BMSCs and Pectin-Based E2-Loaded Microcapsules with Injectable Pectin-Pluronic ® F-127 Scaffolds for Mouse Endometrial Regeneration Application

a


Introduction
Uterine endometrium is a highly dynamic tissue which consists of a basal layer and a functional layer [1,2] .
Severe damage to the basal layer caused by curettage, infections, caesarean section and myomectomy lead to endometrium scar formation or intrauterine adhesion (IUA), which may result in amenorrhea, hypomenorrhea, recurrent pregnancy loss, or infertility [3,4] .Several strategies, such as hysteroscopic surgery to remove adhesions and estrogen therapy have been adopted for the treatment of endometrial brosis [5,6] .However, in severe cases, treatment is di cult and the prognosis is usually suboptimal [7] .Therefore, improving the proliferation and regeneration of endometrial epithelium and stromal cells, which in turn can reconstruct the endometrial structure and restore endometrial functioning, is a radical therapy for endometrial damage meant to improve pregnancy rates [8] .
Several studies have demonstrated that bone marrow-derived mesenchymal stem cells (BMSCs), which could facilitate proliferation of endometrial stromal and epithelial cells and accretion of endometrium, and represent potential progenitor cells in the endometrial basal layer directly differentiating into endometrial epithelial cells, have bene cial effects for the treatment of IUA [7][8][9][10][11] .However, the local microenvironment of damaged endometrium cannot provide the stable three-dimensional structure and necessary bioactive molecules for stem cells.It has yielded disappointing results for transplantation of BMSCs with respect to survival, attachment, differentiation, and proliferation [12][13][14] .The therapeutic effect of direct injection of BMSCs is currently inadequate [8,10] .
Pectin is a natural polymer extracted from the peels of apples and oranges, and has good biocompatibility [15] .Pluronic® F-127 is a copolymer of polyethylene oxide (PEO) and polypropylene oxide (PPO) which has been used in FDA approved medical products as a pharmaceutical ingredient and can generate and maintain the desired shape through the thermal sol-gel transition [16] .Banks A et al. found that Pectin-Pluronic® F-127 hydrogels with three-dimensional loose network structures are more conducive to cell adhesion and survival; Pectin-Pluronic® F-127 hydrogel as a carrier wrap growth factor can signi cantly extend the release time of growth factors [17,18] [19] .The hydrogel can be injected into a uterine cavity which has sustained injury to facilitate clinical operation and endometrial regeneration.
Estrogen, speci cally 17 β-estradiol (E2), is essential for maintaining the microenvironment of endometrium and stimulates regeneration of the endometrium [8,11,20] .In clinical practice, E2 therapy is commonly used as an ancillary treatment in IUA patients [21] .However, both oral and systemic administration and delivery show low concentrations of E2 at the injured site of the uterus, which signi cantly reduces its therapeutic effect.E2 at high levels may increase the risk for thrombosis and malignancy [22] , and reduces the receptivity of endometrium [8,11,20,21] .In situ administration of E2 has many disadvantages including limited half-life period, rapid diffusion into extracellular uids, and poor solubility in aqueous solutions [5] .The limitations of E2 usage were addressed in this study by encapsulating E2 into W/O/W microspheres to construct pectin-based E2-loaded microcapsules (E2 MPs), which could provide sustained E2 release serving as a long-term reliable source of E2 for endometrial regeneration.
Although it has been recognized that E2 plays an indispensable role in promoting the repair of endometrial injury, its mechanism is not well understood.Previous studies have demonstrated that the proliferative and differential responses of epithelial cells and endometrial stem cells to E2 is mediated by endometrial stromal cells (ESCs), which synthesize and transmit paracrine mediators to surrounding cells under the direction of E2 [11,23,24] .Exosomes (exos) are considered the biological mediators of intercellular communication, and play an important role in regulating cell proliferation and differentiation [25] .Exos, one form of extracellular vesicle with a diameter of about 30-150nm, act as carriers of bioactive proteins, lipid bilayer, and genetic material and transfer them to the surrounding cells in the form of paracrine [26] .Zhang et al. found that MSCs could be differentiated in the direction of endometrial epithelial cells when they were co-cultured with ESCs and E2 [11,24,27] .It has been reported that ESCs also have the capacity to release exos to modify endometrial microenvironments in a paracrine fashion [28] .Therefore, we consider that exos derived from ESCs play paracrine roles in endometrial regeneration stimulated by E2, potentially modulating the differentiation of BMSCs.
The objectives of this study were to investigate the bene cial effects of BMSCs/E2 MPs/scaffolding grafts on promoting the recovery of injured endometrium, and to investigate the mechanism of BMSCs and E2 in mouse endometrial regeneration.

BMSCs isolation and culture
The animal study was performed in strict accordance with the guidelines of the Animal Care and Use Committee in the Xinhua A liated Hospital of Shanghai Jiao Tong University, and was approved by Xinhua Hospital Research Ethics Committee, Shanghai Jiao Tong University School of Medicine (reference number: XHEC-F-NSFC-2018-122).BMSCs were isolated and cultured as previously described [8] .In brief, ve-week-old female C57BL/6 mice were killed.The femora were isolated and rinsed with phosphate-buffered saline (PBS, Gibco, Grand Island, NY, USA).The marrow cells in the bone cavities were ushed out with Dulbecco's Modi ed Eagle Medium /Nutrient Mixture F-12 (DMEM/F-12, Gibco, USA).The resultant lavage was passed through 100µm cell strainer (BD Bioscience, San Jose, CA, USA) and centrifuged at 1200 rpm for 4 min, which were then re-suspended in DMEM/F-12 supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% Penicillin-Streptomycin (P/S, Gibco).The isolated cells were seeded in 100 mm Petri dish and maintained at 37°C in a humidi ed incubator supplied with 5% CO 2 .The non-adherent cells were removed after 72 h of culture.The medium was changed every other day and passage was conducted when cells reached con uence (80-100%).The BMSCs of passage 3-5 were used for the following experiments.

Flow cytometric analysis
The expressions of cell surface markers on BMSCs were evaluated by ow cytometry.BMSCs were detached from the culture dish using 0.05% trypsin-EDTA (Gibco), centrifuged, rinsed and re-suspended in PBS at a concentration of 10 5 cells/mL.Re-suspended cells were incubated with 5 mL of CD44-FITC antibody (eBioscience, San Diego, CA, USA), CD45-BV510 antibody (eBioscience), CD29-APC antibody (eBioscience) and CD105-PE antibody (eBioscience) in the dark at 4 ℃ for 30 min.After being washed twice with PBS, cytometric analysis was performed was performed with FACSDiva (Canto, BD Bioscience, San Jose, CA, USA) and data were analyzed with FlowJo software (Tree Star, Ashland, OR, USA) (Supplement gure).

Pectin-based E2-loaded microcapsules fabrication
A double emulsion system (W/O/W) was used to generate the microspheres, as shown in previous publications [29] (Figure 1).A solution of 2.5 mL of 10 nM 17β-estradiol (E2, Sigma) in ethanol was added to 7.5 mL of pre-warmed 13.33% (w/v) gelatin (Sigma, St. Louis, MO, USA).The rst emulsion of gelatin sphere-in-oil (W/O) phase was prepared by adding 1 mL of E2-gelatin solution into 18 mL of pre-warmed (45 ℃) olive oil containing 0.5 mL Tween20 (Sigma) and 0.5 mL Span80 (Sigma) as stabilizers.The W/O solution was stirred at 450 rpm for 20 minutes at room temperature to allow for cooling resulting in gelation of the gelatin to produce W/O emulsion.Then, 0.77 mL of the W/O emulsion was added dropwise to 22.23 mL of 3.25% (w/v) low-methoxyl pectin (WillPowder, Miami Beach, FL, USA) solution stirred at 550 rpm to form the double emulsion (W/O/W).An electrospray set-up (designed by the University of Shanghai for Science and Technology, Shanghai, China) was then used to spray the W/O/W solution into a 0.15 M CaCl 2 solution to induce pectin gelation forming hydrogel microspheres.Pectinbased E2-loaded Microspheres (E2 MPs) were collected through centrifugation and incubated in 50 kD chitosan (Zhejiang Golden-Shell Pharmaceutical Co. Zhejiang, China) solutions for 10 minutes to form the cationic coating.After rinsing in double-distilled water, the E2 MPs were stored in PBS at 4 ℃.

Pectin-Pluronic® F-127 scaffolds fabrication
A mixture composed of 3% pectin (w/v) and 20% (w/v) Pluronic® F-127 (Sigma) was used as the basic solution for constructing the scaffold (stored at 4°C).The 2×10 7 cells and E2 MPs were then mixed with 10 mL basic solution at 4 ℃ of nal concentration 0.5 mg/mL to construct the scaffold uid.The mixtures were transferred into a 10 mL plastic syringe and the matched needle (BD) was attached.The mixtures were dropped slowly onto a petri dish heated at 37 ℃.When the temperature reached 37°C, the uid solidi ed, due to the thermal sol-gel transition of Pluronic® F-127, forming the hydrogel scaffold.

Pectin-Pluronic® F-127 scaffolds degradation
The scaffold samples with diameter 11 ± 0.5 mm and height 7 ± 0.4 mm were prepared and incubated in DMEM/F12 with 10% FBS and 1% P/S at 37°C with 5% CO 2 for 14 days and weighed every day to determine its degradation.The degradation rate was determined as weight remaining (ratio %) [30] , as given by the following equation: where W a is the dry gel weight after degradation at different time intervals and W b is the dry gel weight before the start of degradation experiment.
The morphology of the lyophilized gel was examined via scanning electron microscopy (SEM).The scaffolds were immersed in liquid nitrogen and lyophilized at room temperature.The cross-section of the hydrogels was gold-coated and viewed using a microscope (PHENOM, Eindhoven, Netherlands).
2.6.Enzyme-linked immunosorbent assay for E2 detection E2 MPs/scaffolds/BMSCs were cultured in DMEM/F-12 with 10% FBS and 1% P/S and maintained at 37°C with 5% CO 2 .Medium was sampled very day in rst 10 days and every other day after that and stored at -80 ℃ for further analysis.The concentration of E2 released from microcapsules was measured using an E2 enzyme-linked immunosorbent assay (ELISA) kit (LDN, Nordhorn, Germany) according to the manufacturer's protocols.In brief, indicated medium was centrifuged to remove cell debris.25 µL of supernatant was dispense into each 96-well plate, which then underwent 1 h of incubation at room temperature on a plate shaker.After adding 100 µL of enzyme conjugate, the plate was incubated for 1h at room temperature on a plate shaker.After complete aspiration of the content, wells were thoroughly washed for four times.After adding 200 µL of substrate solution into each well, the reaction underwent 30 min incubation in dark, and 50 µL stop solution was used to terminate the reaction.TECAN In nite 200 PRO (TECAN, USA) was used to record absorbance at 450 nm with OD 570 nm as the reference.

Acridine orange/ethidium bromide double uorescence staining for cell viability determination and imaging
The BMSCs were incubated with E2-loaded microcapsules in DMEM/F-12 with 10% FBS and 1% P/S for 2 h at 37 ℃, using 6-well ultralow attachment plates.The cells and microcapsules were then mixed with Pectin-Pluronic® F-127 solution at 4 ℃ of nal concentration 0.5 mg/mL.The mixtures were transferred into a 10 mL plastic syringe and the matched needle was attached.The mixtures were dropped slowly onto a petri dish heated at 37 ℃.After the mixtures formed the scaffolds, 1 mL of pre-warmed medium was added into each well.Cell viability was evaluated on day 7. Equal amounts of 100 µg/mL acridine orange (Sigma) and 100 µg/mL ethidium bromide (Sigma) were used to prepare the AO/EB dye.The AO/EB dye was added into the medium at the ratio of 1:25, and the cells were observed by a uorescence microscope (Olympus, Tokyo, Japan) after 15 minutes.

Mouse model of endometrial injury and treatment
In total, 72 nine-week-old females C57BL/6 mice with average body weight 25-30 g were randomly and equally divided into six groups: sham-operated group, spontaneous repair group, BMSCs group, E2/BMSCs group, E2 MPs/BMSCs group, E2 MPs/scaffolds/BMSCs group.For estrous cycle studies, vaginal smears were obtained daily between 8:00-10:00 AM.Only rats with four consecutive 4-day estrus cycles were selected for study.To establish uterine horn damage model, the mice were rst anaesthetized, and the skin was sterilized with 10% povidone-iodine.The uterine horns were exposed using a low abdominal midline incision.After the distal ends of the bilateral uterine horns near the ovaries were incised, an 18 G stainless steel needle with manually grinding rough surface was used to enter the uterine cavity from the incision of the uterine horn and scraped the endometrium repeatedly towards the cervix until the uterine horns were obviously congested.The opened skin and uterine horn were sutured using absorbable suture.The 0.1-0.2mLBMSCs with or without E2 MPs and scaffolds solution were injected into and lled uterine cavity.For spontaneous repair, the uterine horns after modeling were allowed healing without further treatments.For sham operation, after exposure by an abdominal midline incision, the uterine horns were left intact in the abdominal cavity without scratch.All animals were intramuscularly administrated with penicillin twice a day for three consecutive days after surgery to prevent systematic infection.

Histological analysis
The uterine horn at 4 weeks postoperatively was xed with 4% paraformaldehyde overnight, dehydrated in graded alcohols and embedded in para n.Sections of 5 µm were prepared transversally and stained with Hematoxyline-eosin (HE) and Masson staining.For immunohistochemistry, tissue sections were immunolabeled with anti-von Willebrand factor antibody (vWF, Abcam, Cambridge, MA).The thickness of the cross-sectional area of the uterus containing endometrium was quanti ed.The percentage of the endometrium containing brosis, overall endometrium of total epitheliums and secretory glands content in stroma part, total capillary vessels quantity was all semi-quanti ed using the Image-Pro Plus software.

Functional testing of uterine horns
The function of the regenerated uterine horns was assessed by testing whether it was receptive to a fertilized ovum and whether embryo development was supported to the late stages of pregnancy.Four weeks after the procedure, 6 female mice per group were mated with 2 male C57BL/6 mice.The impregnanted mice were euthanized at gestation day 15-19 after the presence of a vaginal plug, and uterine horns were examined for the presence of embryos.

ESCs isolation and culture
Isolation of ESCs was based on the protocol in a previous report [31] .Five-week-old female C57BL/6 mice were injected subcutaneously with 100 µL of the E2 solution (100 ng/100 µL) for 3 consecutive days to stimulate the proliferation of endometrium and killed on the fourth day.The uterine horns were isolated.Uterine horns were transferred to the 0.25% Trypsin-EDTA (Gibco) for 60 min at 4°C, and then incubated for 30 min at 37°C.After incubation, uteri were transfer into a petri dish containing cold DMEM/F12 medium to inactivate trypsin activity, and then vortex for 20 s to release the epithelial sheets.The uteri were transferred to 1mg/mL collagenase I (Sangon Biotech, Shanghai, China) for 30 min at 37°C while shaking (200 rpm).After inhibiting collagenase activity, the stromal cell suspension were centrifuged at 500 x g for 7 min.Cells were re-suspended in DMEM/F-12 with 10% FBS and 1% P/S and then plated on culture dish maintained at 37°C with 5% CO 2 .The non-adherent cells were removed after 6 h of culture.
The medium was changed every other day and passage was conducted when cells reached con uence (80-100%).

Differentiation of BMSCs into EEC-like cells by co-culture system
BMSCs and ESCs were co-cultured in Transwell system (12 mm Transwell with a 0.4-µm pore polycarbonate membrane insert, Corning) [24] .In brief, BMSCs were seeded in the bottom of six-well plate at a density of 6×10 5 cells/well and ESCs were seeded on the Transwell membrane at a density of 6×10 4 cells/well to separate the cells but allow soluble factors to pass freely between them.Six experimental groups were set up.In group A (control group), BMSCs were cultured in the bottom of the co-culture system alone without ESCs in differentiation medium (DMEM/F-12 with 2% Charcoal-Stripped FBS, BI, USA); in groups B, C, D, E, and F, BMSCs were co-cultured with ESCs and different concentrations of E2 (0, 1, 10, 100, 1000 nM, respectively) in differentiation medium.The medium was changed every two days over a period of 4 weeks.

Exosome isolation
ESCs at 70% of con uence were incubated in DMEM/F12 containing 10% exosome-depleted FBS (System Biosciences, CA, USA) for 48 h.Subsequently, the medium of ESCs was collected and centrifuged at 2000 g for 10 min at 4 ℃ to eliminate cell debris.The resulting supernatants were ltered through a 0.22-µm lter (Millipore, Billerica, MA, USA) to remove microvesicles, and then concentrated through an Amicon Ultra-15 100 kDa centrifugal lter (Millipore, Billerica, MA, USA) at 4000 rpm for 30 min.Exosomes were then isolated using the Total Exosome Isolation kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol.Brie y, the supernatant was mixed with Total Exosome Reagent and incubated for 24 h at 4°C.Samples were centrifuged at 10,000 x g for 2 h and the supernatant removed.The exosome pellet was re-suspended in 100 µL of PBS and stored at -80 ℃.

Identi cation of ESCs-derived exosomes
The size distribution of ESCs-derived exosomes (ESCs-Exos) was measured by nanoparticle tracking analysis (NTA) with a NanoSight NS3000 instrument (Malvern Instruments, Malvern, UK), and the exosomes morphologies were observed by transmission electron microscope (Hitachi, Tokyo, Japan).The characteristic surface marker proteins of exosomes were analyzed by Western blot.

Exosome uptake
Exosomes were labeled with a membrane-labeling dye (1,1'-Dioctadecyl-3,3,3',3'-Tetramethylindocarbocyanine Perchlorate, Dil) according to the manufacturer's protocol (Invitrogen).Brie y, 15 µg exosomes diluted in DMEM/F12 medium were added to 0.5 µL Dil and incubated for 20 min at 37°C with 5% CO 2 .Excess dye was removed from the labeled exosomes by ultracentrifugation at 100,000 g for 70 min at 4 ℃, and the exosome pellets were puri ed on Exosome Spin Columns MW 3000 (Invitrogen) according to the manufacturer's instructions.The nal pellets were re-suspended in 100 µL PBS.Dil-Labelled exosomes were added to the culture medium of BMSCs, which were seeded at a density of 250 cells/mm 2 the previous day and incubated at 37°C for 2 h.Subsequently, BMSCs were xed with 4% paraformaldehyde for 15 min at 4°C, and then counterstained with 1 µg/mL DAPI for 5 min.The uorescence was detected under a confocal laser scanning microscope (Leica TCS SP8, Leica Microsystems, Buffalo Grove, IL, USA).At last, co-localization was performed using the Image J Fiji software.

Exosome mediated stem cell differentiation
BMSCs were plated at a density of 150 cells/mm 2 .After 24 h, BMSCs were rinsed with PBS and then maintained in a differentiation medium, differentiation medium containing ESCs-Exos with 50 µg/mL, or differentiation medium containing exosomes released from ESCs by E2 stimulation (E2-ESCs-Exos) with 50 µg/mL.The medium was changed every 2 days for up to 4 weeks.

Statistical analysis
All data in this study were analyzed with SPSS 23.0 and presented as mean ± SD.Signi cance of difference between two groups was tested by Student's t-test or ANOVA.Fisher's exact test (Freeman-Halton) was employed to assess the outcome of treatment.x2 tests were performed to compare pregnancy rates.The statistical signi cances were presented as p values, and p < 0.05 was considered statistically signi cant.

Fabrication and characterization of Pectin-Pluronic® F-127 scaffolds
When temperature increases from below to above (37 ℃ for this study) its lower critical solution temperature (LCST), Pluronic® F-127 with a sol-gel transition behavior in aqueous solution can help generate and maintain the desired shape.Figure 2A shows the images of the Pectin-Pluronic® F-127 solution (Left) and scaffolds (Right).Then, BMSCs were suspended in Pectin-Pluronic® F-127 solution to reach a 1 × 10 6 cells/mL concentration for forming the scaffolds.As revealed in the SEM images, Pectin-Pluronic® F-127 scaffolds exhibit uniform pores with a mean diameter of 150∼400 µm, which indicates excellent water absorption capacity and a fair improvement in the three-dimensional network structure of the Pectin-Pluronic® F-127 scaffolds (Figure 2B).These continuously interconnected porous structures are also conducive to the nutrient supply for cell growth in tissue engineering.BMSCs could attached to the scaffold and distributed the inner space of the scaffold (Figure 2C).As shown in Figure 2D, the scaffold began to degrade after day 5, and approximately 28% of scaffold degraded at a consistent and sustained pace at day 14.
The cytotoxicity of scaffolds was further explored.After 7 days of culture, the even distribution of BMSCs within the scaffold can be viewed in focus and out of focus as they are at different planes by AO/EB staining (Figure 2E).BMSCs had a high cellular viability rate of 95.98 ± 1.30% within the bioprinted scaffold compared to 98.00 ± 0.50% for control, indicating that the scaffold could provide 3D architecture for the attachment of BMSCs and exhibit negligible toxicity of BMSCs for the culture in vitro (p=0.0656)(Figure 2E and F).

Preparation and characterization of Pectin-based E2loaded microcapsules
Figure 3A and B present that the double emulsion (W/O/W) microspheres were successfully created with the diameters averaging around 100 µm.Then, BMSCs were co-incubated with E2 MPs for 72 h.The cell counting kit-8 assay revealed that BMSCs incubated with E2 MPs proliferated faster than those without E2 MPs (Figure 3C).Furthermore, Pectin-based E2-loaded microcapsules (E2 MPs) (equivalent concentration 10 nM) were incubated in medium, which was harvested for 30 days to assay for the release of E2 using ELISA.As shown in Figure 3D, the E2 in medium would degrade rapidly, and the effective concentration of E2 was reduced to 19.7% of the original concentration in the rst 24 h and almost completely degraded within 72 h.However, the microcapsule could maintain the effective concentration of E2 in medium for 30 days by controlling the release of E2.These results supported that E2 was shown to be successfully encapsulated into the W/O/W microspheres, which provided a sustained E2 release and served as a reliable source of E2 to promote endometrial regeneration.
The prepared E2 MPs and BMSCs were mixed with Pectin-Pluronic® F-127 solution to generate the hydrogel scaffold for investigating the in uence of E2 MPs on micro-vascularized 3D tissue formation.The presence of the W/O/W microspheres does not impact the scaffold formation process.Without microcapsule, the BMSCs are evenly distributed within the scaffolds.In the presence of E2 MPs, the BMSCs self-assembled around the E2 MPs and interacted to form a three-dimensional network (Figure 3E).

BMSCs/E2 MPs/scaffolds system for in vivo endometrial regeneration in mice
To investigate the potential therapeutic effect of BMSCs/E2 MPs/scaffolds transplantation in endometrial regeneration, BMSCs/E2 MPs/scaffolds were injected into the uterine cavity of mouse endometrial injury model.At 28 days after surgery, E2 MPs and scaffolds were completely degraded.Sections of all groups were lined by simple columnar epitheliums that were similar to the Sham group.

Pregnancy within the regenerated uterine horns
Four weeks after surgery, pregnancy was observed in some of the regenerated uterine horns, which were maintained to a late, viable stage of pregnancy.Implanted embryos were present in some of the regenerated uteri after grafting with BMSCs, BMSCs with E2 or scaffolds, but embryos were found in all regenerated uterine horns in BMSCs/E2 MPs/scaffolds group.The pregnancy rate of BMSCs/E2 MPs/scaffolds group (100%) was much higher than those of spontaneous regeneration group (25%), BMSCs group (41.7%),BMSCs/E2 group (50%), BMSCs/scaffolds group (66.7%), and comparable with that of sham operated group (100%) (Figure 5).This suggested a nearly full uterine recovery from injury in pregnancy rate.These results supported the bene cial effect of BMSCs/E2 MPs/scaffolds in promoting the recovery of injured endometrium to functional endometrium.

Culture and identi cation of endometrial cells
Endometrial cells were isolated from mice.ESCs were spindle-shaped broblast-like cells, while EECs were polygonal cells rounder than stromal cells (Figure 6A).A vimentin/cytokeratin double staining was performed and indicated that stromal cells were positive for vimentin (Vim), and negative for cytokeratin (Pan-CKs), whereas epithelial cells do express Pan-CKs but not Vim.These immuno uorescence staining showed that endometrial stromal cell culture has a purity of up to 90% (Figure 6B).

Evaluation of the effects of E2 on epithelial differentiation of BMSCs
BMSCs were co-cultured with ESCs in Transwell system, the system is characterized by separating the cells but allowing soluble factors to pass freely between them.After 6 weeks of culture, the expression of endometrial epithelial cell markers, CK19, CK18, CK13, ER-α and PR on BMSCs were tested by Western blot.Notably, these markers expression were up-regulated in BMSCs co-cultured with ESCs compared to that in BMSCs cultured alone (Figure .6C).This suggests that endogenous factors secreted by ESCs provide an environment for BMSCs to differentiate in the direction of epithelial cells.
After 6 weeks in culture, the expression levels of the endometrial epithelial cell markers except ER-α were highest in cells cultured with 1×10 −8 mol/L E2 (p < 0.05).This data suggested that 10 nM E2 provided the optimal microenvironment for the differentiation of BMSCs towards epithelial lineages (Figure 6D).

Characterization of ESCs-Exos
Exosomes derived from ESCs (ESCs-Exos) were characterized in terms of size, morphology, and surface markers.NTA revealed that the diameters of these particles predominantly ranged from 30 nm to 150 nm (Figure 7A).ESCs-Exos exhibited a cup-or sphere-shaped morphology (Figure 7B), as shown by TEM.The identity of these particles was further con rmed as exosomes by Western blot, which showed the presence of exosomal surface markers including CD63, CD9 and TSG101 (Figure 7C).All this data suggested that these nanoparticles were indeed exosomes.

Responses of BMSCs to ESCs-Exos in vitro
The confocal microscope observed that exosomes were internalized into the cytoplasm of BMSCs and were partly co-localized within the nuclei (Figure 7D), which implied that ESCs-Exos can be transferred into stem cells with high cellular uptake e ciency.Furthermore, BMSCs were co-incubated with exosomes released from ESCs with or without E2 stimulation.After 21 days of culture, the expression of the endometrial epithelial cell markers, CK19, CK18, CK13, ER-α and PR signi cantly augmented in BMSCs treated with 50 µg/mL of ESCs-Exos.However, exosomes derived from E2-stimulated ESCs (E2-ESCs-Exos) could further promote the expression level of these markers in BMSCs (Figure 8A).Higher proportion of Pan-CKs staining positive cells were also observed in this group according to immuno uorescence analysis (Figure 8B).These results indicated that ESCs-Exos provided biochemical cues to stimulate differentiation of BMSCs towards endometrial epithelial cells, and involvement of E2 could enhance its differentiation-promoting capacity.

Discussion
Intrauterine adhesions (IUA) frequently occur after infectious or mechanical injury to the endometrium, which may lead to infertility and/or pregnancy complications [3,4] .Several strategies, such as hysteroscopic surgery to remove adhesions and hormonal therapy have been adopted for the treatment of endometrial brosis [5,6] .However, due to severe damages to the basilar layer results in loss of resident stem cells and injures the microenvironment of endometrium followed by the failure of endometrial functional layer regeneration [3,4] , treatment in severe cases is di cult and the prognosis is usually suboptimal [7] .BMSCs have been recognized as new candidates for treating serious endometrial injuries [32,33] .However, low differentiation e ciency and poor proliferation ability remain as the major hurdles BMSCs transplantation [7,8,10] .Thus, a therapeutic strategy, by modifying endometrial microenvironment with promotion of BMSCs survival, attachment, proliferation and differentiation, is needed for the restoration of endometrium.
Both Pectin and Pluronic® F-127 show well-proved biocompatibility and bioactivity [15] .In addition, Pluronic® F-127 exhibits a sol-gel transition behavior in aqueous solution when temperature increases from below to above (37 ℃ used in this study) its lower critical solution temperature [16] .Thus, the Pectin-Pluronic® F-127 solution can completely ll the uterine cavity and reach the injured region and form the hydrogel scaffold in vivo.In the present study, it was found that a Pectin-Pluronic® F-127 hydrogel with slow degradation time and three-dimensional loose network structure is more conducive to BMSCs adhesion and survival.The low degradation rate promotes their applications for long-term delivery.
BMSCs/Scaffolds group shows better integration into adjacent tissues and more effective tissue regeneration at 4 weeks after surgery than BMSCs groups, such as marked proliferation of cells and abundant neovascularization.
E2 is essential for maintaining the microenvironment of endometrium.In clinical practice, E2 therapy after adhesiolysis is considered an effective ancillary treatment to restore endometrium function for IUA patients [21] .Based on our and other previous studies, E2 is responsible for uterine endometrial regeneration through stimulating the proliferation and differentiation of stem cells and forming new capillaries after injury [8,11,20] .However, the usage of E2 is limited by its low concentration at the injured uterus, the risk for thrombosis and malignancy in systemic administration, a limited half-life period, rapid diffusion into extracellular uids, and poor solubility in aqueous solutions.Hence, it is challenging to retain E2 at the injured endometrium when administered through simple injections [8,11,20,21] .The aim of this study was to develop a new sustained release system intended for the in situ administration of E2.In this project, E2 was encapsulated into the W/O/W microspheres to construct Pectin-based E2-loaded microcapsules (E2 MPs), which could maintain the effective concentration of E2 in the medium within 30 days by the controlled release E2 to promote the differentiation of BMSCs and endometrial regeneration.
The therapeutic effect of BMSCs/E2 MPs/scaffolds group in promoting recovery of functional endometrium was also proven.At four weeks after the transplantation, BMSCs/E2 MPs/scaffolds system increased proliferative abilities of uterine endometrial and vascular endothelial cell.Furthermore, Zhang et al. demonstrated the potential of microsphere incorporated scaffolds for generating vascularized tissue when co-incubated with HUVECs, which self-assembled around the MPs and formed a 3D vascularlike network [34] .In our mouse endometrial injury model, the BMSCs/E2 MPs/scaffolds group also showed an increased expression of the blood vessel marker vWF and more vessels within regenerated endometrium.BMSCs/E2 MPs/scaffolds that promoted restoration of the endometrial construction may, to some extent, have resulted from increased vascularity that improves the access to nutrients, oxygen and hormones of injured tissues.Pregnancy rate of the BMSCs/E2 MPs/scaffolds group after the transplantation was signi cantly higher than other groups, indicating well developed endometrial glands, uterine vasculature, and a normal reproductive hormone response of the regenerated tissue in the BMSCs/E2 MPs/scaffolds group, close to normal tissue.Therefore, BMSCs/E2 MPs/scaffolds system is a promising strategy for endometrial tissue regeneration.
Although it has been recognized that E2 plays an indispensable role in promoting the repair of endometrial injury, its mechanism remains poorly understood.Based on previous studies, the proliferative and differential response of epithelial cells and endometrial stem cells to E2 is mediated by endometrial stromal cells (ESCs), which synthesize and transmit paracrine mediators to surrounding cells under the direction of E2 [23,24,35] .In the present study, BMSCs were co-cultured with ESCs in Transwell system.After 6 weeks of culture, the expression of endometrial epithelial cell markers was up-regulated in BMSCs cocultured with ESCs, and E2 is an effective inducer for facilitating the differentiation of BMSCs towards epithelial lineages.It has been noted that the exogenous stem cells implanted into the endometrium are surrounded by endometrial stroma which provides a local environment for BMSC survival and motivates cell differentiation in the direction of epithelial cells to reconstruct the injured uterine endometrium via the autocrine and paracrine pathways [36] .Several studies reported that endometrial stromal cells can secrete a large number of vesicles, i.e., exosomes, that in uence the growth, function, and development of endometrial tissue [37][38][39][40][41] .Exosomes, which act as carriers of bioactive proteins, lipid bilayer and genetic material, are considered the biological mediator of intercellular communication and play a key role in regulating cell proliferation and differentiation [25,26,42,43] .In the present study, we found that ESCs could secrete exosomes, which are successfully delivered into BMSCs with high cellular uptake e ciency and signi cantly affect cell proliferation in a dose-dependent manner.ESCs-derived exosomes provide biochemical cues to direct differentiation of BMSCs towards endometrial epithelium, and their functional effects were more signi cant than that of ESCs itself.In addition, exosomes derived from E2-stimulated ESCs could enhance the differentiation e ciency of BMSCs.These results suggest that exosomes derived from ESCs play paracrine roles in endometrial regeneration stimulated by E2 through inducing the differentiation of BMSCs into endometrial epithelial cells.

Conclusion
The BMSCs/E2 MPs/scaffolds therapeutic strategy may be bene cial in the treatment of severely damaged endometrium.The BMSCs/E2 MPs/scaffolds system could provide a three-dimensional architecture for the attachment, growth and migration of BMSCs and vascularization promotion, while serving as a reliable source of E2 to promote endometrial regeneration.Furthermore, exosomes derived from ESCs play paracrine roles in endometrial regeneration stimulated by E2, potentially modulating the differentiation of BMSCs.

Figures Figure 1 A
Figures