Studies have demonstrated that mesenchymal stem-like cells can be isolated from endometrium. However, the potential of endometrial-derived stem cells to differentiate into insulin-positive cells and functionally secrete insulin remains undetermined. We isolated endometrial mesenchymal stem-like cells (EMSCs) from human endometrial tissue from six donors. The insulin-secreting function of EMSCs was further analyzed in vitro and in transplanted grafts in vivo. We successfully isolated EMSCs from human endometrium, and our results showed that EMSCs expressed high levels of stemness genes (Nanog, Oct-4, Nestin). Under specific induction conditions for 2 weeks, EMSCs formed three-dimensional spheroid bodies (SBs) and secreted C-peptide. The high insulin content of SB-EMSCs was confirmed by enzyme-linked immunosorbent assay, and glucose responsiveness was demonstrated by measuring glucose-dependent insulin secretion. Using cDNA microarrays, we found that the expression profiles of SB-EMSCs are related to those of islet tissues. Insulin and C-peptide production in response to glucose was significantly higher in SB-EMSCs than in undifferentiated EMSC controls. Furthermore, upon differentiation, SB-EMSCs displayed increased mRNA expression levels of NKx2.2, Glut2, insulin, glucagon, and somatostatin. Our results also showed that SB-EMSCs were more resistant to oxidative damage and oxidative damage-induced apoptosis than fibroblasts from the same patient. It is noteworthy that SB-EMSCs xenotransplanted into immunocompromised mice with streptozotocin-induced diabetes restored blood insulin levels to control values and greatly prolonged the survival of graft cells. These data suggest that EMSCs not only play a novel role in the differentiation of pancreatic progenitors, but also can functionally enhance insulin production to restore the regulation of blood glucose levels in an in vivo transplantation model.
As the number of patients with diabetes increases worldwide (Ryan et al., 2002), diabetes mellitus has been recognized as the most prevalent and serious metabolic disease. Type 1 diabetes is caused by progressive degeneration of β cell function in the islets of Langerhans and the resulting loss of capacity to generate insulin. Type 2 diabetes results from insulin resistance and/or insufficiency of insulin secretion. Transplantation of islets of Langerhans is a promising therapy for the treatment of type I diabetes (Ryan et al., 2002). However, numerous obstacles remain, such as the limited number of available donor cells, the difficult process of islet isolation, and the cost of the procedure, all of which have largely discouraged the use of this therapy in the treatment of diabetic patients to date (Robertson et al., 2003, 2004).
The human endometrium is responsive to sex steroid hormones, undergoes extraordinary growth in a cyclic manner, and is shed and regenerated throughout a woman's lifetime. It has been proposed that the human endometrium may contain a population of stem cells that is responsible for its remarkable ability to regenerate (Chan et al., 2004; Gargett, 2007). Data have demonstrated that adult mesenchymal stem-like cells (Schwab and Gargett, 2007) and side-population cells have been identified in normal endometrial tissues obtained from hysterectomy (Tsuji et al., 2008). In addition, in a study by Schwab and Gargett (2007), endometrial mesenchymal stem-like cells (EMSCs) differentiated into mesoderm-derived lineages, including adipogenic, chondrogenic, osteogenic, and myogenic lineages, in vitro. A well designed study by Hida et al. (2008) showed that menstrual blood-derived mesenchymal cells began beating spontaneously after induction, exhibiting cardiomyocyte-specific action potentials. Furthermore, Meng et al. (2007) isolated a stem cell-like cell from menstrual blood that they termed the “endometrial regenerative cell” that was capable of differentiating into nine lineages, including a pancreatic lineage that stained positive for insulin. However, the capability of the endometrial regenerative cell to differentiate into endodermal tissues and its connection with putative endometrial mesenchymal stem-like cells are still unclear. Adult stem cells have yielded controversial results with regard to their ability to secrete insulin in vitro and normalize hyperglycemia in vivo. Several in vitro studies have shown that insulin-producing cells can be generated from adult pancreatic ductal tissues (Noguchi et al., 2005; Lin et al., 2006, 2007). Bone marrow-derived mesenchymal stem cells (BMSCs), which possess pluripotent differentiation capabilities, are a candidate for stem cell therapy in diabetic islet cell replacement (Lee et al., 2006). However, other studies have failed to support the ability of BMSCs to differentiate into islet cells (Taneera et al., 2006). Because of issues of donor shortages and the required use of immunosuppressants, the clinical application of using BMSCs or other adult stem cells to differentiate directly into insulin-positive cells or β-cells is still in question.
Endometrial-derived multipotent stem cells from human endometrium have been isolated and demonstrated to possess a multilineage differentiating capacity (Meng et al., 2007). However, the insulin production function of endometrial-derived multipotent stem cells was not detected in this study. Our previous data demonstrated that multipotent stem cells derived from human term placenta were able to be isolated, and these cells could also be induced to differentiate into well functioning, insulin-positive cells (Chang et al., 2007). In the study presented here, we isolated EMSCs as a follow-up to our previous study (Chang et al., 2007). We then investigated the potential of spheroid-like body (SB)-EMSCs for differentiation into the pancreatic lineage in a modified serum-free medium for pancreatic induction. Under culture in the serum-free modified pancreatic selection medium, spindle-like EMSCs gradually formed three-dimensional (3D) SB-EMSCs. Using cDNA microarrays, we found that the expression profiles of SB-EMSCs are related to those of islet tissues. Quantitative reverse transcription-polymerase chain reaction (Q-PCR) demonstrated that the expression of pancreatic progenitor-related genes, insulin, and β-islet-related genes were significantly up-regulated in the SB-EMSCs. The high insulin content of SB-EMSCs was further confirmed by ELISA, and glucose responsiveness was demonstrated by measuring glucose-dependent insulin secretion. In addition, we investigated the therapeutic potential of SB-EMSC in streptozotocin (STZ)-induced diabetes in mice. Our results demonstrate that SB-EMSCs have a strong potential for transdifferentiation into insulin-secreting cells in vitro and provide a novel characterization of the functional recovery of STZ-induced diabetes in mice in vivo.
Materials and Methods
Isolation of EMSCs
This research follows the tenets of the Declaration of Helsinki, and informed consent was obtained from donor patients. Endometrial tissue samples (cases 1–12) were obtained from hysterectomy specimens from women with uterine fibroids who had not undergone hormone therapy (n = 12, ages 37–47; menstrual cycle phase: six proliferative phase and six secretory phase). The isolation procedure for EMSC isolation has been described previously (Chen et al., 2010).
Endometrial tissue samples were obtained by scraping the endometrium with a curette from hysterectomy specimens from women with uterine fibroids who had not undergone hormone therapy. The presence of endometrium was confirmed by frozen section and histopathological examination.
Endometrial cells were placed in a Petri dish and minced into small pieces (1–2 mm3) in the presence of Ca2+/Mg2+-free PBS (Chen et al., 2010). The human endometrium tissues were digested by collagenase P (Roche Molecular Biochemicals, Mannheim, Germany) in HEPES-buffered saline for 7 h at 37°C and then incubated for 60 min at 37°C with gentle pipetting every 15 min. Cell suspensions were filtered through a 40-μm sieve (BD Biosciences, San Jose, CA) to separate single cells from debris and undigested myometrial tissue fragments (Taneera et al., 2006; Chang et al., 2007; Chen et al., 2010). The sieves were backwashed to obtain endometrial fragments, which were further dissociated for 45 to 60 min as described above. Endometrial fragments were checked microscopically at 15-min intervals until all obvious penetrating endometrial glands were dissociated. The solution predominantly containing endometrial glands was centrifuged, and the supernatant was discarded. The pellet was treated with 0.25% trypsin/0.03% EDTA (Sigma-Aldrich, St. Louis, MO) at 37°C for 10 min, and the reaction was stopped by adding cold Dulbecco's modified Eagle's medium with 1 g/l of glucose (DMEM-LG; Invitrogen, Carlsbad, CA) with 10% fetal bovine serum (FBS; Invitrogen). The cell suspensions were filtered as above and combined. Single endometrial cell suspensions were obtained by negative selection using magnetic Dynabeads (MACS; Stem Cell Technologies, Vancouver, BC, Canada) coated with specific antibodies to remove hematopoietic cells (glycophorin A) and leukocytes (CD45) (Taneera et al., 2006; Chang et al., 2007). Isolated single endometrial cells (CD45− and glycophorin A− cells) were seeded on a culture dish at a density of 104 cells/cm2. Approximately 20% of purified endometrial cells attached within the first 24 h of culture, with approximately eight to nine cells observed per field of view (10× magnification). Of the attached cells, only a small number actually initiated colonies. Expansion medium consisted of DMEM-LG (Invitrogen) supplemented with 10% FBS (Invitrogen), 10 ng/ml basic fibroblast growth factor (bFGF), 10 ng/ml epidermal growth factor (EGF), 10 ng/ml platelet-derived growth factor-BB(R&D Systems, Minneapolis, MN), 100 units/ml penicillin, 1000 μg/ml streptomycin, and 2 mM l-glutamine (Invitrogen). Once the adherent adipose tissue-derived mesenchymal stem cells (AMSCs) were more than 50% confluent, they were detached with 0.25% trypsin-EDTA (Sigma-Aldrich) and replated at a 1:3 dilution under the same culture conditions (Chiou et al., 2005).
Isolation of Endometrial Fibroblasts
Endometrial fibroblasts were derived directly from the patients' endometrium. Endometrial tissue samples were obtained by scraping the endometrium with a curette from hysterectomy specimens from women with uterine fibroids who had not undergone hormone therapy (n = 12). The endometrium biopsy was rapidly washed in PBS in a Petri dish, cut into small fragments, and transferred to a flask. The endometrial tissue was then digested by collagenase P in HEPES-buffered saline for 7 h at 37°C. These cells were plated in a 6-cm tissue culture dish. Expansion medium consisted of DMEM-LG and 10% FBS supplemented with 10 ng/ml bFGF, 10 ng/ml EGF, 10 ng/ml platelet-derived growth factor-BB, 100 units/ml penicillin, 1000 μg/ml streptomycin, and 2 mM l-glutamine.
Protein Extraction and Western Blot Analysis
The procedures of protein extraction from EMSCs and AMSCs were performed as described previously (Chen et al., 2010). The cell lysates were prepared by RIPA buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM Na2EDTA, 1 mM EGTA, 1% NP-40, 1% sodium deoxycholate, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na3VO4, and 1 μg/ml leupeptin). Fifteen microliters of sample was boiled at 95°C for 5 min and separated on a 10% SDS-polyacrylamide gel electrophoresis. The proteins were transferred to Hybond-ECL nitrocellulose paper (GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK) by a wet-transfer system. Primary antibodies were as follows: rabbit anti-human Nanog, Oct 4, and Nestin (Cell Signaling Technology, Danvers, MA) and mouse anti-β actin (Millipore Corporation, Billerica, MA). The reactive protein bands were detected by the ECL detection system (GE Healthcare). Data are representative of two or more experiments from independent cell cultures.
Identification of Cell Phenotypic Markers by Flow Cytometry Analyses
EMSCs of passage 10 were used for phenotypic marker identification by flow cytometry (Lin et al., 2007). A total of 105 cells were resuspended in 100 μl of PBS and incubated with primary antibodies (anti-human CD9, CD13, CD14, CD29, CD31, CD34, CD44, CD45, CD49, CD54, CD73, CD90, CD117, CD133, and STRO-1) (1:100 dilutions) at 4°C for 1 h. After washing twice with PBS, labeled cells were resuspended in 100 μl of PBS with 1 μl of the fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG antibody (AP124F; Millipore Corporation) at 4°C for 1 h. Cells were then examined with a FACSCalibur apparatus (BD Biosciences) (Lin et al., 2007).
Induced Differentiation of EMSCs into Adipogenic, Chondrogenic, and Neuronal Cells
First, 105 EMSCs at the fifth to eighth passages were treated with adipogenic medium for 14 days. Adipogenic medium consisted of Iscove's modified Dulbecco's mediumsupplemented with 0.5 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich), 1 μM hydrocortisone (Sigma-Aldrich), 0.1 mM indomethacin (Sigma-Aldrich), and 10% rabbit serum (Sigma-Aldrich). For the evaluation of adipocytes, cells were fixed with 4% formaldehyde and stained with Oil-red O (Sigma-Aldrich) (Gargett et al., 2009).
To induce chondrogenic differentiation, 105 EMSCs at the fifth to eight passages were transferred into a 15-ml polypropylene tube and centrifuged at 1000 rpm for 5 min to form a pelleted micromass at the bottom of the tube, then treated with chondrogenic medium for 3 weeks. Chondrogenic medium consisted of high-glucose DMEM (Bio-Fluid, Rockville, MD) supplemented with 0.1 μM dexamethasone, 50 μg/ml ascorbic acid, 100 μg/ml sodium pyruvate (Sigma-Aldrich), 40 μg/ml proline (Sigma-Aldrich), 10 ng/ml transforming growth factor-β1, and 50 mg/ml ITS+ premix (6.25 μg/ml insulin, 6.25 μg/ml transferrin, 6.25 ng/ml selenious acid, 1.25 mg/ml bovine serum albumin, and 5.35 mg/ml linoleic acid; BD Biosciences). Medium changes were performed twice weekly. The specimens were analyzed by Alcian blue immunohistochemistry staining 14 days later (Wolff et al., 2007; Gargett et al., 2009).
The 105 EMSCs at the fifth to eighth passages were treated with osteogenic medium for 3 weeks. Osteogenic medium consists of DMEM-LG (Invitrogen) supplemented with 50 lg/ml ascorbate-2 phosphate, 10−8 M dexamethasone, and 10 mM β-glycerophosphate (all from Sigma-Aldrich). Osteogenesis was assessed by alizarin red staining. For evaluation of the mineralized matrix, induced EMSCs and AMSCs were treated with 2% silver nitrate (Sigma-Aldrich) under UV radiation for 1 h. Cells were fixed with 2.5% sodium thiosulfate (Sigma-Aldrich) for 5 min and washed in distilled H2O. Cells were then counterstained with 1% neutral red (Sigma-Aldrich) for 1 min (Chen et al., 2010).
First, 105 EMSCs at the fifth to eighth passages were treated with serum-free modified neurogenic selection medium for 2 weeks. The serum-free medium for the selection of neural precursor cells contained DMEM/F12 1:1 (Invitrogen) and was supplemented with 0.6% glucose, 25 μg/ml insulin, 100 μg/ml transferrin, 20 nM progesterone, 60 μM putrescine, 30 nM selenium chloride, 2 mM glutamine, 3 mM sodium bicarbonate, 5 mM HEPES, 2 μg/ml heparin, 20 ng/ml EGF, and 20 ng/ml bFGF (all from Sigma-Aldrich). For further neural differentiation, EGF was removed from the medium, and the medium was supplemented with 20 ng/ml Sonic hedgehog (R&D Systems), 10 ng/ml brain-derived neurotrophic factor (R&D Systems), and all-trans retinoic acid (100 nM) for another 7 days (Kao et al., 2010).
Serum-Free Modified Pancreatic Selection Medium
First, 105 EMSCs (n = 11, passages 5–8) were washed twice with HEPES-buffered saline solution, then plated onto a 10-cm dish and cultured with DMEM (25 mM glucose)/F12 serum-free medium [a 1:1 ratio of DMEM to F12 containing 0.6% glucose, 25 μg/ml insulin, 100 μg/ml transferrin, 20 nM progesterone, 60 μM putrescine, 30 nM selenium chloride, 2 mM glutamine, 3 mM sodium bicarbonate, 5 mM HEPES buffer, 2 μg/ml heparin, 20 ng/ml EGF, 20 ng/ml bFGF, and 20 ng/ml hepatocyte growth factor (PeproTech, Rocky Hill, NJ), 10 mM nicotinamide (Invitrogen), and 100 ng/ml Activin A]. The medium was changed twice, and cells were subcultured once a week with the ratio of 1:3.
Microarray Analysis and Bioinformatics
Total RNA was extracted from cells by using TRIzol reagent (Invitrogen) and the Qiagen RNAeasy column (QIAGEN, Valencia, CA) for purification. Total RNA was reverse-transcribed with Superscript II RNase H reverse transcriptase (Invitrogen) to generate Cy3- and Cy5-labeled (GE Healthcare) cDNA probes for the control and treated samples, respectively. The labeled probes were hybridized to a cDNA microarray containing 10,000 gene clone immobilized cDNA fragments. Fluorescence intensities of Cy3 and Cy5 targets were measured and scanned separately by using a GenePix 4000B Array Scanner (Molecular Devices, Sunnyvale, CA). Data analysis was performed with GenePix Pro 22.214.171.124 (Molecular Devices) and GeneSpring GX 7.3.1 software (Agilent Technologies, Palo Alto, CA) (Yang et al., 2008b). Classical multidimensional scaling was performed by using the standard function of the R program to provide a visual impression of how the various sample groups are related.
Quantitative Reverse Transcription-Polymerase Chain Reaction
SYBR green was used for Q-PCR detection. In brief, total RNA (1 μg) of each sample was reverse-transcribed in 20 μl of 0.5 μg of oligo(dT) and 200 U Superscript II RT (Invitrogen). Amplification was carried out in a total volume of 20 μl containing 0.5 μM of each primer, 4 mM MgCl2, 2 μl of LightCycler–FastStart DNA Master SYBR green I (Roche Diagnostics), and 2 μl of 1:10 diluted cDNA. All PCRs were performed in triplicate. The transcript levels of genes were standardized to the corresponding glyceraldehyde-3-phosphate dehydrogenase level, and for each candidate gene, mRNA levels relative to the highest candidate gene level were estimated in percentages (Supplemental Table 1).
Immunofluorescent Staining and Immunohistochemistry
An avidin–biotin complex method was used for the immunohistochemical staining of the neurogenically differentiated EMSCs. After washes with 3% hydrogen peroxide and sodium azide, antigens were retrieved by using a microwave. Each slide was then treated with antibodies for MAP-2 (Millipore Bioscience Research Reagents, Temecula, CA) and glial fibrillary acidic protein (Millipore Bioscience Research Reagents). Immunoreactive signals were detected with a mixture of biotinylated rabbit anti-mouse IgG and Fluorsave (Calbiochem, San Diego, CA) under a fluorescent microscope (FV300; Olympus, Tokyo, Japan) (Chiou et al., 2006). To evaluate insulin and glucagon expression, the aggregated EMSC sphere bodies and solid tissue from transplanted SB-EMSCs were embedded in O.C.T. (Sakura Finetek Japan Co., Tokyo, Japan) for frozen sectioning. Sections were fixed with ice-cold acetone (50%) for 2 min at 4°C and then blocked with 5% skim milk at room temperature for 2 h. The sections were then incubated in antiglucagon (1:500; Abcam Inc., Cambridge, MA) and anti-insulin (1:500; Millipore Bioscience Research Reagents) antibodies in skim milk at 37°C for 2 h, followed by a secondary antibody incubation [rhodamine-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) for insulin detection; FITC-conjugated anti-rabbit IgG (Millipore Bioscience Research Reagents) for glucagon detection].
Cell Survival Analysis by Methyl Thiazol Tetrazolium Assay and Detection of Apoptotic Cells with Annexin V Staining
For evaluation of cell survival, cells were seeded on 24-well plates at a density of 2 × 104 cells/well. At the end of cell culture, methyl thiazol tetrazolium (MTT; Sigma-Aldrich) was added. The amount of MTT formazan product was determined by using a microplate reader detecting absorbance at 560 nm (SpectraMax 250; Molecular Devices). Annexin V staining was used to determine the percentage of apoptotic cells. Cells were harvested and stained with FITC-labeled annexin V (BD Biosciences) and propidium iodide (Sigma-Aldrich) in the dark at room temperature for 30 min. The percentage of annexin V-positive and propidium iodide-negative cells was determined by flow cytometry.
Measurement of Insulin and C-Peptide Secretion
Cells were washed twice with PBS and incubated for 3 h in DMEM-LG (5.6 mM glucose). After 3 h, the medium was collected and stored at −20°C. Cells were then washed twice with PBS and incubated for 3 h in DMEM-HG (25 mM glucose). Again, the medium was collected and stored at −20°C. C-peptide content of culture medium was assessed by using insulin and C-peptide ELISA kits (Mercodia, Uppsala, Sweden) according to the manufacturer's instructions. In brief, 25 μl of sample was added to 50 μl of assay buffer and incubated for 1 h at room temperature on a shaker. After washing six times with washing buffer, 100 μl of enzyme conjugate was added and incubated for 1 h on a shaker. Tetramethylbenzidine substrate (200 μl) was then added and incubated for 15 min. Finally, 50 μl of stop solution was added to the wells and mixed for 5 s on a shaker. The samples were analyzed after absorbance at 450 nm, and the values were normalized to the DNA content. For DNA estimation, equal numbers of cells were seeded simultaneously and cultured in the same treatment, and DNA was isolated by using the DNeasy tissue kit (QIAGEN) following the manufacturer's instructions.
Transplantation into Streptozotocin-Induced Hyperglycemic Mice
All procedures involving animals were performed in accordance with the institutional animal welfare guidelines of Taipei Veterans General Hospital. Eight-week-old severe combined immunodeficiency (SCID) mice were treated with STZ (200 mg/kg; Sigma-Aldrich) freshly dissolved in 0.025 M sodium citrate, pH 4.0. EMSCs and differentiated EMSCs (2 × 107 total) were injected into the subcapsule of the left kidney of the SCID mice following the previously described protocol (Chang et al., 2007). Blood sampling from the retro-orbital plexus was performed every 2 days and measured with a OneTouch SureStep (LifeScan Inc., Milpitas, CA) (Chang et al., 2007).
The results are reported as the mean ± S.D. Statistical analysis was performed by one-way or two-way analysis of variance followed by Student's t test, as appropriate. A p < 0.05 was considered to be statistically significant.
Isolation of EMSCs.
We successfully isolated EMSCs from 11 women after negative immunoselection to deplete CD45- and glycophorin A-positive cells (Table 1). The appearance of EMSCs was different from that of endometrial fibroblast cells (Fig. 1A). EMSCs usually appeared as spindle-shaped cells with scant cytoplasm and granules around the nuclei (Fig. 1A). Compared with the primary culture of endometrial fibroblast cells derived from the same patients, high expression levels of Oct-4, Nanog, and nestin were detected in EMSCs by Q-PCR and Western blot (Fig. 1A; Supplemental Fig. 1A) with sustained expression of all genes detected through 10 passages. Confocal immunofluorescence analysis showed the protein expression of Oct-4 (green) and Nestin (red fluorescence) in EMSCs (Fig. 1B). Consistent with our previous report (Chang et al., 2007), flow cytometry analyses revealed that EMSCs (passages 5–8) were strongly positive for CD44, CD29, CD105, and CD81, but negative for endometrial fibroblast cells (Fig. 1C). The spheres body formation of the endometrial stem-like cells was significantly more than that of endometrial fibroblast (Fig. 1D).
Potential for Adipogenic, Chondrogenic, Osteogenic, and Neurogenic Differentiation in EMSCs.
EMSCs possessed multilineage potential for differentiation into mesodermal (adipocyte, chondrocyte, and osteogenic) and neuroectodermal lineage cells (Fig. 2). In previous studies, including ours, MSCs were successfully isolated from bone marrow and induced to differentiate into mesodermal lineages (Chiou et al., 2005; Yang et al., 2008a). Following the protocols used in the aforementioned studies, we were able to differentiate EMSCs into adipocytes after 14 days of culture in adipogenic medium, as confirmed by positive staining with oil red O and up-regulated expression of PPARγ2 mRNA (Fig. 2A). To test the potential for chondrogenic differentiation, EMSCs were cultured in chondrogenic medium. After chondrogenic induction for 14 days, the EMSCs exhibited chondrocyte morphology, as evidenced by Alcian blue staining (Fig. 2B). As measured by Q-PCR, the expression of Col2a1 (a chondrogenic marker) increased significantly after 7 and 14 days of chondrogenic induction in EMSCs, compared with undifferentiated EMSCs (Fig. 2B). To test the potential of osteogenic differentiation of EMSCs, EMSCs were cultured in the presence of a high concentration of dexamethasone and β-glycerophosphate. These cells exhibited osteocyte phenotypes as evidenced by alizarin red staining. As measured by Q-PCR, the expression of Runx2 (a osteogenic marker) increased significantly after 7 and 14 days of osteogenic induction in EMSCs, compared with undifferentiated EMSCs (Fig. 2C). In contrast, Oct-4 and Nanog were highly expressed in undifferentiated EMSCs, but their expression was significantly reduced in EMSCs after 7 and 14 days of osteogenic induction (Fig. 2D).
Neuroectodermal differentiation in EMSCs was also investigated. After being cultured in serum-free DMEM/F12 medium for 2 weeks, EMSCs aggregated and formed spheroid-like bodies called neurospheres. By using immunofluorescent staining, MAP-2-positive neurons (MAP-2: neuron marker; Fig. 2E) and Nestin-positive cells (Nestin: neural progenitor marker; Fig. 2E) were detected in differentiated EMSCs, and merged images were created (Fig. 2E). The neurosphere-like EMSCs were further cultured in induction medium. After 7 days of neurogenic induction, neuron-like cells exhibiting neurite formation were observed. Immunofluorescent staining confirmed the presence of neurofilament protein in the neurites of these EMSC-derived neuron-like cells (Fig. 2E). In contrast, after treating the induction media, adipogenic, chondrogenic, osteogenic, and neurogenic differentiation in endometrial fibroblasts was not present.
Direction of SB-EMSCs toward the Pancreatic Lineage.
To further evaluate the potential of pancreatic lineage in EMSCs, the serum-free modified pancreatic selection medium was used. After 7 days of induction, EMSCs gradually aggregated and developed a cluster-like morphology (Fig. 3A; Table 1). Furthermore, EMSCs formed 3D formation and floating spheroid bodies (SB-EMACs; pancreatic-like progenitors) (n = 10) at day 14 under the serum-free modified pancreatic select ion medium (Fig. 3A). The culturing successful rate of SB-EMACs (pancreatic-like progenitors) was 83.33% (10 cases in total of 12 cases) (Table 1). Some 24.51 ± 3.09% cells of the EMSCs were differentiated into pancreatic-like cell lineage (SB-EMSC).
We next examined the expression profile of EMSC day 0, EMSC day 7 (cluster-like cells, 7 days postpancreatic selection medium), and SB-EMSC (14 days postpancreatic selection medium) using microarray analysis (Fig. 3B). Microarray analysis showed that the expression of probe sets was significantly altered in SB-EMSCs (n = 3) compared with EMSC day 0 (n = 3) when compiled with the hierarchical clustering method (Fig. 2B). A total of 716 genes (Fig. 3B) differed significantly in their expression levels between SB-EMSCs and EMSC day 0 by more than 10-fold (up-regulation) or less than 0.1-fold (down-regulation) (p < 0.05). The up-regulated and down-regulated genes are shown in Supplemental Tables 2 and 3. Based on the microarray findings, the gene expression in the subset of SB-EMSCs resembled pancreatic progenitor-related expression patterns more than the EMSC day 0 (Fig. 4A). Multidimensional scaling analysis further suggested that the expression pattern of SB-EMSCs was closer to the gene signature of pancreas and pancreatic islet tissue than the expression patterns of EMSC day 0 (p < 0.001; Fig. 4B).
To further investigate the process of pancreatic and insulin-positive cell differentiation, EMSCs were changed to serum-free modified pancreatic selection medium. After 14 days of culture in modified pancreatic selection medium, Q-PCR showed a significant increase in the mRNA expression levels of insulin, Glut2, Pax4, Nkx2.2, NeuroD, Isl-1, somatostatin, and glucagon in differentiated SB-EMSCs compared with EMSCs and endometrial fibroblast cells (p < 0.001; Fig. 4C).
Differentiation into Pancreatic-Lineage Cells and Increased C-Peptide /Insulin Secretion in SB-EMSCs.
Using a dual immunofluorescence assay, we showed that the percentages of insulin-positive and glucagon-positive cells in differentiated SB-EMSCs (Fig. 5A) were also significantly higher than those in the EMSCs and endometrial fibroblast cells (p < 0.001; Fig. 5, B and C).
Insulin secretion in response to glucose stimulation is the most important function of pancreatic β-cells, which are essential for maintaining glucose homeostasis (Yaney and Corkey, 2003; Odegaard et al., 2010). To examine the regulation of insulin secretion in differentiated 3D SB-EMSCs in response to glucose, cells were exposed to different glucose levels and insulin was detected. In the presence of 5.5 and 25 mM glucose, secretion of insulin into the culture medium from differentiated SB-EMSCs was significantly higher than secretion from EMSCs and endometrial fibroblast cells (p < 0.001; Fig. 5D). Moreover, the secreted levels of C-peptide (a product of de novo insulin formation) in the culture medium of differentiated SB-EMSCs were significantly higher than those in EMSCs and endometrial fibroblast cell medium (p < 0.001; Fig. 5E).
SB-EMSC-Derived Insulin-Positive Cells Resist Oxidative Stress Damage and Are Resistant to Interleukin-1β-Induced Apoptosis.
Moreover, interleukin-1β (IL-1β) plays a critical role in the pathophysiology of diabetics and severely interferes with insulin production (Chiou et al., 2010). To evaluate the antioxidant activity and cytoprotective potential of EMSCs and SB-EMSCs, we examined whether EMSCs and SB-EMSCs were protected from IL-1β-induced apoptosis. Endometrial fibroblast cells, EMSCs, and SB-EMSCs were cultured at 22.2 mM glucose and exposed to 50 U/ml of IL-1β. Under these conditions, the viability of EMSCs and SB-EMSCs was significantly higher than that of control cells (endometrial fibroblast cells) as measured by MTT assay (p < 0.05; Fig. 6A). The percentage of annexin V-positive apoptotic cells in EMSCs and SB-EMSCs was dramatically decreased compared with that of control cells (p < 0.05; Fig. 6B).
To further explore the possible mechanism involved in the antioxidant effects of EMSCs and SB-EMSCs, the glutathione (GSH) level and reactive oxygen species (ROS) production were measured in IL-1β-treated cells. Noticeably, treatment of 50 U/ml of IL-1β in EMSCs and SB-EMSCs significantly inhibited ROS production (Fig, 6C; p < 0.001) and increased intracellular GSH level compared with control cells (endometrial fibroblast cells) (Fig. 6D; p < 0.001).
SB-EMSCs Improve Glucose Control and Prolong the Survival of STZ-Treated SCID Mice.
We used SCID mice pretreated with STZ to examine the restoration of normoglycemia by xenotransplantation of differentiated, insulin-positive cells derived from EMSCs. The renal subcapsular space (Fig. 7A) in SCID mice has been demonstrated to provide a microenvironment suitable for endocrine cell differentiation (Chang et al., 2007). The diets were controlled between groups. A total of 2 × 107 endometrial fibroblast cells, EMSCs, or SB-EMSCs were implanted into the subcapsular space of the left kidney (n = 6 in each group). After 4 weeks, in vivo histology revealed that transplanted SB-EMSCs could proliferate and grow in solid tissues in the subrenal site (Fig. 7A).
The histology survey demonstrated that diffuse aggregated islet-like clusters were found in the transplanted grafts of SB-EMSCs (Fig. 7A). Immunofluorescence experiments further confirmed that more strong signals for insulin-positive (red fluorescence) and glucagon-positive (green fluorescence) staining were detected in the islet-like cluster of SB-EMSC-derived graft than in the endometrial fibroblast cells and EMSC-derived graft in the subrenal site in SCID mice (Fig. 7, B and C; p < 0.05). Because no teratoma formation was found in the H&E staining of transplanted grafts of SB-EMSCs at 12 weeks, further immunohistochemical examinations were performed. As shown in Fig. 7D, the expression of c-Myc and Oct-4, two markers of teratoma formation, was absent in the solid tissue formed by transplantation of SB-EMSCs.
Although blood glucose levels were reduced in transplanted EMSC and SB-EMSC groups of transplanted animals in comparison with the untransplanted STZ control group and transplanted endometrial fibroblast cells group, a significantly lower blood glucose level was observed in the animals transplanted with SB-EMSCs (p < 0.05; Fig. 8A). The serum level of human insulin was also measured. The serum level of human insulin was nearly undetectable before STZ treatment. As early as 1 week after transplant, the insulin levels in STZ-treated mice transplanted with SB-EMSCs were significantly higher than those in mice transplanted with endometrial fibroblast cells and EMSC groups (p < 0.05; Fig. 8B), and this difference was maintained 8 weeks after the transplantation. It is noteworthy that no teratoma formation was observed in SB-EMSC-derived xenografts 12 weeks after transplantation (Fig. 7A). These results suggest that SB-EMSCs can aid in the restoration of insulin to a nearly normal level in STZ-pretreated SCID mice.
In this study, we successfully isolated EMSCs from human endometrial tissues. The Q-PCR results show that endometrial mesenchymal stem-like cells preserved the genetic characteristics of a primitive embryonic stage as evidenced by the expression of the genes Oct-4, Nanog, and Nestin. By culturing cells for 2 weeks in serum-free induction medium supplemented with essential growth factors, a monolayer of spindle-like endometrial mesenchymal stem-like cells gradually formed 3D SBs. Moreover, compared with EMSCs, SB-EMSCs could efficiently differentiate into both insulin- and glucagon-positive cells, expressing high levels of these molecules. It is noteworthy that our results suggest that the SB-EMSCs were responsive to stimulation with different concentrations of glucose for the production of insulin and C-peptide. Finally, our in vivo xenotransplantation study confirmed that SB-EMSC-derived β cells helped to stably restore normoglycemia in diabetic SCID mice. To our knowledge, this is the first report to demonstrate that EMSCs and SB-EMSC possess the potential to differentiate into β-islet progenitors and insulin-secreting cells and can also participate in the functional secretion of insulin to effectively control blood glucose levels.
The phenotype of EMSCs from human endometrium in this study (Table 2) is similar to the phenotype of MSCs isolated from full thickness endometrium attached to 5-mm myometrium (Schwab and Gargett, 2007). Side-population cells have been identified in human endometrium (Kato et al., 2007; Tsuji et al., 2008). Tsuji et al. showed that side-population cells expressed a stromal cell marker (CD13), endothelial cell markers (CD31 and CD34), an epithelial cell marker, and mesenchymal stem cell markers (CD105 and CD146). Surface marker analysis indicated that endometrial side-population cells formed a heterogeneous population (Kato et al., 2007; Tsuji et al., 2008). Menstrual blood-derived mesenchymal cells have also been isolated (Noguchi et al., 2005; Hida et al., 2008). In a study by Meng et al. (2007), menstrual blood-derived mesenchymal cells expressed an epithelial cell marker (CD9) and a mesenchymal stem cell marker (CD105). This result indicated that menstrual blood-derived mesenchymal cells also make up a heterogeneous population. As shown by the aforementioned studies, there are some different characteristics in the endometrial stem cells derived from different sources and isolation methods. Because the applications of endometrial stem cells with different characteristics may differ, it is worth comparing the expression profiles and functional features of endometrial stem cells from various means of preparation.
The fact that hyperglycemia itself can decrease insulin secretion has led to the concept of glucose toxicity, which implies the development of irreversible damage to cellular components of the insulin production pathway over time (Robertson et al., 2003, 2004; Busik et al., 2008). The effects of the proinflammatory cytokine IL-1β have been conclusively shown to impair glucose-stimulated insulin production in mouse, rat, and human islets and increase β-cell death (Montolio et al., 2007; Téllez et al., 2007). In this study, we found that 3D SB-EMSCs were more resistant to hyperglycemia-induced or hyperglycemia and oxidative stress-induced toxicity and damage in comparison with control cells (fibroblasts) (Fig. 6). Moreover, in streptozotocin-pretreated nude mice, normoglycemia was restored in the group transplanted with SB-EMSC. Our findings indicate that SB-EMSC-derived pancreatic-lineage progenitors possess the ability to fight against proinflammatory- and high glucose-induced oxidative stress.
In the current study, blood glucose levels were gradually normalized after transplantation of SB-EMSC, whereas insulin was kept at a constant level followed by a steep elevation. The possible reason is that insulin was infused via subrenal implanted cells that release insulin into circulation. Despite plasma insulin levels that were elevated, the hyperglycemia induced by STZ was gradually normalized. The time dependent to normalize plasma glucose levels was consistent with an important role for hepatic insulin delivery in glucose homeostasis, because subrenally administered insulin does not reproduce the preferential increase of insulin concentrations within the hepatic portal vein that occurs when insulin is secreted from the pancreas (Havel et al., 2000). STZ diabetes induce a state of insulin resistance in addition to insulin deficiency such that relative high insulin levels are needed to restore normal glucose homeostasis (Nishimura et al., 1989).
In conclusion, this is the first study to show that EMSCs not only possess multipotent capabilities, but also are able to differentiate into insulin-secreting cells (SB-EMSCs) to functionally recover insulin insufficiency in vivo. On the other hand, allograftability of endometrial cells may be important especially for males or patients who have a specific genetic background (i.e., diabetes mellitus type 1). To avoid rejection in allograft, establishment of stem cell bank and donor-recipient matching of human leukocyte antigen may be important. Our results indicate that SB-EMSCs possess cytoprotective properties and are protected from oxidative stress-induced damage, glucotoxicity, and IL-1β-induced apoptosis. Therefore, EMSCs are an excellent source of adult stem cells. This approach to diabetic islet cell replacement should overcome the ethical and immunologic concerns associated with the use of fetal tissues and embryonic stem cells.
We thank Dr. P. H. Wang (Department of Obstetrics and Gynecology, Taipei Veterans General Hospital, Taipei, Taiwan) for discussion during the preparation of the manuscript.
This study was supported by the National Science Council of the Republic of China [Grants 97-3111-B-075-001-MY3, 97-2320-B-075-003-MY3], Taipei Veterans General Hospital [Grants V97B-004,V97B1-006, V98A-093, V99C1-166, V99A-065, E1-008, F-001, B99A-065], the Joint Projects of the Veterans General Hospitals University System of Taiwan [Grant VGHUST 98-G6-6], Yen-Tjing-Ling Medical Foundation [Grants 95/96/97/98, CI-97-6, CI-99-7], National Yang-Ming University (Ministry of Education, Aim for the Top University Plan) and Genomic Center Project, Technology Development Program for Academia, Department of Industrial Technology, Ministry of Economic Affairs [Grant 98-EC-17-A-19-S2-0107], and the Center of Excellence for Cancer Research at Taipei Veterans General Hospital [Grant DOH99-TD-C-111-007].
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- endometrial mesenchymal stem-like cell
- bone marrow-derived mesenchymal stem cell
- adipose tissue-derived mesenchymal stem cell
- spheroid-like body
- quantitative reverse transcription-polymerase chain reaction
- Dulbecco's modified Eagle's medium
- DMEM with 1 g/l of glucose
- fetal bovine serum
- severe combined immunodeficiency
- hemoatoxylin and eosin
- phosphate-buffered saline
- fluorescein isothiocyanate
- methyl thiazol tetrazolium
- reactive oxygen species
- epidermal growth factor
- basic fibroblast growth factor
- peroxisome proliferator-activated receptor γ2
- microtubule-associated protein 2.
- Received April 24, 2010.
- Accepted September 17, 2010.
- Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics