Isolation and culture of human UC-MSCs

The two human umbilical cord samples from full-term newborns were collected from the Shanghai First Maternal and Infant Hospital, and informed consent was obtained (ethical certification number: KS1956) with permission from the Medical Ethics Committee of Shanghai First Maternity and Infant Hospital. Human UC-MSCs were isolated using the adherent method. The umbilical cord was rinsed with DPBS (Gibco, USA) and cut into approximately 5 mm3 tissue fragments after removing the veins and arteries. Subsequently, fragments were dispersed in a 10 cm dish coated with 0.1% gelatin (Millipore, USA). After 6 h, α-MEM (Gibco) complete culture medium containing 5% UltraGRO-advanced (AventaCell BioMedical, USA), penicillin/streptomycin (Gibco) and heparin (Anhui, China) was added, and the dishes were placed in a 37 °C, 5% CO2 incubator until human UC-MSCs migrated out from the tissue. During the passage 0 (P0) culture period, an appropriate amount of culture medium was added to the dishes every two days to maintain a good condition. Approximately 10 days later, UC-MSCs were digested using 0.05% trypsin–EDTA (Gibco), and they were passaged in a new cell culture dish at a ratio of 1:3. UC-MSCs were cryopreserved using a serum-containing freezing medium containing 50% α-MEM, 40% FBS (Gibco), and 10% DMSO (Sigma-Aldrich, USA) and stored in liquid nitrogen for long-term preservation. The human UC-MSCs used for single-cell sequencing in this study were from the sixth passage (P6), while human UC-MSCs from passages earlier than P9 were used for cellular and animal experiments.

Culture of the KGN human ovarian granulosa cell tumor line

The human ovarian granulosa cell tumor cell line KGN was obtained from the American Type Culture Collection (ATCC, Manassas, USA). The cells were cultured in DMEM/F12 (Gibco) with 10% FBS and penicillin/streptomycin and maintained in a 37 °C, 5% CO2 incubator. KGN cells were digested using 0.05% trypsin–EDTA and cryopreserved using serum-free freezing medium.

For experiments, KGN cells were seeded at a density of 5 × 104 cells per 100 μL in 6-well plates. After adherence, the cells were pretreated with complete medium containing 1 mg/mL cyclophosphamide (CTX, C0768, Sigma-Aldrich, USA) for 3 h. Subsequently, each subpopulation of human UC-MSCs was seeded at a density of 3 × 104 cells per 100 μL in the upper chamber, followed by coculture for 24 h. Quantitative reverse transcription PCR (qRT‒PCR) was performed to assess the expression of target genes.

Flow cytometry analysis and sorting

Human UC-MSCs were collected and evenly distributed and then resuspended in 100 μL of FACS buffer (DPBS + 5% BSA (Sigma-Aldrich, USA)). Antibody staining was performed at a concentration of 5 μg/106 cells using the following antibodies: CD73-FITC (Biolegend, USA, 344015), CD90-APC (Biolegend, USA, 328113), CD13-APC (Biolegend, USA, 301705), CD34-FITC (Biolegend, USA, 343503), CD29-PE (Biolegend, USA, 303004), CD105-PE (Biolegend, USA, 800503), CD45-PE/Cy7 (Biolegend, USA, 304015), CD19-PE/Cy7 (Biolegend, USA, 302215), PLAU-APC (Biolegend, USA, 369004), KRT19-Alexa Fluor 647 (BD Biosciences, USA, 563648), and CD91-PE (BD Biosciences, USA, 550497). The staining was carried out on ice protected from light for 30 min. After staining, the cells were washed with FACS buffer to remove excess antibodies, followed by filtration. Flow cytometry analysis was performed within 30 min to identify the primary UC-MSCs. For the sorting of specific subpopulations of human UC-MSCs, 1–2 mL of complete culture medium was added to a centrifuge tube to receive human UC-MSCs expressing high levels (LRP1high) or low levels (LRP1low) of LRP1. The criteria for categorizing into LRP1high and LRP1low are as follows: the top 25% of LRP1-positive cells are designated as the LRP1high subpopulation, while the bottom 25% of LRP1-positive cells are classified as the LRP1low subpopulation. UC-MSCs between these two subpopulations were subsequently excluded. The collected cells were then plated in a culture dish and incubated overnight. The next day, the cells were subjected to experimental analysis or transplanted into the ovarian bursa of mice.

Enzyme-linked immunosorbent assay (ELISA)

Following cell counting, human UC-MSCs were seeded in 6-well plates at the same cell density. After 24 h of incubation, the supernatant was collected and centrifuged at 1000×g for 10 min. The target proteins in the supernatant were detected using human CXCL8 and IL-1β ELISA kits (ABclonal, Shanghai, China), following the manufacturer's instructions. The remaining supernatant was stored at -80 °C to prevent repeated freeze‒thaw cycles.

Total RNA extraction, reverse transcription, and quantitative PCR

For cell line cells, a chemical extraction method was used. The cells were collected and lysed by adding 1 mL of TRIzol (TaKaRa, Japan) followed by vigorous pipetting. Then, 200 μL of chloroform was added, mixed by vortexing, and allowed to separate into layers. The mixture was centrifuged at 4 °C and 12,000 rpm for 15 min, and the supernatant was transferred to a new 1.5 mL centrifuge tube. An equal volume of isopropanol (Shanghai, China) was added, mixed by inversion, and left undisturbed for 30 min. After centrifugation at 4 °C and 12,000 rpm for 10 min, the supernatant was discarded, and the pellet was washed twice with 1 mL of 75% ethanol (Shanghai, China), followed by repeated centrifugation. The pellet was then dissolved in 9 μL of RNase-free water (ABM, Canada), and the RNA concentration was determined. For mouse oocytes, a PicoPure RNA Isolation Kit (Thermo Fisher Scientific, USA) was used for extraction of trace amounts of oocyte RNA according to the manufacturer's instructions. Subsequently, 1 μg of total RNA was reverse transcribed using the 5 × All-in-one kit (ABM, Canada) according to the manufacturer's instructions. The resulting cDNA was appropriately diluted and subjected to quantitative PCR using Premix Ex Taq™ (TaKaRa, Japan), and ΔCT values were calculated for expression analysis. The sequences of the primers used in this study are shown in Additional file 7: Table S6.

H&E staining of tissues

Mouse organ tissues were fixed with 4% PFA (Servicebio, Wuhan, China) and dehydrated using a graded ethanol series, followed by clearing with xylene (Shanghai, China). The pretreated tissues were embedded in paraffin (Sigma-Aldrich, USA), and once the paraffin blocks solidified, they were trimmed and sectioned into 5 μm thick slices. The paraffin sections were then deparaffinized using xylene and a graded ethanol series, stained with hematoxylin–eosin (Servicebio, Wuhan, China), and finally mounted with neutral resin (Servicebio, Wuhan, China). The morphology of the tissues was observed under a microscope, and images were captured using a slide scanner.

Measurement of mitochondrial membrane potential

For the detection of mitochondrial membrane potential in KGN cells, the cells were seeded in 35 mm confocal dishes. Mitochondrial membrane potential was assessed using the JC-1 Mitochondrial Membrane Potential Assay Kit (Beyotime, Shanghai, China) following the instructions provided. The cells were then observed, and images were captured using a laser confocal microscope.

Animal experiments

5-week-old female and male ICR mice and 12-month-old female ICR mice were obtained from Shanghai Bikai Laboratory Animal Co., Ltd. All experimental mice were kept in SPF-grade animal facilities. 12-month-old mice were housed in SPF-level mouse facilities and raised until 18 months of age for experimental use. The experiment on mice was approved by the Animal Ethics Committee of Tongji University.

Establishment of POF mouse model. The method for establishing the CTX-induced POF mouse model is shown in Additional file 1: Fig. S1A. Female ICR mice at 6–8 weeks of age were intraperitoneally injected with 50 mg/kg CTX for 14 consecutive days, while the control group received an equivalent dose of DPBS. Vaginal smears were collected daily for one week to monitor the estrous cycle and assess ovarian function. After a week, peripheral blood was collected from the mice to measure serum levels of FSH, E2, AMH, LH, and P4 using ELISA kits (Mlbio, Shanghai, China), which indicated ovarian hormone function and ovarian reserve depletion, confirming the successful establishment of the POF mouse model.

Transplantation of human UC-MSCs under the ovarian capsule in mice. After the mice were anesthetized with an intraperitoneal injection of 2,2,2-tribromoethanol (T48402, Sigma-Aldrich, USA), the mice were placed ventral side down on a sterile operating table. Following disinfection of the surgical site, a 2–3 cm skin incision and a 0.5–1 cm incision of the muscle layer were made. Under a stereomicroscope, the ovaries were gently pulled out from the abdominal cavity. Subsequently, the ovarian capsule was lightly punctured with a 26 G needle, and 5 μL of human UC-MSCs suspended in DPBS at a concentration of 106 cells was slowly injected beneath the ovarian capsule. The ovarian capsule was sealed using an electrocoagulation pen, and the ovaries were gently placed back into the abdominal cavity. The incisions in the muscle layer and skin were sutured with a thread.

Intraorbital venous blood collection in mice. The mice were immobilized to expose the eye for blood sampling. A 3 cm length 3 mm capillary tube was gently inserted into the intraorbital venous plexus. Blood was collected in a 1.5 mL centrifuge tube and left to stand for 15 min at 4 °C. The tube was then centrifuged at 3000 rpm for 20 min to separate the serum. The upper layer of serum was transferred to a new centrifuge tube and stored at − 80 °C for subsequent hormone analysis.

Vaginal smears and estrous cycle determination in mice. During the three weeks following stem cell transplantation, vaginal smears were collected from the mice every morning at 8 o'clock. The vaginal cavity was flushed with 20 μL of DPBS and gently blown 2–3 times, and the lavage fluid was spread onto a glass slide. Cell morphology was observed under a microscope to determine the stage of the estrous cycle. The proestrus stage is characterized by numerous small leukocytes, the estrus stage by irregularly shaped nucleated epithelial cells, the metestrus stage by cornified epithelial cells, and the diestrus stage by a few cornified epithelial cells and leukocytes. The complete number of cycles within the three-week period was recorded, and the number of days required for each estrous cycle was calculated.

Ovarian follicle counting in mice. After H&E staining, follicle counting was performed on mouse ovarian sections. The morphological criteria for follicles at different stages were as follows: (1) primordial follicle: surrounded by a layer of flattened granulosa cells or a mixture of flattened and cuboidal granulosa cells, with a total cell count of less than 7; (2) primary follicle: ≥ 7 cuboidal granulosa cells surrounding the oocyte; (3) secondary follicle: ≥ 2 layers of granulosa cells surrounding the oocyte; (4) early antral follicle: ≥ 2 layers but < 4 layers of granulosa cells surrounding the oocyte, with a follicular cavity diameter < 20 μm; and (5) antral follicle: a follicle with a clearly visible follicular cavity.

Cohousing of mice and collection of E12.5 embryos. Female and male mice were housed at a 1:2 ratio. The next morning, vaginal plugs were observed in female mice. The presence of a milky white solid plug at the vaginal orifice indicated successful mating. The mated female mice were placed in separate cages, and the day of plug observation was recorded as Day 1. On Day 12, the female mice were euthanized by cervical dislocation. Open the mouse abdomen to expose the uterus. From both sides of the uterine horns, incise the uterine muscle layer, sequentially retrieve the embryos, and place them in DPBS. The placentas attached to the embryos were removed, and the amniotic membranes were torn open to separate the E12.5 embryos. The well-developed embryos were counted and photographed.

Collection of mouse oocytes and granulosa cells. The ovaries of mice were dissected and placed in a 100 μL droplet of DPBS in a 10 cm dish on a heated stage. The surface of the ovaries was gently scraped with a 26 G needle to rupture the follicles and release oocytes and granulosa cells. Granulosa cells were isolated by digesting the cumulus-oocyte complexes with hyaluronidase (H1115000, Sigma-Aldrich, USA). The granulosa cells were then washed three times with 0.5% PBS-BSA, while the oocytes were transferred to a droplet of pronase solution (PRON-RO, Roche, Switzerland). After complete zona pellucida digestion, the oocytes were washed three times with 0.5% PBS-BSA. Finally, the granulosa cells and oocytes were separately collected in low-adsorption 200 μL centrifuge tubes for subsequent library construction.

Single-cell sequencing and data analysis

Library construction and sequencing for single-cell RNA-Seq. Human UC-MSCs cultured up to the 6th passage were digested into a single-cell suspension, and the generation of single-cell gel beads (GEMs) was performed rapidly using the automated Chromium Controller system. The GEMs underwent reverse transcription, amplification, adapter ligation, and other steps, followed by 10 × Genomics single-cell transcriptome sequencing on the NovaSeq 6000 platform (Illumina) at Berry Genomics Co., Ltd., to analyze the heterogeneity of human UC-MSCs.

Data processing and analysis for single-cell sequencing. CellRanger (https://www.10xgenomics.com/) was used to align and annotate the raw sequencing data (fastq files) to obtain expression data. The cloupe file exported from the output data folder can be directly viewed using the professional software Loupe Cell Browser provided by 10X Genomics (https://www.10xgenomics.com/). Furthermore, the filtered_feature_bc_matrix.h5 file from the output data folder was imported into R software (version 4.0.5), and the Seurat (https://satijalab.org/) package's FindNeighbors and FindClusters functions were used to perform subpopulation clustering on the single-cell data. The RunUMAP function was used to visualize the dimensionally reduced results using UMAP. The FindAllMarkers function in the Seurat package was used to identify marker genes for each cell cluster, which serve as potential marker genes for cell type identification. Based on differentially expressed genes (DEGs) in each cluster and known cell markers reported in the literature, the cell types of human UC-MSCs were distinguished and annotated. The limma package was used for conventional screening of DEGs between different groups, with the following criteria: |FC|> 1.5 and adj.P.Value < 0.05. The Clusterprofile package was used for functional enrichment analysis of DEGs to elucidate the main functional differences between different groups. The analysis primarily focused on Gene Ontology (GO) terms, including biological process (BP), cellular component (CC), and molecular function (MF), as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Additionally, enrichment analysis of biological molecular pathways in the Reactome database and Wikipathways database was conducted to further explore relevant functions. The criteria for selecting biologically significant pathways were adj.P.Value < 0.05.

RNA sequencing and data analysis

Library construction and sequencing for RNA-Seq. RNA sequencing of mouse oocytes and granulosa cells was performed using the Smart-seq2 method for library construction. The cells to be sequenced were lysed using cell lysis buffer, and the lysate was reverse transcribed into single-stranded DNA. The reverse transcription product was pre-amplified, purified using Ampure XP beads (Beckman, USA), and fragmented using a Covaris S220 instrument. The fragmented DNA was then recovered using the QIAquick PCR Purification Kit (QIAGEN, Germany), followed by end repair, A-tailing, and adapter ligation. Subsequent to two rounds of purification, the purified DNA was amplified after measuring its concentration. The amplified product was loaded onto a DNA gel for electrophoresis, and a gel block with band sizes ranging from 200 to 500 bp was selected. Gel extraction of the recovered DNA library was performed using the PCR Clean-up Gel Extraction Kit (Macherey–Nagel, Germany). The recovered DNA library was then subjected to high-throughput sequencing on the HiSeq 2500 platform (Illumina) at Berry Genomics Co., Ltd.

Processing and analysis of RNA-Seq data. The raw fastq data from the transcriptome sequencing were subjected to quality control and then aligned and annotated using HISAT2 and StringTie to generate an expression matrix. The transcriptomic expression data were imported into R software, and principal component analysis (PCA) was performed on the samples using the prcomp function. The distribution differences among the sample groups were visualized using the ggplot2 package. Differential expression analysis of genes between different groups was conducted using the limma package, and DEGs were selected based on the following criteria: |FC|> 1.5 and adj.P.Value < 0.05. Functional enrichment analysis of DEGs was performed using the clusterProfiler package, and functional terms with adj.P.Values < 0.05 were selected. The results were visualized using the ggplot2 package. DEGs obtained from the transcriptome sequencing of young and aged mouse oocytes were used to construct a scoring system for oocyte aging. This scoring system was then applied to evaluate the degree of oocyte aging in mice transplanted with different subtypes of UC-MSCs.

Statistical analysis

The statistical data were analyzed and visualized using Microsoft Excel and GraphPad Prism 9 (GraphPad Software, Inc.). The analysis was performed by unpaired Student's t tests between two groups. All experiments were repeated at least three times.