated induction of SGK1 was due to the transcriptional activation of ERα or ERβ, we treated endometriotic stromal cells with agonists selective for ERα (PPT) or ERβ (DPN). We observed that compared to E2 and PPT, DPN more strongly induced the protein expression of SGK1 ( Figure 2C ); however, densitometric analyses showed that the differences induced by E2, PPT, and DPN were not significantly different ( Figure 2D ). These results indicated that in E-Osis, E2 induces the protein expression of SGK1. We performed chromatin immunoprecipitation of ERβ followed by quantitative PCR (ChIP-qPCR) to confirm that ERβ was enriched at the SGK1 promoter. Previous studies showed that ERβ is frequently recruited to DNA regions rich in SP1 consensus sites ( 29 ). Transcription factor binding analysis indicated that several half EREs and SP1 sites were located in the region 2.5kb upstream of the SGK1 transcription start site ( Figure 3A ). We performed ERβ ChIP in E-Osis cells treated with vehicle or 10 −7 M E2 for 45 minutes. We then amplified the promoter region of SGK1 up to −2.5kb upstream of the transcription start site. The SGK1 promoter showed significant ERβ enrichment after E2 treatment at two of the sites analyzed, at −2kb (3.59-fold, p<0.001) and −2.5kb (2.60-fold, p<0.05) upstream of the SGK1 transcription start site ( Figure 3A ). This confirmed that in fact, ERβ was enriched at the SGK1 promoter. To further demonstrate that ERβ transcriptionally regulates SGK1 , we performed siRNA knockdown of ERβ in E-Osis cells. As shown in Figure 3B , we observed decreased SGK1 expression following ERβ knockdown. Densitometric analyses of the immunoblots were performed in cells isolated from different subjects (n=3, Figures 3C–D ). These results showed that a significant decrease in SGK1 expression (0.54-fold, p=0.003) was obtained following ERβ knockdown ( Figure 3D ). We concluded that in endometriotic stromal cells, ERβ contributes to the transcriptional regulation of SGK1. Based on previous studies indicating that SGK1 promotes cell survival ( 30 , 31 ), we hypothesized that SGK1 contributes to endometriotic stromal cell survival. To test the effects of decreased SGK1 expression in endometriosis, we performed siRNA-mediated knockdown of SGK1 in E-Osis cells. Following knockdown of SGK1, we detected increased expression of cleaved PARP, an apoptosis-related protein that is cleaved by caspase 3 ( Fig. 4A ). Densitometric analysis of E-Osis cells from 4 additional subjects ( Fig. 4B–C ) demonstrated that after SGK1 knockdown (0.22-fold, P = 0.001) there was a statistically significant increase in cleaved PARP expression (2.22-fold, P = 0.03). These results indicate that in endometriosis, SGK1 may contribute to increased endometriotic cell survival by inhibiting apoptosis. Previous studies showed that oxidative and pro-inflammatory stress activate SGK1 ( 24 , 32 ). SGK1 then confers resistance to apoptosis by phosphorylating and inactivating FOXO3a, a transcription factor that induces apoptosis ( 33 , 34 ). When phosphorylated by SGK1, FOXO3a is excluded from the nucleus, thereby inhibiting its pro-apoptotic function ( 24 ). We hypothesized that SGK1 confers E-Osis survival by phosphorylating FOXO3a, leading to its inactivation and loss of apoptotic function. We treated E-Osis cells with DPN and the pro-inflammatory factor, PGE2, which induces survival signaling in E-Osis ( 35 – 37 ). In Figure 4D , we showed that DPN and PGE2 significantly induced SGK1 expression compared to vehicle-treated cells. Furthermore, PGE2 synergized with DPN to further increase SGK1 in E-Osis. We then measured the phosphorylated and total levels of FOXO3 following DPN and PGE2 treatment. We observed that DPN and PGE2 increased the phosphorylation of FOXO3a, correlating with the increased levels of SGK1. These results indicate that in E-Osis, SGK1 may contribute to cell survival by phosphorylating and inactivating FOXO3a ( Figure 4E ).

We found that SGK1 is transcriptionally regulated by ERβ, and that it may contribute to increased E-Osis cell survival through inactivation of pro-apoptotic processes. Previous research showed that ERβ is increased in endometriosis and that signaling via ERβ is important for establishing and maintaining the disease ( 16 , 38 , 39 ). Inflammation is a major contributor to the pain associated with endometriosis, and recent studies demonstrated that inflammation is also involved in the pathogenesis of the disease. For example, ectopic and eutopic endometrium express a proteolytically-modified isoform of the steroid receptor co-activator protein (SRC-1) ( 39 ). Cleaved SRC-1 renders the ectopic endometrium resistant to the inflammatory and apoptotic signals elicited by TNF. In pre-clinical studies, compounds that selectively target ERα and ERβ effectively block lesion growth, angiogenesis, and neurogenesis associated with endometriosis ( 40 ). Recently, ERβ gain-of-function was shown to contribute to endometriotic lesion establishment and progression by evading the immune cell surveillance in mice ( 41 ). Furthermore, the inflammatory milieu of the disease, which is characterized by elevated PGE2, activates E2 synthesis in ectopic endometriotic stromal cells via SF1 and CYP19A1 activation ( 25 , 36 , 42 , 43 ). In this study, we demonstrated that ERβ and PGE2 activate SGK1, leading to inhibition of pro-apoptotic factors that may support increased survival of endometriotic stromal cells. SGK1 is a substrate for various kinases, including BMK1, ERK5 and mTOR ( 44 – 46 ). When SGK1 is phosphorylated by mTOR, SGK1 then phosphorylates p27, a kinase inhibitor protein (KIP) that phosphorylates and inactivates the cyclin E/cdk2 complexes and inhibits cell cycle progression ( 46 ). SGK1 phosphorylation of p27 prevents its nuclear import and renders it non-functional. Thus, SGK1 regulates pathways that ultimately result in cell proliferation. SGK1 also activates kinases that promote cell survival, such as GSK3β and B-Raf ( 22 , 47 ) and phosphorylates and inactivates the pro-apoptotic transcription factor FOXO3a. FOXO3a activates the transcription of apoptosis-related genes, such as TRAIL, IGFBP3 , and STK11 ( 24 , 48 ). Studies conducted in human endometrial stromal cells also demonstrate a role for FOXO1/FOXO3a during stromal cell decidualization ( 49 ). In these studies, the lack of FOXO3a activation in decidualized endometrial stromal cells confers resistance to the high levels of oxidative stress exerted during placental trophoblast invasion. Thus, in the absence of FOXO3a, decidualized endometrial cells evade apoptosis induced by elevated free radicals ( 49 ). Together, SGK1 effects on cell proliferation and resistance to apoptosis indicate its importance in cell survival. Aberrant regulation of SGK1 in the endometrium is associated with infertility or with recurrent pregnancy loss, depending on the expression level ( 32 , 50 ). SGK1 is highly upregulated in the endometrium of women with unexplained infertility ( 32 ). It is highly upregulated in response to cyclic-AMP and progesterone and is necessary for endometrial stromal cell decidualization ( 50 ). Implantation failure occurs in mice with ectopic expression of constitutively active SGK1 (S422D) during the window of implantation. Likewise, Sgk1 −/− mice experience spontaneous pregnancy loss despite having normal implantation rates. In human endometrial stromal cells, SGK1 confers resistance to oxidative stress during in vitro decidualization. These results indicate that SGK1 expression is finely regulated in the uterus to confer normal reproductive function. Here we show that in endometriosis, elevated expression of SGK1 contributes to the survival of the endometriotic stromal cells. Our results suggest that the local microenvironment of the endometriotic lesion may affect the expression levels of SGK1. For example, Eyster et al, ( 51 ) demonstrated that relative to the normal endometrium, the expression of SGK1 was elevated in both endometriomas and peritoneal lesions. However, the expression of SGK1 was higher in endometriomas than in the peritoneal lesions. As we showed in our studies, SGK1 expression is sensitive to estradiol and PGE2, suggesting that the factors present in the microenvironment of the lesion may account for the differences in expression. In this study, we present data that demonstrate another mechanism by which ERβ signaling contributes to the pathology of endometriosis. We show that ERβ transcriptionally regulates the expression SGK1, a kinase that supports endometriotic stromal cell survival through inhibition of pro-apoptotic pathways. SGK1 contributes to cell survival through the phosphorylation and inactivation of FOXO3a. The ERβ agonist DPN and PGE2 synergize to activate SGK1 expression. Future studies testing the efficacy of SGK1 inhibitors for the treatment of endometriosis are needed.

The study was approved by the Northwestern University Institutional Review Board and informed consent was obtained from all participants (Reproductive Tissue Registry STU00018080). Normal endometrium (NoEM) and ovarian endometrioma cyst walls (E-Osis) were used for all experiments. We used freshly isolated tissues or stromal cells within three passages after initial culture. Tissues were obtained during the follicular phase of the menstrual cycle, and women on oral contraceptives or other hormonal therapies were excluded from the study. NoEM was obtained from women undergoing hysterectomies for benign conditions other than endometriosis and E-Osis tissues were obtained from women undergoing surgical removal of endometriosis. For E-Osis samples, confirmation of endometriosis was obtained from clinical diagnosis and by pathology analysis. For the studies performed in stromal cells, we established primary human cell cultures as previously described ( 25 , 26 ). Briefly, tissues were digested using a 0.2 mg/ml DNAse and 5 mg/ml Collagenase solution followed by a second digestion in 0.2 mg/ml DNAse, 5 mg/ml Collagenase, 1 mg/ml Pronase, and 2 mg/ml Hyaluronidase solution prepared in Hanks Buffered Saline Solution. The primary were cultured in DMEM/F-12 (Life Technologies) supplemented with 10% FBS and antibiotic/anti-mycotic solution (VWR). E2, diarylpropionitrile (DPN), and propyl pyrazole triol (PPT) were purchased from Tocris (resuspended in ethanol) and used at a final concentration of 100 nM. PGE2 was purchased from Cayman, resuspended in ethanol, and used at a final concentration of 50–100 nM. Phosphatase Inhibitor Cocktail (cat # P5726) and Protease Inhibitor Cocktails (Cat # P8340) were purchased from Sigma (St. Louis, MO) and diluted 1:100 in the appropriate lysis buffer. The following antibodies were used for immunoblotting: SGK1 (Cell Signaling, Danvers, MA, Cat # 3272), cleaved PARP (Cell Signaling, Cat# 9542), pFOXO3a (Cell Signaling, Cat # 9464), FOXO3a (Cell Signaling, Cat# 2497), and β-actin (Sigma, Cat # A1978). Anti-SGK1 (Enzo, Farmingdale, NY, Cat# ADI-KAP-PK015-D) antibody was used for immunohistochemistry. RNA was prepared from the samples according to RNeasy kit (Qiagen, Valencia, CA). 1 μg of RNA was reverse transcribed into cDNA using qScript cDNA Supermix (Quanta Biosciences, Gaithersburg, MD). SYBR Green (Life Technologies, Grand Island, NY) and primers were used to amplify genes of interest. Gene expression data were normalized to the GAPDH or TBP genes. Samples were processed in the 7900HT Fast Real-Time PCR System and data were collected with SDS 2.3 software from ABI. Primer sequences used were, SGK1 , Fwd-5′-CAGCATACGCCGAGCCGGTC-3′ Rev-5′-ATGAAAGCGATGAGAATTGCCACCA-3′, GAPDH , Fwd-5′ GAAGGTGAAGGTCGGAGTC-3′, Rev-5′-GAAGATGGTGATGGGATTTC-3′. Reverse transfections were used for all E-Osis and NoEM transfections. Specifically, 40 nM ON-TARGET Plus Non-Targeting siRNA #1 (Thermo Scientific, Waltham, MA, D-001810-01-05), ON-TARGET Plus Human SGK1 (6441) siRNA (Thermo Scientific, J-003027-14-0010), 100 nM Silencer Select siESR2 (Ambion, Waltham, MA, s4827) or Silencer Select Non-Targeting siRNA #1 (Ambion, 4390843) were diluted in Opti-MEM Reduced Serum Media with RNAiMAX (Life Technologies) in a 1:5 siRNA:RNAiMAX ratio. Complex formation was carried out for 15 minutes at room temperature and then added to serum-free/antibiotic-free DMEM/F-12 media overnight. Media was changed the following day to complete FBS- and antibiotic-containing media. The cells were lysed for mRNA or protein collection 48–96 hours after transfection. We performed ERβ ChIP in E-Osis cells following the protocol published by Lee et al ( 27 ) and using the conditions for ERβ ChIP used by Charn et al ( 28 ). Briefly, cells were grown until 70–80% confluency and serum starved overnight, followed by a 45-minute treatment with 100 nM E2. Protein lysates were incubated with 4 μg each of an ERβ antibody mixture (CWK-F12 ( 28 ), Pierce, Waltham, MA, PA1-311; GeneTex, Irvine, CA, GTX70182; Calbiochem, Billerica, MA, GR-40) or IgG (Sigma) overnight at 4°C. Dynal Beads (Life Technologies) were used to capture the protein/DNA/antibody complexes. After DNA purification, qRT-PCR was performed on INPUT and ChIPed DNA using primers spanning 2.5kb upstream and downstream of the SGK1 transcription start site. Data was quantified as fold enrichment relative to IgG and normalized to vehicle treatment. Experiments were replicated in E-Osis stromal cells from at least 3 subjects. Primers used were, SGK1 Fwd −2.5kb 5′ agcagacatgggccagttac-3′, Rev 5′-gcgagactccgtctcaaaac-3′; SGK1 Fwd −2kb 5′-ttgcaacaaagcaaaccaag-3′, Rev 5′-catgtgaaacgccttttcct-3′; SGK1 Fwd −1.5kb −5′ atgacctgcagggttttcag-3′, Rev 5′-ccaagaacacgtgaggaggt-3′; SGK1 Fwd −1kb 5′-tttcagccttgcttggtttt-3′, Rev 5′-aagatttcctgccccgagt-3′; SGK1 Fwd −0.5kb 5′-gagggtatctgcagggacag-3′, Rev 5′-cggggtagttttccacctct-3′; SGK1 Fwd +1kb 5′-TCCTCCTTCATCCACAGCTT-3′, Rev 5′-ttcctcaaatccggtcaaac-3′; SGK1 Fwd +1.5kb 5′-agtggcgagctggattctaa-3′, Rev 5′-atgcacggcacatacaaaaa-3′; SGK1 Fwd +2kb 5′-agCCAAGTCCTTCTCAGCAA-3′, Rev 5′-TTCCAAAACTGCCCTTTCC-3′; SGK1 Fwd +2.5kb 5′-ggcggtagacactccttgaa-3′, Rev 5′-CAGGAAAGGGTGCTTCACAT-3′; SGK1 Fwd ChIP BS 5′-ttggccaaaagcacaaaaa-3′, Rev 5′-gagcactgacgtttccttga-3′. We obtained frozen or paraffin-embedded sections of endometriotic tissue from women with pathologically confirmed ovarian endometriosis and of normal endometrium tissue from women without endometriosis. Paraffin-embedded sections were deparaffinized and re-hydrated with serial washes in xylene and 100%, 90%, 80%, and 60% ethanol followed by washing in deionized water. Sections underwent antigen retrieval, blocking in 5% normal donkey serum, and incubation with SGK1 (Enzo, Cat# ADI-KAP-PK015-D) antibody for 1 hour at room temperature in a humidified chamber. Secondary antibody was added for 1 hour at room temperature in the dark, followed by incubation with DAPI in 1× TBS (5 ng/ml) and 2 washes in TBS. Tissue sections were placed under coverslips using ProLong Gold antifade mounting media (Life Technologies). Image J was used to calculate the mean intensity of the SGK1 immunofluorescence. Briefly, the SGK1 (red) intensity values were calculated by designating a region on interest (ROI) and plotting the mean signal intensity per outlined area. Results were plotted as mean ± standard deviation. Protein from stromal cells was performed using M-PER (Pierce) lysis buffer and quantified using the BCA Protein Assay (Pierce) as indicated by the manufacturer’s instructions. At least 20μg of protein were diluted with reducing 4× LDS Sample Buffer (Life Technologies), electrophoresed on 4–12% Novex Bis-Tris Polyacrylamide Pre-Cast gels (Life Technologies) and transferred onto PVDF or nitrocellulose membranes. The membranes were blocked with 5% milk in TBST and probed for each specific antibody shaking overnight at 4C. HRP-conjugated secondary antibodies (Cell Signaling) were diluted in 5% milk at 1:5000 and incubated for 1 hour shaking at room temperature. The membranes were then washed 4 times in TBST and once in TBS for 10 minutes each time followed by incubation with chemiluminscence reagent for 5 minutes (Femto from Pierce or Luminata Crescendo from Millipore). Film was used to develop the western blots in a Konika Minolta developer. All experiments were performed in tissues from three or more subjects and analyzed using Student’s t-tests or ANOVA with Tukey’s multiple comparison post-tests using GraphPad versions 5–6.