Human Endometrial Stromal Stem Cells Differentiate into Megakaryocytes with the Ability to Produce Functional Platelets

Human endometrium is a high dynamic tissue that contains endometrial stromal stem cells (hESSCs). The hESSCs have been differentiated into a number of cell lineages. However, differentiation of hESSCs into megakaryocytes (MKs) has not yet been investigated. The aim of this study was to investigate the feasibility of MK generation from hESSCs and subsequent production of functional platelets (PLTs). In our study, hESSCs were cultured from endometrial stromal cells as confirmed by positive stromal cell specific markers (CD90 and CD29) and negative hematopoietic stem cell markers (CD45 and CD34) expression. Then, hESSCs were differentiated in a medium supplemented with thrombopoietin (TPO) for 18 days. The MK differentiation was analyzed by flow cytometry and confocal microscopy. The differentiation medium was collected for PLT production analysis by flow cytometry, transmission electron microscopy and functional measurements. Our results show: 1) MKs were successfully generated from hESSCs as identified by expression of specific markers (CD41a: 1±0.09% and 39±3.0%; CD42b: 1.2±0.06% and 28±2.0%, control vs. differentiation) accompanied with reduction of pluripotent transcription factors (Oct4 and Sox2) expression; 2) The level of PLTs in the differentiation medium was 16±1 number/µl as determined by size (2–4 µm) and CD41a expression (CD41a: 1±0.4% and 90±2.0%, control vs. differentiation); 3) Generated PLTs were functional as evidenced by the up-regulation of CD62p expression and fibrinogen binding following thrombin stimulation; 4) Released PLTs showed similar ultra-structure characteristics (alpha granules, vacuoles and dense tubular system) as PLTs from peripheral blood determined by electron microscopic analysis. Data demonstrate the feasibility of generating MKs from hESSCs, and that the generated MKs release functional PLTs. Therefore, hESSCs could be a potential new stem cell source for in vitro MK/PLT production.


Introduction
Platelets (PLTs) play a key role in haemostatic plug formation [1] and arterial thrombosis [2]. PLT transfusion is a mainstay therapy of various PLT-related diseases, such as thrombocytopenia and in cases of menorrhagia that are refractory to medical therapy [3][4][5][6]. In the United States, the total number of PLT transfusions is over 10 million units per year and is increasing annually. Traditional source of PLTs from blood donors has several limitations, including short storage life of PLT concentrates and risk of pathogenic contamination [7,8]. Thus, developing a reliable, safe alternative approach to producing large numbers of PLTs is intensely appealing.
Stem cell based generation of megakaryocytes (MKs), the parent cells of circulating PLTs, would provide significant advantages over the current donor-supply approach. So far, embryonic stem cells [9][10][11][12] and hematopoietic stem cells (HSCs) harvested from bone marrow, peripheral blood and cord blood [13][14][15] have been explored for MK production. However, there are several wellestablished limitations of these sources, including achieving adequate amounts of HSCs and flawed maturation, as well as political and ethical ramifications.
The endometrium of reproductive-age women is highly dynamic with its remarkable capacity for self-renewal attributed to stromal progenitor or stem cells [16]. Human endometrial stromal stem cells (hESSCs) can be isolated relatively simply from endometrium and maintained in cell culture. The cells can be cultured for more than 15 passages (.20 months) [17]. In addition, successful differentiation of hESSCs into targeted cells has clinical significance because it can provide an autologous cell source that avoids the complications associated with allogeneic transplantation. Indeed, these characteristics render hESSCs a potentially good cell source for regenerative medicine. So far, hESSCs have been reported to be capable of differentiating into mesodermal and ectodermal lineages, including smooth muscle cells [18], adipocytes [17,18], chondrocytes [18,19], osteoblasts [18], pancreatic b-cells [20] and neurons [21]. However, to our knowledge, there is no report of differentiating hESSCs into MKs.
In this study, we investigated whether hESSCs could be differentiated into MKs and the generated MKs could release functional PLTs in an in vitro culture system.

Morphology Observation of hESSC Culture
Three days after plating endometrial cells, a significant number of cells attached to the tissue culture plates and grew to small colonies (Fig1-A). Then, the small colonies gradually formed large colonies at day 5-6 (Fig1-B) and finally achieved 90% confluence at approximately day 8-10 (Fig1-C).

Characterization of hESSCs
The passage 4 and 6 (P 4, 6 ) cultured cells were characterized by flow cytometry and immunocytochemistry methods. Flow cytometric results showed that the hESSCs positively expressed stromal cell specific markers CD90 and CD29 (Fig 1-D, E) and negatively stained with hematopoietic lineage markers CD45 and CD34 (Fig 1-F, G). Immunocytochemistric results revealed that the hESSCs positively stained CD90 (Fig 1-H), but negatively stained an epithelial specific marker cytokeratin (Fig 1-I).
These results showed that the P 4, 6 hESSCs was characterized by a homogenous cell population with depletion of hematopoietic and epithelial cells.

Flow Cytometric Analysis of MKs Differentiated from hESSCs
Differentiated MKs were also detected by flow cytometric analysis of CD41a and CD42b expression. Flow cytometric results revealed that the expression rates of CD41a and CD42b were significantly increased (CD41a: 160.09% and 3963%, In addition, MK differentiation was identified using DNA ploidy study. We found that cells in the MK differentiation group displayed polyploidy ranged from 2 N to 16 N (Fig 3-G), whereas, the cells in control group were diploidy and the DNA ploidy was 2 N (Fig 3-F).

Differentiated MKs Release PLTs into Culture Medium
The differentiation medium was collected for examining the PLTs released from MKs by flow cytometric analysis. The diameter of normal PLTs is approximately 2-3 mm. The gate was fixed in the forward scatter and side scatter histogram by using of 2 mm and 4 mm calibration beads (Fig 5-A). Our data indicate a population of cells presented within the gate and had an expression rate of CD41a (160.4% and 9062%, control versus differentiation, P,0.01, n = 4, Fig 5-B, C). These data demonstrate that PLTs were released from generated MKs. The calculated number of PLTs produced was approximately 3610 5 6360 PLTs from 1.8610 5 6642 hESSCs. The level of PLT in the differentitation medium supernatant was 1661 number/ml.

Generated PLTs Bind to Fibrinogen and Express CD62p after Thrombin Stimulation
The function of the released PLTs was determined by thrombin (5 U/ml, 10 min) stimulation in the presence of 488-labeled fibrinogen. We found the fibrinogen binding rate was increased (160.4% and 3263%, control versus thrombin, P,0.01, n = 4) after thrombin stimulation (Fig 6-A, B). These data are summarized in  The morphology and ultra-structural composition of PLTs released from generated MKs was examined by transmission electron microscopy. A representative image of a generated PLT, shown in Fig 7-A, displayed a discoid shape and was similar to the discoid shape characteristic of peripheral blood PLTs (Fig 7-B). Magnified views show the presence of alpha granules, vacuoles and dense tubular system in PLTs released from differentiated MKs (Fig 7-C). These intracellular structures are similar to PLTs harvested from human peripheral blood (Fig 7-D).

Discussion
In this study, we developed a protocol to effectively induce the differentiation of hESSCs into MKs. Generated MKs were able to release functional PLTs with size and ultra-structure characteristics similar to normal PLTs obtained from peripheral blood. Combined methodologies such as immunocytochemistry, flow cytometry and transmission electron microscopy were applied to characterize the generated MKs or PLTs.
Human endometrium is characterized by its cyclic selfregeneration during the course of menstrual cycle in reproductive-age females. The upper functionalis layer (0.5-1 mm to 5-7 mm thickness), which sheds at menstruation is thought to regenerate from stem cells in the basalis layer of endometrium. The existence of hESSCs has been investigated in several previous studies. Initially, Chan and coworkers [22,23] demonstrated the presence of clonogenic stromal stem cells with high proliferative potential in active cycling and inactive endometrium. And their high proliferation rate was preserved despite their long duration in culture (more than 20 months) [17]. These innate characteristics suggest that hESSCs may have advantages over other adult stem/ progenitor cells and may represent a new resource for cell-based therapies and regenerative medicine. Recent studies have been shown that hESSCs are able to differentiate into multiple cell lineages, including adipocytes, chondrocytes, osteoblasts, neurons and pancreatic b-cells [17][18][19][20][21]. However, differentiation of hESSCs into MKs has not been reported yet.
This study is the first time to describe a culture system in which MKs was differentiated from hESSCs in a serum-free medium supplemented with TPO. In this study, P 4-6 hESSCs were used for the differentiation experiment. These hESSCs was a homogenous cell population as evidenced by highly positive expression of stromal cell markers (CD29 and CD90), negative expression of hematopoietic stem cell markers (CD34 and CD45) and an epithelial marker cytokeratin. After 18 days differentiation, the hESSCs-derived MKs displayed polyploidy with the DNA ploidy ranged from 2 N to 16 N, accompanying with a significantly increased expression profile of characteristic MK specific plasma membrane glycoprotein markers (e.g., CD41a, CD42b). In addition, since Oct4 and Sox2 are pluripotent human stem cell markers detected in endometrium and other adult stem cells [24][25][26], we also verified the differentiation of hESSCs by measuring the expression profile of Oct4 and Sox2. Our data indicated that the expression levels of Oct4 and Sox2 were both significantly reduced after differentiation, indicating their phenotype paralleled the molecular changes associated with differentiation.
In order to further explore whether the differentiated MKs were able to release PLTs, MK differentiation medium were collected to detect PLT production by analysis of specific marker CD41a and size (2-4 mm). Our data showed that the level of released PLTs in the culture medium was around 16 number/ml. Moreover, we functionally evaluated the produced PLTs with two thrombin stimulation studies. One is thrombin stimulation induced activation of fibrinogen receptor (aII b III a ), which is an integrin complex found on PLTs and aids in PLT activation by conformational change. The result showed the fibrinogen binding rate significantly increased from 160.4% to 3263% with the stimulation. The other is thrombin stimulation induced surface expression of antigen P-selectin (CD62p). The result showed the percentage of specific CD62P + PLTs significantly increased from 2.561% to 2664% with the stimulation. Both stimulation studies demonstrate that the released PLTs from generated MKs are functional in our culture system. In addition, we applied transmission electron microscopy (TEM) to examine the ultra-structures of generated PLTs. The TEM examination results revealed that the produced PLTs displayed a discoid shape and contained a normal distribution of organelles such as alpha granules, vacuoles and dense tubular system, which are indistinguishable from the PLTs obtained from peripheral blood. These data are consistent with previous studies on in vitro generated MKs and PLTs from other tissue of origin sources [9,12,[27][28][29]. Despite further studies are required to determine the optimal protocols for MK differentiation from hESSCs, as well as for PLT release from MKs, this study for the first time reports that MKs can be feasibly generated from hESSCs and that the generated MKs can release functional PLTs. PLTs generated in such an approach could be applied to autologous PLT transfusion. Our future work is to examine whether hESSCs induced inducible pluripotent stem cells (iPSCs) could further increase the efficiency of MK differentiation and PLT production.

Conclusion
In conclusion, the results of our study demonstrate that hESSCs can differentiate into MKs which subsequently release functional PLTs in an in vitro culture system. The hESSCs have the potential to serve as a new source of PLT production for clinical application.

Ethics Statement
For this study, ethics approval was obtained from the Institutional Review Boards and Research Committees at the Miami Valley Hospital and Wright State University. Written informed consent was obtained from each patient and the patient confidentiality was protected according to HIPPA guidelines.

Human Endometrial Tissue Samples
Four premenopausal women (30-45 years old) who were undergoing hysterectomy for benign disease indications, leiomyoma, were selected. The full thickness endometrium attached to myometrium sample was collected and put into PBS containing 10% FBS and 2% antibiotic-antimycotic solution (final concentration: 200 mg/ml penicillin G sodium and 200 mg/ml streptomycin sulfate) (Invitrogen, Carlsbad, CA). The sample was processed within 30-60 min.

Isolation and Culture of hESSCs
As soon as the sample was delivered to the laboratory, the endometrial tissue was mechanically and enzymatically digested into a single-cell suspension as previously reported with slight modifications [17,22]. Briefly, the endometrium layer was dissected, minced into 1-2 mm 3 pieces, and digested in enzyme digestion buffer, PBS with 25 mM HEPEs (Invitrogen, Carlsbad, CA), 1 mg/ml collagenase (Roche Applied Science, Indianapolis, IN), 0.1 mg/ml DNAse1(Sigma, Fairfax, VA) and 1% antibiotic-antimycotic solution, in a humidified 5% CO 2 /95% air atmosphere at 37uC for 45 min with a brief shaking once every 5 min. The cell suspension was filtered through a 70 mm wire sieve to remove mucus and undigested tissue. The filtrates were centrifuged at 400 6 g for 10 min to collect cells. Cell pellets were resuspended with hESSCs culture medium, DMEM/F12 (Invitrogen, Carlsbad, CA) supplemented with 10% FBS and 1% antibiotic-antimycotic solution. The cells were seeded in a 6-well tissue culture plate in a humidified 5% CO2/95% air incubator. The medium was changed twice a week during expansion of the cells. When the cells were 90% confluent, they were detached using 0.025% trypsin/0.01% EDTA and passaged at a ratio of 1:3.  Then the cells were mounted with fluorescence mounting medium and imaged under a fluorescence microscope (Leica TCS SP2, Leica, Germany).

Identification of hESSCs
For flow cytometry, the cultured cells were trypsinized and centrifuged. Cell pellets were resuspended and incubated with PEor FITC-conjugated antibodies in a final volume of 100 ml for 30 min at 4uC in the dark, individually. Isotype matched (IgG) non-specific antibodies served as negative controls. Antibodies (PE-conjugated anti-human CD29, PE-conjugated anti-human CD90, FITC-conjugated anti-human CD45, FITC-conjugated anti-human CD34, FITC-conjugated mouse IgG1 isotype control, PE-conjugated mouse IgG1 isotype control) were purchased from eBioscience (San Diego, CA). The concentrations of antibodies were applied according to manufacture instructions. A total of at least 10 000 events were collected and analyzed using Accuri C6 flow cytometer and CFlow Plus Analysis software (Ann Arbor, MI).

Differentiation of MKs from hESSCs
The P 4-6 of hESSCs were plated as 1. , and 1% antibiotic-antimyotic solution for 18 days. Half of the culture medium was replaced in the first 9 days. The culture medium (2 ml) from day 10-18 was collected for PLT isolation. Cells on culture plates were harvested for flow cytometry or western blotting analysis. Cells on coverslips were used for MK immunocytochemistry assay.

Characterization of MK Differentiation by Immunocytochemistry Assay
MK differentiation was determined by analysis of surface specific markers (CD41a and CD42b). For immunostaining, the cultured cells on coverslips were fixed with 2% PFA solution for 30 min at RT, and permeabilized with 0.3% triton X-100 for another 30 min. Then slides were incubated overnight with prediluted primary antibody CD41a (glycoprotein aIIb) or CD42b (glycoprotein aIb, both from eBioscience, San Diego, CA), and FITC-conjugated anti-mouse IgG secondary antibody (1:100, Jackson lmmunoResearch Lab, West Grove, PA) for 2 hrs. Cell nuclei were stained with propidium iodide (PI) solution containing RNAse (PI/RNAse final concentration: 20 mg/ml, Sigma, Fairfax, VA) for 15 min in the dark. Between each step, slides were washed with PBS/0.1% BSA. Cells were examined

Characterization of MK Differentiation by Flow Cytometry
Surface specific markers (CD41a and CD42b) were also used for flow cytometric analysis of MK differentiation. The cells were detached, centrifuged and fixed with 2% PFA. Then, the cells were permeabilized with blocking buffer containing 0.3% triton X-100, incubated with FITC-conjugated antibodies in a final volume of 100 ml for 30 min at 4uC in the dark, respectively. Antibodies (FITC-conjugated anti-CD41a, FITC-conjugated anti-CD42b, FITC-conjugated mouse IgG1 isotype controls) were purchased from BD Biosciences (San Jose, CA). The concentrations of antibodies were applied according to manufacture instructions. Cell nuclei were stained with PI solution containing RNAse (PI/RNAse final concentration: 20 mg/ml) for 30 min in the dark. A total of at least 150 000 events were collected and analyzed using an Accuri C6 flow cytometer and CFlow Plus Analysis software (Ann Arbor, MI).

Characterization of MK Differentiation by Western Blot Analysis of Oct4 and Sox2 Expression
After hESSCs were cultured in MK differentiation or control medium for 18 days, cells were harvested with lysis buffer. Cell lysates were then centrifuged at 10 000 6 g for 5 min. Protein concentrations in the supernatants were determined by the Bradford method using Bio Rad reagent (Bio Rad Laboratories) and the same amount of protein (30 mg) from each was separated by 10% Tris-Glycine SDS-Page gel (Bio Rad Laboratories) and transferred to PVDF membrane. After blocking with 5% non-fat milk for 1 hr at RT, membranes were incubated for primary monoclonal antibodies rabbit anti-Oct4 (1:500, Abcam Inc, Cambridge, MA) and mouse anti-Sox2 (1:2000, R&D systems, Minneapolis, MN) overnight at 4uC, respectively. After 3 times wash, secondary antibodies horse raddish peroxidase (HRP) conjugated goat anti rabbit (1:4000) and HRP-conjugated donkey anti mouse (1:4000, both were from Jackson lmmunoResearch Lab, West Grove, PA) were probed, individually. Then the signal was read with a luminescent analyzer (LAS-3000, Fujifilm, Japan) and the band intensities were analyzed by ImageGauge (Fujifilm, Japan).

Flow Cytometric Analysis of PLT Production in Culture Medium
To determine the number of PLT released, we collected the culture medium (2ml) of the same well of a 6-well culture plate every day from day 10-18 of the differentiation group or control group. The collected culture medium was centrifuged at 150 6 g for 20 min to eliminate any large cells each time. The supernatant was fixed with 2% PFA, centrifuged at 1700 6 g for 30 min to sediment a PLT pellet. The pellet was resuspended with PBS/1% FBS, and stored at 4uC for further analysis. For analysis, the fixed and stored PLT pellets from the same well were pooled together and then divided into two groups. One group was resuspended and incubated with FITC-conjugated anti-CD41a antibody in a final volume of 100 ml for 30 min at 4uC in the dark. The other group was incubated with Isotype matched (IgG) non-specific antibody in the same amount of volume (100 ml) and served as negative control. 2 mm and 4 mm flow cytometry calibration bead solution was added into each tube. The concentrations of calibration beads and antibodies were applied according to manufacture instructions. A total of at least 10 000 events within the gate were collected and analyzed using the Accuri C6 flow cytometer. The concentration of PLT (CD41a + ) in the culture medium supernatant of each well was calculated as: PLT numbers/ml of culture medium = [CD41a + numbers 6 100 ml/ actual run volume (ml)]/total culture medium volume (2 6 9 6 10 3 ml) 6 groups (2).

Functional Analysis of Released PLTs
The functional capabilities of generated PLTs were analyzed for both fibrinogen binding and surface antigen CD62P expression in the presence or absence of thrombin stimulation. To prepare cell samples, PLTs were harvested from the culture medium, centrifuged at 150 6 g for 20 min to remove large cells, resuspended and then centrifuged at 1700 6 g for 30 min to sediment a PLT pellet as above described. The PLT pellets were resuspended in modified Tyrode-HEPEs buffer (138 mM NaCl, 0.36 mM NaH 2 PO4, 2.9 mM KCl, 12 mM NaHCO 3 , 10 mM HEPEs, 5 mM glucose, 1 mM MgCl 2 , and 1 mM CaCl 2 , pH 7.4).
To determine fibrinogen binding, generated PLTs were stimulated with agonist thrombin (5 U/ml, Sigma, Fairfax, VA) for 10 min at 37uC in the presence of Alexa Fluor 488-labeled human fibrinogen (100 mg/ml, Invitrogen, Carlsbad, CA) or the isotype control mouse IgG. For controls, generated PLTs without thrombin stimulation were also stained with Alexa Fluor 488labeled fibrinogen or its corresponding isotype control IgG.
To determine the surface expression of P-selectin (CD62p, eBioscience, San Diego, CA), after thrombin stimulation, the generated PLTs were fixed with 2% PFA and incubated with FITC-conjugated anti-CD62p or the isotype control mouse IgG. For controls, generated PLTs without thrombin stimulation were also stained with FITC-conjugated anti-CD62p or its corresponding isotype control IgG. Final volume per tube was brought to 100 ml. A total of at least 100 000 events were collected each and all samples were analyzed by Accuri C6 flow cytometer. All antibody application concentrations followed manufacture instructions.

Ultra-structure Analysis of Generated PLTs
PLT specimens for ultra-structural study by TEM were prepared as described previously with slight modifications [9,30]. Briefly, generated PLTs in culture medium supernatant were collected as above described. Peripheral blood samples were collected from the health donors using standard phlebotomy techniques into a collection tube containing acid-citrated-dextrose (ACD) solution. The blood samples were centrifuged at 150 6 g for 20 min to obtain PLT-rich plasma (PRP). Blood PLTs were fixed in 2% glutaraldehyde in 0.1 M phosphate buffer for 1 hr at 4uC. The samples were then washed, post-fixed with 1% osmium tetroxide (Electron Microscopy Science, Hatfield, PA) in 0.1 M phosphate buffer for 2 hrs at 4uC, dehydrated with a graded ethanol series and propylene oxide (Sigma, Fairfax, VA), and embedded in spurr resin 812 (Sigma, Fairfax, VA) and baked. Ultrathin sections (60-80 nm) were prepared with MT700, mounted on 300-mesh copper grids, stained with 2% uranyl acetate (Fisher Scientific, Pittsburgh, PA) for 5 min, rinsed with double distilled water for 5 times, and allowed to dry. The specimen were then stained with lead citrate (Fisher Scientific, Pittsburgh, PA) for 10 min, rinsed with double distilled water another 5 times and allowed to air dry. Finally, all specimens were examined with an EM 208 (Philips) transmission electron microscope at an accelerating voltage of 70 KV.

Statistic Analysis
Results were described as mean 6 SEM. The p value ,0.05 was considered to be statistically significant in the paired t-test.