Treatment of Intrauterine Adhesions with Human Amnion Mesenchymal Stromal Cells (hAMSCs) on Promoting Endometrial Regeneration and Repair Via Notch Signaling Pathway Regulation

Background: Human amniotic mesenchymal stem cells(hAMSCs) can repair and improve the damaged endometrium which its aplastic disorder is the main reason for intrauterine adhesions(IUAs). Methods: We conducted in vivo and in vitro experiments. In vivo experiments: 45 female Sprague-Dawley(SD) rats were involved and randomized equally into Sham group, IUA group, Estradiol(E2) group, hAMSCs group, and E2 + hAMSCs group. The effect of hAMSCs and E2 only or combined was evaluated by Hematoxylin-eosin(HE) and Masson staining. The expression of epithelial markers and key proteins of Notch signaling pathway by Immunohistochemistry. In vitro experiments: Firstly, the hAMSCs cells were taken and divided into control group and induced group in which hAMSCs were differentiated into endometrial epithelial cells in induced microenvironment, and extracted their RNA respectively. The expression of epithelial markers and Notch1 messenger RNA (mRNA) was detected by Real-time quantitative polymerase chain reaction(qRT-PCR). and the changes in expression position of Notch intracellular domain(NICD) and expression amount of target gene, hairy enhancer of split 1(Hes1) were detected by Immunouorescence. Then, Activated and inhibited the Notch signaling pathway while induction, and detected mRNA expression of hAMSCs epithelial markers by quantitative real-time polymerase chainreaction (qRT-PCR) respectively and detected hAMSCs cell cycle by ow cytometric. Results:This study showed that hAMSCs alone or combined with E2 could promote endometrial repair, and Notch signaling pathway a great role in it. And otherwise, the activation or habitation of Notch signaling pathway determines whether hAMSCs could differentiate into endometrial epithelial cells or not. Conclusion: we concluded that activate the Notch signaling pathway promote the differentiation of hAMSCs into endometrial epithelial cells, and further treat IUAs. adenovirus hAMSCs, Jaaged1was statistically B). The cells divided into 4 groups, control group(RFP adenovirus transfected hAMSCs), induced group (hAMSCs in the RFP adenovirus transfected and induced environment), and inhibited group (hAMSCs in Adr-dnNotch adenovirus transfected and inducted environment), overexpressed group (hAMSCs in Adr-Jaaged1 adenovirus transfected and inducted environment). And changed the induced and differentiated medium just after hAMSCs adhere to the wall, and simultaneously transfected RFP adenovirus, Adr-dnNotch adenovirus, Adr-Jaaged1 adenovirus). After 3 days of culture, RNA was extracted from each group, and the epithelial markers (CK7, CK8, CK19 (cid:0) E-Cadherin) were detected by qRT-PCR. The results showed that compared with RFP control group, the expression of epithelial markers in induced group was signicantly increased, the expression of epithelial markers (CK7, CK8, CK19 (cid:0) E-Cadherin) in activated group was signicantly increased and the expression of epithelial markers (CK7,


Introduction
IUA is a kind of disease caused by damage in basal layer of the endometrium resulting in partial or complete obliteration of uterine cavity and/or the cervical canal [1][2] , which associated with infertility, oligomenorrhea and recurrent pregnancy loss and others [3][4][5][6] . The treatment aims to reestablish anatomy (removal of adhesions) and restore uterine function [7,8] , but the prognosis remains poor with a recurrence ratio of up to 62.5% [9] in surgery of restoring uterine cavity, because of failure regeneration of functioning endometrial tissue [10] . Some scholars have found that uterine cavity transplantation of mesenchymal stem cells(MSC) can effectively promote endometrial regeneration and treat uterine adhesions.
MSC can be divided into bone marrow-derived, amniotic membrane-derived, fat-derived and human menstrual blood-derived mesenchymal stem cells according to different cell sources. Human amniotic mesenchymal stromal cells (hAMSCs), isolated from discarded tissue and ethically nonproblematic, are a readily available, abundant, immunoprivileged cell source [11] . Moreover, similar to other mesenchymal stromal cells, hAMSCs are capable of differentiation into cells of all three germ layers both in vivo and in vitro, and differentiation into typical mesenchymal lineages and exhibition immunomodulatory properties through paracrine effects [12] . Therefore, hAMSCs have become exciting candidates for transplantation therapy techniques and may present as an attractive choice over other classically established stem cells for xenograft and allograft transplantation [11][12][13][14] .Our previous studies have found that hAMSCs transplantation can effectively treat uterine adhesions and found that hAMSCs have the potential to differentiate into endometrial epithelial cells.
Some studies have reported that they found genes of Notch receptors, NOTCH1, NOTCH2, and NOTCH3, the NOTCH ligand, JAG1, and downstream effectors HEY1, HEY2, and HEYL to be up-regulated in the endometrial MSC (eMSCs). This strongly shows the activation of the Notch system in the eMSC population [15] . And other studies have found that there is a signi cant abnormal expression of Notch molecules Notch1 and Jagged1 in the endometrium of uterine adhesions which indicates that the Notch signaling pathway plays a certain role in uterine adhesions. And this study further discussed the role of Notch signaling pathway in regulating the hAMSCs in IUAs treatment.
There are 4 Notch receptors (Notch1-4) and 5 Notch ligands (Jagged1, 2 and DLL1, 3,4), including intracellular, extracellular and transmembrane domains in vertebrates, which all participate in a series of physiological activities such as regulating cell proliferation and differentiation [16][17] . After the reaction between Notch ligand and receptor in adjacent cells, the outer membrane of Notch is cleaved by metalloprotease ADAM17 (also known as TACE), and the intracellular membrane hydrolytic cleavage by key enzyme γ-secretase, then the the intracellular segment (NICD) of Notch is released and enters the nucleus and binds to the DNA binding protein CSL (RBPjk/CBF1) and MAM protein to regulate the expression of target genes Hes1, 5 and Hey1 [17][18] , and exerts its biological effects.
Notch signaling is activated by the binding reaction of Notch ligand and receptor in adjacent cells. Therefore, the ligand is a good activator of Notch signal transduction. We intended to explore the Notch's role in the differentiation of hAMSCs into endometrial epithelial cells by choosing the overexpression of the ligand Jagged1 and the competitive inhibition of the receptor Notch1 to .In our study, rstly, we aimed to verify the role of Notch signaling pathway in hAMSCs transplantation in the IUA treatment; secondly, speculate that Notch signaling pathway's role in regulating the differentiation of hAMSCs into endometrial cells; and lastly, further explore the molecular mechanism of Notch signaling pathway in the differentiation of hAMSCs into endometrial cells.

Experimental animals
45 female Sprague-Dawley(SD) rats(200-220g), 8 week of age, were purchased from the Animal Experimental Center of Chongqing Medical University. The rats were bred and housed with cages(5 animals per cage) in controlled environment(12-h light-dark cycle), and room temperature 22℃, and allowed free access to chow and water. The study protocol was proved by the ethics committee of Chongqing Medical University (20141230).

hAMSCs isolation and culture
Amniotic membrane specimens were obtained from the University-Town Hospital of Chongqing Medical University and the Obstetrics Department of the First A liated Hospital of Chongqing Medical University.
The isolation and culture methods of hAMSCs followed the previous hAMSCs extraction method of research group. hAMSCs were isolated from the amniotic membranes of healthy pregnant women at day cesarean section after informed constent was obtained from them. The amniotic membrane was minced to fragments of about 1×1 mm in size, mixed them in 0.25% trypsin (Beyotime C0201, China) and digested in a water bath at 37°C with shaking for 30 min, then used 0.1% type I collagenase (Meilun MB2686, China) to digest in a water bath at 37°C with shaking for 1 h. Then, put the separated hAMSCs into Dulbecco's Modi ed Eagle Medium (DMEM) nutrient mixture F-12(DMEM/F12) (Gibco) supplemented with 10% fetal bovine serum (FBS, PAN ES) in an incubator at 37°C with 5% CO 2 (Thermo Scienti c, USA) [18] . Changed half medium after 24h, changed it completely after 72h, and took the 3rd generation cell after the cell fusion reaches more than 80% , identi ed and stored until use (Figure1 A).

Preparation of endometrial conditioned medium
12-week-old female SD rats were fully anesthetized with 5% chloral hydrate (10 ml/kg, Biosharp, China), and the bilateral uteruses were harvested under an aseptic environment and placed on ice. The serosa was carefully removed, the uteruses were dissected along its long axis and were cut into transverse segments of 0.3cm in length. Then, the segments were weighed and added with pre-cooled DMEM/F12 at a concentration of 100 g/L, which was sealed and mixed at 4 ° C with shaking in shaker for 2h, and then placed in the refrigerator at 4 ° C overnight. Next day, the mixture was centrifuged at 18000g for 30 min at 4 ° C. The supernatant was collected by using a sterile lter into a new centrifuge tube and stored at -20 °C in the refrigerator [19] .

Construction of Intrauterine adhesion model
In order to prevent the experiment from being affected by operating, the chemical injury method (95% alcohol injury method) was selected to damage the endometrium, and a stable and repeatable Intrauterine adhesion model was established. According to vaginal smear analysis, rats at diestrus were selected and fully anesthetized with 5% chloral hydrate (10 ml/kg), then their lower abdomen was disinfected with iodophor and was longitudinally incised from 3cm above the vaginal ori ce. The incision was about 2 cm in length. The abdomen was inserted layer by layer to expose the uterus (Figure1 B1). A small round needle with silk thread was used to suture the avascular area of the mesometrium, which was 0.5 cm above the Y-shaped intersection point of the right uterus. After the whole uterus was ligated, an assistant clamped the ovarium of the right uterus with at forceps. Then, 95% alcohol was injected into the uterine cavity using a 1 ml needle until the uterus was full of alcohol ; The change of uterine color was observed. When the uterus became white (about 3min later) (Figure1 B2), the 95% alcohol was drawn off and the uterine cavity was washed with normal saline. Then the assistant released the at forceps and removed the silk thread. Restored the uterus to its normal position and sutured the muscles and skin of rats. We reentered the abdomen at week1, week 2, and week 4 after modeling and observed the general morphological changes of the uterine tissue on the modeling side (Figure1 B4). Lastly, we collected uterine specimens, performed histochemical staining, and evaluated the modeling effect.
2.5 Experimental design and treatment protocol 45 SD female rats were randomized equally into ve groups, including sham group, IUA group, IUA+Estradiol(E2) group, IUA+hAMSCs group, and IUA+Estradiol(E2)+hAMSCs group. In sham group: ligated the uterus at the same position as the modeling group and clamped with at forceps without injecting alcohol. After 3 min, loosened the at forceps, then removed the silk thread to restore the anatomical position of the uterus and closed the abdomen. In IUA group: 9 rats were selected and established the uterine adhesion model, and were performed without any intervention or treatment. In IUA+E2 group: at the end of the second week after severe IUA modeling in SD rats, estrogen (estrogen tablets purchased from the University-Town Hospital of Chongqing Medical University) was ground into powder and dissolved in water for intragastric administration, 0.1 mg/kg, daily once. In IUA+hAMSCs group: at the end of the second week after severe IUA modeling in SD rats, opened abdomen at the original incision in the lower abdomen of rats, took out the right uterus, and injected hAMSCs into the uterus with a 1 ml sterile syringe where the diameter of the uterus became smaller and the texture became hard (resuspend 1*10 7 hAMSCs in 200ul PBS buffer). Then, uterine tissue were welling with the injection of hAMSCs, and closed the abdomen after hAMSCs treatment. In IUA+E2+hAMSCs group: at the end of the second week after severe IUA modeling in SD rats, hAMSCs injection and estrogen gavage in the same operation as in before group.

Sampling
In each group, 3 rats were sacri ced on the end of week 1, week 2, and week 4 after operations, and left and right uterine tissues were taken. Hematoxylin-eosin (H&E) and Masson trichrome staining were used to evaluate the specimens and immunohistochemical to detect the expression of key proteins (Notch1-4, Jagged1, Jagged2, Hes1, NICD) in the E-cadherin and Notch pathways.

Histochemical staining and image analysis
Uterine tissues were collected from the four groups at each prespeci ed period and then were xed, embedded and sectioned in 4% paraformaldehyde (Biosharp, China). The operations of HE staining and staining calculated the number of endometrial glands, Masson staining calculated the percentage of brosis area (the total area of endometrial interstitial brosis divided by the sum of the area of endometrial interstitium and glands), the collagen bers were blue under the microscope. The three-step immunohistochemical method was used to detect the expression of E-cadherin, Notch1, Notch2, Notch3, Notch4, Jagged1, Jagged2, NICD, and Hes1 in endometrial tissue. (immunohistochemistry kit, SP9001, Zhongshan Jinqiao, China). The major procedures are as follows: the sections were de-waxed to water and immersed in Phosphate buffered saline (PBS) for 3 times (5 min each time). EDTA antigen retrieval solution (AR0023,BOSTER China) was used for antigen retrieval and then the sections were immersed in PBS for 3 times (5 min each time). After that, the sections were removed of endogenous peroxidase and immersed in PBS for 3 times (5 min each time). Then, the sections were blocked with the use of goat antirabbit serum, incubated with primary antibody, each section were separately treated with E-cadherin (3195T, CST,USA), Notch1(3608S, CST, USA), Notch2 (5732, CST, USA), Notch3 (ab23426 Abcam USA), Notch4 (ab199295, Abcam, USA), Jagged1 (70109, CST, USA), Jagged2 (2205, CST, USA), NICD (ab52301, Abcam, USA), Hes1 (ab108937, Abcam, USA), which were diluted in the maximum dilution proportion using primary antibody dilution buffer overnight at 4°C and immersed in PBS for 3 times (5min each time). The treated sections were subsequently incubated using secondary antibody (biotin-labeled goat anti-rabbit IgG at 37°C for 60min) and immersed in PBS for 3 times (5min each time). Next, the horseradish peroxidase-labeled streptavidin solution was added dropwise for 15min, after which DBA (Zhongshan Jinqiao, China) staining was conducted and observed for positive staining under microscope. The sections were then rinsed with ultrapure water and counterstained with hematoxylin stain (G1080, Solarbio, China) for 20s. Thereafter, they were washed and alkalized with saturated lithium carbonate and observed under microscope. In the end, regular dehydration, transparency and mounting were conducted. The positive expression is brown-yellow, the nal results were determined according to the ratio of IOD value to area size. Image J software was used to measure the data of the pictures. Three specimens were randomly selected in each group, and three high-power elds of view were randomly selected for each slice to take pictures, and the average value was taken for subsequent comparison.

Immuno uorescence
Immuno uorescence was used to detected the expression changes of NICD and Hes1 before and after hAMSCs induction. The third-generation hAMSCs were inoculated in a 24-well culture plate (Corning, USA), and the experiment was set up in 2 groups, inducted group and control group. The cells were took and cultured in a 24-well plate, climbing piece (801010, NEST) were placed in each well according to the size of the climbing piece, rst several drops of medium were added to each well before placing climbing piece, and then place it on the droplet and press it tightly to make the climbing piece and the culture dish stick together by the tension of the culture medium for the prevention of the oating of slides at the time of cell suspension being added, which could result in cells covering two sides of the slide. Morphological changes and growth of cells were observed every day by an inverted phase contrast microscope (Nikon, Japan). After 5 days of culture, the differences in the expression of NICD and Hes1 were detected by immuno uorescence. Immuno uorescence staining procedures are as follows: the medium was extracted from the 24-well plate and were washed with PBS for 3 times (5min each time); the cells were xed with 4% paraformaldehyde for 15 min and were washed with PBS for 3 times (5min each time); then the cells were xed with 0.1% Triton X-100 (sigma V900502) for 15 min; next, they were blocked with 10% goat serum (AR004) which was added in a drop wise manner and incubated at room temperature for 30 min without washing; thereafter, the cells were added anti-rabbit NICD monoclonal antibody in a 1:100 ratio and anti-rabbit Hes1 monoclonal antibody in a 1:1000 ratio. The control group was added with PBS, incubated overnight at 4 ° C in the dark and washed with PBS for 3 times (5min each time); goat antirabbit IgG Alexa Fluor uorescent secondary antibody(ab150077, Abcam, USA) were then added in a 1:100 ratio and incubated in the dark at 37 ° C for 2h and were washed with PBS for 4 times (5min each time); the nuclei were counterstained with 0.5 μg/ml DAPI uorescent dye (AR1177, Boster Bio, China) and placed in the dark for 10 min, which was washed subsequently with PBS for 3 times to remove the remaining DAPI. After that, 20 μl anti-uorescent quencher (Meilun MA0222) was added for mounting. The slides were observed and photographed under a uorescence microscope. Positive NICD was green uorescence and nuclei were blue uorescence.
2.11 Recombinant adenovirus vector AdR-dnNotch1, Adr-Jag1, Ad-RFP transfection AdR-dnNotch1, Adr-Jagged1, Ad-RFP were purchased from denovo company, Ad-RFP were used as a virus control. In order to determine the relationship between the Notch signaling pathway and the induction and differentiation of hAMSCs, Adr-dnNotch1 (inhibited group), Adr-Jaagged1 (activated group) and Ad-RFP (control group) were added to hAMSCs induction and differentiation medium to irreversibly interfere with the activation of the pathway. The medium was changed every other day, continuously cultured for 36 and 72h. And real-time uorescence quantitative PCR was used to detect total RNA, repeated 3 times.

Flow cytometry analysis
The old medium was aspirated by the cells in inhibited group, activated group and control group , washed the bottom of the bottle with PBS, and digested the cells into single cells with 0.25% trypsin (Beyotime, China). Then, added PBS buffer(pre-cooled), centrifuged at 1000r/min (Thermo Scienti c, USA) for 5min, discarded the supernatant. Next, added PBS to resuspend the cells, centrifuged at 1000r/min for 5min, discard the supernatant, and repeat twice. Finally, 1×10 6 cells were resuspended in PBS 0.1ml PBS buffer, 500ul of pre-cooled 75% ethanol (Chongqing Chuandong Chemical Co., Ltd., China) was slowly added, and shaking when adding to avoid cell aggregation, and then used ow cytometry to analyze the percentage of cells in G0/G1, S and G2/M phases. All reactions were carried out in three stages.

Statistical analysis
Data was analyzed with SPSS 22.0 software and expressed as `x ±s. Normal analysis and homogeneity of variance test were performed. Independent sample t test was used for comparison between two groups, single-factor analysis of variance was used for multiple group comparisons, and multiple comparisons between groups were performed by LSD method. The inspection level p = 0.05.

Establishment of IUA model and evaluation on hAMSCs combined with estrogen therapy
We took uterine tissue for histochemical staining (HE, Masson and immunohistochemistry). The results of HE staining showed that the morphology of endometrium in sham group was normal, the endometrial glands were scattered in submucosa and basal layer, and the surface of endometrium was covered with columnar epithelial cells (Fig. 1C1); while in IUA group, the volume of uterine cavity disappeared, the uterine muscle wall adhesion was dense, the number of endometrial glands was signi cantly reduced, the columnar epithelial cells on the surface of endometrium disappeared, and the thickness of endometrium became thinner (Fig. 1C2). The effect of E2 group was not obvious, with thicker endometrium, but showed not restore in uterine cavity shape, increase in number of glands, or recover in columnar epithelium on endometrial surface (Fig. 1C3). Compared with IUA group, the shape of uterine cavity in hAMSCs group was restored, and the endometrial surface was signi cantly covered with columnar epithelial cells, and the thickness of endometrium and the number of glands were increased (Fig. 1C4). It had the best therapeutic effect with restored uterine cavity morphology in hAMSCs + E2 group, with signi cantly regenerated single-layer tall columnar epithelial cells, thicker in endometrium, increased number of glands (Fig. 1C5).The mean number of endometrial glands was 6.77 ± 0.78 in sham group. And it was 1.14 ± 0.37 and 1.69 ± 0.57 in IUA group and IUA + E2 group, respectively, which was signi cantly lower than that in sham group (P < 0.05). The number of endometrial glands in IUA + hAMSCs group and IUA + hAMSCs + E2 group was 4.96 ± 0.80 and 5.66 ± 0.88, respectively, which was signi cantly increased when compared with IUA group (P < 0.05) (Fig. 1E ). The mean endometrial thickness in sham group was 6.77 ± 0.78. And it in IUA group was 1.69 ± 0.57, which was signi cantly lower than that in sham group (P < 0.05). The endometrial thickness in IUA + E2 group, IUA + hAMSCs group and IUA + hAMSCs + E2 group was 1.14 ± 0.37, which was signi cantly higher than that in IUA group (P < 0.05) (Fig. 1F).
The results of Masson staining showed that even distribution of collagen bers in sham group (Fig. 1D1), but signi cantly darker in the color of collagen ber in IUA group (Fig. 1D2); and signi cantly reduced in degree of endometrial brosis in IUA + E2 group, IUA + hAMSCs group and IUA + hAMSCs + E2 group(P < 0.05) ( Fig. 1D3D4D5).
We evaluated the immunoexpression of epithelial marker E-Cadherin to evaluate the regeneration and recovery of epithelial cells. The results of immunohistochemistry showed that E-Cadherin was signi cantly expressed in uterine cavity epithelial cells and glandular epithelial cells in sham group (Fig. 1H1), and it signi cantly increased in IUA + hAMSCs group and IUA + hAMSCs + E2 group (Fig. 1H4H5), but signi cantly decreased in IUA group and E2 group(P <0.05) (Fig. 1H2H3).

Important role of Notch signaling pathway in hAMSCs for IUA treatment
We detected the expression of NICD and Hes1 to verify the role of Notch signaling pathway in hAMSCs transplantation of the treatment of IUAs, for which NICD is the activated form and Hes1 is the downstream target gene of Notch signaling pathway, so we can evaluate whether the Notch signaling pathway or not through changes of expression volume and position of these two. The expression of NICD and Hes1 in endometrial tissue in IUA + hAMSCs group and IUA + hAMSCs + E2 group (Fig. 2A4A5C4C5), but lower in IUA group and IUA + E2 group (Fig. 2A2A3C2C3). And it also showed that NICD is mainly expressed in nucleus in IUA + hAMSCs group and IUA + hAMSCs + E2 group, but in cytoplasm in sham group and Hes1 was expressed in nucleus in each group(P < 0.05) ( Fig. 2B D).
Then we detected the expression of Notch receptors (Notch1-4) and Notch ligands (Jagged1-2) by using immunohistochemical staining to analyze which receptors and ligands in the Notch signaling pathway affect the differentiation of hAMSCs. And the results showed that receptors Notch1, Notch3, and ligand Jagged1 were expressed in uterine tissues in each group, and all expressed on cell membrane, while Notch2, Notch4, and Jagged2 were not signi cantly expressed in uterine tissues. The expression of Notch1 and Jagged1 was signi cantly increased in sham group, was low in IUA group and E2 group, and the expression in hAMSCs group and the hAMSCs + E2 group was much higher than that in IUA group. The expression of Notch3 existed in sham group, increased in IUA group and IUA + E2 group, and it in IUA + hAMSCs group and IUA + hAMSCs + E2 group, but showed slightly lower than that in IUA group, the difference was statistically signi cant (P < 0.005). The expression trend of Notch1 and Jagged1 in each group was consistent with that of NICD and Hes1, but different from Notch3 (Fig. 2E). Therefore Notch1 and Jagged1 play an important role in hAMSCs transplantation of IUA treatment.

Induction and differentiation of hAMSCs into endometrial epithelial cells under microenvironment of endometrial conditions
We conducted the in vitro experiments to verify whether hAMSCs can induce differentiation into endometrial epithelial cells or not. We rst, simulated the uterine microenvironment and extract the endometrial conditioned medium, and induced hAMSCs to differentiate into endometrial epithelial cells under the induction condition of combing growth factors and estradiol(growth factor (TGF-β1 + EGF + PDGF-BB) + 10 − 6 mol/Lβ-estradiol + endometrial conditioned medium) then cultured it for 3 days and 5 days. From the morphological observation, the hAMSCs in control group were polygonal or long fusiform (Fig. 3A1), and the cells in induced group were round or round (epithelial cell morphology transformation) (Fig. 3A2), and the changes were more clear in days 5 (Fig. 3C1C2). RNA wads extracted in days 3 and days 5 in two groups to detected the expression of mRNA epithelial markers (CK7, CK8, CK18, CK19, E-Cadherin) by qRT-PCR. And the results showed that when compared with control group after days 3, the expression of CK7, CK8, CK19, and E-Cadherin mRNA in inducted group increased signi cantly, and the difference was statistically signi cant (P < 0.05 ), while CK18 was not signi cant ( Fig. 3B). And compared with control group after days 5, the expression of CK7, CK8, CK18 and E-Cadherin mRNA in induced group increased signi cantly, and the difference was statistically signi cant (P < 0.05 ), while CK19 had a downward trend (Fig. 3D).

mRNA and protein expression of key molecules in
Notch signaling pathway during the induction and differentiation of hAMSCs Compared with control group, the expression of epithelial markers in the induced group was signi cantly increased, which indicated that hAMSCs can differentiate into endometrial epithelial cells in a suitable induced microenvironment.We detected the mRNA expression of Notch1 in control group and induced group by qRT-PCR on the day 3, day 5 to clarify the role of the Notch signaling pathway in the differentiation process and detected the expression of NICD and Hes1 by immuno uorescence. The results of qRT-PCR showed that when compared with control group, the expression of Notch1 mRNA in induced group increased on day 3 and day 5, and the difference was statistically signi cant (P < 0.05) (Fig. 4C). The results of immuno uorescence showed that compared with control group, the expression of NICD and Hes1 green uorescence in hAMSCs in induced group was increased, the difference was statistically signi cant (P < 0.05) (Fig. 4A D B E),and NICD was mainly expressed in the cytoplasm in control group, while in induced group it was mainly expressed in the nucleus (Fig. 4A).

Up-regulation or down-regulation of Notch signaling pathway affects the differentiation of hAMSCs into endometrial epithelial cells
In previous experiments, we found that Notch1 mRNA expression changed during the differentiation of hAMSCs into endometrial epithelial cells, and NICD and Hes1 protein expression also changed. But whether the Notch signaling pathway plays a decisive role or not, we used Adr-dnNotch adenovirus to down-regulate the expression of Notch1 to inhibit the Notch signaling pathway, and Adr-Jaaged1 adenovirus to up-regulate the expression of Jagged1 in Notch signal pathway to activated and then detected the expression changes of epithelial markers (CK7, CK8, CK19,E-Cadherin) mRNA.
qRT-PCR was performed to detect mRNA expression of Notch1 and Jaaged1 to verify Adr-dnNotch adenovirus could reduce the expression of Notch1, and Adr-Jaaged1 adenovirus could overexpress Jaaged1. After transfection of Adr-dnNotch adenovirus and Adr-Jaaged1 adenovirus were respectively into hAMSCs, RNA was extracted at the end of day 3 and tested by qRT-PCR. Compared with hAMSCs, the transfection of Adr-dnNotch adenovirus into hAMSCs, the expression of Notch1 mRNA was reduced, and the difference was statistically signi cant (P < 0.05) (Fig. 5A). After transfection of Adr-Jaaged1 adenovirus into hAMSCs, the mRNA expression of Jaaged1was signi cantly reduced, and the difference was statistically signi cant (P < 0.05) (Fig. 5B).
The cells were divided into 4 groups, including control group(RFP adenovirus transfected hAMSCs), induced group (hAMSCs in the RFP adenovirus transfected and induced environment), and inhibited group (hAMSCs in Adr-dnNotch adenovirus transfected and inducted environment), overexpressed group (hAMSCs in Adr-Jaaged1 adenovirus transfected and inducted environment). And changed the induced and differentiated medium just after hAMSCs adhere to the wall, and simultaneously transfected RFP adenovirus, Adr-dnNotch adenovirus, Adr-Jaaged1 adenovirus). After 3 days of culture, RNA was extracted from each group, and the epithelial markers (CK7, CK8, CK19,E-Cadherin) were detected by qRT-PCR. The results showed that compared with RFP control group, the expression of epithelial markers in induced group was signi cantly increased, the expression of epithelial markers (CK7, CK8, CK19 E-Cadherin) in activated group was signi cantly increased and the expression of epithelial markers (CK7, CK8, CK19 E-Cadherin) mRNA in inhabited group was signi cantly reduced, and the difference was statistically signi cant (P < 0.05) ( Figure 5D (Fig. 6A C), the difference was statistically signi cant (P < 0.05) ( Fig. 6B D) .And there was no signi cant changes in cell cycle in each group on the 72 hours (Fig. 6E).

Discussion
IUA is a disease of failure of endometrial regeneration caused by damage in the basal layer. It is still challenging to promote endometrial repair for severe IUAs [20] . Stem cell therapy is a new method to treat tissue damage and promote tissue regeneration. Different studies have con rmed that Mesenchymal Stem Cells (MSCs) play an important role in IUAs treatment [21] , so amniotic mesenchymal stem cells (hAMSCs) ) is a kind of adult stem cell with self-renewal and multi-differentiation potential for its rich source and low cost. In a suitable in vivo or in vitro induced microenvironment, hAMSCs can differentiate into decidua, osteogenesis, brogenesis, adipogenesis, etc [22][23] . In this experiment, hAMSCs were selected as ideal stem cells to treat IUAs.
In this experiment, the chemical damage method (95% alcohol damage method) was selected to establish a IUA model its stable, repeatable, and will not be affected by pro ciency of human operation [24] . the endometrium cycle-to-cycle would necessitate self-renewal as well as migration into the remaining tissue (functionalis and basalis) and differentiation into lineage cells (stromal broblasts and perhaps other lineages) [25][26] .Based on an effective animal model, hAMSCs transplantation combined with estrogen therapy can restore the shape of uterine cavity, promote the recovery of endometrial epithelial cells, promote endometrial thickening and glandular hyperplasia, and reduce the degree of brosis and has a positive effect when compared with estrogen alone and hAMSCs alone (Fig. 1C D).
Epithelial repair may be associated with resurfacing of epithelial cells from glands in the basal layer [27] ,These cells on the luminal surface were positive for E-Cadherin, (Fig. 1H) which indicated better epithelium recovery following hAMSCs transplantation. [28] In addition, the increasing expression of the activated form of Notch receptor NICD in the Notch signaling pathway in hAMSCs and estrogen alone or combined indicates that the Notch signaling pathway was activated during the treatment of IUAs by hAMSCs and estrogen ( Fig. 2A). Then we detected the expression of Notch receptors (Notch1-4) and Notch ligands (Jagged1-2) by using immunohistochemical staining to analyze which receptors and ligands in the Notch signaling pathway affect the differentiation of hAMSCs. And the results showed that Notch1 and Jagged1 play an important role in hAMSCs transplantation of IUA treatment (Fig. 2E).
In this experiment, we speculated the function of hAMSCs in differentiation into endometrial epithelial cells in uterine microenvironment on regeneration and repair of epithelial cells in hAMSCs combined with estrogen therapy group, then the Notch signaling pathway's function on hAMSCs in IUAs treatment because it was activated in hAMSCs combined with estrogen treatment, and last, under simulated in vitro uterine microenvironment, we speculated Notch signaling pathway can induce hAMSCs differentiate into endometrial cells and it is decisive in this process.
To verify our speculation, rstly, we used an effective method, the largest changes in expression of epithelialization markers as induced method, to promote the endometrialization of hAMSCs which was the method by comparing and improving the methods reported by others. This induced method was to simulates the endometrial microenvironment with endometrial conditioned medium + β-estradiol + growth factor (TGF-β1 + EGF + PDGF-BB). And in this condition, the expression of hAMSCs epithelial markers was signi cantly increased, which veri ed that hAMSCs can differentiate into endometrial epithelial cells in in vivo experiments, moreover, the expressions of NICD and Hes1 in hAMSCs were signi cantly higher than others, which proved that the Notch signal pathway was activated and participated in the differentiation process of hAMSCs. And secondly, we interfered the Notch signal pathway (activation and inhibition) when inducing and observed the changes of its epithelial markers. Through the Adr-dnNotch adenovirus, inhibited or activated in the Notch signaling pathway by Adr-Jagged1 overexpression adenovirus, and the decreased or increased expression of hAMSCs epithelial markers, we can prove that the Notch signaling pathway is decisive in differentiation process of hAMSCs into endometrial epithelial cells which inhibited the Notch signaling pathway stop hAMSCs differentiate into endometrial epithelial cells but activated can strengthen this differentiation (Fig. 5C). This paper found that Notch signaling pathway promoted the transformation of hAMSCs from G1 phase to S phase (Fig. 6).Stem cells have high proliferation ability to achieve self-renewal before differentiation, and can differentiate into speci c cells under appropriate induction conditions.
In summary, this study con rmed that activation of the Notch signaling pathway can promote the differentiation of hAMSCs into endometrial epithelial cells, thereby promoting the regeneration and repair of endometrium of IUAs.

Conclusion
This study indicated that estrogen or hAMSCs alone cannot exert the best function in treatment of IUAs, while the combination of the two can effectively treat IUAs and promote endometrial repair. In addition, it studied the differentiation of hAMSCs into endometrial epithelial cells and the molecular mechanism of Notch signaling pathway in this process, and further clari ed the molecular mechanism of hAMSCs in repair of endometrial injury. Therefore, therapeutic effect of the cell could be further explored, when a certain stem cell were transplanted into the uterine cavity after inducing to differentiate in vitro, or transplanted into the uterine cavity after interfering with the Notch signaling pathway. This study provides a reliable molecular biological basis for elucidating hAMSCs transplantation for the treatment of IUAs, and provides new ideas for clinical treatment of endometrial injury .  We took uterine tissue for histochemical staining (HE, Masson and immunohistochemistry). The results of HE staining showed that the morphology of endometrium in sham group was normal, the endometrial glands were scattered in submucosa and basal layer, and the surface of endometrium was covered with columnar epithelial cells (Figure 1 C1); while in IUA group, the volume of uterine cavity disappeared, the uterine muscle wall adhesion was dense, the number of endometrial glands was signi cantly reduced, the columnar epithelial cells on the surface of endometrium disappeared, and the thickness of endometrium became thinner (Figure 1 C2). The effect of E2 group was not obvious, with thicker endometrium, but showed not restore in uterine cavity shape, increase in number of glands, or recover in columnar epithelium on endometrial surface (Figure 1 C3). Compared with IUA group, the shape of uterine cavity in hAMSCs group was restored, and the endometrial surface was signi cantly covered with columnar epithelial cells, and the thickness of endometrium and the number of glands were increased (Figure 1 C4). It had the best therapeutic effect with restored uterine cavity morphology in hAMSCs+E2 group, with signi cantly regenerated single-layer tall columnar epithelial cells, thicker in endometrium, increased number of glands (Figure 1 C5).The mean number of endometrial glands was 6.77±0.78 in sham group.
And it was 1.14±0.37 and 1.69±0.57 in IUA group and IUA+E2 group, respectively, which was signi cantly lower than that in sham group (P<0.05). The number of endometrial glands in IUA+hAMSCs group and IUA+hAMSCs+E2 group was 4.96±0.80 and 5.66±0.88, respectively, which was signi cantly increased when compared with IUA group (P<0.05) (Figure 1 E ). The mean endometrial thickness in sham group was 6.77±0.78. And it in IUA group was 1.69±0.57, which was signi cantly lower than that in sham group (P<0.05). The endometrial thickness in IUA+E2 group, IUA+hAMSCs group and IUA+hAMSCs+E2 group was 1.14±0.37, which was signi cantly higher than that in IUA group (P<0.05) (Figure 1   And it also showed that NICD is mainly expressed in nucleus in IUA+hAMSCs group and IUA+hAMSCs+E2 group, but in cytoplasm in sham group and Hes1 was expressed in nucleus in each group(P<0.05) ( Figure 2 B D). Then we detected the expression of Notch receptors (Notch1-4) and Notch ligands (Jagged1-2) by using immunohistochemical staining to analyze which receptors and ligands in the Notch signaling pathway affect the differentiation of hAMSCs. And the results showed that receptors Notch1, Notch3, and ligand Jagged1 were expressed in uterine tissues in each group, and all expressed on cell membrane, while Notch2, Notch4, and Jagged2 were not signi cantly expressed in uterine tissues. The expression of Notch1 and Jagged1 was signi cantly increased in sham group, was low in IUA group and E2 group, and the expression in hAMSCs group and the hAMSCs+E2 group was much higher than that in IUA group. The expression of Notch3 existed in sham group, increased in IUA group and IUA+E2 group, and it in IUA+hAMSCs group and IUA+hAMSCs +E2 group, but showed slightly lower than that in IUA group, the difference was statistically signi cant (P <0.005). The expression trend of Notch1 and Jagged1 in each group was consistent with that of NICD and Hes1, but different from Notch3 (Figure 2 E). Therefore Notch1 and Jagged1 play an important role in hAMSCs transplantation of IUA treatment.  Compared with control group, the expression of epithelial markers in the induced group was signi cantly increased, which indicated that hAMSCs can differentiate into endometrial epithelial cells in a suitable induced microenvironment.We detected the mRNA expression of Notch1 in control group and induced group by qRT-PCR on the day 3, day 5 to clarify the role of the Notch signaling pathway in the differentiation process and detected the expression of NICD and Hes1 by immuno uorescence. The results of qRT-PCR showed that when compared with control group, the expression of Notch1 mRNA in induced group increased on day 3 and day 5, and the difference was statistically signi cant (P<0.05) (Figure 4 C). The results of immuno uorescence showed that compared with control group, the expression of NICD and Hes1 green uorescence in hAMSCs in induced group was increased, the difference was statistically signi cant (P<0.05) (Figure 4 A D B E),and NICD was mainly expressed in the cytoplasm in control group, while in induced group it was mainly expressed in the nucleus ( Figure 4A). Figure 5 qRT-PCR was performed to detect mRNA expression of Notch1 and Jaaged1 to verify Adr-dnNotch adenovirus could reduce the expression of Notch1, and Adr-Jaaged1 adenovirus could overexpress Jaaged1. After transfection of Adr-dnNotch adenovirus and Adr-Jaaged1 adenovirus were respectively into hAMSCs, RNA was extracted at the end of day 3 and tested by qRT-PCR. Compared with hAMSCs, the transfection of Adr-dnNotch adenovirus into hAMSCs, the expression of Notch1 mRNA was reduced, and the difference was statistically signi cant (P<0.05) (Figure 5 A). After transfection of Adr-Jaaged1 adenovirus into hAMSCs, the mRNA expression of Jaaged1was signi cantly reduced, and the difference was statistically signi cant (P<0.05) (Figure 5 B)