|Home | About | Journals | Submit | Contact Us | Français|
Female mice lacking the transcription factor C/EBPβ are infertile and display markedly reduced estrogen (E)-induced proliferation of the uterine epithelial lining during the reproductive cycle. The present study showed that E-stimulated luminal epithelial cells of a C/EBPβ-null uterus are able to proceed through the G1 phase of the cell cycle before getting arrested in the S phase. This cell cycle arrest was accompanied by markedly reduced levels of expression of E2F3, an E2F family member, and a lack of nuclear localization of cyclin E, a critical regulator of cdk2. An increased nuclear accumulation of p27, an inhibitor of the cyclin E-cdk2 complex, was also observed for the mutant epithelium. Gene expression profiling of C/EBPβ-null uterine epithelial cells revealed that the blockade of E-induced DNA replication triggers the activation of several well-known components of the DNA damage response pathway, such as ATM, ATR, histone H2AX, checkpoint kinase 1, and tumor suppressor p53. The activation of p53 by ATM/ATR kinase led to increased levels of expression of p21, an inhibitor of G1-S-phase progression, which helps maintain cell cycle arrest. Additionally, p53-dependent mechanisms contributed to an increased apoptosis of replication-defective cells in the C/EBPβ-null epithelium. C/EBPβ, therefore, is an essential mediator of E-induced growth and survival of uterine epithelial cells of cycling mice.
The ovarian steroid hormones estrogen (E) and progesterone (P) play critical roles in the maintenance of the mammalian uterus through cyclical rounds of cell proliferation and differentiation during the reproductive cycle (50). In rodents, ovarian E produced during the proestrus stage stimulates uterine luminal and glandular epithelial cell proliferation, preparing the uterus for potential pregnancy. During this growth phase, E causes distinct physiological changes such as an increased uterine wet weight and structural remodeling of the luminal epithelium (LE) cell layer while also accelerating their entry into the S phase of the cell cycle (17, 18, 29, 50). At the onset of pregnancy, increasing levels of P produced from the newly formed corpora lutea in the ovaries suppresses the E-stimulated proliferation of uterine LE cells. The actions of E and P in these epithelial cells are mediated via their respective nuclear receptors, estrogen receptor alpha (ERα) and progesterone receptor (PR) (23, 26). Acting in concert, these hormones control early events that are essential for providing a suitable environment for blastocyst attachment to LE cells and successful implantation. In the adult female mouse, the administration of exogenous E and P to ovariectomized mice faithfully reproduces the uterine epithelial responses seen during the estrous cycle and early pregnancy. Therefore, mouse uterine LE cells offer an excellent model with which one may explore the molecular mechanisms by which steroid hormones control cell proliferation.
Several previous studies documented the mitogenic effects of E on rodent uterine LE cells and examined the mechanisms underlying this steroid-stimulated proliferation (29, 50). The administration of E to ovariectomized mice led to the transcriptional induction of proto-oncogenes such as c-fos, c-myc, and n-myc, epidermal growth factor (EGF) and its receptor EGFR, transforming growth factor α (TGF-α), and insulin-like growth factor 1 (IGF-1) concomitant with uterine epithelial cell proliferation (29, 33, 37, 50). Based on these findings, it was proposed that proto-oncogene and growth factor pathways mediate the E-induced growth response in the uterus (5, 9, 53, 55). However, later studies using knockout mouse models failed to provide unequivocal evidence in support of a role of certain of these factors in the female reproductive tract (24, 25). Interestingly, the administration of P to ovariectomized mice, which downregulated E-induced uterine LE cell proliferation, did not alter the expression profiles of many of these growth-promoting factors, indicating that these factors may not be participating directly in E-induced cell cycle entry. This observation raised the possibility that additional factors are involved in transmitting the E-induced growth response to the cell cycle machinery (5, 50, 55).
In recent studies, we identified and characterized the transcription factor CCAAT enhancer binding protein beta (C/EBPβ) as a steroid-hormone-regulated gene that is critical for uterine functions (28). C/EBPβ is a member of the bZIP family of leucine zipper proteins and has been implicated as a regulator of proliferation and differentiation in diverse tissues (11, 30, 41, 54, 56). The importance of C/EBPβ in female fertility was revealed when C/EBPβ-null females were found to have severely compromised ovarian and uterine functions. In addition to a failure in ovarian follicle rupture (48), the C/EBPβ-null uterus is refractory to embryo attachment to the uterine LE, as demonstrated by embryo transfer experiments performed with ovariectomized mice supplemented with E and P (28). The expression of C/EBPβ is induced rapidly in the uterine LE during the early phases of the E-induced proliferation of this tissue (28). Its importance in uterine growth was highlighted by the markedly diminished level of expression of the Ki67 antigen, a widely used marker of cell proliferation, in E-treated uterine LE cells of C/EBPβ-null mice (28).
In this study, we noted that simultaneous treatment with P, which reduces E-induced uterine LE cell proliferation, also decreased the level of expression of C/EBPβ, thereby positioning this transcription factor as a biologically relevant mediator of E action in this tissue. We postulated that the E-induced expression of C/EBPβ and its downstream targets controls the production and/or activity of one or more major cell cycle regulatory molecules in uterine LE cells. Using stage-specific cell cycle markers, we determined that the loss of C/EBPβ leads to a significant impairment in DNA replication in uterine LE cells. We also identified alterations in the expression and cellular localization of cyclin E, E2F3, and p27, key molecules that modulate the G1-S-phase transition of the cell cycle. We further showed that the proliferation defect in C/EBPβ-null uterine LE cells is associated with the activation of well-known DNA damage response pathways involving ataxia telengiectasia mutated (ATM), ATM-Rad3 related (ATR), the checkpoint kinases Chk1 and Chk2, and p53. Activated p53 helps to maintain the G1-S-phase blockade of the cell cycle by inducing the synthesis of the inhibitor p21. p53 also promotes the apoptosis of replication-defective uterine LE cells. Collectively, these studies identified the molecular pathways by which C/EBPβ mediates E-induced proliferation and the survival of uterine LE cells during the reproductive cycle and early pregnancy.
All experiments involving animals were conducted in accordance with National Institutes of Health standards for the use and care of animals. The animal protocols were approved by the University of Illinois Institutional Animal Care and Use Committee. Heterozygous mice carrying a mutation in the C/EBPβ gene were provided by Peter F. Johnson of the National Cancer Institute. Female mice carrying a null mutation in the C/EBPβ gene were derived from crosses of heterozygous females with nullizygous male mice as described previously (48). Female wild-type (WT) and C/EBPβ-null mice of the 129Sv background were ovariectomized at 10 to 11 weeks of age and rested for 2 weeks. These mice were then treated with a single dose of 17β-estradiol (E) (250 ng in sesame oil) by intraperitoneal (i.p.) injection for various durations. In experiments where progesterone (P) was used, mice were injected with a single dose of P (1 mg in sesame oil) along with E (250 ng). For each treatment group, at least five mice were used at each time point. In some experiments, animals were injected i.p. with bromodeoxyuridine (BrdU) (2 mg/animal; BD Pharmingen) 1 h prior to sacrifice. Uteri were collected and fixed in 10% formalin prior to immunohistochemistry (IHC) analysis. Alternatively, uteri were pooled for the isolation of primary epithelial cells.
Hormone-treated uterine horns were excised, trimmed of fat, and dissected longitudinally to expose uterine lumen. Dissected horns were then cut into 4- to 5-mm-long pieces and washed in Hanks balanced salt solution (HBSS; Gibco). Uterine tissue pieces were then placed into HBSS containing 6 g/liter dispase (Invitrogen) and 25 g/liter pancreatin (Sigma) for 1 h at room temperature (RT), followed by 10 min at 37°C. The tubes were gently agitated to release luminal epithelial cells from the rest of the uterine tissue. The cell suspension was filtered through a 100-μm-pore-size sieve (BD) to remove any tissue debris. The filtrates containing luminal epithelial cells were centrifuged at 2,000 rpm for 5 min to pellet the cells. The cells were washed twice in phosphate-buffered saline (PBS) and repelleted. The pellets were lysed with Trizol for RNA extraction or with radioimmunoprecipitation assay (RIPA) lysis buffer for the preparation of whole-cell protein lysates.
Total RNA was isolated from uterine cells by standard Trizol-based protocols and converted to cDNA as described previously (20). The cDNA was amplified by real-time PCR to quantify gene expression using gene-specific primers and Sybr green (Bio-Rad Laboratories). As a loading control, the expression level of the 36B4 gene, which encodes a ribosomal protein, was determined. For each treatment, the mean threshold cycle (CT) and standard deviation were calculated from CT values obtained individually from 3 to 4 replicates of that sample. Each sample was subjected to three independent real-time PCR trials. The fold change was derived from the mean CT values as described previously (20). Analysis of variance (ANOVA) single-factor analysis was conducted on the grouped means to determine statistical significance at a significance level of a P value of <0.05.
WT and C/EBPβ-null mice were ovariectomized and treated with E, and after 18 h, the LE cell layer was isolated and total RNA was prepared. The RNA was hybridized to Affymetrix mouse arrays (GeneChipMouse Genome 430 2.0 array) containing probes that represented ~14,000 known genes. They were processed and analyzed according to the Affymetrix protocol as described previously (20).
Formalin-fixed uterine pieces were processed for paraffin embedding. Cross sections (5-μm thickness) were mounted onto microscope slides (Fisher Scientific). For immunostaining, uterine sections were deparaffinized in xylene (three times for 5 min each), rehydrated through a graded series of treatment with ethanol (100%, 95%, 85%, and 70% for 5 min each), and rinsed in tap water. For all samples, antigen retrieval was performed by boiling the sections in 0.01 M sodium citrate buffer (pH 6.0) for 20 min, followed by incubation at RT for 30 min. A 5% solution of normal donkey serum (Jackson Immunoresearch) in PBS was used as a blocking buffer. Sections were incubated with the following primary antibodies diluted in blocking solution (0.25% bovine serum albumin [BSA], 0.3% Triton X-100, sterile PBS) overnight at 4°C: BrdU and Ki67 (BD Pharmingen), phospho-Ser10 histone H3 (Upstate Biotechnology); cyclin E, cyclin A, and Rad18 (Abcam); cyclin D1 (LabVision NeoMarkers); E2F3 (Santa Cruz Biotechnology); p27 (BD Transduction Laboratories); and phospho-(Ser1981)-ATM, phospho-(Ser139)-H2AX, caspase 3, and cleaved caspase 3 (Cell Signaling Technology). The sections were washed and incubated with biotinylated secondary antibodies (Jackson Immunoresearch Laboratories Inc.) for 60 min, followed by incubation with streptavidin-conjugated horseradish peroxidase (Histostain kit; Zymed Laboratories Inc.) for 45 min. Sections were stained with 3-amino-9-ethyl carbazole (AEC) solution (Zymed Laboratories Inc.) and counterstained with Mayer's hematoxylin (Sigma). Immunofluorescence staining was performed by incubating sections with the Cy3-conjugated streptavidin complex (Jackson Immunoresearch Laboratories) following incubation with secondary antibodies. Counterstaining was done by using 4′,6′-diamidino-2-phenylindole (DAPI; Invitrogen Inc.). Stained sections were mounted in fluorescence mounting medium [20% glycerol, 8% (wt/vol) polyvinyl alcohol (PVA), diazabicyclo(2,2,2)octane (DABCO), Tris-Cl (pH 8.5), and sterile double-distilled water (ddH2O)]. Negative controls included incubation with donkey serum and omission of the primary antibody for all samples.
To detect apoptotic cells containing fragmented DNA, in situ labeling of free 3′ OH of nicked DNA was carried out by using the In Situ Cell Death Detection kit (Roche Applied Science) according to the manufacturer's protocol. Briefly, deparaffinized and rehydrated sections were incubated at RT for 15 min with proteinase K (20 μg/ml in 10 mM Tris-HCl [pH 8.0]). Sections were washed and incubated for 1 h at 37°C in a terminal deoxynucleotidyltransferase (TdT)-mediated dUTP-biotin nick end labeling (TUNEL) reaction mixture containing TdT enzyme, fluorescein-labeled nucleotide mixture, and diluent buffer (30 mM Tris-HCl [pH 8.0]). For negative controls, TdT enzyme was omitted from the reaction mixture. After rinsing in PBS, sections were counterstained with DAPI and mounted in fluorescence mounting medium.
The images of immunohistochemical staining were captured by using a Leica (Nussloch, Germany) DM2500 light microscope fitted with a Qimaging Retiga 2000R camera (Qimaging, British Columbia, Canada). For counting, at least 5 to 6 individual 40× fields from each sample were captured. The numbers of positively stained uterine LE nuclei in each field were averaged and expressed as a percentage of the total number of these cells. The standard deviation was determined for each averaged total. ANOVA single-factor analysis was conducted on the grouped means to determine statistical significance at a significance of a P value of <0.05. Fluorescent images for phospho-histone H3, cyclin E, p27, and E2F3 were captured by using the Leica DM2500 microscope. For all other antibodies, fluorescence microscopy and image capturing were performed by using a Leica DMR microscope and a Retiga EXI monochrome charge-coupled-device (CCD) camera (Qimaging). All images were processed by using Adobe Photoshop CS, version 8.0.
Whole-cell lysates were prepared from primary LE cells isolated from uteri collected from control (untreated) or E-treated WT and C/EBPβ-null mice. The cells were homogenized using RIPA lysis buffer. Protein concentrations of cell lysates were determined by using the Bradford assay from Bio-Rad. Twenty micrograms of each cell lysate was separated by SDS-PAGE (10%) and transferred onto Hybond-P membranes (Amersham Pharmacia Biotech). The membranes were blocked in Tris-buffered saline with 0.1% (vol/vol) Tween 20 (TBS-T) and 5% (wt/vol) nonfat dry milk for 1 h. The membranes were then incubated overnight at 4°C with primary antibodies diluted 1:1,000 in TBS-T containing 5% (wt/vol) BSA. These antibodies were directed against Chk1, p53, phospho-Thr187-p27, and calnexin (Santa Cruz Biotechnology); p21 (BD Pharmingen); phospho-Ser296-Chk1 and phospho-Ser15-p53 (Cell Signaling Technology); and cdk2 and phospho-Thr160-cdk2 (Abcam). Membranes were then incubated for 1 h at RT with secondary antibodies conjugated to horseradish peroxidase and diluted 1:5,000. After thorough washes with TBS-T, chemiluminescence detection was performed by using SuperSignal Femto and Pico reagents from Pierce (Thermo Scientific).
Chromatin immunoprecipitation (ChIP) analysis was conducted by using an EZ-ChIP kit (Upstate Biotechnology) according to the manufacturer's protocol. Briefly, mouse uterine LE cells were isolated from 11-week-old, ovariectomized WT mice of the 129Sv background that were subjected to E treatment (250 ng) for 12 h. The cells (7.5 × 106 cells) were washed in PBS and cross-linked with 1% formaldehyde at room temperature for 10 min. The cross-linked cells were lysed by using SDS lysis buffer and sonicated 5 times for 10 s each pulse. Lysates were precleared with salmon sperm DNA-protein A, and the DNA-protein complexes were subsequently immunoprecipitated by using antibodies against RNA polymerase II, rabbit IgG, or C/EBPβ (Santa Cruz Biotechnology). The immune complexes were recovered with protein A agarose. The beads were then repeatedly washed, and protein complexes were eluted. The cross-linking was reversed, and proteins were digested by using 0.5 mg/ml proteinase K. Purified DNA were used as templates for real-time PCR using various primer sets to amplify specific regions of the E2F3 promoter.
Putative C/EBPβ binding sites at the E2F3 promoter were determined by in silico analysis of the proximal promoter region (bp −1 to −1900) using TESS (45), TFsearch (15), and Consite (44) software. The following primers were designed to amplify specific regions containing putative C/EBPβ binding sites: TTCATACCCCTCCCACAAGA and TTTTATTGTCCTTCTAGCCATGA for bp −329 to −340, CCAGTGATTCAGCATACATTACA and TTATTGCCCTCACCACCTTC for bp −458 to −473, ACAGTCTTGGGTGAGCTGGT and CCCCTACACACTCGGTTCCT for bp −584 to −595, and AGCCTCTTTGACTGGGACTG and TCAGTATTTTGCTGGGGTCTC for bp −1054 to −1066.
Statistical significance was assessed by ANOVA at a significance level of a P value of <0.05 and is indicated by asterisks in the figures. NS indicates nonsignificant changes (P > 0.05).
It is well known that P antagonizes the E-induced proliferation of uterine LE cells (40, 50, 51). Our previous studies described that C/EBPβ expression is markedly induced in the uterine LE of ovariectomized mice in response to an acute administration of E (28). We report here that the simultaneous administration of P inhibited acute E-induced uterine LE cell proliferation, as measured by Ki67 immunostaining (Fig. (Fig.1A).1A). We also determined the levels of C/EBPβ mRNA and protein in uterine LE cells of ovariectomized mice after treatment with E or E plus P for 2 and 11 h. We noted that treatment with E alone leads to a robust increase in C/EBPβ mRNA and protein levels within 2 h, followed by a decline in its expression to basal levels by 11 h (Fig. 1B and C). When mice were treated with a combination of E and P, the E-induced increase in C/EBPβ mRNA levels at 2 h was strongly suppressed. P therefore counteracts the transcriptional induction of C/EBPβ by E. These results are consistent with our hypothesis that C/EBPβ is a critical downstream mediator of E-stimulated uterine epithelial cell proliferation and that P exerts its antiproliferative effects, at least in part, by suppressing C/EBPβ expression in these cells.
Previous studies demonstrated that the E-induced proliferation of uterine LE cells is significantly reduced in C/EBPβ-null mice compared to wild-type (WT) mice of the same genetic background (28). The goal of the present study was to determine the precise stage at which cell cycle progression is impaired in mutant uterine LE cells. To assess cell cycle entry and progression through the G1 phase, we monitored the expression of two cell cycle markers, minichromosome maintenance molecule 3 (MCM3) and Ki67, by IHC analysis. MCM3 is a major component of the “prereplicative licensing complex” that is highly expressed during G1 phase in preparation for entry into S phase (7). Ki67 is a nuclear antigen that is expressed at all stages of the active cell cycle (7, 14). At 8 and 12 h after E treatment, which correspond to early and late G1 phase, both MCM3 and Ki67 were expressed at comparatively equal levels in WT and C/EBPβ-null epithelial cells (Fig. (Fig.2).2). This observation suggested that both WT and mutant LE cells are competent to enter G1 from G0 phase upon E stimulation and are able to progress through the G1 phase of the cell cycle.
To examine the G1-S-phase transition and S-phase activity, ovariectomized WT and C/EBPβ-null mice were treated with E for durations of 8, 12, 15, and 18 h. A 1-h pulse of BrdU, a thymidine analogue, was administered to these mice immediately prior to the collection of uterine tissue. Nuclear immunostaining of BrdU, which is indicative of S-phase activity, was absent in untreated ovariectomized WT and C/EBPβ-null mice (data not shown). Sporadic staining of BrdU was visible for ~5% of WT LE cells at 8 h following E treatment (Fig. 3Aa and B). This staining was significantly increased by 12 h, indicating the entry of a substantial number (~16%) of uterine LE cells into S phase (Fig. 3Ac and B). The BrdU incorporation increased further at 15 h (Fig. 3Ae). By 18 h of E treatment, almost all uterine LE cells are actively undergoing DNA replication in WT mice (Fig. 3Ag). In comparison, in the uterine LE of C/EBPβ-null mice, the level of BrdU incorporation was drastically reduced at all time points tested following E treatment (Fig. 3Ab, d, f, and h and B). While more than 90% of WT LE cells showed BrdU-positive staining at 18 h following E treatment, only 10% were BrdU positive in mutant epithelial cells (Fig. (Fig.3).3). These results suggested that in the absence of C/EBPβ, E-stimulated uterine LE cells are severely impaired in their ability to undergo DNA replication.
One would predict that the blockade in S-phase activity in C/EBPβ-null uterine LE cells would result in a loss of E-induced mitotic activity in these cells. To test this prediction, we employed an antibody that specifically recognizes phosphorylated serine 10 of histone H3. This marker typically localizes to distinct regions of chromosome condensation during mitosis, appearing as punctuate loci in the nucleus (16). As expected, the levels of phospho-histone H3 staining steadily increased from 18 to 22 h in E-treated WT uterine LE cells (Fig. (Fig.4A).4A). However, in the absence of C/EBPβ, phospho-histone H3 staining was absent or drastically reduced in LE cells compared to WT uteri (Fig. 4A and B). Consistent with the S-phase impairment that we observed, the C/EBPβ-null mice displayed virtually no evidence of mitotic activity throughout the LE cell layer. The mitotic block was further confirmed by examining the mRNA and protein levels of cyclin B, a well-defined regulator of mitosis. As expected, the levels of expression of cyclin B2 mRNA (Fig. 4Ca) and protein (Fig. 4Cb) were greatly reduced in C/EBPβ-null LE cells compared to WT tissue (Fig. (Fig.4C4C).
The progression of the cell cycle from G1 to S phase is orchestrated by stage-specific cyclins, cyclin-dependent kinases, and other cell cycle regulatory factors (14). Cyclin D1 is typically expressed early in the G1 phase and associates with cdk4 to promote S-phase entry. Cyclins E and A then associate with cdk2 and directly regulate S-phase progression. cdk2 achieves full functionality through the Cdk-activating kinase (CAK)-dependent phosphorylation of a threonine residue (Thr160) (13, 32). To analyze the basis of the lack of E-induced proliferation in C/EBPβ-null epithelial cells, we therefore examined the expression levels of each of these S-phase-regulatory cyclins and cdk's in WT and C/EBPβ-null LE cells at 15 h following E treatment.
When the expression of cyclin D1 and cyclin A mRNAs in WT and C/EBPβ-null uterine LE cells was examined, no significant difference in their levels in WT and mutant epithelial cells was seen (Fig. 5Aa and c). IHC of uterine sections at 15 h following E treatment revealed that the levels of the cyclin D1 protein and its nuclear localization were unaltered (Fig. 5Ab). Interestingly, we noted a modest decrease in the levels of the cyclin A protein in the nuclei of mutant LE cells (Fig. 5Ad). We did not observe any significant difference in the expression levels of cdk2 and cdk4 mRNAs in WT and C/EBPβ-null uterine epithelial cells at 15 h following E treatment (Fig. 5Ae and data not shown). In contrast, we noted an approximately 50% reduction in the expression of cyclin E mRNA in C/EBPβ-null uterine epithelial cells compared to WT cells (Fig. (Fig.5B,5B, top). This decrease in cyclin E levels in mutant epithelial cells was further confirmed by IHC analysis of tissue sections collected at 15 h following E treatment (Fig. (Fig.5B,5B, bottom). Most notably, we failed to detect a nuclear localization of cyclin E in mutant epithelial cells. While the presence of this cyclin was prominent in the nuclei of WT epithelial cells (Fig. 5Ba and c), its immunostaining was markedly reduced and predominantly cytoplasmic in C/EBPβ-null epithelial cells (Fig. 5Bb and d). Coincident with the reduced nuclear presence of cyclin E, we observed that the activation of cdk2 via the phosphorylation of its Thr160 residue was diminished in C/EBPβ-null epithelial cells (Fig. 5Af). The reduced level of expression of cyclin E, combined with a lack of its nuclear localization in epithelial cells, therefore contributes to the impaired activation of cdk2, leading to the defect in E-induced S-phase activity in C/EBPβ-null uterine LE cells.
E2Fs 1, 2, and 3 belong to a family of six known E2Fs and have been implicated in several studies as being the most important regulators of the G1-S-phase checkpoint (35). Upon mitogenic stimulation, activated E2F factors transactivate downstream genes, including cyclin E, promoting the progression of the cell cycle into S phase. The mRNA levels of these three E2Fs in WT and C/EBPβ-null LE cells were investigated. While the expression levels of E2F1 and E2F2 displayed no statistically significant difference between WT and mutant epithelial cells, the E2F3 mRNA level was diminished by approximately 50% in LE cells lacking C/EBPβ (Fig. 6Aa to c). This observation was further confirmed by IHC analysis showing reduced nuclear expression of the E2F3 protein in LE cells from WT and C/EBPβ-null tissue sections following 15 h of E treatment (Fig. (Fig.6B).6B). This finding is consistent with previous reports that E2F3 plays an essential role in promoting S-phase activity (19, 22). It is of interest that the steady-state expression level of E2F3 mRNA is considerably higher than that of E2F1 or E2F2 mRNA in E-treated uterine LE cells (Fig. 6Ad). Therefore, the reduced levels of E2F3 may have a strong impact, resulting in a diminished level of expression of cyclin E mRNA and the consequent impaired ability of C/EBPβ-null uterine epithelial cells to progress through the S phase of the cell cycle.
C/EBPβ manifests its gene regulatory role by binding to the promoters of its target genes. To explore the possibility that C/EBPβ directly mediates the E-induced expression of E2F3 and/or cyclin E, we examined the promoter regions of these genes by in silico analysis. We identified four putative binding sites for C/EBPβ in the E2F3 promoter and two such sites in the cyclin E promoter, as indicated in Materials and Methods. We then performed a ChIP analysis to test for C/EBPβ occupancy at these sites. While we did not detect C/EBPβ recruitment to any of the putative binding sites in the cyclin E promoter sites (data not shown), we found a robust recruitment of this transcription factor to the −584 region of the E2F3 promoter following E treatment (Fig. (Fig.6C6C).
The cyclin-dependent kinase inhibitors (CKIs) block cell cycle progression by negatively regulating the actions of cyclin-cdk complexes (14, 47). The predominant CKIs belong to the INK4 and Cip/Kip families of inhibitors, the latter of which includes p21 and p27, two well-characterized G1-S-phase regulators (4). Both p21 and p27 directly inhibit the cyclin E-cdk2 and cyclin A-cdk2 complexes, thereby controlling S-phase activity. We found that the levels of p27 mRNA were only modestly elevated in C/EBPβ-null LE cells compared to WT cells at 0, 12, and 15 h after E treatment (Fig. (Fig.7A).7A). Strikingly, however, a prominent nuclear localization of the p27 protein in C/EBPβ-null epithelial cells was observed at 15 and 18 h of E treatment (Fig. 7Bb and d). This nuclear accumulation of the p27 protein coincided with the block in S-phase progression and the lack of a G2-M-phase transition of these cells. The phosphorylation of p27 at threonine 187 by cdk2 serves as a signal for the ubiquitin-mediated degradation of this inhibitor (46). Consistent with the lack of cdk2 activity in the C/EBPβ-null LE cells, we observed that the level of phosphorylation of p27 at Thr187 was reduced in these cells compared to WT cells, as indicated by IHC and Western blotting (Fig. 7C and D).
We also detected an increased level of p21 mRNA in C/EBPβ-null LE cells relative to WT cells at all times following E treatment (Fig. (Fig.7E).7E). Of particular note was the more-than-2-fold-higher level of expression of p21 mRNA (Fig. (Fig.7E)7E) and a marked enhancement in the p21 protein level in the mutant epithelium compared to WT tissue at 18 h after E treatment (Fig. (Fig.7F).7F). This increased level of expression of p21 in C/EBPβ-null epithelial cells is likely to contribute to the maintenance of the G1-S-phase arrest.
To further investigate the gene networks underlying C/EBPβ function during the E-induced proliferation of uterine LE cells, we compared the gene expression profiles of WT and C/EBPβ-null uterine LE cells. Briefly, ovariectomized mice were treated with E, and after 18 h, the LE cell layer was isolated and total RNA was subjected to DNA microarray analysis using Affymetrix murine arrays as described in Materials and Methods. The results indicated that the expressions of 620 genes were upregulated and that the expressions of 116 genes were downregulated greater than 2-fold in C/EBPβ-null uterine LE cells compared to WT cells (C. R. Ramathal and M. K. Bagchi, unpublished results). Using gene ontology analysis, we identified a subset of genes with well-established roles in the DNA damage response among those markedly upregulated in LE cells lacking C/EBPβ. These genes encoded factors regulating cell cycle checkpoints (such as Atm, Atr, Birk5, Skp2, and Cdc20), DNA repair (such as Rad18, Rad51, Rad54B, Rpa1, Mre11A, Exo1, and Xrcc5), and apoptosis (such as Casp3, Cdh2, and Tp73l).
A potential consequence of compromised DNA replication is the collapse of replication forks or the creation of aberrant structures at the replication forks, making the DNA prone to damage, such as single-strand or double-stand break formation (38, 57). These DNA lesions are detected by the ATM and ATR proteins, which are primary sensors and transducers of the DNA damage response (38, 57). These proteins signal downstream of the checkpoint kinases Chk1 and Chk2 and the tumor suppressor p53 in an attempt to correct the defect and remove damaged cells by triggering apoptosis (Fig. (Fig.8A).8A). An additional mechanism involves the recruitment of the E3 ubiquitin ligase Rad18 to stalled replication forks. Rad18, in partnership with Rad6, is thought to assist in the disassembly of the aberrant fork (36). Consistent with the results of our microarray analysis, real-time PCR confirmed the upregulation of mRNAs corresponding to Atm and Atr in C/EBPβ-null uterine LE cells (Fig. (Fig.8B).8B). The activation of the ATM and ATR kinases via unique phosphorylation events signals to the cell cycle checkpoint molecules that govern the G1-S-phase transition (38, 57). As shown in Fig. Fig.8C,8C, in E-treated WT uterine LE cells, ATM exists in an inactive nonphosphorylated form, indicating the absence of any DNA damage response. In contrast, prominent nuclear staining for the active phosphorylated (Ser1981) form of ATM was detectable throughout C/EBPβ-null uterine LE cells at 18 h following E treatment (Fig. (Fig.8C).8C). In addition, an accumulation of phosphorylated (Ser139) histone H2AX, a well-established marker of the ATM-dependent DNA damage response (8, 43), was observed at 18 to 20 h in C/EBPβ-null uterine epithelial cells (Fig. 8Ea and b). As further evidence of replication fork lesions, we observed an increased nuclear expression of Rad18 in C/EBPβ-null epithelial cells compared to WT cells (Fig. 8Ec and d).
We next assessed the occurrence of the DNA damage response downstream of ATM/ATR activation by monitoring other markers such as checkpoint kinase 1 (Chk1) and p53 in C/EBPβ-null epithelial cells. The phosphorylation of Chk1 at Ser296 by ATM indicates that the DNA damage response and checkpoint activation are initiated (38). The phosphorylation of p53 at the Ser15 residue by the activated ATM-Chk1 pathway causes p53 transcriptional activation (38). Western blotting experiments were performed by using soluble lysates prepared from LE cells obtained from uteri collected from WT and C/EBPβ-null mice following E treatment for 18 h. The data revealed a marked elevation in the levels of phosphorylated forms of Chk1 (Ser296) and p53 (Ser15) (Fig. (Fig.8D).8D). There was also a significant increase in the level of total p53 protein in the C/EBPβ-null LE cells relative to WT cells at 18 h of E treatment (Fig. (Fig.8D).8D). It was previously reported that a consequence of the upregulation and activation of p53 is an elevated level of expression of its transcriptional target, p21 (3, 38). We indeed saw a significant rise in the level of the p21 protein (Fig. (Fig.7E),7E), which presumably helps maintain cell cycle arrest in mutant epithelial cells. Collectively, our results established that the lack of C/EBPβ, which leads to a block in the E-mediated proliferation of uterine epithelial cells, triggers the ATM/ATR-dependent DNA damage response pathway and activates p53-mediated signaling that helps maintain the cell cycle block.
The phosphorylation of p53 at Ser15 by the ATM-Chk1/Chk2 pathway promotes its stabilization by preventing ubiquitination and degradation (1, 38). p53 is known to promote the expression of several proapoptotic genes, which in turn stimulate the mitochondrial pathways leading to caspase activation (43). We therefore investigated the possibility that the activation of an ATM/ATR/Chk1/Chk2 DNA damage response mechanism in C/EBPβ-null uterine LE cells, which results in the activation and accumulation of p53, commits these cells to enter programmed cell death. In order to assess the initiation of apoptosis in E-treated uterine LE cells of WT and C/EBPβ-null mice, we first examined the activation of caspase 3 (Casp3), a well-known marker of this process (27, 43). We performed an IHC analysis to determine the levels of Casp3 and its proteolytically cleaved active form (cl-Casp3) in these cells. Examination of the levels of full-length Casp3 at 48 h following E treatment showed a higher level of expression of this protein in the cytosolic compartment of C/EBPβ-null epithelial cells than in WT epithelial cells (data not shown). When we assayed for the presence of cl-Casp3 at the 48-h time point, an increased presence of this apoptosis marker was evident in C/EBPβ-null epithelial cells (Fig. (Fig.9A9A).
We next performed TUNEL staining to assess the extent of DNA degradation in apoptotic cells of uterine epithelial cells of WT and C/EBPβ-null mice. We detected a notable increase in TUNEL staining at discrete foci in LE cells of mutant mice compared to cells of WT mice at 48 h of E treatment (Fig. 9B and C). These data suggested that the arrest of E-stimulated C/EBPβ-null LE cells at the S phase of the cell cycle triggers the activation of p53-dependent mechanisms that increase their susceptibility to programmed cell death and affect their survival.
Our previous studies revealed that C/EBPβ is a major downstream target of E regulation in the uterus (28). The discovery that the uterine epithelium in C/EBPβ-null mice is nonreceptive to embryo implantation (28) prompted us to utilize this animal model to investigate the functional role of C/EBPβ and its downstream pathways in uterine epithelial biology. The studies presented here demonstrated that the absence of C/EBPβ results in a significant defect in the ability of uterine LE cells to proliferate in response to E stimulation. This is evidenced by a major impairment in DNA replication activity and multiple dysregulated cell cycle components in C/EBPβ-null uterine LE cells. We determined that C/EBPβ-null uterine LE cells are able to enter the cell cycle and progress through the G1 phase before being arrested in the S phase (Fig. (Fig.22 and and3).3). The ablation of the C/EBPβ gene therefore led to a complete loss of E-induced DNA replication and subsequent mitotic activity in uterine LE cells. We therefore uncovered a role for C/EBPβ as an essential mediator of E-induced DNA replication in uterine epithelial cells.
Previous reports documented that E administered to ovariectomized mice promotes the G1-to-S-phase transition of uterine LE cells (29, 50, 51). Previous studies by Tong and Pollard indicated that the E-induced nuclear localization of cyclin D1 is an essential event that permits the S-phase entry of uterine LE cells (50, 51). It was shown that in the presence of E, the activation and nuclear translocation of cyclin D1 occur via the inhibition of glycogen synthase kinase 3β (GSK-3β) activity. It was further proposed that P opposes the proliferative actions of E by allowing the phosphorylation of cyclin D1 by GSK-3β and the subsequent nuclear export of this cyclin. In the present study, we noted that the nuclear presence of cyclin D1 remained unaltered in S-phase-arrested LE cells of C/EBPβ-null uteri (Fig. (Fig.5A).5A). These results suggested that the C/EBPβ regulation of uterine LE cell proliferation occurs via a mechanism that is unrelated to the nuclear translocation of cyclin D1. Our findings are corroborated by data from previous reports showing that the nuclear entry of cyclin D1 occurs in the absence of C/EBPβ in the liver and mammary epithelium (11, 12).
A major finding of our study is that, upon E treatment, the level of cyclin E mRNA is notably lower in C/EBPβ-null LE cells than in WT LE cells and that the cyclin E protein failed to localize to the nucleus (Fig. (Fig.5B).5B). It is likely that the C/EBPβ-mediated entry of LE cells into the S phase of the cell cycle involves primarily the regulation of cyclin E-cdk2 function. During the S phase of the cell cycle, the active cyclin E-cdk2 complex is localized in the nucleus of WT LE cells and coordinates the assembly of DNA replication factors into the origins of replication (14, 47). In contrast, in C/EBPβ-null uteri, the combination of a reduced level of expression of cyclin E and its absence from the nucleus of LE cells would prevent its partnership with cdk2 and its S-phase-promoting activities. Consequently, inadequate cdk2 activation in C/EBPβ-null LE cells would also contribute to an S-phase blockade (Fig. (Fig.5A).5A). Tong and Pollard previously reported that P abolishes E-induced cyclin E-cdk2 activity in uterine epithelial cells (51). Data from that report are consistent with our observation that the simultaneous administration of P suppresses E-induced C/EBPβ expression, which in turn controls the expression and nuclear presence of cyclin E. Our finding is also consistent with data from a previously reported microarray study conducted using whole uteri of mice (17). It was reported that the acute administration of E stimulated the expression of cyclin E and cdk2 in this tissue, although the uterine cell type(s) in which these molecules are expressed remained unclear.
Previous reports indicated that the loss of C/EBPβ results in reduced cyclin E expression levels, reduced cdk2 functionality, and decreased proliferative activity in mammary epithelial cells and in hepatocytes during liver regeneration (11, 12). The precise mechanism by which C/EBPβ regulates the expression and nuclear localization of cyclin E is presently unknown. Several studies proposed that cyclin E is a target of transcriptional and posttranslational regulation by the E2F transcription factors, which are important regulators of the G1-S-phase transition and DNA replication (19, 21, 22, 34, 35, 39). Interestingly, among the E2F family members that we have examined, E2F3 is the most highly expressed in WT LE cells (Fig. 6Ad). It displayed reduced expression levels in C/EBPβ-null uterine epithelial cells (Fig. (Fig.6A),6A), while the levels of E2F1 and E2F2 remained unaltered. ChIP analysis demonstrated the recruitment of C/EBPβ to the proximal promoter region of E2F3, strongly suggesting that E2F3 is a primary target of C/EBPβ in uterine LE cells. We postulate that E2F3 regulates cyclin E expression, which in turn controls the entry of uterine LE cells into the S phase of the cell cycle.
We found that the CDK inhibitors p21 and p27 also play important roles in blocking E-induced cell cycle progression in C/EBPβ-null uterine LE cells. In G0 and early G1 phases, a high level of p27 is present in hormone-withdrawn uterine LE cells. p27 binds to the cyclin E-cdk2 complex and inhibits its activity. Upon E stimulation, p27 protein levels decline via ubiquitin-mediated degradation. In late G1 phase, p27 becomes a substrate of phosphorylation by the cyclin E-cdk2 complex. Phosphorylated p27 is released from the cyclin E-cdk2 complex, exits its nuclear location, and undergoes degradation, resulting in cyclin E-cdk2 activation (14, 42, 46, 47). Consistent with this scenario, we observed that E-stimulated WT uterine LE cells efficiently targeted p27 for ubiquitin-mediated degradation by phosphorylation, thereby depleting the nuclear p27 protein upon S-phase entry (Fig. 7C and D). In contrast, a robust nuclear accumulation of p27 was detected in C/EBPβ-null epithelial cells, coincident with the drastically reduced cyclin E level in the nuclear compartment and the cell cycle blockade in S phase (Fig. (Fig.7B).7B). Our results supported the view that the accumulation of the p27 protein arises from an inefficient phosphorylation of p27 by the cdk2-cyclin E complex and the consequent lack of a ubiquitin-mediated degradation of p27 in mutant LE cells (Fig. 7C and D). The nuclear retention of p27 suggests that it may interact with and inactivate any residual cyclin E-cdk2 or cyclin A-cdk2 complex that might form in the nuclei. We also noted a remarkable rise in the level of the p21 inhibitor at later stages of cell cycle arrest. The sustained presence of p27 and p21 inhibitors during early and late S phase maintains the blockade of cyclin-cdk2 activities, preventing uterine LE cells from completing DNA replication and entering mitosis.
The dysregulation of cyclin E was previously reported to cause the destabilization of prereplication complexes necessary for DNA synthesis (6). We postulate that a defect in cyclin E-cdk2-controlled replication fork assembly is encountered when E-stimulated C/EBPβ-null LE cells traverse through G1 phase and attempt to progress through S phase. We favor the hypothesis that the onset of a replication defect in E-stimulated C/EBPβ-null uterine LE cells triggers DNA damage checkpoint activation. In mammalian cells, stalled replication forks or replication-blocking lesions are typically recognized by the ATM/ATR kinase checkpoint pathways as DNA damage signals, triggering a series of well-defined molecular events culminating in cell cycle arrest and DNA repair (1, 10, 38, 43, 57). Stalled replication forks also recruit the ubiquitin ligase Rad18, a key component of the replication repair pathway (2, 36). In our studies, the ATM-dependent checkpoint pathway is activated in C/EBPβ-null uterine LE cells, as evidenced by the enhanced accumulation of active, phosphorylated forms of ATM kinase and its downstream targets, histone variant H2AX, Chk1, and p53 (Fig. (Fig.8).8). The increased presence of Rad18 in the nuclei of replication-arrested C/EBPβ-null epithelial cells further confirms the replication fork lesions in mutant LE cells. The activation of Chk1 via phosphorylation at Ser296 is known to affect S-phase progression (1, 10). Furthermore, the ATM/ATR-dependent phosphorylation of p53 at Ser15 leads to its activation and stabilization. Transcriptionally active p53 in turn promotes the synthesis of p21, which directly inhibits the cdk2-cyclin E complex (4, 14, 38, 47). In this manner, the activation of p53-dependent mechanisms downstream of the DNA damage response helps maintain the DNA replication arrest in hormone-stimulated C/EBPβ-null epithelial cells.
In addition to its mitogenic effects, E is also known to promote cell survival by acting as an antiapoptotic agent (31, 52). A block in E-induced uterine LE cell proliferation in the absence of C/EBPβ may trigger apoptosis, which is the culmination of a cascade of molecular steps, including cell cycle withdrawal and DNA fragmentation (27). We also considered apoptosis as a potential biological end point for S-phase-arrested C/EBPβ-null uterine LE cells since there is a strong connection between the ATM/p53-dependent checkpoint response and the removal of damaged cells via programmed cell death (38, 43). One of the essential players in apoptosis is the serine-protease Casp3, which plays a central role in coordinating a variety of downstream events leading to cell death (27, 43). Our studies revealed that a subpopulation of E-treated C/EBPβ-null LE cells blocked in S phase entered the apoptotic pathway, as evidenced by the increased presence of the cleaved, active form of Casp3 in these cells compared to WT LE cells (Fig. (Fig.9A).9A). Additionally, we noted a significant increase in cell death due to DNA fragmentation in C/EBPβ-null epithelial cells compared to WT tissue (Fig. 9B and C). The genetic deletion of C/EBPβ was previously linked to increased apoptosis and reduced tumorigenesis in skin keratinocytes (49, 58). It is likely that the majority of the C/EBPβ-null uterine LE cells undergo DNA repair and recovery, thereby preventing widespread apoptosis in the epithelium. In support of this concept, the levels of expression of several proteins associated with the DNA repair pathway such as Rad18, Rad51, Rad54B, Rpa1, Mre11A, Exo1, and Xrcc5 were elevated in uterine epithelial cells lacking C/EBPβ.
In conclusion, C/EBPβ is an early and critical mediator of the E-controlled proliferative response in uterine LE cells. The functional link between E and C/EBPβ provides a novel mechanism by which this hormone controls the expression, localization, and activity of specific cell cycle regulatory molecules, such as E2F3, cyclin E, p27, and p21, to influence DNA synthesis in LE cells (Fig. (Fig.10).10). The downstream actions of this key transcription factor are also indispensable for DNA repair and cell survival, allowing the preparation of a functional uterine epithelium for the establishment of pregnancy. Furthermore, the role of C/EBPβ as a mediator of proliferative as well as antiapoptotic effects of E in uterine epithelial cells presents it as a potential target for anticancer therapeutics for preventing E-dependent endometrial cancers.
We acknowledge Elizabeth Hunt, Katya Dribinsky, and Jarrad Marcell for genotyping.
This work was supported by the Eunice Kennedy Shriver NICHD/NIH through cooperative agreement U54 HD055787 as part of the Specialized Cooperative Centers Program in Reproduction and Infertility Research.
Published ahead of print on 19 January 2010.