|Home | About | Journals | Submit | Contact Us | Français|
The mouse endometrium harbours stem/progenitor cells that express the stem cell marker mouse telomerase reverse transcriptase (mTert).
We used a mouse carrying a transgenic reporter for mTert promoter activity to identify rare endometrial populations of epithelial and endothelial cells that express mTert.
Stem/progenitor cells are hypothesized to be responsible for the remarkable regenerative capacity of the endometrium, but the lack of convenient endometrial stem/progenitor markers in the mouse has hampered investigations into the identity of these cells.
A mouse containing a green fluorescent protein (GFP) reporter under the control of the telomerase reverse transcriptase promoter (mTert-GFP) was used to identify potential stem/progenitor cells in the endometrium. mTert promoter activity was determined using fluorescence microscopy and flow cytometry to identify GFP+ cells. GFP+ cells were examined for epithelial, stromal, endothelial, leucocyte and proliferation markers and bromodeoxyuridine retention to determine their identity. The endometrium of ovariectomized mice was compared to that of intact cycling mice to establish the role of ovarian hormones in maintaining mTert-expressing cells.
We found that mTert-GFP is expressed by rare luminal and glandular epithelial cells (0.3% of epithelial cells by flow cytometry), rare CD45− cells in the stromal compartment (0.028 ± 0.010% of stromal cells by microscopy) and many CD45+ leucocytes. Ovariectomy resulted in significant decrease of mTert-GFP+ epithelial cells (P = 0.029 for luminal epithelium; P = 0.034 for glandular epithelium) and a decrease in the percentage of mTert-GFP+ CD45+ leucocytes in the stromal compartment (P = 0.015). However, CD45− mTert-GFP+ cells in the stromal compartment were maintained in ovariectomized mice. This population is enriched for cells bearing the endothelial marker CD31 (10.3% of CD90− CD45− and 97.8% CD90+ CD45− by flow cytometry). CD45− mTert-GFP+ cells also immunostained for the endothelial marker von Willebrand factor. These results suggest that the endometrial epithelium and vasculature are foci of stem/progenitor activity and provide a system to investigate molecular mechanisms involved in endometrial regeneration and repair.
The stem/progenitor activity of endometrial mTert-GFP+ cells needs to be experimentally verified.
The identification and characterization of mTert-expressing progenitor cells in the mouse will facilitate the identification of equivalent populations in the human endometrium that are likely to be involved in endometrial function, fertility and disease.
This study was funded by National Health and Medical Research Council (NHMRC) of Australia grants (1085435, C.E.G., J.A.D.), 1021127 (C.E.G.), NHMRC Senior Research Fellowship (1042298, C.E.G.), the Victorian Infrastructure Support Program, U.S. National Institutes of Health grant R01 DK084056 (D.T.B.) and the Harvard Stem Cell Institute (D.T.B.). The authors have no conflicts of interest to declare.
The endometrium is the highly regenerative lining of the uterus and undergoes ~400 cycles of shedding and regeneration during a woman's reproductive years. Each month 7–10 mm of endometrial tissue is generated, making the endometrium one of the most cyclically regenerative tissues in the human body. The endometrium is essential for implantation and insufficient regeneration results in infertility (Gargett et al., 2012). Conversely, over exuberant or inappropriate regeneration of the endometrium can cause endometrial hyperplasia, cancer and endometriosis (Gargett, 2007). The regenerative ability of the endometrium has been attributed to endometrial mesenchymal stem/stromal cells (eMSCs) and epithelial progenitors (Gargett et al., 2012).
Evidence for endometrial stem/progenitor cells was demonstrated in cell cloning studies of human endometrial single cell suspensions. Individual colonies generated in vitro at very low seeding densities (5–20 cells/cm2) undergo differentiation, and possess the ability to reconstitute endometrium in vivo (Masuda et al., 2012; Gargett et al., 2009). The most studied of these is the eMSC population that expresses a range of MSC markers, including CD90 and CD44 (Gargett et al., 2009). Studies of mouse endometrium have identified CD44+ tissue reconstituting epithelial progenitors (Janzen et al., 2013) and label retention studies have identified slow cycling epithelial and stromal cells that participate in endometrial regeneration (Chan and Gargett, 2006; Chan et al., 2012; Cervello et al., 2007). However, the lack of a directly detectable genetic marker for mouse endometrial stem/progenitor cells has hindered progress in understanding the cell biology of endometrial regeneration. To address this shortcoming, we have investigated mouse telomerase reverse transcriptase (mTert) expression as a marker for stem/progenitor cells in the mouse endometrium.
mTert is part of the telomerase complex that maintains telomere length and allows cells to undergo repeated rounds of division without succumbing to senescence triggered by loss of telomeres (Blackburn, 1990). Telomerase reverse transcriptase is the catalytic and rate limiting component of telomerase and is expressed by stem/progenitor cells (Blackburn, 2005). Transgenic reporter constructs driven by the mTert promoter mark stem/progenitor cells in the bone marrow, the intestine and testis (Breault et al., 2008; Montgomery et al., 2011). Slow cycling label-retaining cells in the intestine also express mTert-green fluorescent protein (GFP) (Breault et al., 2008). Telomerase activity has been reported in the endometrium (Hapangama et al., 2008) and the successful use of mTert expression as a stem cell marker in the examples cited above suggests that mTert expression may be a useful marker of stem/progenitor cells in the endometrium. Here, we examine whether mTert promoter activity identifies stem/progenitor cells in the mouse endometrium.
mTert-GFP reporter mice on a C57BL/6J background were bred in the Monash Medical Centre Animal Facility. These mice contain a transgene with the mTert promoter driving the expression of green fluorescent protein (GFP). mTert-GFP expression identifies hematopoietic stem cells and intestinal epithelial stem cells and mice carrying the transgene display normal behaviour and fertility (Breault et al., 2008). C57BL/6J mice were used as a control in flow cytometry experiments. All mouse studies were approved in advance by the Monash Medical Centre Animal Ethics Committee, according to the National Health and Medical Research Council of Australia guidelines.
Uteri from mTert-GFP mice were immersion fixed in 4% w/v paraformaldehyde in phosphate-buffered saline (PBS) at room temperature for 3–5 h or at 4°C overnight, cryoprotected with 30% w/v sucrose in PBS overnight at 4°C, frozen in Optimal Cutting Temperature medium (Sakura Finetek, Netherlands) and cryosectioned with a Leica CM1850 Cryostat (Leica Microsystems GmbH, Wetzlar, Germany). GFP was detected by its endogenous fluorescence unless stated otherwise. Sections were stained with primary antibodies (Table (TableI)I) for 1 h and visualized either directly or with secondary antibodies (Supplementary data, Table S1). Isotype control antibodies were used to confirm the specificity of CD44, CD45, CD90 and EpCAM immunostaining (Supplementary data, Fig. S1). Nuclei were counterstained with 5 µg/ml Hoechst 33258 (Molecular Probes) and images captured using a Nikon C1 confocal microscope (Nikon, Tokyo, Japan) with a 20× air or 60× oil objective. Captured images were adjusted for brightness and contrast in a linear manner using FIJI (Schindelin et al., 2012).
Post-natal bromodeoxyuridine (BrdU) loading was as previously described (Chan and Gargett, 2006). Briefly, female mTert-GFP pups were injected twice daily with 100 µg of BrdU (Sigma-Aldrich, St. Louis, MO) in 50 µl PBS on post-natal Days 3–5 via an intraperitoneal route. Mice were killed at 9 weeks of age and uteri prepared for fluorescence microscopy as described above. To preserve GFP fluorescence, BrdU was detected using treatment with 500 U/ml DNaseI (Worthington Biochemical, Lakewood, NJ, USA) in 40 mM Tris/HCl pH 7.9, 10 mM NaCl, 6 mM MgCl2 and 10 mM CaCl2 for 10 min at room temp, then detected with sheep anti-BrdU (1:100, M20105S, Meridian Life Science, Memphis, TN USA) with chicken anti-sheep AlexaFluor 647.
Vaginal smears were stained with toluidine blue and the oestrous stage determined on the basis of epithelial cell morphology and the absence/presence of leucocytes (Byers et al., 2012). For ovariectomy, mice were anaesthetized with ketamine/xylazine, ovariectomized via a dorsal incision and the endometrium allowed to regress for at least 2 weeks on a soy-free diet to ensure no dietary phytoestrogens contributed to observed responses (Specialty Feeds, Glen Forrest, WA, Australia).
Longitudinal sections of mTert-GFP uterine horns were examined for GFP+ cells in the epithelial layer or immunostained with CD45-allophycocyanin (APC) (eBioscience) or CD45-phycoerythrin (PE) and CD44-APC as described above. Ten 461 × 461 µm fields sampling the length of the uterus were captured for each mouse. For quantification of GFP and CD45, FIJI (Schindelin et al., 2012) was used to calculate luminal epithelial length and stromal area in each section by manual tracing. The number of nuclei in the defined stromal area was calculated in FIJI using threshold, watershed and analyse particle functions. The area occupied by CD45 immunostaining in the stromal regions was calculated using the threshold and measure area functions. Epithelial mTert-GFP+ and CD44+ cells, mTert-GFP+ CD45− stromal cells, and mTert-GFP+ CD45+ stromal cells were counted manually. Data were analysed using GraphPad PRISM (GraphPad Software, La Jolla, USA).
Fresh uteri were dissected from female mTert-GFP mice of mixed oestrous cycle (8–12 weeks old, n = 12 as indicated in legends) or age matched GFP negative C57BL6 wild type controls. Uteri for each group were pooled, finely minced and enzymatically digested with 0.5% w/v collagenase Type I (Worthington Biochemical) in PBS with 5 mM glucose and 20 µl/ml deoxyribonuclease type I (Worthington) at 37°C for 1 h on a rotator. Dissociated tissues were then filtered through a 40-µm cell strainer (BD Bioscience), centrifuged at 230g for 5 min at 4°C and resuspended in 1% v/v fetal bovine serum (FBS) (Life Technologies) in PBS (FBS/PBS).
Cells were incubated with directly conjugated primary antibodies (Table (TableII).II). Antibody stained cells were washed and resuspended in 1% FBS/PBS. Fluorochrome-conjugated isotype controls and unlabelled controls were included for each antibody and were used for flow cytometry gating (Table (TableII).II). Flow cytometry analysis was performed using a FACSCanto II flow cytometer with FACSDiva Software (BD Biosciences, Le Pont-de-Claix, France).
Statistical analysis was performed with GraphPad Prism v6 (GraphPad Software) and data were analysed with an unpaired two-tailed t-test. Differences were considered statistically significant when p < 0.05. Data are presented as mean ± SEM.
In adult female cycling mice, mTert-GFP marked minor subpopulations in the endometrial stromal compartment and rare epithelial cells in luminal (Fig. (Fig.1A1A and B) and glandular epithelium (not present in Fig. Fig.1,1, but shown in Fig. Fig.4E).4E). Immunostaining with anti-GFP colocalized with GFP fluorescence and verified the GFP+ status of these stromal (Fig. (Fig.1C)1C) and epithelial cells (Fig. (Fig.1D).1D). The morphology of stromal mTert-GFP+ cells ranged from rounded to flattened with multiple projections and the intensity of GFP detected varied greatly suggesting that the stromal mTert-GFP+ population is heterogeneous (Fig. (Fig.1C).1C). Flattened mTert-GFP+ cells were observed lining voids at the junction of the endometrium with the myometrium (Fig. (Fig.11A).
A BrdU retention assay was used to identify endometrial label-retaining cells as previously reported (Chan and Gargett, 2006) and determine their relationship to mTert-GFP+ cells. Post-natal administration of BrdU and an 8 week chase produced label-retaining cells in the stromal compartment of the endometrium and the myometrium (Fig. (Fig.2A2A and B). BrdU retention and mTert-GFP were mutually exclusive in stromal cells as judged by the examination 309 mTert-GFP+ cells and 119 BrdU-retaining cells from the endometrium of three mice. Epithelial BrdU-retaining cells were not detected, possibly due to the length of the chase period, and the reduced sensitivity of the DNase-mediated BrdU detection protocol employed to preserve GFP fluorescence, as opposed to HCl used previously (Chan and Gargett, 2006; Cervello et al., 2007; Chan et al., 2012).
We examined whether mTert-GFP+ cells were actively proliferating in adult cycling endometrium, a property that might account for their failure to retain BrdU. The nuclear proliferation marker Ki67 immunostained stromal cells (Fig. (Fig.2C)2C) and many epithelial cells (Fig. (Fig.2D2D and E). mTert-GFP expression did not correlate with Ki67 immunostaining in epithelial or stromal cells (Fig. (Fig.22C–E).
Epithelial mTert-GFP+ cells were detected based on their expression of the epithelial marker EpCAM which is not expressed by leucocytes, vascular or stromal cells. Overlapping immunostaining of EpCAM and GFP was detected infrequently in the luminal epithelium by immunofluorescence microscopy (Fig. (Fig.3A).3A). Flow cytometry of dissociated uterus was used to examine EpCAM and GFP co-expression. A population negative for both CD31 and CD90 was examined to exclude endothelial cells, stromal cells and many leucocytes (black box, Fig. Fig.3B).3B). A small population of epithelial mTert-GFP+ cells (0.3% of EpCAM+) was detected in mTert-GFP uterus (Fig. (Fig.3C3C and D).
The influence of ovarian hormones on endometrial epithelial mTert-GFP expression was examined in hormonally intact mice with an oestrous cycle and hormonally deprived ovariectomized mice. No differences in epithelial mTert-GFP expression were detected between diestrus, pro-oestrus, oestrus and metestrus phases (data not shown). Mice undergoing an oestrous cycle had significantly higher numbers of mTert-GFP+ luminal and glandular epithelial cells than ovariectomized mice (Fig. (Fig.3E3E and F), demonstrating a role for ovarian hormones in promoting epithelial mTert-GFP expression.
mTert-GFP+ epithelial cells in the luminal epithelium did not immunostain for the putative epithelial progenitor marker CD44 (Fig. (Fig.4A).4A). The absence of CD44 in luminal epithelium was confirmed by carefully examining ten 461 × 461 µm fields from CD45 co-immunostained (to exclude intraepithelial leucocytes) sections of the uterus from six mice in oestrus stage. All CD44+ cells detected in the luminal epithelium were CD45+ intraepithelial leucocytes (Fig. (Fig.4B–D).4B–D). mTert-GFP+ and mTert-GFP− glandular epithelial cells both immunostained for CD44 (Fig. (Fig.4E).4E). Examination of mice in oestrus stage (n = 6) showed that glands were composed entirely of either CD44+ or CD44− cells, or a combination of CD44+ and CD44− cells. On average, 1.34±0.75 epithelial cells per gland immunostained for CD44. CD44 was abundant in the endometrial stroma (Fig. (Fig.4A,4A, B and E) as examined in more detail below (Fig. (Fig.7C7C and Fig. Fig.88D).
mTert-GFP expression marks bone marrow haematopoietic stem cells and is retained in myeloid and lymphoid lineages present in the endometrium (Breault et al., 2008). Immunofluorescence labelling for the pan-leucocyte marker CD45 revealed that the majority of stromal mTert-GFP+ cells were CD45+ and a small proportion were CD45− (Fig. (Fig.55A–D).
The endometrium is an immunologically active site and the recruitment of immune cells and accompanying stromal oedema are key components of endometrial regeneration during the oestrous cycle. We examined stromal mTert-GFP+ cell abundance by fluorescence microscopy throughout the oestrous cycle but did not detect any differences between the diestrus, pro-oestrus, oestrus and metestrus phases (data not shown).
We then assessed the role of ovarian hormones in the recruitment of mTert-GFP+ and CD45+ cells in the endometrium by comparing mice at the oestrus stage of the oestrous cycle and hormonally deprived ovariectomized (non-cycling) mice (Fig. (Fig.6).6). Compared with oestrus phase endometrium, hormonally deprived endometrium had a significantly higher number of nuclei per unit area of stroma (mean 9500 ± 1260 cells/mm2 versus 15 900 ± 1370 cells/mm2, P = 0.009; Fig Fig6A–C)6A–C) and a reduced area of CD45 immunostaining (mean 24 ± 3.4% versus 10 ± 1.1%, P = 0.005; Fig. Fig.6A,6A, B and D). These results indicate reduced oedema and reduced leucocyte infiltration, respectively, in hormonally deprived endometrium. Compared with oestrus phase endometrium, hormonally deprived endometrium had a lower percentage of mTert-GFP+ CD45+ leucocytes in the total stromal population (mean 0.35 ± 0.071% versus 0.12 ± 0.027%, P = 0.015; Fig. Fig.6E).6E). No significant difference was detected in the percentage of mTert-GFP+ CD45− non-leucocytes in oestrus versus ovariectomized (mean 0.03 ± 0.010% versus 0.02 ± 0.009%, P = 0.639; Fig. Fig.66F).
CD45− mTert-GFP+ cells that immunostained for the endometrial stromal marker CD90 were observed (Fig. (Fig.7A)7A) as well as CD45− mTert-GFP+ cells that did not immunostain for CD90− (Fig. (Fig.7B).7B). CD90 was also expressed by CD45+ leucocytes (Fig (Fig7A).7A). The MSC marker CD44 immunostained CD45+ leucocytes and some glands as described by Janzen et al. (2013), but was not detected on CD45− cells in the stromal compartment (Fig. (Fig.77C–F).
Endometrial mTert-GFP+ cells were further examined by flow cytometry for the presence of CD90 and CD44 (Fig. (Fig.8).8). Characterization of the non-epithelial (EpCAM−) mTert-GFP+ stromal population (Fig. (Fig.8A8A and B) confirmed that the majority of mTert-GFP+ cells in the stromal compartment were CD45+ leucocytes, many of which were also CD90+ or CD44+ (Fig. (Fig.8C8C and D). A subpopulation of mTert-GFP+ EpCAM− CD45− cells was also detected. These were either CD90+ (5.8% of mTert-GFP+ EpCAM−) or CD90− (2.2% of mTert-GFP+ EpCAM−) (Fig. (Fig.8C).8C). CD44 was detected on CD45+ mTert-GFP+ cells but not on CD45− mTert-GFP+ cells (Fig. (Fig.88D).
The presence of GFP+ cells lining voids at the endometrial/myometrial junction (Fig. (Fig.1A)1A) raised the possibility that mTert-GFP marks a subpopulation of endothelial cells in the uterus. CD45− cells with a flattened morphology immunostained for the endothelial marker von Willebrand factor (Fig. (Fig.9A–C).9A–C). The endothelial cell surface marker CD31 was used to further explore the phenotype of rare CD90+ mTert-GFP+ and CD90− mTert-GFP+ cells by flow cytometry (Fig. (Fig.9D–F).9D–F). A small proportion of mTert-GFP+ CD90− CD45− cells were CD31+ (10.3%) (Fig. (Fig.9E),9E), but the majority of mTert-GFP+ CD90+ CD45− cells were CD31+ (97.8%) (Fig. (Fig.9F).9F). A substantial mTert-GFP+ CD90− population were neither CD45+ nor CD31+ (Fig. (Fig.9E),9E), indicating that these cells are not leucocytes or endothelial cells, and may be epithelial cells described in Fig. Fig.33.
For the first time, we have demonstrated that a transgenic mTert-GFP reporter allows the detection and analysis of candidate endometrial stem/progenitor cells at the single cell level. This system overcomes the technical shortcomings of immunolocalizing mTert (Wu et al., 2006) and facilitates analysis using flow cytometry. Our study shows that mTert promoter activity identifies a heterogeneous endometrial population, including cells from epithelial, haematopoietic and endothelial lineages. Details of the mTert-GFP+ populations identified in the mouse endometrium are summarized in Fig. Fig.1010.
mTert-GFP expression did not correlate with BrdU retention or the proliferative marker Ki67 in the endometrial epithelium and stroma, indicating that endometrial mTert-GFP+ cells are distinct from previously described slow cycling (label-retaining) cells (Chan and Gargett, 2006). This suggests that mTert-GFP+ endometrial cells do not incorporate BrdU during the post-natal labelling window and indicates a limitation of the label retention method of identifying putative stem/progenitor cells.
Rare mTert-GFP+ cells immunostained with EpCAM were observed in the luminal and glandular epithelium of the endometrium. The presence of epithelial mTert-GFP+ cells is in line with reports of epithelial progenitor cells with stem-like properties such as clonogenicity and tissue regeneration (Janzen et al., 2013; Gargett et al., 2009). CD44 has been identified as a marker of mouse endometrial epithelial cells with an enhanced ability to reconstitute glands in an in vivo assay (Janzen et al., 2013). However, we did not observe a relationship between mTert-GFP and CD44 in endometrial epithelial cells. The immunolocalization of CD44 in mTert-GFP+ epithelial cells appeared to depend on whether they were luminal or glandular. Using the same CD44 antibody as Janzen et al. (2013), we found no evidence for CD44 immunolocalization in the luminal epithelial layer aside from CD45+ intraepithelial leucocytes. In contrast, there were many CD44+ glandular epithelial cells including some that were mTert-GFP+. Unlike CD44+ epithelial progenitor cells (Janzen et al., 2013), epithelial mTert-GFP+ cells were depleted by hormonal deprivation. Thus, mTert-GFP+ epithelial cells do not appear to overlap with the CD44+ putative progenitor population described by Janzen et al. (2013) and may represent a different stage of the endometrial stem/progenitor hierarchy.
CD45+ mTert-GFP+ cells from the haematopoietic lineage were the most abundant non-epithelial GFP-labelled population in the endometrium. This result is not unexpected as the endometrium is an immunologically active tissue and haematopoietic stem cells and their myeloid and lymphoid progeny are known to express mTert-GFP (Breault et al., 2008). Bone marrow-derived cells, including those of the CD45 lineage, have been reported to give rise to endometrial stroma in the mouse and human. The concept that bone marrow stem cells can cross lineage barriers to form non-haematopoietic cells types (transdifferentiation) has been studied in a range of organs and is contentious. If there is a genuine bone marrow-derived contribution to the endometrial stroma and epithelium, it is small and accounts for only a fraction of the profound regenerative potential of the endometrium (reviewed in Gargett et al., 2015). Thus, we suggest that CD45+ mTert-GFP+ haematopoietic lineage cells detected in the endometrium are predominantly transient immune cells recruited from the bone marrow rather than an intrinsic part of the endometrium involved directly in regeneration. Expression of the mTert-GFP reporter used in this study accurately reflects telomerase activity in CD45+ bone marrow cells (Breault et al., 2008) and human telomerase reverse transcriptase expression and telomerase activity are also elevated upon T cell activation (Hathcock et al., 1998; Chebel et al., 2009). As such, bulk measures of mTert expression and telomerase activity in the endometrium probably reflect levels of leucocyte infiltration and possibly activation rather than intrinsic endometrial stem cell or progenitor activity (Kyo et al., 1997; Kim et al., 2007; Hapangama et al., 2010). Leucocyte abundance in the endometrium, and probably overall endometrial telomerase activity, is modulated by ovarian hormones (Salamonsen and Lathbury, 2000). Thus, we raise the possibility that changes in endometrial telomerase activity observed at the whole tissue level during the human menstrual cycle may reflect leucocyte abundance and type rather than endometrial stem/progenitor cell expansion or activity.
While bone marrow-derived leucocytes are the most likely source of CD45+ mTert-GFP+ cells in the endometrium (Givan et al., 1997; Lee et al., 2015), haematopoietic stem cell activity has been reported in the mouse endometrium (Sun et al., 2015). Uterine resident haematopoietic stem cells could also contribute to the population of CD45+ mTert-GFP+ cells in the endometrium.
mTert-GFP+ cells that are not of the haematopoietic lineage (CD45−) are rare in the stromal compartment and may represent an intrinsic part of the endometrium with roles in tissue maintenance and regeneration. This is the case in the mouse intestine where rare mTert-GFP+ cells can give rise to a range of differentiated cell types (Montgomery et al., 2011). Further studies are required to determine whether a similar system operates in the endometrium. Unlike mTert-GFP+ leucocytes, the abundance of non-haematopoietic mTert-GFP+ cells in the endometrium was resistant to depletion in the absence of ovarian hormones. The atrophied inactive endometrium of ovariectomized mice and post-menopausal women retains its regenerative potential indicating the presence of persistent stem/progenitor populations (Paulson et al., 2002; Gargett and Chan, 2006; Gargett et al., 2012; Ulrich et al., 2014). The observation that the CD45− mTert-GFP+ population is not dependent on ovarian hormones is consistent with the concept that it contains a reserve population of endometrial stem/progenitor cells that survives hormonal deprivation in inactive endometrium (Paulson et al., 2002; Gargett and Chan, 2006; Janzen et al., 2013; Ulrich et al., 2014).
Studies of human endometrium have identified mesenchymal stem/stromal cells (MSC) (Schwab and Gargett, 2007; Schwab et al., 2008; Masuda et al., 2012; Gargett et al., 2015). MSC have been reported to express telomerase reverse transcriptase and have telomerase activity, although levels are low compared to hematopoietic stem cells (Liu et al., 2004; Serakinci et al., 2008). We sought evidence of mTert-GFP+ MSC in the mouse endometrium using the endometrial stromal/MSC markers CD90 and CD44. While CD90 marks endometrial stroma, it does not do so exclusively and CD90 is also expressed by some endothelial cells and leucocytes. Non-leucocyte mTert-GFP+ CD90+ cells are also CD31+ and von Willebrand factor positive, suggesting that they are endothelial cells. The only non-epithelial mTert-GFP+ CD44+ cells detected were CD45+ leucocytes. A lack of endogenous CD44+ stromal cells may also be due to this MSC marker only being expressed on MSC as a consequence of in vitro culture (Qian et al., 2012). Thus, we conclude that mTert expression does not identify an endometrial MSC population in the mouse. This finding does not exclude the possibility that mTert-expressing progenitor cells give rise to an endometrial MSC population.
The existence of CD31+ and von Willebrand factor positive mTert-GFP+ endothelial cells in the endometrium is similar to reports of endothelial mTert expression in endothelial cells of the mouse heart (Richardson et al., 2012). Telomerase activity has been linked to the proliferative capacity of endothelial cells and their ability to avoid senescence (Erusalimsky and Skene, 2009). A high prevalence of endothelial cells has also been demonstrated in tissue reconstituting side population cells with telomerase activity from the human endometrium (Masuda et al., 2010; Cervello et al., 2011; Miyazaki et al., 2012). Angiogenesis is an essential component of endometrial repair and endothelial mTert-GFP+ cells may represent a progenitor population involved in this process (Fan et al., 2008). Telomerase reverse transcriptase expressing endothelial progenitor cells from the human endometrium may also play a role in the angiogenesis that facilitates the establishment and persistence of endometrial growth outside the endometrium as endometriotic lesions (Laschke et al., 2011). Human telomerase reverse transcriptase expression is up-regulated in cultured endothelial cells by estrogen (Grasselli et al., 2008). However, we did not observe a decline in the predominantly endothelial CD45− mTert-GFP+ population in the stroma of ovariectomized mice, suggesting that these mTert-GFP+ endothelial cells are a reserve population that is not reliant on ovarian hormones for survival.
The existence of mTert-GFP+ vascular and epithelial putative stem/progenitor populations in the uterus raises questions as to the relationship between these cell types. There is evidence of a stromal/mesenchymal origin for epithelial cells during epithelial repair and regeneration following parturition or induced endometrial shedding (Cousins et al., 2014; Huang et al., 2012; Patterson et al., 2013). However, this process does not occur during the normal oestrous cycle examined in this study (Huang et al., 2012; Patterson et al., 2013). Thus, a direct link between the endometrial mTert-GFP+ vascular and epithelial cells we observed is unlikely. Transient reactivation of telomerase activity has been reported in luminal progenitors of mammary epithelium and is thought to represent a telomere salvage mechanism (Kannan et al., 2013). A telomere salvage mechanism may account for the mTert-GFP+ epithelial cells we observe in the endometrial epithelium.
In summary, we have shown that mTert reporter expression identifies rare epithelial and endothelial populations in the mouse endometrium. This reporter system will allow the role of these previously uninvestigated putative progenitor populations to be examined in endometrial repair, regeneration and experimental models of endometriosis. Understanding the properties of mTert-expressing progenitor cells in the mouse will facilitate the identification of equivalent populations in the human endometrium likely to be involved in endometrial function, fertility and disease. Another important observation is that mTert expression, and most likely telomerase activity, is predominantly due to infiltrating immune cells in the endometrium. This shows that telomerase activity needs to be considered at the cellular level in several discrete endometrial subpopulations with distinct properties, rather than by whole tissue-based measures.
C.E.G. and D.T.B. conceived the project, C.E.G., D.T.B. and J.A.D. obtained funding; J.A.D., Y.R.O., C.E.G. and D.T.B. designed experiments; J.A.D. and Y.R.O. performed the experiments; J.A.D. and Y.R.O. analysed and assembled data wrote the paper with C.E.G. and D.T.B.; J.E.C., W.S.N.K., D.L.C., D.N.W. provided advice on experimental design and techniques and essential reagents; A.T. advised on experimental design for flow cytometry and contributed to analysis of data. All authors provided feedback on drafts of the manuscript and approved the final version.
This study was funded by National Health and Medical Research Council (NHMRC) of Australia grants (1085435, C.E.G., J.A.D.), 1021127 (C.E.G.), NHMRC Senior Research Fellowship (1042298, C.E.G.), the Victorian Infrastructure Support Program, U.S. National Institutes of Health grant R01 DK084056 (D.T.B.) and the Harvard Stem Cell Institute (D.T.B).
The authors have no conflicts of interest to declare.
The authors acknowledge Dr Camden Lo (Monash Micro Imaging) for assistance with microscopy and quantification, Lesley Wiadrowski (Hudson Institute Histology) for assistance with histology and Dr Fiona Cousins for critically reading the manuscript.