Third passage hUCMSCs were purchased from Qilu Cell Therapy Technology (Shandong, China) and were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, US), which was supplemented with 10% exosome-depleted fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco), at cellular incubation conditions (37°C and 5% CO₂). The MSCs were identified as previously described (25). First, we confirm their morphology by using a light microscope (Imager.D2; ZEISS, Germany). Then, Alizarin Red staining was used for osteogenic identification; and Oil Red O staining was conducted for adipogenic differentiation. Finally, Flow cytometry (FCM; FACSCanto II, BD, US) was used to detect MSC-related surface markers (CD34, CD44, CD45, CD73, CD90, CD105, and HLA-DR; Biolegend, US).

Culture medium of hUCMSCs at the fourth to sixth passage was obtained and centrifuged at 2,000 × g for 15 min at 4°C to filter cells. Then, the supernatant was centrifuged at 10,000 g for 30 min at 4°C to filter debris after which the supernatant was centrifuged at 120,000 g for 70 min at 4°C to remove the supernatant. The precipitate was washed using phosphate-buffered saline (PBS; Servicebio, China) after which we centrifuged the supernatant at 120,000 g for 70 min at 4°C. Then, the precipitate was suspended in pre-cooled PBS (Servicebio), which led to the isolation of exosomes at a concentration of 1 mg/mL (8.69×10¹⁰ particles/ml). The hUCMSC-Exos were experimentally identified (26). First, we used Flow NanoAnalyzer (U30; NanoFCM, China) to measure exosomal size distribution and concentration. Then, we observed their morphologies by using transmission electron microscopy (TEM; Talos F200C; Thermo Scientific, US). Finally, surface markers of hUCMSC-Exos, including CD9, CD63, and CD81 (Biolegend, US), were detected by FCM (BD).

Ethical approval for the use of animal models was obtained from the Ethical Committee of Second Hospital of Hebei Medical University. Three-week-old healthy female C57BL/6 mice were acquired from SPF Biotechnology (China). Stimulation of follicle growth was done by intraperitoneal injection of pregnant mare serum gonadotropin (Solarbio, China). Mice ovarian tissues were obtained after 48h of the injection of pregnant mare serum gonadotropin. Then, we isolated GCs under the anatomical microscope (SMZ-10A; Nikon, Japan). Next, digestion of the was conducted in 0.1% hyaluronidase (Sigma, US) at 37°C for 10 min. After terminated the digestion and washed the tissues, we filtered the suspension through a cellular strainer. After that, we cultured single-cell suspensions in DMEM containing Ham’s F12 medium (DMEM: F12; Gibco) at cellular incubation conditions. Lastly, immunofluorescence staining of follicle stimulating hormone receptors (FSHR; 1:800; Servicebio) was performed to identify GCs.

The first passage GCs were seeded on the six-well plate with a density of 1×10⁵ cells/well. We divided GCs into five groups. To confirm the protective role of exosomes against cyclophosphamide (CTX)-induced GC injury, in the CTX group, 30 μM CTX (Sigma, US) was co-cultured with GCs for 24 h. In the Exos group, we added 20 μg/mL hUCMSC-Exos and CTX (Sigma) to the cellular medium. Verteporfin (Bioss antibodies, China) was used as a YAP inhibitor that disrupts YAP-TEAD interactions. In the Exos+Verteporfin group, in addition to culturing GCs with hUCMSC-Exos and CTX, we added 10 μM verteporfin into the cellular medium. In the CTX +Verteporfin group, we added CTX and verteporfin to the medium of GCs. Lastly, in the Normal group, we cultured GCs with the common cellular medium.

We conducted EdU and CCK-8 assays to assess the proliferative ability of GCs. After drug interventions and exosome-mediated treatment for all groups, we used an EdU Kit (Beyotime, China) to detect cell proliferation according to the manufacturer’s instructions. After completed the operation, we observed GCs under a fluorescence microscope (AxioCam HRc; ZEISS, Germany). The CCK-8 assay was also performed. Briefly, we seeded first passage GCs into 96-well plates with a density of 4000 cells/well. After drug interventions and exosome-mediated treatment for all groups, 10 μl of CC K-8 solution (Servicebio, China) was added into wells and incubated at cellular incubation conditions for 4 h. Absorbance was measured at 490 nm using a microplate reader (Synergy-H1; BioTek, US). The same procedure was performed after 1-5 days to evaluate GCs proliferation.

Eight-week-old healthy female C57BL/6 mice were acquired from SPF Biotechnology (China), and all procedures were ethically approved by the Ethics Committee of Second Hospital of Hebei Medical University. Mice were equally assigned into three groups. The POI (n=12) and Exos (n=12) groups were intraperitoneally administered with CTX (120 mg/kg) two times (Once every seven days; according to the results of our pilot study) to establish the POI model. The Normal group (n=12) received non-operate. We confirmed the establishment of POI mice models by constantly detecting vaginal smears and measuring weight for two weeks after the injection of CTX. Then, we administered mice in the Exos group by intraperitoneal transplantation of 150 μg hUCMSC-Exos two times (Once every seven days). At the same time, we administered mice in the POI group by intraperitoneal injection of 150 μl PBS. Two weeks after the first treatments (27), 18 mice (n=6 per group) were sacrificed, and blood serum and ovaries were harvested for subsequent experiments.

Reproductive tests were conducted to assess the efficiency of hUCMSC-Exos involved functional improvement of fertility. Two weeks after the first treatment, we mate female mice in the experimental group (n=6 per group) with sexually mature male mice at a 1:1 ratio continuing eight weeks. Fertility levels of the three groups, including pregnancy rate, number of offspring, and time of birth were recorded (28). Pregnancy rate was respectively calculated in four- and eight-week after mating. The number of offspring and the time of birth were calculated in eight-week.

After the completing treatment, we obtained ovaries from sacrificed mice. Next, we fixed ovaries in 4% paraformaldehyde (Servicebio) for 48 h. After completing paraffin embedding, we sliced the ovaries into 4 μm sections. Then, we selected more than three sections of every ovary for hematoxylin-eosin (HE) staining. Lastly, we observed morphological characteristics of every ovarian section and counted all stages of follicles (primordial, primary, secondary, antral, and atretic follicles) using light microscopy (ZEISS).

TUNEL analysis of ovarian sections was performed by using the Tunel Cell Apoptosis Detection Kit (Servicebio, China). Briefly, 0.3% Triton X-100 (Solarbio, China) was used for permeabilization of ovarian sections. Then, the sections were incubated with TUNEL detection solution. The DAPI solution (SouthernBiotech, US) was used to label the nucleus. Finally, sections were analyzed by fluorescence microscopy (ZEISS). The DAPI-labelled nuclei appeared blue while positive apoptosis cells appeared green.

After mice were sacrificed, blood was collected after which serum was obtained by centrifuging at 5000 rpm for 10 min. Anti-Mullerian hormone (AMH), estradiol (E₂), and FSH levels in serum were measured by ELISA Kit (MEIMIAN, China) according to the manufacturer’s instructions. In brief, we added 50 μl of each serum to each coated well for incubation. After 30 min, we washed these wells five times using a wash buffer. Next, we added HRP-conjugated antibodies into the wells. After the same wash steps of wells, we added the color development solution into the wells for 10- minute incubation at 37°C. Finally, we added the stop buffer and measured the optical density (OD) at a 450 nm wavelength. The concentration of these hormones was calculated based on the standard curve.

Paraffin-embedded mouse ovarian sections were immersed in sodium citrate buffer (Servicebio, China) at a sub-boiling temperature for 5 min, cool for 5 min at room temperature, followed by another sub-boiling temperature for 5 min for antigen retrieval. Then, 0.3% Triton X-100 was used for permeabilization after which the objective area was covered with 5% goat serum (Servicebio) for 1 h. Then threw away the blocking solution slightly and incubated the slides at 4°C overnight in the presence of primary antibodies, including FSHR (1:400; Servicebio), Ki67 (1:800; Servicebio), and proliferating cell nuclear antigen (PCNA; 1:800; Servicebio). On the next day, slides were washed with PBS (Servicebio) three times after which the sections were incubated for one hour in the presence of a secondary Alexa Fluor 488-conjugated goat anti-rabbit antibody (1:400; Servicebio, China) or Cy3 conjugated goat anti-rabbit antibody (1:400; Servicebio, China). Finally, cell nuclei were labeled with the DAPI solution (SouthernBiotech). We observed these slides by using a fluorescence microscope (ZEISS).

Ovaries from each group of mice were fixed, sectioned (4 μm), and incubated in the presence of primary mouse antibodies, including FSHR (1:400; Servicebio) and Ki67 (1:800; Servicebio), at 4°C overnight. Subsequently, the slides were incubated with biotinylated secondary antibodies at room temperature for one hour. Reaction products containing diaminobenzidine (ZSGB-BIO, China) were used to color the slides, which were counterstained with hematoxylin (Servicebio).

Firstly, we used the RIPA lysis buffer (Servicebio) to extract protein from mouse ovaries in vivo and GCs in vitro. Protein concentrations were measured by using a BCA protein assay kit (Solarbio, China). Then, we made protein samples denatured at 100°C for 10 min. SDS-PAGE (Servicebio) was used to separate proteins, and then proteins were transferred into 0.22 μm PVDF membranes (Sigma, US). After being blocked with 5% skim milk for 1.5 hours, membranes were incubated with primary antibodies overnight at 4°C. The next day, we washed them three times and incubated them with an HRP-conjugated goat anti-rabbit antibody (Servicebio, China) for one hour. Lastly, we detected the proteins levels using the ChemiDoc MP Imaging System (Bio-Rad, USA).

Each test was performed at least three times. All results were analyzed by using SPSS 26.0 software. Student’s t-test was used for comparisons between two groups. A one-way analysis of variance was used for the distribution of data. Results are presented in the form of mean ± SD. A p value of < 0.05 was considered to be significantly different.

After inoculation, the MSCs isolated from the human umbilical cord grew into long fusiform cells with fibroblast-like morphologies. In passage four to ten, cells exhibited a consistent spindle shape ( Figure 1A ). Immunophenotypes of MSCs were characterized by flow cytometry using specific biomarkers. The results showed a high expression of CD73 (98.8%), CD105 (98.1%), CD44 (99.9%), and CD90 (99.9%), and a low expression of CD34 (0.4%), CD45 (0.7%), and HLA-DR (0.2%) in hUCMSCs ( Figure 1B ). In addition, we induced hUCMSCs to develop into different cell types in a conditional culture system. The positive results of Alizarin Red and Oil Red O staining for induced hUCMSC demonstrated the multi-lineage differentiation potential of them ( Figure 1C ). These characterizations meet the criteria for defining human MSC ( Table 1 ) (25).

Subsequently, hUCMSC-Exos were identified according to minimal experimental requirements for definition of extracellular vesicles (26). The exosomes exhibited a cup-shaped morphology ( Figure 1D ) with their size distributions ranging from about 50 to 100 nm at a concentration of 8.69×10¹⁰ particles/mL ( Figure 1E ). Additionally, the surface marker proteins were positive in exosomes by conducting FCM, including CD9, CD81, and CD63 ( Figure 1F ). These results confirmed that exosomes had been isolated from hUCMSCs ( Table 1 ).

GCs were isolated from mouse ovaries and cultured in vitro to investigate the possible mechanisms of hUCMSC-Exos in POI. After 48 h of initial plating, GCs exhibited oval or polygonal shapes with a single layer. And after 72 h, cells exhibited significant proliferation under light microscopy ( Figure 2A ). In the ovarian tissue, FSHR was a molecular marker for the identification of GCs (27). Therefore, immunofluorescence staining of FSHR was used to identify ovarian GCs. Most cells exhibited a green fluorescence, which indicated that they are GCs ( Figure 2B ). These findings proved that ovarian GCs had been successfully isolated and cultured ( Table 1 ).

To verify the therapeutic effect of exosomes in vitro, they were co-cultured with CTX-damaged GCs. Briefly, proliferation of GCs was measured through EdU and CCK-8 assays. In the EdU assay, red fluorescence with Alexa Fluor 555 exhibited proliferative GCs, which were more in the Exos group compared to the CTX group ( Figure 2C ). Moreover, the EdU assay revealed that the proliferative ratio of GCs was significantly increased in the Exos group compared with the POI group (p < 0.01) ( Figure 2D ). The CCK-8 assay showed that, compared to the CTX group, the proliferation of GCs was significantly enhanced when they were co-cultured with hUCMSC-Exos for 5 days ( Figure 2E ). Then, proteins associated with GCs proliferation and function were analyzed by western blotting. In the Exos group, proliferative proteins (PCNA) and functional proteins (FSHR) were elevated when compared to the CTX group ( Figures 2F–H ). These findings showed that hUCMSC-Exos promote GC proliferation and expression of FSHR in vitro.

Animal experiments were performed according to the experimental timeline ( Figure 3A ). Specifically, hUCMSC-Exos transplantation we performed following the POI model was established by injecting CTX twice. Seven days after the last treatment, six mice of each group were sacrificed, and six mice were mated with male mice (1:1). To confirm the establishment of POI mice models, body weights for mice in each group were measured, and vaginal smear was performed every morning at 8 am. In the Normal group, a regular estrous cycle was four to six days. In the POI model group, almost all mice stayed in the estrous phase and have no periodic change. Compare with the Normal group, HE staining of ovarian sections revealed that follicle numbers were significantly decreased, the arrangement of GCs was disordered, and the interstitial fibrosis was more severe in the POI group ( Figure 3B ). Furthermore, TUNEL staining was performed to detect apoptotic cells in ovarian tissues. In the POI group, there were more TUNEL-positive ovarian GCs in follicles ( Figure 3C ). These results confirmed that the POI models had been successfully established ( Table 1 ).

Then, the therapeutic effects of hUCMSC-Exos on restoration of ovarian function were analyzed. After treatment of hUCMSC-Exos, we observed gradual restoration of the estrous cycle, but not in the POI group. Compared to the POI group, HE staining of ovarian sections revealed that the numbers of follicles in all stages were significantly increased in the Exos group, and were nearly close to normal levels ( Figure 3D ). Meanwhile, these pathological results showed that no inflammatory cells and immunological reactions appeared in the ovarian tissue after exosomes treatment, indicating low immunogenicity of MSC-derived exosomes ( Figure 3D ). After ovaries of different groups were collected, we counted the number of follicles in each section to further analyze the therapeutic effect. Compared to the Normal group, the numbers of primordial (p < 0.05), primary (p < 0.05), secondary (p < 0.05), mature (p < 0.05), and atretic follicles (p < 0.01) were significantly decreased in the POI group. However, after conducting exosomal transplantation, the numbers of primordial (p < 0.05), primary (p < 0.05), mature (p < 0.01), and atretic follicles (p < 0.01) were significantly increased than they were in the POI group ( Figure 3E ). Taken together, the number of different stages of follicles in the Exos group exhibited an increasing trend compared to the POI group. Before CTX injection, the body weights of mice were almost at the same levels among the three groups. After injection of CTX, mice in the Exos and POI groups lost weight gradually within seven days compared to them in the Normal group. However, After the first hUCMSC-Exos transplantation, body weights of mice in the Exos group exhibited a gradual increase following the first exosomal transplantation, while mice in the POI group only exhibited a slow weight gain ( Figure 3F ). Then, we measured AMH, E₂, and FSH levels in the serum of POI models. Compared with the Normal group, the AMH (p < 0.001) ( Figure 3G ) and E₂ (p < 0.001) ( Figure 3H ) levels were significantly decreased, and FSH (p < 0.001) ( Figure 3I ) levels were remarkably elevated in the POI group. The change of hormones levels implying the successful establishment of POI mice models. After exosomal administration, AMH (p < 0.001) ( Figure 3G ) and E₂ (p < 0.001) ( Figure 3H ) levels were significantly increased, while FSH (p < 0.001) ( Figure 3I ) levels were notably suppressed when compared to the POI group. However, AMH (p < 0.05) and E₂ (p < 0.05) levels in the Exos group had not returned to normal levels after two weeks of treatment when compared to the normal group. These findings showed that exosomal transplantation promoted the recovery of ovarian functions and physiological functions of POI mice.

Fertility results of the three groups, including pregnancy rate, number of offspring, and time to birth, were recorded. Representative figures of fertility results showed that the transplantation of exosomes significantly improved the reproductive functions of POI mice ( Figure 4A ). Four weeks after exosomal transplantation, pregnancy rate of mice in the Exos group was higher (66.67%) when compared to that in the POI group (16.67%) ( Figure 4B ). In eight weeks, the rate was 83.33% and 33.33%, respectively ( Figure 4C ). Moreover, the number of offspring was significantly suppressed in the POI group compared with the Normal group (p < 0.001). In contrast, treatment with hUCMSC-Exos significantly increased the number of offspring compared with the POI group (p < 0.05) ( Figure 4D ). Additionally, time to birth in the POI group was significantly prolonged (p < 0.001), which was reduced by exosomal transplantation (p < 0.01) ( Figure 4E ). These findings showed that exosomal administering could greatly improve reproductive ability of POI mice.

FSHR is a crucial molecular for GCs to perform their roles in regulating follicular function and activity. Therefore, immunological methods for ovarian sections were used to detect the expressions of FSHR in each group. Immunofluorescence staining of FSHR showed that more functional follicular GCs were observed in the Exos and Normal groups when compared to the POI group ( Figure 5A ). Furthermore, immunohistochemical staining revealed the same results ( Figure 5B ). High expression levels of FSHR indicated that hUCMSC-Exos transplantation improved folliculogenesis and follicular function.

PCNA and ki67 are the most conventional molecular markers for detecting proliferative cells. Accordingly, the detection of them was performed to evaluate the efficiency of exosomes in promoting ovarian cells proliferation. Immunofluorescence staining of PCNA showed more follicular proliferative GCs in the Exos and Normal group compared with the POI group ( Figure 6A ). Similarly, Immunohistochemical ( Figure 6B ) and immunofluorescence staining ( Figure 6C ) of Ki67 indicated that there were more proliferative cells in the Exos than in the POI group. Therefore, treatment with hUCMSC-Exos promoted ovarian cells proliferation, which contributed to the restoration of follicular and ovarian functions.

The Hippo signaling pathway plays a vital role in folliculogenesis and activation of primordial follicles (21). To determine the regulatory role of Hippo pathway in the hUCMSC-Exos-mediated therapeutic effects on POI mice models, proteins associated with key Hippo molecules and ovarian function (FSHR and PCNA) were simultaneously detected by western blotting. In the Exos group, proteins expression levels of YAP1, TAZ, and TEAD1, and ovarian functional proteins (FSHR and PCNA) were significantly elevated compared to the POI group. Conversely, proteins levels of p-YAP and MST1 were suppressed in the Exos group ( Figures 6D–K ). These findings preliminary indicate that hUCMSC-Exos promoted ovarian proliferation and function by regulating the Hippo pathway.

Verteporfin was used as an inhibitor of YAP in cultured GCs with/without hUCMSC-Exos to determine whether the pathway is involved in GCs proliferation and function in vitro. The proliferative ability of GCs was evaluated in different groups by using EdU and CCK-8 assays. In EdU assay, the effect of hUCMSC-Exos on promoting GCs proliferation was disappeared after co-culture with verteporfin ( Figure 7A ). Additionally, the proliferative rate was significantly increased in the Exos group compared with the CTX group (Exos group vs CTX group; p < 0.01), and the rate was notably decreased when verteporfin was added to the cell culture medium (Exos group vs Exos-Verteporfin group; p < 0.01) ( Figure 7B ). In the CCK-8 assay ( Figure 7C ), similar results were observed. These findings suggest that proliferative ability of GCs was significantly enhanced by co-culture with hUCMSC-Exos after CTX pretreated, but this therapeutic effect disappeared following the addition of a YAP inhibitor.

Furthermore, proteins associated with the Hippo pathway, and proliferation (PCNA) and function (FSHR) of GCs, were detected by Western blot analysis ( Figure 7D ). After co-cultured with hUCMSC-Exos, expression levels of proteins of YAP1 (p < 0.01), TAZ (p < 0.001), TEAD1 (p < 0.001), PCNA (p < 0.001), and FSHR (p < 0.001) in GCs were significantly elevated compared to the CTX group ( Figures 7E–G, I, J ). Meanwhile, proteins levels of p-YAP (p < 0.01) and MST1 (p < 0.001) were decreased ( Figures 7H, K ). However, contrasting results were found when the YAP inhibitor was added to the medium (Exos group vs Exos-Verteporfin group; YAP1, p < 0.01; TAZ, p < 0.001; TEAD1, p < 0.001; PCNA, p < 0.001; FSHR, p < 0.001; p-YAP, p < 0.001; MST1, p < 0.001) ( Figures 7D–K ).

In summary, these findings show that hUCMSC-derived exosomes promote the proliferation and function of ovarian cells by regulating the Hippo pathway ( Figure 8 and Table 2 ). These therapeutic effects were suppressed when the key Hippo molecule (YAP) was blocked, further confirming the necessity and indispensability of the Hippo pathway in exosome-mediated treatment for POI.