SEARCH

SEARCH BY CITATION

Keywords:

  • Mouse myometrium;
  • Stem cell niche;
  • Label-retaining cells;
  • Mesenchymal stem cell;
  • Myometrial stem cell

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Conditional deletion of β-catenin in the Müllerian duct mesenchyme results in a degenerative uterus characterized by replacement of the myometrial smooth muscle with adipose tissue. We hypothesized that the mouse myometrium houses somatic smooth muscle progenitor cells that are hormonally responsive and necessary for remodeling and regeneration during estrous cycling and pregnancy. We surmise that the phenotype observed in β-catenin conditionally deleted mice is the result of dysregulation of these progenitor cells. The objective of this study was to identify the mouse myometrial smooth muscle progenitor cell and its niche, define the surface marker phenotype, and show a functional response of these cells to normal myometrial cycling. Uteri were labeled with 5-bromo-2′-deoxyuridine (BrdU) and chased for up to 14 weeks. Myometrial label-retaining cells (LRCs) were observed in the myometrium and stroma throughout the chase period. After 12 weeks, phenotypic analysis of the LRCs by immunofluorescence demonstrated that the majority of LRCs colocalized with α-smooth muscle actin, estrogen receptor-α, and β-catenin. Flow cytometry of myometrial cells identified a myometrial Hoechst 33342 effluxing “side population” that expresses MISRII-Cre-driven YFP. Functional response of LRCs was investigated by human chorionic gonadotropin stimulation of week 12 chase mice and demonstrated sequential proliferation of LRCs in the endometrial stroma, followed by the myometrium. These results suggest that conventional myometrial regeneration and repair is executed by hormonally responsive stem or progenitor cells derived from the Müllerian duct mesenchyme.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

The embryonic Müllerian ducts (also known as the paramesonephric ducts) are derived from the coelomic epithelium in the bilateral urogenital ridges during differentiation of the bipotential gonad. In female embryos, the absence of Müllerian inhibiting substance (MIS) allows the Müllerian ducts to persist and to differentiate into the internal female reproductive tract structures [1]. After birth, the mesenchyme of the primitive uterine tube differentiates into two layers of the adult uterus: the endometrial stroma and the myometrial muscle layers [2]. By postnatal day 15, the myometrium is well developed and the endometrial glands are visibly coiled, comparable to those observed in the adult uterus.

The adult uterus undergoes repeated cycles of cellular proliferation and degeneration in response to hormonal signals during a normal mammalian reproductive life span. Although this process of remodeling is essential for reproduction, the molecular mechanisms involved are largely unknown, and little is known about the effects of cyclic hormonal regulation on the uterine myometrium. We have shown that conditional deletion of β-catenin in the Müllerian duct mesenchyme during early embryogenesis using Cre recombinase knocked into the MIS type II receptor locus results in progressive replacement of the uterine myometrium with adipose tissue [3]. β-Catenin has two roles: as an intracellular transcriptional cofactor of the canonical Wnt signaling cascade and as a structural adaptor protein linking cadherins to the actin cytoskeleton in cell-cell adhesion [4]. We speculate that the observed switch from a myogenic to an adipogenic cell fate is due to a disruption of β-catenin signaling in the lineage commitment pathway of a myometrial or stromal somatic stem/progenitor cell, resulting in transdifferentiation, or to a disruption of the cell-cell adhesion function of β-catenin, resulting in loss of niche architecture and subsequent adipogenesis. The progressive smooth muscle atrophy and resultant adipogenesis in the conditional β-catenin-deleted myometrium suggests that this cell fate switch is likely due to the dysregulation of a previously uncharacterized intrinsic myometrial maintenance and repair mechanism, presumably through perturbed stem or progenitor cell activity.

Somatic stem cells are a subset of normal tissue cells that, through asymmetric division, have the ability to self-renew and produce lineage-committed daughter cells responsible for tissue regeneration and repair [5]. Somatic stem cells within normal tissue niches remain quiescent until activation by injury or other stimuli, as described in skin and hair follicle [6, [7], [8]9], mammary gland [10, 11], intestine [5, 12, [13], [14], [15], [16]17], and other organs [18, [19], [20]21], including skeletal muscle [22, [23]24]. The somatic stem cell defining property of quiescence has been used to identify candidate stem cells through their ability to retain the nucleotide analog 5-bromo-2′-deoxyuridine (BrdU) (or 3H-thymidine) for long periods, whereas asymmetrically derived lineage-committed daughter cells dilute the BrdU label during rapid proliferation. These so called label-retaining cells (LRCs) have been demonstrated to correlate with somatic stem cells in various tissues [25, [26], [27], [28], [29], [30], [31], [32], [33]34], including the uterine endometrial epithelium and stroma [35], and have been used as a means of isolating somatic stem cells in tissues where stem cell surface markers have yet to be characterized. Another method used to identify stem or precursor cells is exemplified by Goodell et al. [36] and Welm et al. [11], who demonstrated that hematopoietic and mammary gland stem cells may be isolated based on their ability to efflux Hoechst 33342 dye through the ABC transporter Abcg2/Bcrp1. Identification of these “side population” (SP) cells has since been used to identify somatic and cancer stem cells from various tissues [37, [38], [39], [40], [41], [42], [43], [44], [45], [46]47]. BrdU retention and Hoechst dye efflux are two distinct techniques that can be used individually to identify candidate somatic stem cells.

We hypothesized that the observed β-catenin knockout phenotype of progressive muscle wasting to fat is the result of either a lineage commitment switch or a breakdown in the niche architecture of uterine myometrial progenitor cells. We speculate that these uterine cells may represent an intrinsic reservoir for myometrial tissue repair and regeneration during the repeated injury associated with the estrous cycle. Here, we show evidence that there is a stem/progenitor cell population in the mouse myometrium and stroma that is responsive to hormonally stimulated injury and repair.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Animal Housing and Estrous Staging

All protocols involving animal experiments were approved by the Institutional Animal Care and Use Committee at Massachusetts General Hospital. Animals were housed in a controlled environment at 24°C, 12-hour dark/light cycle, with food and water ad libitum. Vaginal smear cytology was used to determine stage of estrous cycle. Thirty microliters of sterile phosphate-buffered saline (PBS) was injected into the introitus of each mouse and used to lavage the vaginal cavity. Lavage contents were then placed on a slide and viewed by light microscopy by two independent observers to determine estrous stage of each mouse. The generation of mice with MISRII-Cre-driven conditional deletion of β-catenin has been reported previously [3].

BrdU Labeling and Human Chorionic Gonadotropin Stimulation

Pulse-chase experiments were repeated at least twice with similar results using postpubertal (20–25 g) virgin female CD1 mice from Charles River Laboratories (Wilmington, MA, http://www.criver.com) that were injected i.p. once daily with 250 μl of 1 mg/ml BrdU for 7 days during the pulse period. On day 7, two mice were sacrificed to determine the initial BrdU labeling at chase day 0. During the chase period, mice were sacrificed weekly for a total chase period of 14 weeks. Similar experiments were also performed in β-catenin conditionally deleted mice for a total chase period of 8 weeks. During chase weeks 11 and 12, animals in proestrus were injected with 10 IU human chorionic gonadotropin (hCG) i.p. and subsequently followed for 5 days to capture each stage of a complete estrous cycle. For short-term labeling experiments, mice were injected with 250 μl of 1 mg/ml BrdU i.p. and sacrificed 2 hours postinjection. Tissues were harvested and analyzed for BrdU-labeled nuclei by immunofluorescence.

BrdU Immunofluorescence

Mice were euthanized, and tissues were harvested. Uteri were fixed in 4% paraformaldehyde at room temperature for 1 hour, washed three times with PBS for 5 minutes, and incubated overnight in 15% sucrose solution. Sections were embedded in 7.5% gelatin/15% sucrose solution and frozen at −60°C in isopentane, and 8–12 μm sections were made. Sections were then dried for 30 minutes, and the gelatin was removed by washing the sections in warm PBS (37°C) for 5 minutes. Slides were next incubated in 2 N HCl for 30 minutes at 37°C. Acid was then neutralized with two 5-minute washes of 0.1 M sodium borate and two 2-minute washes of PBS. Sections were subsequently incubated for 30 minutes with normal horse serum blocking solution (2% horse serum, 1% bovine serum albumin [BSA], 0.1% Triton X-100, 0.05% Tween 20) to reduce nonspecific binding. Slides were then stained with Alexa Fluor 488-conjugated mouse monoclonal primary antibody (dilution, 1:20; Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) for 1 hour at room temperature in a humidified chamber. Nuclei were counterstained using 4′6′-diamidino-2-phenylindole (DAPI) (1:20,000)-impregnated Vectashield (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) mounting medium for microscopy using a Nikon 80i microscope with epifluorescence attachments (Nikon, Tokyo, http://www.nikon.com). Images were captured using the SPOT RT-KE Camera and Spot Advance software (Diagnostic Instruments, Sterling Heights, MI, http://www.diaginc.com). BrdU-labeled nuclei quantification was performed by setting the threshold level of each grayscale image to 60 using Adobe Photoshop software (Adobe Systems Inc., San Jose, CA, http://www.abode.com). Subsequently nuclei above that threshold were counted using ImageJ software (NIH) set to count particle sizes from 2 to 500 pixels and K-means clustering through the nucleus-counting plugin.

Multicolor Immunofluorescence

After gelatin removal as described above, antigen retrieval was performed by incubating the sections for 5 minutes in 1% SDS followed by two 5-minute PBS washes. For BrdU detection, slides were incubated in 2 N HCl for 15–20 minutes and then neutralized and blocked with horse blocking serum as described above. Primary antibodies were purchased from the following vendors and titered before use: CD24 (eBiosciences, San Diego, CA, http://www.ebioscience.com), c-Kit (R&D Systems Inc., Minneapolis, http://www.rndsystems.com), α-smooth muscle actin (αSMA), peroxisome proliferator-activated receptor γ2 (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), β-catenin (Abcam, Cambridge, MA, http://www.abcam.com), CD31/PECAM, Sca-1, CD45.1, CD34, CD44, CD45R/220, CD90.1, GR-1, Mac-1, Ter119, Pan-NK, and Tie2 (BD Biosciences, Palo Alto, CA, http://www.bdbiosciences.com); all primary antibodies were diluted in 1% BSA in PBS and incubated in a humidified chamber for 1 hour. Sections were then washed three times in PBS with 0.05% Tween 20. Secondary antibodies were as follows: Alexa Fluor 488 donkey-anti-goat and Alexa Fluor 568 goat-anti-rabbit (1:800; Molecular Probes) diluted in 1% BSA/1× PBS and incubated for 1 hour.

Confocal Microscopy

For the purpose of studying the colocalization of BrdU+ LRCs with other surface markers, immunofluorescent sections were examined and image files acquired with a Nikon TE2000 inverted scope and a PerkinElmer LCI spinning disk confocal system (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com).

Flow Cytometry

Uteri were dissected and washed in Hanks' balanced saline solution (HBSS)/2% fetal calf serum (FCS) while excess fat was trimmed under the dissecting scope. Each horn of the uteri was sliced open to reveal the endometrium, at which time the endometrium was removed by scraping with a scalpel. Myometria were then digested in 0.2% collagenase type II (in Dulbecco's modified Eagle's medium [DMEM]; Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com) with mild agitation in a 37°C water bath for 90 minutes. After digestion, myometria were washed in 1× PBS and poured into a 100 × 15 mm Petri dish, and the collagenase was inactivated with Ham's F-12 medium + 20% fetal bovine serum (FBS). After aspiration, 5 ml of clean 1× PBS was added, and the myometria were triturated to disperse the cells. Cells were then collected, spun down, resuspended in PBS, filtered through a 70-μm mesh (BD Falcon; Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com), spun down, and resuspended in 2 ml of ACK lysis solution (Invitrogen) on ice for 3 minutes. Ten milliliters of HBSS + 2% FBS was added, and cells were spun down, resuspended in DMEM + 2% FCS, counted, and subjected to flow cytometry. Cell sorting and analysis of side population cells was performed in the Flow Cytometry Laboratory of the Department of Pathology and the Center for Regenerative Medicine, Massachusetts General Hospital, according to its published protocols [46]. In brief, single-cell suspension myometrial cells were stained with 5 μg/ml of the preferential A-T intercalating Hoechst 33342 dye for 2 hours at 37°C, washed, and resuspended in PBS containing 2% fetal calf serum. Prior to cell sorting, 2 μg/ml propidium iodide was added; cells demonstrating uptake of this dye were deleted as nonviable cells. SP cells were identified and electronically gated based on their characteristic light-scatter properties and singular Hoechst 33342 red versus blue fluorescence emission pattern after excitation with 100 mW of 350–360 nm ultraviolet light on a Digital-Vantage Cell Sorter (Becton Dickinson). SP fluorescence emissions were directed toward a 610 nm dichroic filter and then captured simultaneously through both a 450 nm band-pass and a 675 nm long-pass filter on a linearly amplified fluorescence scale. Immunophenotypic characterization of SP cells was done with titered fluorescein isothiocyanate, phycoerythrin, or allophycocyanin monoclonal antibody conjugates of CD45 (LCA, Ly-5, Ptprc), CD117 (c-Kit, Steel Factor), and Ly-6A/E/Sca-1 (Becton Dickinson). Up to 106 cells were collected for each analysis. Both cell sorting and immunophenotypic characterization of SP cells involved computer gating on light scatter to omit debris 0.01%–0.2% of the characteristic SP tail presented in a Hoechst 33342 red versus blue bivariate plot.

MISRII-YFP Isolation and Differentiation

MISRII-Cre/YFP mice were generated as previously described [3]. Uteri were harvested, and single-cell suspensions prepared as described above. Cells were plated at equal density in 12-well plates on glass coverslips. On days 1 and 3, cells were stained for αSMA immunofluorescence as described above. YFP immunofluorescence was detected with an Alexa Fluor 647 antibody to green fluorescent protein from Molecular Probes. Z-Series images were made by confocal microscopy as described above.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Identification of Label-Retaining Cells in Myometrium and Stroma

Short-term (2-hour) BrdU labeling was performed to determine whether myometrial cells are mitotically active before embarking on long-term pulse-chase experiments. BrdU was detected throughout the uterus (Fig. 1A), including the longitudinal and circular smooth muscle layers of the myometrium during this short pulse period. BrdU label was also occasionally observed to colocalize with α-smooth muscle actin (arrowheads), indicating that these were newly differentiated muscle cells. Pulse-chase BrdU labeling of adult female mice was performed to identify a population of slowly replicating uterine LRCs. Mice were injected daily for 7 days in an effort to label the maximum number of cells possible over 1–2 mouse estrous cycles. At the end of the pulse period (chase day 0), we observed significant labeling of the endometrial epithelium and stroma, with scattered labeling in both the circular and longitudinal muscle layers of the myometrium (Fig. 1B). The chase phase of the experiment was carried out for 14 weeks, and two mice were sacrificed each week to evaluate BrdU label retention by uterine cells. Figure 1C shows that there is a dramatic reduction in number of BrdU-labeled nuclei in the epithelium after 19 days of chase. On chase day 33, continued retention of BrdU was observed in the myometrium and endometrial stroma, with loss of labeling in the endometrial epithelium (Fig. 1D). At chase day 68, we observed LRCs in both the outer longitudinal and inner circular muscle layers of the myometrium, whereas the endometrial stroma showed occasional scattered LRCs and the endometrial epithelium showed no retention of label (Fig. 1E). During this late phase of the chase, we observed that many of the LRCs in the myometrium were adjacent to another LRC (Fig. 1E, arrowheads), suggesting that cell division could have occurred recently in one of the paired cells. At the end of the chase period, LRCs persisted in the myometrium and stroma, but the intensity level of detectable BrdU was much lower, with visually dimmer nuclei (Fig. 1F). Quantitative analysis of the number of BrdU-labeled nuclei (Fig. 1G–1I) shows that the total number of BrdU-labeled nuclei at the end of the chase period was only 10% of the number found before the chase and that the relative decrease in labeled nuclei was greater in the stromal and epithelial compartments. Figure 1J shows that whereas 60% of BrdU-labeled cells before the chase were found in the stromal and epithelial compartments, by the end of the chase period, only 30% of the BrdU-labeled cells remain there.

thumbnail image

Figure Figure 1.. Label-retaining cells (LRCs) in the mouse myometrium. Mice were injected daily with BrdU in short-term experiments for 2 hours or in long-term experiments for 7 days with a 14-week chase. The uteri were collected, fixed, and analyzed for BrdU labeling by immunofluorescence in frozen sections. (A): Short-term BrdU labeling in the myometrial muscle layers, which are shown stained with αSMA. Arrowheads show cells that colocalize with αSMA. Long-term BrdU labeling is shown in (B–F). (B): BrdU label after the pulse or chase day 0. The majority of the labeling is in the ES and the E, whereas there are a few BrdU-labeling (green) cells in the myometrium. (C–E): The BrdU LRCs on chase days 19, 33, and 68. The LRCs appear to be segregating in pairs, a few of which are indicated with arrowheads, and may represent a LRC and an asymmetrically derived daughter cell. (F) demonstrates a few remaining LRCs in the smooth muscle layers of the uterus by chase day 96. Scale bar = 100 μm. Nuclei were stained with 4′6′-diamidino-2-phenylindole. The total number of nuclei labeled with BrdU for the indicated days is shown in (G). The number of BrdU-labeled nuclei in the myometrial (H) and stromal/epithelial (I) compartments are also shown. After day 19, LRCs were not observed in the epithelial cells. (J): Percentage of total BrdU-labeled nuclei in the stroma and epithelial compartments. Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; E, epithelium; ES, endometrial stroma; MC, myometrium circular; ML, myometrium longitudinal; αSMA, α-smooth muscle actin.

Download figure to PowerPoint

Phenotypic Characterization of Label-Retaining Cells

To characterize the phenotype of the BrdU+ LRCs and distinguish smooth muscle LRCs from other known cell types in the myometrium, we analyzed LRCs 8–12 weeks into the chase for a series of known somatic stem cell, hematopoietic lineage, and skeletal muscle progenitor markers. In the myometrium, LRCs were largely found in cells expressing estrogen receptor-α (ERα) (Fig. 2A), suggesting that LRCs may be hormonally responsive, as would be expected if they played a role in myometrial regeneration during estrous cycling. We next assessed whether LRCs might have characteristics of stem or progenitor cells by testing for the presence of three common stem cell markers: Abcg2, a multidrug resistance transporter, the c-Kit receptor, and stem cell antigen-1 (Sca-1). Abcg2 was found throughout the myometrium (Fig. 2B) and stroma, often adjacent to an LRC, examples of which are shown in the higher resolution inset. Cells expressing c-Kit were also often found adjacent to an LRC but, unlike Abcg2, c-Kit+ cells were far fewer, often contained dim BrdU labeling (Fig. 2C, inset), and were largely found in the periphery of the smooth muscle bundles (Fig. 2C). Sca-1 expression was observed throughout the myometrium and stroma but was largely absent in LRCs (Fig. 2D). We also observed that cells expressing Sca-1 appeared to encircle the smooth muscle bundles in the longitudinal muscle layer.

thumbnail image

Figure Figure 2.. Colocalization of label-retaining cells (LRCs) with known stem cell and hematopoietic lineage markers. All images are from uteri after 8–12 weeks of chase. (A, B) demonstrate LRCs colocalize with ERα and Abcg2, respectively. Insets show higher-resolution images. (C, D) demonstrate the localization of two common stem cell markers c-Kit and Sca-1, respectively. The arrowheads in (C) indicate that c-Kit+ cells, although much fewer in numbers compared with Sca-1, were routinely observed adjacent to LRCs. The inset in (C) shows a confocal image of a BrdUdim c-Kit+ cell adjacent to a bright LRC. (E, F) demonstrate that LRCs do not normally colocalize with the hematopoietic cell markers CD45.1 and CD31. Scale bar = 100 μm. Nuclei were stained with 4′6′-diamidino-2-phenylindole. Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; ERα, estrogen receptor-α; ML, myometrium longitudinal.

Download figure to PowerPoint

Bone marrow-derived stem cells may also retain BrdU label for long periods [48] and have been found to contribute cells during injury and repair of target tissues [23]. The contribution of bone marrow-derived hematopoietic cells, if any, to smooth muscle cells of the myometrium has not been reported. To investigate whether LRCs were of endothelial or hematopoietic origin, we also analyzed sections for CD45.1 and CD31 (PECAM), markers commonly used to identify leukocytes and endothelial cells, respectively. We found that LRCs do not colocalize with CD45.1 (Fig. 2E) and rarely colocalize with CD31 (Fig. 2F) in cells that appeared to be part of the uterine vasculature. To rule out hematopoietic lineage further, we also evaluated for other hematopoietic markers, including pan-NK, Gr-1, Ter-119, Mac-1, CD34, Thy-1/CD90.1, CD44, and Tie-2, none of which colocalized with BrdU+ LRCs during this stage of the chase (Table 1).

Table Table 1.. Cell surface markers on uterine cells
Thumbnail image of

Label-Retaining Cells Are Found in Fat Cells of β-Catenin-Deficient Uteri

We hypothesize that aberrant regulation of myometrial stem/progenitor cells or the cells that form their niche, when conditionally deleted for β-catenin, results in the observed muscle-to-fat phenotype. Therefore, we investigated whether the presence of BrdU-label-retaining cells could be correlated with the switch to fat in mice conditionally deleted for β-catenin. In Figure 3A, we show a low-power-magnification image of a normal uterus stained with αSMA and BrdU antibodies to delineate the myometrium and identify the LRCs. Comparison with a β-catenin-deficient uterus stained for αSMA and BrdU (Fig. 3B) shows a dramatic loss of myometrial smooth muscle, consistent with our previous observations, and a concomitant decrease in LRCs. Higher magnification of the mutant uteri (Fig. 3C, 3D) shows that LRCs were observed in both smooth muscle cells and adipocytes, with a preponderance of dim BrdU nuclei in the adipocytes (arrowheads), suggesting that the absence of β-catenin perturbs the self-renewal mechanisms of these mutant stem or progenitor cells and releases them to divide at a greater frequency along a terminally differentiated adipocyte lineage.

thumbnail image

Figure Figure 3.. Comparison of BrdU labeling of αSMA cells in normal and β-catenin-deficient uteri. Analyses of BrdU labeling and αSMA expression by immunofluorescence were performed as described in Materials and Methods. BrdU-labeled cells were found in the endometrial stroma and in the cells expressing αSMA, both in the normal uterus (A) and in the β-catenin-deficient uterus (B). At higher magnification (C, D), dim BrdU-labeled cells, as well as brightly labeled BrdU cells, were readily observed in the fat cells (arrowheads). Nuclei are shown with 4′6′-diamidino-2-phenylindole. Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; αSMA, α-smooth muscle actin.

Download figure to PowerPoint

MISRII-Expressing Cells Are Enriched for Side Population

In an effort to confirm that a potential stem cell population may exist in the wild-type uterus, we turned to the well-accepted technique of flow cytometry SP analysis, which identifies cells that express multiple drug resistance (MDR) genes, such as the Bcrp1/Abcg2 locus. Stem cells are thought to harbor relatively lower cellular concentrations of Hoechst 33342 dye compared with other, similarly treated cells [40, 41, 43] because they express MDR genes with the ability to efflux lipophilic compounds. In Figure 2B, we show that BrdU-label-retaining cells are often found adjacent to cells expressing Abcg2, suggesting that it may be a marker for a post-LRC population. Thus, we next investigated whether the myometrium contains a population of cells with the stem cell-ascribed SP phenotype to confirm functional Abcg2 expression. Single-cell suspensions of myometrium were analyzed by flow cytometry for their ability to efflux Hoechst 33342 dye, and approximately 0.3% ± 0.2% of the total population were classified as SP cells (Fig. 4A). To remove confounding hematopoietic cell populations from our analysis, CD45+ and CD31+ cells were electronically gated out (Fig. 4B), revealing that only 0.17% ± 0.03% of the remaining cells were SP cells (Fig. 4C), all of which were verapamil-sensitive (Fig. 4D). Multiple analyses showed that approximately 0.2% of the cells could be classified as functional SP cells by flow cytometry.

thumbnail image

Figure Figure 4.. Mouse uterus myometrium contains a side population (SP) of cells. Myometrial cells were collected from the uteri of adult mice, separated with collagenase, and analyzed by flow cytometry. In (A), 0.3% ± 0.2% of the total cells were classified as SP by Hoechst dye efflux. (B) demonstrates that approximately 60% of the total population of myometrial cells is CD45/CD31 by fluorescence analysis. If the CD45+/Cd31+ cells are eliminated from the analysis, only 0.17% ± 0.03% of the remaining cells are SP (C). The sensitivity of the cells to verapamil was assessed to confirm the SP cells (D). MISRII-driven YFP-expressing cells were found in an easily gated population and are shown outlined in red in (E). Of the total YFP/CD45/CD31 cells (shown gated in black in [E]), approximately 0.2% ± 0.08% were found in the SP (F). Analysis of only the YFP+/CD45/CD31 cells shows that 0.45% ± 0.23% of those cells are found in the SP (G) and that these cells are verapamil-sensitive (H). Analysis of the total number of SP cells shows that 29% ± 8.4% are YFP+(I). All percentages are given ±SD from a minimum of three experiments. Abbreviations: FITC, fluorescein isothiocyanate; PE, phycoerythrin.

Download figure to PowerPoint

The progressive loss of smooth muscle and replacement with adipocytes in β-catenin-deficient MISRII-expressing myometrial cells [3] prompted us to speculate that with continuous normal estrous cycling, the lack of β-catenin resulted in stem/progenitor cell differentiation into fat. Therefore, we also investigated whether MISRII could be a marker for post-LRC SP cells. We isolated myometrial cells from mice expressing a floxed [stop] YFP gene controlled by MISRII-driven Cre expression and analyzed them for Hoechst 33342 exclusion by flow cytometry as described above. In Figure 4E, the YFP-expressing cells are found in an easily gated population of cells, which is shown outlined in red. SP analysis of YFP/CD31/CD45 cells (gated in black in Fig. 4E) demonstrated that 0.2% ± 0.08% of these cells could be classified as SP (Fig. 4F). SP analysis of YFP+/CD45/CD31 cells (Fig. 4G) demonstrated a greater than twofold enrichment (0.45% ± 0.22%) in the percentage of verapamil-sensitive (Fig. 4H) SP cells over that observed in the YFP/CD31/CD45 population. In addition, analysis of only the CD31/CD45 SP cells shows that nearly 30% are YFP+ (Fig. 4I), indicating that a subpopulation of SP cells of nonhematopoietic origin either express the MIS type II receptor or are derived from a cell that once expressed the MIS type II receptor.

MISRII-Cre-Driven YFP Cells Differentiate into Smooth Muscle Cells

We next tested whether the myometrial cells expressing YFP under the control of MISRII-driven Cre could be grown in tissue culture and induced to differentiate into smooth muscle cells. Cells were collected from YFP-expressing uteri and plated in 12-well plates. The following day, the cells were fixed in some of the wells and stained with DAPI to detect nuclei and αSMA antibody to detect smooth muscle (Fig. 5A). A few of the YFP-expressing cells were detected in a given field, but these cells did not colocalize with αSMA expression. After 3 days (Fig. 5B–5D), the remaining wells were nearly confluent, and most of the YFP-expressing cells were shown to colocalize with αSMA. These findings suggest that myometrial cells expressing MISRII-Cre, as visualized by YFP, and initially lacking αSMA differentiate into smooth muscle cells, confirmation that MISRII-Cre YFP is a myometrial smooth muscle precursor cell marker.

thumbnail image

Figure Figure 5.. YFP+ cells differentiate into αSMA-expressing cells. Myometrial cells were harvested from the uteri of mice expressing YFP under the control of the MISRII promoter and plated. The following day (A), some of the cells were fixed and stained for αSMA by immunofluorescence and YFP expression. After 3 days in culture, another set of wells was analyzed for YFP expression (B) and αSMA (C). (D): Merged images from (B) and (C) by confocal microscopy. Abbreviation: αSMA, α-smooth muscle actin.

Download figure to PowerPoint

hCG Stimulation Causes Proliferation of LRCs

To address whether LRCs, which express ERα (Fig. 2A), demonstrate a functional response to hormonal manipulation, we injected mice in proestrus that were in week 12 of the chase with 10 units of hCG to hyperstimulate their estrous cycle and characterized the changes in BrdU LRCs. A control mouse was sacrificed to determine the level of prestimulation BrdU (Fig. 6A), which was consistent with that shown in Figure 1C. Thereafter, a mouse was sacrificed daily for 5 days, and the stage in the estrous cycle was analyzed. BrdU immunofluorescence revealed that in estrus there is an initial proliferation of endometrial stromal cells (Fig. 6B) followed by proliferation of the myometrial LRCs during metestrus (Fig. 6C) and diestrus (Fig. 6D). By proestrus (Fig. 6E), the number of BrdU-labeled cells was visibly fewer in the endometrial stroma. As in earlier periods of the chase (Figs. 1D, 3B), we observed that after 24 hours, many of the BrdU-labeled cells of the hyperstimulated myometrium were adjacent to another, dimmer BrdU-labeled cell, suggesting that these LRCs were proliferating in response to hCG stimulation (Fig. 6F).

thumbnail image

Figure Figure 6.. Label-retaining cells (LRCs) proliferate with hCG-induced stimulation of the estrous cycle in week-12 mice. Mice were injected with BrdU daily for a 7-day pulse. On week 12 of the chase, mice in proestrus were injected with hCG and analyzed by immunofluorescence for BrdU during the following 5 days. (A) (−hCG), (B) (estrus), (C) (metestrus), (D) (diestrus), and (E) (proestrus) demonstrate that there is proliferation of BrdU LRCs in response to hyperstimulation induced by hCG. (B) demonstrates that during estrus there is proliferation in the ES. (C, D) demonstrate proliferation in both ES and in the myometrial layers. (E): BrdU is diminished in the ES, whereas in the myometrium, BrdU persists. Twenty-four hours after treatment with hCG, pairs of LRCs were observed, with one of the pair noticeably dimmer than the other (arrowheads). Scale bar = 100 μm (A–E) and 25 μm (F). Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; ES, endometrial stroma; h, hour(s); hCG, human chorionic gonadotropin; MC, myometrium circular; ML, myometrium longitudinal.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Cell lineage and early differentiation markers in the mouse myometrium are largely uncharacterized. Although much effort has focused on the events surrounding the endometrium during the estrous cycle and at pregnancy, little is understood about the normal mechanisms of repair and remodeling in the myometrium. Here, the identification of cells in the normal myometrium with stem/progenitor cell properties contributes to our understanding of the innate regenerative mechanisms used in the differentiation and maturation of myometrial cells throughout the reproductive life span of an individual.

We observed that cells expressing αSMA incorporate BrdU during a relatively short 2-hour labeling period. The fact that there are αSMA-expressing, BrdU-labeled cells indicates that the smooth muscle cells of the uterus or their terminally differentiating progenitors divide during the normal estrous cycle. Myofibroblasts or protomyofibroblasts, which express αSMA, have been proposed as such cells in other tissues that are involved in mechanical stress and force generation [49]. Our data suggest that these observed αSMA+/BrdU are not myofibroblasts, as they are contained within the bundles of smooth muscle that make up the uterine myometrium. Thus, we have shown for the first time that mouse myometrial smooth muscle layers contain cells that are mitotically active during the normal estrous cycle and that the normal uterus requires routine maintenance to preserve its structure and function.

To identify the precursor cells responsible for this observed intrinsic uterine maintenance and repair that we hypothesize is involved in the previously described muscle-to-fat switch, independent modalities for stem cell identification were used to identify two candidate populations of cells in the normal mouse myometrium and stroma. Although similar in that they both posses somatic stem cell-ascribed properties, there were some differences between the identified LRCs and SP cells. Founded on the somatic stem cell property of quiescence, a facet of asymmetric division, we used the LRC technique to identify small populations of myometrial and stromal LRCs in the mouse uterus. Stromal LRCs were observed near the endometrial epithelium, near the myometrial-stromal junction, and in a perivascular distribution, which has been described previously [35]. Myometrial LRCs were observed cells and at the periphery of the longitudinal muscle bundles, a location similar to that where Pax7-expressing satellite or stem cells in skeletal muscle are found [22]. Interestingly, myometrial BrdUhigh LRCs were observed by confocal microscopy to be located directly adjacent to BrdUdim c-Kit expressing cells (Fig. 3B). The “immortal strand” hypothesis, which states that newly synthesized DNA from parental stem cells segregates preferentially in the daughter cells during asymmetric division [50, 51], suggests that these BrdUdim c-Kit+ cells might be asymmetrically derived daughter cells of the LRCs. Thus, we speculate that the c-Kit+ cells may be originating from the LRCs after asymmetric division as a transient amplifying cell population for subsequent terminal differentiation. In short-term BrdU experiments, we did not observe colocalization of αSMA and c-Kit (data not shown), which suggests that the αSMA-expressing LRCs could be a differentiation product of the c-Kit-expressing LRCs or that these two different LRCs may represent separate progenitor cell populations. Likewise, we did not observe c-Kit colocalizing with adipocyte LRCs in the β-catenin knockout mice (data not shown), suggesting that adipocyte LRCs could be a differentiation product of the c-Kit-expressing LRCs and that these cells are lost in the absence of MISRII+/β-catenin+ cells. In addition, the observation that both the stromal and myometrial LRCs proliferated (Fig. 6) in response to estrous cycle stimulation strongly suggests that both of these LRC populations are functionally active putative stem/progenitor cells; however, additional studies are required to make definitive conclusions.

The SP technique, which has been used to identify both somatic and cancer stem cells, is based on the stem cell property of indefinite self-renewal and uses multidrug resistance transporters to efflux potentially noxious substances. Here, we demonstrated that the uterine myometrium harbors a functional nonhematopoietic verapamil-sensitive SP (Fig. 4). Phenotypic analysis of these SP cells demonstrated that they are Sca-1+/c-Kit/CD45/CD31 and directed us to the conclusion that these SP cells were perhaps downstream to the previously identified BrdUDim/c-Kit+ cells found adjacent to LRCs. Based on our original work with the MISRII-Cre β-catenin knockout mice, we investigated whether MISRII may serve as a stem/progenitor cell marker in the myometrium. Here, we demonstrate that there is enrichment of SP cells based on the presence of MISRII-YFP, suggesting that MISRII-YFP is marking a subset of SP cells and that these YFP+/SP+ cells may represent putative stem/progenitor cells that, in the absence of β-catenin, have the potential for adipogenesis. Based on the combination of findings from the short-term BrdU labeling and the LRC and SP techniques, we propose that the lineage differentiation of uterine myometrial cells may represent a continuum, with LRCs being the most primitive, followed by c-Kit+ and SP cells as transient amplifying cells, and finally terminal differentiation into myometrial cells that express αSMA (supplemental online Fig. 1).

There is in vitro and in vivo evidence that either of the two known biological functions of β-catenin, Wnt signaling and cell-cell junctions, may be involved in the muscle-to-fat cell fate switch [52, 53]. Disruption of the Wnt canonical signaling pathway induces adipogenesis and causes the transdifferentiation of myoblast into adipocytes by a mechanism that could involve the disruption of β-catenin-mediated inhibition of adipogenic transcription factors [53]. The other possibility is that β-catenin cell-cell junctions could play a role in the maintenance of the myometrial stem cell niche, and disruption of this niche may result in a cell fate switch from smooth muscle myogenesis to adipogenesis.

Based on our observations, we speculate that the LRCs found in the periphery of the smooth muscle bundles as described in this study function as myogenic “satellite” precursor cells for myometrial repair and that it is in these satellite cells in which the lack of β-catenin expression results in adipocyte differentiation. Although the myogenic satellite cells in skeletal muscle appear to be derived from the same somitic origin as embryonic skeletal muscle [54], we hypothesize that in the uterus, similar precursor cells are probably derived from an intrinsic source in uterine smooth muscle or its progenitor, the Müllerian duct mesenchyme, emanating from the intermediate mesoderm and urogenital ridge.

We have identified candidate smooth muscle precursors in the uterus; functional proof will require their isolation and differentiation into either smooth muscle or fat depending on their culture conditions. Since there is a smooth muscle differentiation deficit in the β-catenin conditionally deleted uteri, another possibility that should be considered is that the uteri contain latent preadipocytes that may also be derived from the Müllerian duct mesenchyme. It may be that these preadipocytes are now able to differentiate because the smooth muscle cell precursors required to make replacement muscle cells are no longer fully functional in the β-catenin mutant. At this point, we also cannot rule out the possibility that the myometrial cells involved in hormonally regulated regeneration or repair are derived from a stromal precursor, particularly since we have previously shown that the endometrial stroma is also derived from the Müllerian duct mesenchyme [3], and we show stromal LRC populations here. To continue investigating our hypothesis that abnormal activation of the myometrial stem cell population may play a role in uterine fibroid development, additional studies are needed to determine which of these possibilities might be functioning in this system.

Disclosure of Potential Conflicts of Interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

The authors indicate no potential conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Dr. Dennis Brown for assistance with confocal microscopy, Lankai Guo for instruction in staging mouse estrous, and Drs. Allan Goldstein and Nelson Arango for critical reviews of the manuscript. P.P.S. is supported by National Research Service Award T32 Training Grant in Cancer Biology T32CA071345. P.K.D. is supported by NIH Grants CA17393 and HD32112. This work was supported by generous contributions to the Pediatric Surgical Research Laboratories from the McBride and Austen funds.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
Supp_Fig_1.pdf103KSupplemental Figure 1
Supp_Fig_2.pdf171KSupplemental Figure 2

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.