Stem cells implanted with nanofibrous mats for injured endometrial regeneration and immune-microenvironment remodeling.

Lina Zhou, Hao Wang and Danpeng Shen: Conceptualization, Methodology, Investigation, Resources, Writing - Original Draft; Jiangdong Xiang, Na Yu, Xiaoying He, and Weiguang Zhao: Investigation, Writing - Original Draft, Data Curation, Formal analysis; Renjie Wang, Hongwei Wang, Hongliang Yu, and Xueying Ding: Formal analysis, Data Curation; Yinyan He, Zhihong Liu and Lina Zhou: Conceptualization, Funding acquisition, Writing - Review & Editing, Visualization.

Not applicable.

The study was reviewed and approved by the Institutional Review Board and the Research Ethics Committee of Shanghai General Hospital.

The present study was supported by the Clinical Research Innovation Plan of 10.13039/501100013103 Shanghai General Hospital (No. CTCCR-2019D01, to Yinyan He), the Clinical Research Plan of SHDC (SHDC2020CR4091, to Yinyan He), the Integrated Traditional and Western Medicine Hospitals Foundation of SHDC For Compositive Hospital (ZHYY-ZXYJHZX-201912, to Yinyan He), the 10.13039/501100001809 National Natural Science Foundation of China (82072823, to Zhihong Liu), the National Natural Science Foundation of Shanghai (20ZR1444900, to Zhihong Liu) and the Clinical Characteristic Medical Technology Cultivation Plan of Shanghai General Hospital (No. 02.DY12.06.22.07 to Lina Zhou and No. 02.DY12.06.22.12 to Jiangdong Xiang).

Both the extraction method and direct contact method showed that various SF/PCL nanofibrous mats had no toxicity to ADMSCs at different times and did not interfere with the cells (detailed in Supplementary material Fig. S1). The distribution and morphology of ADMSCs on the mat were observed by SEM. Many ADMSCs were located on the surface of the SF/PCL nanofibrous mats, and some cells could even be observed in the deep layer of the mats. Thus, ADMSCs not only adhered to the surface of the mats but also migrated into the interior of the mats. More ADMSCs could be found on the (SF/PCL) 3:1 mats compared with those with other ratios, indicating that the (SF/PCL) 3:1 mats were more suitable for the growth of ADMSCs. The above results confirm that (SF/PCL) 3:1 could recruit more stem cells in the period of 3–24 h, which was in line with the requirement of the clinical application of stem cell therapy. In brief, these results suggested that the SF/PCL mats and ADMSCs can be effectively loaded together (ADMSCs-SF/PCL system), and the system was expected to facilitate ADMSC proliferation and viability ( Fig. 1 A–C). Fig. 1 SEM images of ADMSCs-SF/PCL. A: (SF/PCL) 2:1 ; B: (SF/PCL) 3:1 ; C: (SF/PCL) 4:1 Fig. 1 SEM images of ADMSCs-SF/PCL. A: (SF/PCL) 2:1 ; B: (SF/PCL) 3:1 ; C: (SF/PCL) 4:1 HE staining was used to evaluate the regeneration of the endometrium. On Day 14 post-surgery, the sham group (Group S ) exhibited a star-shaped or polygonal irregular uterine cavity, while the model group (Group M ) showed regeneration of the endometrium without luminal structures. However, the three treatment groups (Groups E, AE and A ) exhibited regeneration of the endometrium with apparent luminal structures ( Fig. 2 , Line 1). On Day 28 post-surgery, the endometrium of Group S was continuous with ordered glands, while in Group M, the endometrial epithelium was discontinuous without luminal structures or ordered glands. The endometria of the three treatment groups were well organized, and the secretory glands were abundant ( Fig. 2 , Line 2). On Day 60 post-surgery, Groups S, E, AE and A showed normal-appearing endometrium, while Group M still exhibited severe IUAs ( Fig. 2 , Line 3). Fig. 2 HE and Masson’s trichrome staining of harvested uteri. Line 1: Restoration of the endometrium at Day 14 post-surgery. Line 2: Restoration of the endometrium at Day 28 post-surgery. Line 3: Restoration of the endometrium at Day 60 post-surgery. Line 4: The collagen remodeling at Day 14 post-surgery. Line 5: The collagen remodeling at Day 28 post-surgery. Line 6: The collagen remodeling at Day 60 post-surgery. Fig. 2 HE and Masson’s trichrome staining of harvested uteri. Line 1: Restoration of the endometrium at Day 14 post-surgery. Line 2: Restoration of the endometrium at Day 28 post-surgery. Line 3: Restoration of the endometrium at Day 60 post-surgery. Line 4: The collagen remodeling at Day 14 post-surgery. Line 5: The collagen remodeling at Day 28 post-surgery. Line 6: The collagen remodeling at Day 60 post-surgery. Masson’s trichrome staining was used to assess collagen remodeling in the reconstructed endometrium. On Day 14 post-surgery, Groups S, E and AE showed lower collagen deposition than Group A and Group M ( Fig. 2 , Line 4), reflecting the mild fibrosis caused by estrogen treatment. On Day 28, the collagen deposition in Group AE was the most similar to that in Group S (p > 0.05) and much lower than that in Group M (p < 0.01, Fig. 2 , Line 5). Collagen deposition was more obvious in Group M from Day 14 to Day 60, especially on Day 60 post-surgery. Extremely dense endometrial fibrosis was found in Group M ( Fig. 2 , Line 6), which indicated accompanying endometrial damage. As shown in Table S3, the gland number in the regenerated endometrium of Groups E, AE and A was substantially greater than that of Group M (p < 0.05). These results support the notion that the three different treatments (E, AE and A) exhibited substantial efficacy in endometrial regeneration. Most importantly, on Day 60 post-surgery, Group AE showed the strongest reversal of collagen deposition compared to the other two treatment groups and Group M. HOXA11, HAND2 and FOXO1 are involved in the proliferation, differentiation, and decidualization of the endometrium. After endometrial damage, the expression of HOXA11, HAND2 and FOXO1 was substantially downregulated in Group M ( Fig. 3 ). After treatments, the expression of HOXA11 was significantly upregulated in Groups E and A. Compared with that of Group E, the expression of HAND2 and FOXO1was dramatically upregulated in Groups AE and A. There was no difference in the expression of PRL-R after injury and treatment, which suggested that endometrial injury did not change the expression of PRL-R. Fig. 3 Expression of endometrial specific markers(Smad3, TGF-β1, HOXA11, HAND2, FOXO1 and PRL-R). Fig. 3 Expression of endometrial specific markers(Smad3, TGF-β1, HOXA11, HAND2, FOXO1 and PRL-R). TGF-β1/Smad3 signaling plays a pivotal role in tissue fibrosis. The highest expression of TGF-β1/Smad3 in the endometrium was observed in Group M ( Fig. 3 ), which was attributed to endometrial damage. In the three treatment groups (E, AE, A), the expression of TGF-β1/Smad3 was drastically downregulated, indicating that the three treatments could restrain the TGF-β1/Smad3 pathway and then block fibrosis. On Day 14, CK19, a marker of the epithelium, was not expressed in Group M, however, it was observed in the three treatment groups without typical glandular epithelium. Then, on Day 28, in the three treatment groups, not only did the expression of CK19 increased but also the glandular epithelium was more apparent ( Fig. 4 A–C top panel). Interestingly, CD31 was highly expressed in the three treatment groups with vascular morphology on Day 28 compared to Day 14 or Day 60 ( Fig. 4 A–C middle panel). These results indicated that 28 days may be the best recovery time because both CK19 and CD31 were expressed at high levels. Fig. 4 Expression of markers of endometrial vascular proliferation and fibrosis. A: Day 14 post-surgery; B: Day 28 post-surgery; C: Day 60 post-surgery. A-C top panel: The expression of CK19. A-C middle panel: The expression of CD31. A-C bottom panel: The expression of α-SMA. Fig. 4 Expression of markers of endometrial vascular proliferation and fibrosis. A: Day 14 post-surgery; B: Day 28 post-surgery; C: Day 60 post-surgery. A-C top panel: The expression of CK19. A-C middle panel: The expression of CD31. A-C bottom panel: The expression of α-SMA. α-Smooth muscle actin (α-SMA) is a marker for a subset of activated fibrogenic cells and myofibroblasts, mediating the process of fibrosis in IUAs. On Day 14, the expression of α-SMA in Group C was lower than that in the other four groups, but there was no significant difference on Day 28 and Day 60 among the five groups ( Fig. 4 A–C bottom panel), which suggested an increase in α-SMA expression after endometrial damage. Generally, in the endometrium, the predominant lymphocyte population was uterine NK (uNK) cells (50–75%). CD4 + and CD8 + T cells and macrophages also infiltrated. Rat NKR-P1 (CD161a) receptors were primarily expressed in NK cells. The proportion of NK cells in Group C was much higher than that in the surgery groups (Group S and Group M), and Group M had the lowest expression of CD161a (P < 0.05, Fig. 5 A). Additionally, an increase in CD68 + macrophages (Møs) was observed in Group S compared to Groups C and M. These results documented a dramatic decrease in the proportion of NK cell infiltration and a rapid increase in the proportion of T cells after endometrial injury. Fig. 5 The proportion of NK cells, T cells and macrophages, CD8 + and CD4 + T cells in the endometrial immune microenvironment of each group. A–B: The proportion of NK cells, T cells and macrophages. C–D: The proportion of CD8 + T cells and CD4 + T cells. *: P < 0.05, vs. the same cells of Group C; $: P < 0.05, vs. the same cells of Group S; #: P  0.05, vs. the same cells of Group C; Δ: P > 0.05, vs. the same cells of Group M; Ω: P  0.05 vs. the same cells of Group AE; a: P  0.05 for any two groups. Fig. 5 The proportion of NK cells, T cells and macrophages, CD8 + and CD4 + T cells in the endometrial immune microenvironment of each group. A–B: The proportion of NK cells, T cells and macrophages. C–D: The proportion of CD8 + T cells and CD4 + T cells. *: P < 0.05, vs. the same cells of Group C; $: P < 0.05, vs. the same cells of Group S; #: P  0.05, vs. the same cells of Group C; Δ: P > 0.05, vs. the same cells of Group M; Ω: P  0.05 vs. the same cells of Group AE; a: P  0.05 for any two groups. Interestingly, the highest proportion of CD3 + T cells along with a reduction in CD161a+ NK cells was identified in Group M compared to Group C ( Fig. 5 B). In Group AE, the proportion of CD68 + macrophages was higher than that in Group C and Group M, and the proportion of CD161a+ NK cells increased to the highest level. The proportions of T cells, CD68 + cells and NK cells in the immune microenvironment of Group A were the most consistent with those of Group C. The treatments putatively favored the recovery of NK cell, T-cell and Mø infiltration in the endometrial microenvironment. In detail, for T-cell subsets, the main population was CD8 + T cells, which are twice as abundant as CD4 + T cells (66% vs. 33%) [ 44 ]. In Groups S and M, the proportion of CD8 + T cells was considerably higher than that of CD4 + T cells, which was consistent with the results in Group C. However, the lowest proportion of CD4 + T cells and the highest proportion of CD8 + T cells were found in Group M. Group M exhibited disparate change patterns of both CD4 + and CD8 + T cells after injury ( Fig. 5 C). After the three treatments (E, AE, and A), the proportion of CD8 + T cells decreased, which was most consistent with the results of Group C, although the proportion of CD4 + T cells in Group A was much lower than that in Group C. All these findings demonstrated that the restoration of CD8 + T cells in the treatment groups, especially in Groups A and AE achieved the same degree of the normal or control group, yet CD4 + T cells did not recover very well in Group A ( Fig. 5 D). No difference in the proportion of NKG2A + and NKG2D + NK cells was observed between Groups S and M, however, NKG2A + NK cells in both groups were higher than those in Group C, following the significant reduction in NKG2D + NK cells in both groups compared to Group C. These results displayed a distinct pattern of change in which NKG2A + NK cells increased and NKG2D + NK cells decreased after endometrial injury ( Fig. 6 A and B). Generally, as shown in Fig. 5 , Fig. 6 , all treatments contributed to the recovery of the whole NK cell population, but they were not sufficient to restore the NKG2D + NK-cell subset to the same degree as that of the control group. Fig. 6 The proportion of NKG2A+ and NKG2D+ NK cells, CD86 + and CD163+ macrophages, and Th1 and Th2 cells in the endometrial immune microenvironment. A–B: The proportion of NKG2A+ and NKG2D+ NK cells. C–D: The proportion of CD86 + and CD163+ macrophages. E–F: The proportion of Th1 and Th2 cells. *: P < 0.05, vs. the same cells of Group C; $: P  0.05, vs. the same cells of Group S; #: P  0.05, vs. the same cells of Group C; Δ: P > 0.05, vs. the same cells of Group M; Ω: P  0.05 vs. the same cells of Group AE; a: P  0.05 for any two groups. Fig. 6 The proportion of NKG2A+ and NKG2D+ NK cells, CD86 + and CD163+ macrophages, and Th1 and Th2 cells in the endometrial immune microenvironment. A–B: The proportion of NKG2A+ and NKG2D+ NK cells. C–D: The proportion of CD86 + and CD163+ macrophages. E–F: The proportion of Th1 and Th2 cells. *: P < 0.05, vs. the same cells of Group C; $: P  0.05, vs. the same cells of Group S; #: P  0.05, vs. the same cells of Group C; Δ: P > 0.05, vs. the same cells of Group M; Ω: P  0.05 vs. the same cells of Group AE; a: P  0.05 for any two groups. Type 2 macrophages (M2) are the dominant macrophages in the normal endometrial immune microenvironment [ 45 ]. CD86 + macrophages (M1) were apparently increased in Group M after injury compared with both Groups C and S ( Fig. 6 C). However, no substantial alteration was found in the proportion of CD163 + macrophages (M2) among Groups C, S and M, which were still the main macrophage population ( Fig. 6 C). These findings recapitulated the increase in CD86 + macrophages after injury, in line with NKG2A + NK cells and CD8 + T cells. The changes were more likely due to the increase in M1 Møs, not the changes in M2 Møs. The proportions of M1 and M2 cells among the three treatment groups were significantly different. The proportion of CD86 + cells in the three treatment groups was notably lower than that in Groups C and M, but there was no difference in CD163 + cells between Groups C and M. Therefore, although the treatments in Groups E, A and AE contributed to maintaining the predominance of M2 populations, they were not sufficient to restore the ratio of M1 to M2 ( Fig. 6 D). Researchers have suggested that Th1 cells predominate in the normal endometrium, especially in the proliferative phase, while Th2 cells are the predominant Th cells in the decidua of early pregnancy [ 46 ]. Similarly, the proportions of IFN-γ + T cells (Th1 cells) and IL-4 + T cells (Th2 cells) in Groups C and S indicated the dominance of Th1 cells in the normal endometrial immune microenvironment. However, there was a much higher proportion of Th2 cells and, inversely, a much lower proportion of Th1 cells in Group M than in Group C. Endometrial injuries occurred in parallel with a decrease in Th1 cells and an increase in Th2 cells. After the treatments, Th1 cells maintained the predominant populations in Groups AE and A, which was consistent with both Groups C and S. Our data support the notion that ADMSCs-SF/PCL or E2+ADMSCs-SF/PCL contributed to shaping the restoration of the normal ratio of Th1 cells to Th2 cells ( Fig. 6 E and F). ELISA was carried out to measure the concentrations of IL-1β, IL-10, TNF-α and TGF-β, and the results are detailed in Supplementary material Fig. S2. On Day 28, the expression level of IL-1β in Group AE was markedly lower than that in Groups E and A (P  0.05), but the expression levels were much lower than that in Group E (P < 0.05, Fig. S2A). On Day 28 and Day 60, the production of IL-10 in Group AE was significantly higher (P < 0.05, Fig. S2B). Estradiol could promote the secretion of TNF-α in the process of endometrial repair, and it peaked on Day 28 (Fig. S2C). On Days 28 and 60, the levels of TGF-β in the two groups containing ADMSCs (Groups AE and A) were considerably lower than those in the groups without ADMSCs (Fig. S2D).

Adipose-derived mesenchymal stem cells (ADMSCs, CPR198), which were purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China), were cultured in ADMSC-specific complete medium (CM-R198) in a 5% CO 2 incubator with constant humidity at 37 °C. The anti-rat antibodies used for western blotting and immunofluorescence staining are listed in Table S1, and the antibodies for flow cytometric analysis (FCA) are listed in Table S2. ELISA kits for the cytokines IL-1β (AB1832P), IL-10 (RAB0247), and TNF-α (RAB0480) were from Sigma‒Aldrich, and TGF-β (BMS623-3) was from Thermo Fisher Scientific. Specific pathogen-free female Sprague‒Dawley rats (220–250 g, 9–10 weeks) were purchased from SLAC Laboratory Animal Co., Ltd. (Shanghai, China). The study was approved by the Institutional Review Board and the Research Ethics Committee of Shanghai General Hospital. SF/PCL electrospun nanofibrous mats were fabricated as described previously [ 38 ]. First, SF and PCL were dissolved in hexafluoroisopropanol (HFIP) at mass ratios of 4:1, 3:1, and 2:1 to obtain an 8 wt% electrospinning solution. Then, the prepared solution was placed in a syringe for electrospinning under the following conditions: the injection rate was 1 mL/h, the needle tip of the syringe was 12 cm away from the collecting drum, and the voltage was set to 16 kV. After that, the resulting nanofibrous mats were dried under vacuum at 25 °C for one day to eliminate the residual organic solvent. Subsequently, the mats were crosslinked with 75% ethanol for 30 min. Finally, the crosslinked mats were washed with deionized water and vacuum-dried again. All samples were examined with a scanning electron microscope (SEM, SU8010, Hitachi, Japan) to determine the morphology of the membranes. The cytotoxicity of the SF/PCL nanofibrous mats was determined by the extraction method and direct contact method. For the extraction method, SF/PCL nanofibrous mats were inoculated in DMEM/F12 with 5% FBS, 100 U/mL penicillin, and 100 U/mL streptomycin in a 5% CO 2 atmosphere for 24 h, 48 h and 72 h at 37 °C to acquire the corresponding extracts. Then, ADMSCs were seeded onto 96-well culture plates and incubated in each extract (200 mL) for 24, 48 and 72 h. Thereafter, the extracts were gently removed, and the cells were rinsed with PBS three times. Subsequently, the culture medium and CCK-8 solution were added to each well for 1 h according to the protocol. Finally, the optical density was determined at 450 nm using an ELx800 microplate reader (BioTek Instruments, VT, USA). For the direct contact method, ADMSCs were directly planted on nanofibrous mats and cocultured for 24 h and 48 h. Acridine orange and propidium iodide (AO/PI) cell staining were used to evaluate cell viability. Briefly, the cocultured ADMSCs were washed and dyed with a fluorescent dye mixture (AO/PI [1:1]). Then, the cell viability was observed by a fluorescence microscope (Eclipse Ti–S, Nikon, Japan). First, when the cell confluence of ADMSCs reached 80%–90%, a single-cell suspension was prepared. Then, the mats were cut into a size of 2.5 × 0.5 cm 2 and placed in a 12-well culture plate with the loose surface up. After that, the mats were soaked in complete stem cell culture medium for 30 min and placed in a 37 °C and 5% CO 2 incubator. Next, the culture medium was aspirated, the residual liquid on the SF/PCL fibers was absorbed with sterile absorbent paper, and then, ADMSCs (1 × 10 6 ; 3.5 × 10 7 /mL in 30 μl) were added onto the mats and cultured for 6 h. After the ADMSCs were planted on SF/PCL mats, they were fixed with 2.5% glutaraldehyde at 3 h, 12 h, 24 h, 48 h and 72 h and detected under an electron microscope (SEM, SU8010, Hitachi, Japan). For establishment of a rat model for IUAs [ 39 ], thirty-six rats were fed in a temperature-limited room (20–22 °C) with a 12 h light and dark cycle. First, 2% sodium pentobarbital was intraperitoneally injected into rats at a dose of 50 mg/kg, and all the following operations were performed under anesthesia. Then, the right uterine horn was exposed. An incision approximately 0.5 cm long was made at 1/3 of the bifurcation of the uterus, and a 2.5-mm-diameter micro-curette was used to fully scratch the uterine cavity. Subsequently, cotton thread soaked in 10 mg/mL LPS normal saline for 24 h was placed in the right uterine cavity, and the left horn was treated similarly. The LPS cotton thread was fixed to the skin through the abdominal incision and extracted from the uterine cavity after 48 h. The animal ethics committee of Shanghai General Hospital approved the current animal experiments. Thirty-six female rats (72 uterine horns) were randomly assigned to six groups: 1) control group, rats without abdominal incisions (n = 12 uterine horns; Group C ); 2) sham group, rats with abdominal incisions but without endometrial damage (n = 12 uterine horns; Group S ); 3) model group, rats with endometrial damage but without any treatment (n = 12 uterine horns; Group M ); 4) estradiol (E2) + SF/PCL group, rats with endometrial damage treated with E2+SF/PCL (n = 12 uterine horns; Group E ); 5) E2+ADMSCs-SF/PCL group, rats with endometrial damage treated with E2+ADMSCs-SF/PCL (n = 12 uterine horns; Group AE ); and 6) ADMSCs-SF/PCL group, rats with endometrial damage treated with ADMSCs-SF/PCL (n = 12 uterine horns; Group A ). First, the abdominal cavity was opened seven days after the rat model was constructed. Then, the bilateral uterine horns were explored 0.5 cm away from the bifurcation of the uterus, and the incision was recut. Subsequently, SF/PCL (2.5 × 0.5 cm 2 ) or ADMSCs-SF/PCL (2.5 × 0.5 cm 2 , containing 1 × 10 6 ADMSCs) was transplanted into one uterine cavity after soaking in normal saline for 5–10 min, and both ends of the mats were fixed on the uterine wall to prevent falling out. The opposite uterine angle was treated in the same way. Finally, 6–0 absorbable sutures were used to stitch the uterus. All the above operations were performed under routine anesthesia and asepsis. For E2 administration, the rats were intramuscularly injected daily with 2 mg/kg estradiol (estradiol benzoate, 2 mg/mL, Ningbo Second Hormone Factory, China). The rats were sacrificed, and the bilateral uterine horns were resected at the appointed time. Samples were stained with hematoxylin-eosin (HE) and Masson [ 40 ]. The mean endometrial thickness was measured in the HE-stained cross-section of the uterus, and the endometrial thickness was defined as the vertical distance between the two endometrial-myometrial interfaces under 10–40 × magnification. Moreover, the number of glands was recorded in detail. The fibrotic area ratio was considered the percentage of positive staining for collagen (blue area) to the total tissue area in sections in the images of Masson’s trichrome staining. At least five regions were analyzed by ImageJ to calculate the endometrial thickness, number of glands and fibrotic area ratio. Total endometrial tissue protein of different groups was extracted, and the concentration was examined with a Bradford assay (Bio-Rad Laboratories, USA). Protein samples and the β-actin control were loaded for SDS-polyacrylamide gel electrophoresis, and the electrophoresis results were blotted onto PVDF membranes (EMD Millipore, USA). The primary antibodies anti-TGF-β1, anti-Smad3, anti-HOXA11, anti-FOXO1, anti-HAND2, and anti-β-actin were used according to the manufacturer’s instructions. An enhanced chemiluminescence (ECL) detection system (Pierce, USA) was used to visualize the bands. We performed densitometric analysis using Image-Pro Plus (version 6.0; Media Cybernetics, USA). Immunofluorescence staining was used to detect markers of uterine vasculature and fibrosis at different times (Day 14, Day 28, and Day 60). After deparaffinization and rehydration, tissue sections were incubated in 3% H 2 O 2 and then with primary antibodies against CD31, CK19, and α-SMA at 4 °C overnight. Next, the sections were incubated with secondary antibodies for 20 min. The nuclei were counterstained with 4ʹ,6-diamidino-2-phenylindole (DAPI; Sigma‒Aldrich). ImageJ was used to set the threshold and remove background signals. Images were acquired sequentially on a fluorescence microscope (Eclipse Ti–S, Nikon, Japan). Single-cell suspensions from the rat uteri for flow cytometry were isolated according to previously described methods [ 41 , 42 ]. Briefly, uteri were finely minced into ∼1-mm 3 pieces in a small volume of MACS buffer and pushed through a 60-mm mesh screen (Sigma-Aldrich, St. Louis, MO) using a syringe plunger. The resulting cell suspension was then passed through a 70-mm filter. All cells were stained with Fixable Viability Dye (eBioscience) before blocking Fc receptors. Cells were stained for 30 min in PBS with 1% FCS on ice for surface staining and then fixed and permeabilized by a BD Cytofix/Cytoperm Kit (BD Biosciences) for intracellular staining (IL-4, IFN-γ). The results were analyzed by BD LSR II (BD Biosciences, USA). The percentages of positive cells were calculated against the background set on a fluorescence-minus-one negative control. All flow cytometric data were analyzed by FlowJo 10.0 software. The endometrial tissue was lysed with precooled RIPA buffer (Solarbio) on ice for 30 min and centrifuged according to literature reports [ 43 ]. The total protein concentration of the supernatant sample was determined with a BCA kit (Beyotime), and ELISA kits (Sigma Life Sciences, USA) were used to measure the concentrations of four cytokines (IL-1β, IL-10, TNF-α, and TGF-β) in the supernatant sample. Finally, the concentrations of cytokines were quantified according to the user guide. Data are presented as the means ± standard deviations. Multiple comparisons were conducted using a one-way analysis of variance (SPSS 22.0). P < 0.05 was considered statistically significant.

Because the endometrium plays an important role in embryo implantation and pregnancy, the treatment of IUAs has crucial clinical significance and value. In this study, we found that both treatments involving ADMSCs-SF/PCL (E2+ADMSCs-SF/PCL or ADMSCs-SF/PCL; Group AE or Group A) restored the endometrial thickness and glands, promoted blood vessel growth and inhibited fibrogenic factors (TGF-β1, Smad3, α-SMA, and IL-1β) to alleviate excessive fibrosis caused by endometrial injury. In addition, both treatments containing ADMSCs-SF/PCL downregulated some proinflammatory factors while upregulating some anti-inflammatory factors, restored Th1 immune superiority and maintained the primary proportion of Møs and NK cells in the immune microenvironment. E2 induced an increase in the proportions of IL-4 + and IFN-γ + cells. Taken together, these two treatments containing ADMSCs-SF/PCL could be a promising strategy for patients suffering severe IUAs and infertility.

Intrauterine adhesion (IUA) caused by endometrial injury is a condition that seriously affects the pregnancies of women of childbearing age and produces severe pathogenic conditions, such as abortion, curettage, surgery and infection. Fibrosis is the most obvious feature of endometrial injury and IUAs, which is consistent with our data. Stem cell therapy has recently been described as a potential new approach for the treatment of injury and fibrosis in response to tissue damage in regenerative medicine, including the regeneration of endometrial disorders. Transplantation of stem cells in the endometrial zone exhibited an impact on endometrial repairs, such as decreasing the fibrotic area, elevating the gland number, enhancing the endometrial thickness, protecting gestation, and finally improving the pregnancy rate [ 16 ]. ADMSCs are currently recognized as the most promising stem cell sources in regenerative medicine because adipose tissue is considered one of the most conspicuous sources of stem cells due to its easy accessibility, abundant mesenchymal progenitor cells, and good immunomodulatory properties. Studies have shown that ADMSCs from adipose tissue secrete high levels of cytokines, such as IGF-1 and VEGF, which are needed for cyclic endometrial regeneration [ 47 ]. To promote ADMSC viability and ability, we developed ADMSCs-SF/PCL for IUA treatment. Our findings indicated that all of the treatments based on SF/PCL, including estrogen, with or without ADMSCs, and ADMSCs alone, exhibited potency in recovering the injured endometrium by decreasing fibrosis, enhancing gland development and angiogenesis, and restoring the immune environment. These results were in line with others, which showed that the combination of stem cells and estrogen had a curative effect and a proliferative effect on epithelial tissues in disrupted endometrial tissue after AS [ 48 ]. Although many studies have suggested that the repair efficacy time of estrogen combined with mesenchymal stem cells is limited to the treatment period of approximately 1 month [ 13 , [49] , [50] , [51] ], our results indicated that the ADMSCs-SF/PCL system could still maintain the treatment effect even after 60 days. This result indicated that the repair effect of our experimental treatment was better than that in previous studies. The pathological feature of IUAs is endometrial fibrosis. Cytokines, including TGF-β1, TNF-α, IL-10 and IL-1β, are important regulators of fibrosis. The TGF-β1/Smad3 signaling pathway is one of the classical signaling pathways regulating fibrotic diseases [ 52 , 53 ]. In addition, TNF-α is a powerful inflammatory cytokine that can participate in the regulation of repair processes and is positively correlated with organ fibrosis [ 54 ]. Higher levels of IL-10 and lower levels of IL-1β in both the lung and serum are beneficial for lung regeneration from pulmonary fibrosis [ 55 ]. Our findings revealed that the expression of TGF-β1 or Smad3 shown by WB was markedly upregulated after injury. After treatment, TGF-β1 and Smad3 decreased, which indicated that these three treatments could inhibit the fibrosis-related TGF-β1/Smad3 pathway, even on Day 60, ADMSCs-SF/PCL can still substantially reduce the level of TGF-β1. Others found that after treating chronic endometrial injury in rats with human umbilical cord-derived MSCs, the expression of TNF-α in the endometrium was significantly reduced [ 13 , 56 ], however, in our research ADMSCs took over the function of E2 and continued to secrete TNF-α after Day 28. For the proinflammatory factor IL-1β, our results showed that ADMSCs might play a role after Day 28 and inhibit the secretion of IL-1β. Regarding IL-10, our experiments indicated that E2-mediated promotion of IL-10 secretion was the most significant at Day 28 and almost no longer existed on Day 60. However, E2 might continue to play a role in promoting IL-10 secretion under the ADMSCs treatment even after more than 28 days, indicating that ADMSCs could maintain and prolong the promoting effect of E2, which was important for endometrial regeneration. However, the mechanisms of these cytokines remain incompletely understood and need further research. Previous studies confirmed that endometriosis-related infertility may be caused by reduced expression of HOXA11 [ 57 ]. Our results showed that HOXA11 and HAND2 were highly expressed in the normal endometrium. The injury of the endometrium was followed by the downregulation of HOXA11, FOXO1 and HAND2. The treatments containing ADMSCs-SF/PCL (Groups AE and A) upregulated the expression of HOXA11, FOXO1 and HAND2, indicating that treatment containing ADMSCs-SF/PCL could regenerate the endometrium by increasing the expression of these three markers. Immune cells, including predominant NK, T and Mø cells infiltrating in the endometrium, account for 10–20% of all endometrial cells, interact with many cytokines and are beneficial to menstruation, implantation and pregnancy [ 58 ]. According to our data, the immune cell population was altered after endometrial injury by increasing the proportion of CD3 + T cells, reducing that of NK cells, and maintaining the proportion of macrophages. NK cells are mainly activated receptor NKG2D + NK cells in the normal endometrium [ 59 ]. After endometrial injury, the total proportion of NK cells decreased, and NKG2A+ NK cells were the dominant NK population. Surprisingly, the proportions of CD3 + T cells, NK cells and CD68 + Mø cells were all restored to those proportions in the control group after treatment with ADMSCs-SF/PCL (Group A). Macrophages (Møs) are typically classified into classically activated Møs (M1) and alternatively activated Møs (M2) [ 60 , 61 ]. M1 macrophages are proinflammatory cells with potent antimicrobial activities, while M2 macrophages play a major role in wound repair and tissue remodeling [ 62 , 63 ]. We found that the proportion of M1 macrophages decreased after endometrial injury without changing the proportion of M2 macrophages. However, the three treatments could not recover the normal proportion of M1 or M2. We speculated that the unrestored proportion of M1 and M2 or activated and inhibited NK cells may have an effect on the immune environment in endometrium regeneration and then may affect endometrial function. Therefore, these results also guided us to seek to remodel the immune environment by restoring the proportions of NK cells and macrophages, which is our primary goal in the future. Moreover, the maintenance of normal endometrial function requires the joint participation of Th1 and Th2 cells. The disrupted balance of Th1 and Th2 cells promotes endometrial fibrosis [ 64 ]. Activated Th2 cells can produce a variety of cytokines, such as IL-4, to promote the occurrence and development of tissue fibrosis [ 65 ]. In contrast, as a kind of Th1 cytokine, IFN-γ is considered antifibrotic [ 66 ]. To understand the changes in T-cell cytokines in the endometrial microenvironment, we detected the changes in IL-4 + and IFN-γ + cells in CD3 + T cells in the endometrial microenvironment. Our results indicated that a Th1-type advantage was present in the normal endometrium, the treatments containing ADMSCs-SF/PCL (Group AE and Group A) could restore the Th1 immune superiority, and E2 could increase the proportion of IL-4 + and IFN-γ + cells.

Endometrial injury caused by repeated invasive intrauterine operations and endometritis may result in severe intrauterine adhesions (IUAs), namely, Asherman’s syndrome (AS) [ 1 , 2 ]. IUAs, characterized by endometrial fibrosis [ 3 , 4 ], are a serious threat to women’s reproductive health. Over the past decades, several therapies have been established for the treatment of IUAs, including stripping adhesions by hysteroscopy, intrauterine device placement and hormone therapy [ [5] , [6] , [7] ]. Unfortunately, physical methods simply alleviate the degree of uterine cavity stenosis and fail to thoroughly repair the endometrium. The efficacy of pharmacotherapy is poor in severe cases [ 8 ], because hormone therapy mainly has strong effects on the proliferation of the normal endometrium but does not exhibit sufficient in vivo effects to reduce fibrosis and promote injured endometrial regeneration [ 9 , 10 ]. Therefore, there is an extremely urgent need to develop a more effective treatment regimen for IUAs to inhibit fibrosis with stimulating endometrial regeneration or repair. Stem cell-based therapy has been considered a promising approach to reconstruct impaired endometrium [ 11 ]. Studies have indicated that intravenous infusion of bone marrow-derived MSCs (BMMSCs) or umbilical cord-derived MSCs (UCMSCs) results in reconstruction of the endometrium and inhibition of endometrial fibrosis [ 12 , 13 ]. In addition to their differentiation potential, MSCs can secrete extracellular vesicles and molecules such as growth factors, cytokines, and chemokines to exert antifibrotic and proangiogenic effects [ [14] , [15] , [16] ]. However, some issues seriously hinder the clinical translation of MSCs. First, limited by the sources of BMMSCs and UCMSCs, it is indeed difficult to harvest sufficient stem cells, which will therefore complicate the culture and expansion process. In addition, injected MSCs generally suffer from a low rate of retention and survival in vivo, thus weakening the sustained reparative capacity of stem cells [ [17] , [18] , [19] ]. Moreover, stem cells contribute to the growth of not only normal or impaired endometrium but also ectopic endometrium owing to their nonspecific mode of function [ 20 ]. In other words, intravenous systemic administration of MSCs may result in unexpected side effects. In these cases, another kind of MSC with a more efficient and safer delivery method is needed. Adipose-derived MSCs (ADMSCs) might be a better alternative because they are easy to sample and can be extracted in large quantities [ 21 , 22 ]. The investigation of ADMSCs in IUAs has rarely been reported, and further studies are required to evaluate their safety and reliability. For improved retention of MSCs, several types of biomaterials have been fabricated [ [23] , [24] , [25] ]. Among available carriers, the electrospun nanofibrous mat is one of the most attractive candidates since it shares similarity to the natural extracellular matrix and shows good mechanical strength. These nanofibrous mats are promising for potential use in tissue regeneration and repair [ [26] , [27] , [28] ]. More importantly, the local application of MSCs with nanofibers can avoid the side effects of systemic administration. This strategy has been verified in the reconstruction of bone [ 29 ], cartilage [ 30 ] and myocardial tissues [ 24 ]. Consequently, it was speculated that nanofibrous mats loaded with ADMSCs could become a useful method to treat severe IUAs. In this study, a system of silk fibroin/polycaprolactone (SF/PCL) electrospun nanofibrous mats loaded with ADMSCs (ADMSCs-SF/PCL) was developed to enhance the functional survival of grafted stem cells. SF possesses favorable biocompatibility and biodegradability, which can promote the attachment and proliferation of seeding cells [ [31] , [32] , [33] , [34] , [35] ]. PCL was introduced to overcome the short absorption times and poor mechanical properties of SF [ 36 , 37 ]. The therapeutic effects of ADMSCs-SF/PCL alone or in combination with estradiol were evaluated. The results suggested that ADMSCs-SF/PCL with estradiol overcome the short-lasting effect of estradiol and showed prolonged efficacy compared with previous studies. Based on the analysis of immune cells and immune factors, the proposed strategy was proven to improve the immune microenvironment of the endometrium, which has rarely been previously studied. In conclusion, this study provides a potential option for the treatment of IUAs and brings novel insight into the treatment mechanism of ADMSCs and estradiol.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Not applicable.