HucMSC-mediated stromal metabolic reprogramming in reactivating aged ovaries: a single-cell transcriptomic perspective

Abstract

Background

Aging-induced decline in ovarian function and oocyte quality contributes to female infertility. However, the mechanisms underlying human umbilical cord-derived mesenchymal stem cell (HucMSC)-mediated rejuvenation of aged ovaries remain poorly understood. This study aimed to systematically investigate whether and how HucMSCs restore ovarian function and oocyte quality and elucidate the potential pathways involved.

Methods

Aged mice received in situ ovarian injections of HucMSCs. Ovarian follicular development and fertility outcomes were assessed. Low-input RNA-seq and single-cell RNA sequencing (scRNA-seq) were applied to evaluate transcriptomic heterogeneity in oocytes and somatic cells separately. Additionally, the molecular change and function of HucMSC-primed stromal cells (SCs) in aged ovaries were assessed to validate SCs’ functional roles in ovarian microenvironment improvement.

Results

HucMSC treatment enhanced follicular development, increased antral follicle numbers, and partially restored fertility in aged mice. Oocyte transcriptomes in HucMSC-treated mice resembled those of young mice, with 75% of aging-dysregulated genes (notably mitochondrial respiratory chain complex assembly-related genes) reverting to youthful expression patterns. scRNA-seq revealed upregulated transcription and glycolysis in granulosa cells (GCs), alongside stromal cell fate redirection toward steroidogenesis and folliculogenesis instead of fibrosis. Transplantation of HucMSC-primed SCs replicated these restorative effects. Mechanistically, HucMSCs promoted pregnenolone synthesis in stromal cells, facilitating follicular development.

Conclusions

HucMSCs reactivate aged ovaries by inducing metabolic reprogramming in both oocytes and somatic cells, enhancing mitochondrial function in oocytes, and redirecting stromal cells toward steroidogenic and folliculogenic pathways. These findings underscore stromal cell functional modulation as a critical mechanism in counteracting ovarian aging and revealing the potential of stromal cells as therapeutic targets.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s11658-026-00919-7.

Background

Ovary is a pivotal organ for female fertility and hormone secretion. Ovarian aging is a multifaceted process driven by many factors, resulting in significant detrimental consequences for women’s health, particularly female infertility. Advanced maternal age is a major cause of ovarian aging and the decline of female fertility, clinically recognized as diminished ovarian reserve (DOR). Ovarian aging is characterized by reductions in both ovarian follicle number and oocyte quality [1]. With the increasing trend of delayed childbearing [2], the prevalence of infertility linked to ovarian aging is rising, underscoring the urgent need for scientific exploration into effective therapeutic strategies to mitigate ovarian aging and restore reproductive potential.

Women of advanced maternal age experience decreased pregnancy rates and increased risk of aneuploidy, in both natural cycles and assisted reproductive programs. This decline in fertility is primarily due to reduced follicle reserves, hormonal imbalances, and diminished oocyte quality [3]. Oocyte meiotic maturation, which involves precise spindle assembly and chromosome segregation, requires substantial energy expenditure. Mitochondria, the organelles responsible for adenosine triphosphate (ATP) production, are crucial for maintaining oocyte quality. Mitochondrial dysfunction is considered a significant factor contributing to poor oocyte quality [4], and recent studies suggest that supplementation with autologous mesenchymal stem cell (MSC)-derived mitochondria can improve oocyte quality and embryo development [5, 6]. Oocyte quality and developmental competence are closely linked to follicular development, and recent studies indicate that the follicular microenvironment plays a critical role in regulating oocyte quality during oogenesis [7]. Aging appears to impact granulosa cells before affecting oocytes, as indicated by transcriptomic alterations [8]. Consistently, replacing aged granulosa cells with young ones could restore high-quality oocytes in a mouse model [9]. In the advanced maternal age, dysfunction of stromal cells contributes to ovarian fibrosis and diminished ovarian reserve [10]. A recent study demonstrated that transplantation of young mouse ovarian stromal cells into aged mice not only improved ovarian function but also extended the reproductive lifespan [11]. These findings highlight the potential of stromal cells as therapeutic targets for addressing age-related declines in ovarian performance.

MSC transplantation has emerged as a promising approach for addressing reproductive disorders [12]. Human umbilical cord-derived MSCs represent a particularly promising subtype of MSCs owing to their ease of acquisition, noninvasive collection process, abundance, and low immunogenicity [13]. Previous studies have shown that HucMSCs can partially restore ovarian function in mouse models following chemotherapeutic exposure or natural aging through paracrine mechanisms [14]. Notable effects include increased primordial follicle activation, enhanced angiogenesis, reduced follicle apoptosis, and decreased ovarian fibrosis [15–18]. Moreover, several clinical studies have shown that HucMSCs can stimulate follicle development in some patients with premature ovarian insufficiency (POI) and premature ovarian failure (POF) with a subset achieving clinical pregnancy and live birth [19, 20]. While much of the focus has been on the effects of HucMSCs on follicle development and their paracrine factors, the cellular landscapes of various cell types in aged ovaries following HucMSC treatment remain poorly understood. scRNA-seq is a powerful tool to investigate cellular heterogeneity within the ovary. Recent studies using scRNA-seq have revealed numerous cell-specific alterations in murine and primate ovaries during aging [21, 22]. However, the single-cell transcriptomic landscape and cell-specific regulatory changes in aged ovaries after HucMSC transplantation are still unclear.

In our study, we observed that HucMSCs predominantly localized in the ovarian stroma following in situ ovarian injection. HucMSC treatment increased the number of antral follicles and improved oocyte quality in aged mice. On the basis of these observations, we hypothesize that HucMSCs exert their beneficial effects by acting on stromal cells to ameliorate the ovarian microenvironment, thereby promoting follicular development. To test this hypothesis, we employed scRNA-seq to analyze transcriptomic changes across different cell types (oocytes, granulosa cells, and stromal cells) in the ovaries of aged mice treated with either HucMSCs or saline. Our findings revealed a metabolic reprogramming axis that included enhanced steroidogenesis in stromal cells, glycolysis in granulosa cells, and mitochondrial ATP production in oocytes, in response to HucMSC treatment. This study provides novel insights into ovarian microenvironment remodeling by HucMSCs at the single-cell transcriptomic level, advancing the potential clinical application of HucMSC therapy for ovarian aging rejuvenation.

Materials and methods

Experimental animals

All experimental mice were procured from the Animal Center of Sir Run Run Shaw Hospital and maintained in a specific pathogen-free (SPF) facility with controlled environmental conditions of 23–25 °C, a 12 h light/dark cycle, and relative humidity ranging from 40% to 70%. Sample sizes for our studies were not determined by a priori statistical calculations but were based on established conventions within the field. Throughout the experiments, “young”' mice denote 6- to 8-week-old C57BL/6 female mice, while “aged” mice correspond to 8- to 10-month-old C57BL/6 female mice. Fertility tests utilized 10-week-old C57BL/6 male mice in our study.

Preparation and culture of HucMSCs

Clinical-grade HucMSCs at passage 4 were obtained from GENS STEM CELL Biotech Co., Ltd (Zhejiang, China). Prior studies have extensively characterized the phenotype and multipotent differentiation potential of HucMSCs [23]. For our subsequent experiments, we utilized cells from passages 5–6. Lentivirally transduced HucMSCs at passage 3 harboring the green fluorescent protein (GFP) gene (CTCC-009-265) were purchased from MeisenCTCC (Zhejiang Meisen Cell Technology Co., Ltd., Zhejiang, China).

The cell culture methodology employed in this study aligns with our previous research [24]. In brief, the complete culture medium was composed of Dulbecco’s modified Eagle medium (DMEM)/F-12 (Meilunbio, China), supplemented with 10% fetal bovine serum (FBS, Cellmax, China) and 1% penicillin–streptomycin solution (Solarbio, China). Cells were cultured in an incubator maintained at 37 °C with a 5% CO₂ atmosphere. Upon reaching a confluence of approximately 80%, the cells were harvested by digestion with a 0.25% trypsin–ethylenediaminetetraacetic acid (EDTA) solution (Solarbio, China).

Intraovarian cell administration

Aged or young mice were anesthetized via intraperitoneal injection of 200 mg/kg avertin (Sigma-Aldrich, USA). A dorsal incision was made to expose the ovarian tissue, and 29G-insulin syringes were used to inject into the ovaries. The mice were randomly assigned to two equal groups: the Aging-con group, which received in situ injection of saline into both ovaries (25 µl per side); and the Aging-HucMSC group, which received in situ injection of HucMSCs into both ovaries (25 µl per side containing 1 × 10⁶ cells). Mice were then kept in the original environment after surgery and humanly executed after 4 weeks to obtain serum and ovaries for further analysis.

Hormone assays of mouse serum

Before euthanasia, mice were anesthetized, and blood was rapidly collected into serum separator tubes (yellow-top) (BD Vacutainer, USA). After blood collection, all mice were humanely euthanized by cervical dislocation. Bilateral ovaries were then harvested from the abdominal cavity of each mouse. Mouse serum was obtained by centrifuging the blood from the collection tubes at 3000g for 10 min under 4 °C. The concentrations of estradiol and progesterone in the serum were determined using the Access Immunoassay System (UniCel DxI 800, Beckman Coulter, USA) by the Clinical Laboratory Department at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University. Additionally, the level of anti-Müllerian hormone (AMH) was determined according to the manufacturer’s protocol for the enzyme-linked immunosorbent assay (ELISA) kit (Elabscience, China).

H&E staining and follicle counting

Ovarian tissues were fixed in 4% paraformaldehyde (PFA, Meilunbio, China), subsequently embedded in paraffin, and sectioned into continuous 3-µm-thick slices using a microtome. The sections were then stained with hematoxylin and eosin (H&E) to assess the morphology of the ovary. These procedures were conducted by the Department of Pathology at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University. Masson staining was performed by Shanghai Ruiyu Biotechnology Co., Ltd. To prevent any duplication in counting, only follicles containing visible oocyte nuclei were included in the analysis.

Western blot analysis

We used five ovaries from five individual mice per group from three separate experiments for Western blotting analysis, with five unilateral ovaries from five mice constituting one sample, and three such samples constituting one group, resulting in three technical replicates. Western blot procedures have been previously published in our studies [23, 25]. Briefly, proteins were extracted using radioimmunoprecipitation assay (RIPA) lysis buffer (R0010, Solarbio, China) containing a mixture of protease inhibitors (P8340, Sigma-Aldrich, USA). Cellular samples were extracted directly, while ovarian samples were homogenized with a 60 Hz shaking for 1 min, followed by a 15-min digestion at 4 °C and centrifugation at 15,000g for 15 min to obtain the protein supernatant. Proteins were then diluted with 5 × loading buffer (P1040, Solarbio, China). After denaturation at 95 °C for 10 min, proteins were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) (Epizyme, China) and transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, USA). Membranes were blocked in Tris-buffered saline with Tween20 (TBST) buffer containing 5% nonfat dry milk for at least 1 h, followed by an overnight incubation at 4 °C with the primary antibodies. Supplementary Table S3 lists the primary antibodies used to detect the proteins in our study. After three washes with TBST buffer, membranes were incubated with secondary antibodies and visualized using a ChemiDoc Touch imaging system (Bio-Rad, USA) with enhanced chemiluminescence (WBKLS0500, Millipore, USA).

Immunofluorescence and confocal microscopy

Following fixation in 4% PFA, ovaries were subjected to a gradient dehydration process using 10% and 30% sucrose solutions. They were then embedded in OCT compound (Tissue-Tek, Sakura) for cryosectioning. The 5-µm sections were permeabilized and blocked simultaneously at room temperature with a blocking buffer containing 5% bovine serum albumin (BSA, Sigma-Aldrich, USA) and 0.3% Triton X-100 for 1 h. Follicles and oocytes were fixed with 3.7% PFA at room temperature for 30 min, permeabilized with a 0.2% Triton X-100/phosphate-buffered saline (PBS) solution for 20 min, and subsequently blocked with a 1% BSA/PBS solution for 30 min. Samples were incubated overnight at 4 °C with the primary antibody diluted 1:200 in a blocking solution.

After washing three times with PBS, the samples were incubated with the corresponding fluorescent secondary antibodies conjugated with 4′,6-diamidino-2-phenylindole (DAPI) or Hoechst at room temperature for 30 min. Following another three washes, the samples were processed with antifade mounting medium (Beyotime, China), and the fluorescence signals were captured using a laser scanning confocal microscope (LSM800, Zeiss, Germany). A list of primary antibodies used is provided in Supplementary Table S3.

Superovulation and morphologic evaluation of oocytes

Aged and young female mice were administered intraperitoneal injections of 15IU and 8IU of pregnant mare’s serum gonadotropin (PMSG, Sansheng, China), respectively. After 48 h, oocytes at the germinal vesicle (GV) stage were retrieved through mechanical dissociation and cultured in pre-equilibrated M2 medium for subsequent experimental procedures. To acquire in vivo ovulated metaphase II (MII) oocytes, each mouse received an intraperitoneal injection of human chorionic gonadotropin (hCG, San Sheng, China) at the same dosage as the PMSG, 48 h post-PMSG. Another 14–16 h later, MII oocyte-cumulus cell complexes were extracted from the ampulla of the fallopian tube using M2 medium supplemented with 3 mg/mL hyaluronidase (H4272, Sigma Aldrich).

Oocyte RNA sequencing (RNA-seq) and data analysis

GV-stage oocytes were obtained as described above. Five oocytes were transferred to PCR tubes containing lysis buffer with dNTPs and oligo-dT primer, 0.2% Triton X-100, and 2 IU/μL RNase inhibitor [26]. ERCC (External RNA Controls Consortium) spike-in RNA (0.1 µl, 4456740; Invitrogen, USA) was used as an external reference. Complementary DNA (cDNA) was then acquired using the SMART-seq2 method as previously described [24, 27]. Sequencing libraries were constructed by Annoroad Gene Technology Co., Ltd. (Beijing, China). Then, the cDNA production was checked by Qubit 3.0 Flurometer and Agilent 2100 Bioanalyzer to ensure the expected production with length around 1–2 kbp. Then, the cDNA was sheared randomly by ultrasonic waves for Illumina library preparation protocol including DNA fragmentation, end repair, 3′ ends A tailing, adapter ligation, PCR amplification, and library validation. After library preparation, the PerkinElmer LabChip GX Touch and Step OnePlus™ Real-Time PCR System were introduced for library quality inspection. Qualified libraries were then loaded on the Illumina Hiseq platform for PE150 sequencing.

All the raw reads were filtered and clean reads in FASTQ format were obtained. Subsequently, reads were aligned to the mouse reference genome (mm10) by using STAR (version 2.5.2a). The DESeq2 package [28] was employed to analyze the raw transcriptomic data and identify differentially expressed genes (DEGs). Genes with a baseMean over 100 were considered for differential expression analysis, applying a stringent P-value threshold of < 0.05 and a log₂ fold change (FC) criterion of > 0.5 or < –0.5. Gene Ontology (GO) analysis was conducted to elucidate the biological significance of these DEGs, utilizing the DAVID online tool (https://david.ncifcrf.gov), a comprehensive resource for functional annotation and bioinformatics analysis. Visualization of the data analysis results, including volcano plots, heatmaps, and scatter plots, was accomplished through an online platform (https://www.bioinformatics.com.cn).

Mitochondrial function measurement of oocytes and follicles

For reactive oxygen species (ROS) staining, oocytes were incubated with 10 μM 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA, S0033S, Beyotime, China) in M2 medium at 37 °C in the dark for 30 min, followed by three washes in fresh M2 medium. Mitochondrial membrane potential was assessed using a JC-1 detection kit (C2003S, Beyotime, China); briefly, oocytes were cultured in M2 medium containing JC-1 working solution (1:200) at 37 °C in the dark for 20 min, then washed three times with fresh M2 medium. Active mitochondria were stained with 500 nM cell-permeable MitoTracker Red CMXRos (M7512, Invitrogen, USA) in M2 medium at 37 °C in the dark for 30 min, followed by three washes with fresh M2 medium. Mitochondrial ROS were detected using MitoSOX Red (M36007, Invitrogen, USA), with follicles incubated in follicle culture medium containing 1 μM MitoSOX at 37 °C in the dark for 30 min, followed by three washes. After washing, samples were mounted on confocal dishes and immediately examined under a laser-scanning confocal microscope.

Sample processing and scRNA-seq library construction and sequencing

A sterile RNase-free culture dish containing an appropriate amount of calcium- and magnesium-free 1 × PBS was placed on ice. Fresh ovarian tissue was transferred to the culture dish and minced into pieces of approximately 0.5 mm². Tissues were washed with 1 × PBS to remove as much nontarget tissue as possible, including blood stains and fatty layers. Tissues were dissociated into single cells using a dissociation solution composed of 0.35% collagenase IV, 2 mg/mL papain, and 120 U/mL DNase I, in a 37 °C water bath with shaking at 100 rpm for 20 min. Digestion was halted by adding 1 × PBS containing 10% FBS, followed by gentle pipetting 5–10 times with a Pasteur pipette. The resulting cell suspension was filtered through a 70-30-μm stacked cell strainer and centrifuged at 300g for 5 min at 4°C. The cell pellet was resuspended in 100 μL 1 × PBS (0.04% BSA) and treated with 1 mL 1 × red blood cell lysis buffer (MACS 130-094-183, 10 ×) for 2–10 min on ice to lyse remaining red blood cells. After incubation, the suspension was centrifuged at 300g for 5 min at room temperature. The suspension was then resuspended in 100 μL Dead Cell Removal MicroBeads (MACS 130-090-101), and dead cells were removed using the Miltenyi Dead Cell Removal Kit (MACS 130-090-101). The suspension was resuspended in 1 × PBS (0.04% BSA) and centrifuged at 300g for 3 min at 4 °C (repeated twice). The cell pellet was resuspended in 50 μL of 1 × PBS (0.04% BSA). Cell viability was confirmed using trypan blue exclusion, requiring a viability of above 85%. Single-cell suspensions were counted using a hemocytometer or Countess II Automated Cell Counter and adjusted to a concentration of 700–1200 cells/μL.

For library preparation and sequencing with 10× Genomics, single-cell suspensions were loaded onto the 10× Chromium platform to capture single cells, following the manufacturer’s instructions for the 10× Genomics Chromium Single-Cell 3' kit (V3). Subsequent cDNA amplification and library construction steps were performed according to the standard protocol. Libraries were sequenced on an Illumina NovaSeq 6000 sequencing system (paired-end multiplexing run, 150bp) by LC-Bio Technology Co., Ltd. (Hangzhou, China), ensuring a minimum depth of 20,000 reads per cell.

Analysis of scRNA-seq data

Sequencing results were demultiplexed and converted to FASTQ format using Illumina bcl2fastq software (version 2.20). Sample demultiplexing, barcode processing, and single-cell 3′ gene counting were achieved by using the Cell Ranger pipeline (https://support.10xgenomics.com/single-cell-geneexpression/software/pipelines/latest/what-is-cell-ranger, version 3.1.0) and scRNA-seq data were aligned to Ensembl GRCh38/GRCm38 reference genome. The Cell Ranger output of each sample was loaded into Seurat (version 4.3.2) for further analysis.

After creation of raw UMI count matrices, quality control was performed on each cell step by step. First, cells with less than 800 genes were filtered. Then potential doublets detected by DoubletFinder (version 2.0.3) as well as cells with more than 10% mitochondrial genes were excluded. Overall, 20,019 cells passed the quality control threshold.

To correct for batch effects and integrate datasets, we employed the Harmony package (version 1.2.0) within the R statistical environment. To analyze differential abundance of cell types across conditions, we used the miloR package (version 1.10.0) in R. CellChat package (version 2.1.0) was used to investigate the cell–cell communication signal following the online tutorial. To perform trajectory analysis and infer cellular differentiation trajectories, we utilized Monocle2 (version 2.30.0) and Monocle3 (version 1.3.7). These tools allow the reconstruction of pseudotime trajectories to explore dynamic changes in cell states across biological processes.

Isolation and culture of primary ovarian GCs and SCs

Ovarian SCs were isolated according to previously reported literature [29]. To purify the stromal cells, 29G-insulin syringes were used under a dissecting microscope to puncture the ovaries, releasing GCs and oocytes. The released GCs were filtered through a 40-μm sieve, followed by centrifugation at 800g for 5 min to obtain a pellet of GCs. The corpus luteum were then separated and removed, and the remaining ovarian tissue was transferred to α-MEM solution containing 1 mg/mL collagenase IV (17104019, Gibco, USA) and 1 U/mL DNase I (EN0521, Thermo Scientific, USA). Samples were incubated at 37 °C for 1 h, and gently mixed with a pipette every 10 min. Digestion was halted with α-MEM solution containing 10% FBS, and SCs were purified using a 40-μm filter, followed by centrifugation at 800g for 5 min to obtain a pellet of SCs. The expression of GCs- and SCs-specific mRNA showed significant differences, confirming the purity of the two cell populations (Supplementary Fig. S6).

The complete growth medium for GCs was DMEM-F12 medium containing 10% FBS, and for SCs, it was Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco, USA) containing 10% FBS. Cells were resuspended in their respective complete growth media, plated, and after 4–6 h, the medium was changed to remove red blood cells, followed by subsequent experiments.

Real‑time reverse transcription‑quantitative polymerase chain reaction (RT‑qPCR)

Total RNA was obtained from cells using the RNA-Quick Purification Kit (RN001, ESScience, China) according to the manufacturer’s instructions. The procedure of RT-qPCR referred to the protocol of SYBR Green Mix (Vazyme, China) on a CFX-Connect platform (BioRad, USA). Individual mRNA level was expressed relative to Actin mRNA level. The transcriptional levels of relative genes were compared with those in the control, and fold changes are shown as the results. The primer sequences of the targeted genes are listed in Supplementary Table S4.

Targeted steroid hormone LC/MS-based metabolome of the cell supernatant

To obtain the supernatant from the SCs of two groups of aged mice, we isolated and cultured primary stromal cells and continued the culture for an additional 48 h before collecting the supernatant. The supernatant was then flash-frozen in liquid nitrogen and stored at −80°C. Targeted steroid hormone metabolomic analysis using liquid chromatography-tandem mass spectrometry (LC–MS/MS) was conducted by Applied Protein Technology Biotechnology Co., Ltd.

Thawing the samples at room temperature, 200 μL of each sample was mixed with 200 μL of cold methanol and 20 μL of isotope internal standards, vortexed adequately for 1 min, followed by the addition of 200 μL of water. The mixture was centrifuged for 10 min at 12,000g at 4°C. The supernatant (450 μL) was then passed over Cleanert PEP 96 Well Microplates (2mg), which had been sequentially activated with 200 μL of methanol and 200 μL of water, and rinsed with 200 μL of acetonitrile and 200 μL of n-hexane. Methanol (60 μL) was used to elute the target compounds. All elution solutions were collected in a 2.0-mL 96-well plate, to which 75 μL of water solution was added and vortexed adequately. The supernatants were prepared for LC–MS/MS analysis. Analyses were performed using an UHPLC system (LC-30AD, Shimadzu) interfaced with a QTRAP mass spectrometer (AB Sciex 6500+). The separation conditions for the C18 column were as follows: mobile phase A was an ammonium fluoride aqueous solution, and B was methanol. Samples were maintained at 4 °C in the automatic sampler, and the column temperature was held constant at 40 °C. The gradients were run at a flow rate of 600 μL/min, with a 25-μL aliquot of each sample injected. The gradient elution procedure was as follows: B was increased from 45% to 95% and then returned to 45% over 15 min. Quality control (QC) samples were employed to assess and ensure the stability and repeatability of the analytical system. In electrospray ionization (ESI), the conditions were set as follows: source temperature 600 °C, Ion Source Gas1 (Gas1): 60, Ion Source Gas2 (Gas2): 60, curtain gas (CUR): 35, ionSpray Voltage (IS): +5500 V or −4500 V in positive or negative mode, respectively. Multiple reaction monitoring (MRM) mode was utilized for compound monitoring. Analyst 1.7.1 software was utilized to extract chromatographic peak areas and retention times, and standard retention times were employed to identify the metabolites.

Untargeted LC/MS-based metabolomics of SC

Cells were gently scraped off by a scraper on ice and frozen in liquid nitrogen. Samples were thawed on ice, and metabolites were extract using 80% methanol buffer. Sample (100 μL) was mixed with 400 μL precooled methanol, stored at −20 °C for 30 min, and centrifuged at 20,000g for 15 min, and the supernatants were vacuum dried. After redissolution in 100 μL 80% methanol, they were stored at −80 °C for LC–MS analysis. Chromatographic separations were performed on an UltiMate 3000 UPLC System with an ACQUITY UPLC T3 column. The mobile phase consisted of solvent A (5 mM ammonium acetate and 5 mM acetic acid) and solvent B (acetonitrile). The gradient elution conditions were: 0–0.8 min, 2% B; 0.8–2.8 min, 2% to 70% B; 2.8–5.6 min, 70% to 90% B; 5.6–6.4 min, 90% to 100% B; 6.4–8.0 min, 100% B; 8.0–8.1 min, 100% to 2% B; 8.1–10 min, 2% B. Detection of metabolites was carried out using a Q-Exactive mass spectrometer in positive and negative ion modes, with precursor spectra collected at 70,000 resolution and fragment spectra at 17,500 resolution. Data preprocessing and annotation were carried out using XCMS software, including peak picking, alignment, and integration. Metabolites were annotated by matching exact molecular mass data with online databases (KEGG and HMDB) and validated by isotopic distribution measurements and an in-house fragment spectrum library. Analysis was performed in R (version 4.0.0) using various packages for normalization, hierarchical clustering, PCA, PLSDA, and correlation analysis. Significance was determined by P value < 0.05, difference multiple > 1.2 (t-test), and VIP values from PLSDA analysis. Untargeted LC/MS-based metabolomic analysis was conducted by LC-Bio Technology Co., Ltd.

In vitro co-culture of preantral follicles and SCs

Following our previously established protocol for the isolation and culture of follicles [24], we isolated preantral follicles from the ovaries of 2-week-old female mice and cultured in growth medium with ratio of 1:1 mixed with supernatant from AC-SC and AM-SC.

SCs preparation and intraovarian injection

After administering saline or HucMSC to aged mice for 4 weeks, the mice were humanely euthanized. SC pellets were isolated as described above and resuspended in red blood cell lysis buffer (C3702, Beyotime, China) for 5–8 min on ice to lyse the remaining red blood cells. Following centrifugation at 300g for 5 min, the cells were resuspended and washed with cold PBS. Isolated primary ovarian SC, at a density of 1 × 10⁶ cells per side, were then injected in situ into the ovaries of aged mice using the same method as for HucMSC injection. An additional 4 weeks later, phenotypic assessments of ovarian function and oocyte quality were conducted in these two groups of mice that received SC injections.

Statistics and reproducibility

Sample sizes were determined on the basis of established standards in reproductive biology and are consistent with those reported in previous publications with similar experimental designs [25, 30, 31]. We ensured transparent reporting of all sample sizes (n numbers) and specific statistical significance throughout the manuscript. However, note that sample sizes were not determined using a priori power calculations but were based on “established conventions.” This sample size strategy is appropriate for exploratory discovery, but future confirmatory studies would require formal power analysis. Animals were randomly assigned to experimental groups using randomization scheme after unique identification through toe coding. Researchers were blinded to group assignments during sample processing and experimental procedure, with group identities only revealed until statistical analyses. While complete blinding during surgical procedures was not possible owing to visible differences between treatment preparations, this approach substantially reduced potential biases in data collection and interpretation. Statistical analyses were performed using GraphPad Prism software (version 10.3.1) or R software (version 4.3.2). For comparisons between two groups, statistical significance was determined using a two-tailed Student’s t-test; for comparisons among three groups, one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test was utilized, unless otherwise indicated in the figure legends or methods. All quantitative results are presented as mean ± standard error of the means (s.e.m.). Data distribution was assumed to be normal, although this was not formally tested; therefore, data distribution is visualized in each figure. A significance level of P < 0.05 was used for all tests. Each measurement was performed at least in three independent experiments. In the figures, significance is denoted as follows: ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, and ns, P > 0.05.

Results

Intraovarian HucMSC injection boosted follicle growth in an aged murine model

HucMSCs were successfully isolated and characterized as our previously established protocols [24]. Consistent with previous findings [32], the follicle number in aged murine models (8–10 months old) was significantly reduced compared with their younger counterparts. Four weeks after in situ ovarian transplantation of HucMSCs (Fig. 1A), we observed that HucMSCs potently stimulated a significant increase in the total follicle number across all developmental stages (Fig. 1B, C). Notably, there was a marked enhancement in the number of primordial and antral follicles, indicating the therapeutic potential of HucMSCs in mitigating age-related follicular decline. In addition, HucMSC treatment effectively alleviated the age-dependent decrease in estradiol (E2) and AMH levels in elderly murine models (Fig. 1D, E), aligning with previous research outcomes [17]. Compared with the aged control (Aging-con) group, HucMSC administration tended to elevate progesterone levels (Fig. 1F), though did not yield a significant difference. The fertility assay revealed that the litter size in the group receiving HucMSC injections (Aging-HucMSC group) was significantly greater compared with that of the Aging-con group (Fig. 1G, H), highlighting the potential of HucMSCs to enhance reproductive outcomes in aged mice.

Fig. 1.

HucMSC infusion improved follicular development and oocyte quality in aged mice. A Schematic of the HucMSC administration workflow. B H&E-stained ovarian sections from the Young-con, Aging-con, and Aging-HucMSC groups, showcasing tissue morphology. Scale bar, 100 μm. C Quantification of ovarian follicle counts at various developmental stages—primordial, primary, preantral, and antral follicles—among the three groups (n = 3). D–F Serum levels of estradiol (D) (n = 21–25), AMH (E) (n = 5–13), and progesterone (F) (n = 24–30), were quantified. G Schematic of the fertility test time flow. H Litter size of the first offspring after cohabitation with male mice (n = 3–4). I Western blot analysis of AMH, SIRT1, PCNA, and β-ACTIN protein expression in ovaries. J Quantification of AMH, SIRT1, and PCNA protein expression with β-ACTIN as a loading control (n = 3). K Photomicrographs of oocytes from aged mice with and without HucMSC treatment, highlighting abnormal morphology with yellow triangles. L Percentage of abnormal oocytes relative to total ovulated oocytes per experiment for the Aging-con group (n = 93) and Aging-HucMSC group (n = 106). M Spindle morphology in MII-stage oocytes from Aging-con and Aging-HucMSC mice, with scale bars set at 25 μm. N Percentage of aberrant spindles in MII oocytes for Aging-con (n = 78) and Aging-HucMSC (n = 78) mice. Data presented as the mean ± s.e.m. P value was determined by unpaired two-tailed t-test between the two groups or by one-way ANOVA in three groups

Subsequently, we performed western blot analysis and immunofluorescence on ovary samples to evaluate the impact of HucMSC injection on the expression of specific proteins related to follicle growth and antioxidation (Fig. 1H, I). Our results revealed a significant upregulation of AMH following HucMSCs administration. Sirtuin 1 (SIRT1) protein, implicated in ovarian antioxidant capacity and cellular senescence [33], showed improved expression following HucMSC treatment. Furthermore, the expression of proliferating cell nuclear antigen (PCNA) and Ki67, two biomarkers indicative of cell proliferation [34, 35], was notably elevated in the Aging-HucMSC group (Fig. 1I and Supplementary Fig. S1A, B). These indicate that HucMSCs promoted follicular growth and development. Conversely, the fluorescence signals of cleaved caspase-3, a well-established marker of cell apoptosis [36], were significantly reduced in ovarian sections from aged mice treated with HucMSCs (Supplementary Fig. S1C, D). Collectively, HucMSC treatment promoted follicular growth, enhanced antioxidation capacity, and inhibited apoptosis in the ovaries of aged mice.

HucMSCs improved oocyte quality associated with advanced maternal age

Ovarian senescence is marked by a significant decline in oocyte quantity and quality, which are pivotal determinants of female fertility [3]. To investigate the role of HucMSCs on mature (metaphase II, MII) oocyte quality, we conducted in vivo superovulation (IVO) in aged mice treated with either HucMSCs or saline. As anticipated, the number of abnormal oocyte morphology (fragmentation and degradation) was markedly diminished following HucMSC administration (Fig. 1K, L). Typically, aged oocytes exhibit spindle morphological aberrations characterized by the presence of fragmentation, fractures, and improper segregation [37] (Fig. 1M). We observed a significant reduction in the percentage of abnormal spindle morphology in the Aging-HucMSC group (Fig. 1N), suggesting that HucMSC treatment ameliorated oocyte meiotic defects and improved oocyte quality.

To investigate the molecular heterogeneity among oocytes, we isolated fully grown oocytes at the germinal vesicle (GV) stage [38] from three groups of mice, 48 h following the administration of PMSG, and performed single oocyte transcriptomic profiling (RNA-seq) using the SMART-seq2 method (Fig. 2A). Principal component analysis (PCA) of the RNA-seq data revealed significant differences in the transcriptional profiles of oocytes from young mice compared with those from the Aging-con group. Notably, following HucMSC treatment, the transcriptomic signatures of most aged oocytes shifted toward the young cluster (Fig. 2B). Consistently, differentially expressed genes (DEGs) delineated pronounced transcriptional differences between oocytes from young and aged mice (Fig. 2C–E). Importantly, HucMSC treatment partially reversed the age-associated changes (Fig. 2C). Furthermore, the volcano plots indicated a robust transcriptional difference in both young and HucMSC-treated groups compared with the Aging-con group, with over 1000 genes showing as either upregulated or downregulated in their oocytes (Fig. 2D, E). Aging was associated with the downregulation of 1012 genes in oocytes. Remarkably, HucMSCs treatment rescued 768 of these genes, restoring their expression to levels closer to those observed in young oocytes (Fig. 2F). Besides, aging caused the upregulation of 1022 genes in oocytes while HucMSCs declined 794 of them (Supplementary Fig. S2A).

Fig. 2.

HucMSCs alleviated oocyte quality decline in aged mice. A Schematic of the single-oocyte SMART-seq workflow, with oocytes collected 48 h post-PMSG injection. B. PCA plot of single-oocyte RNA transcriptomes from Young-con group (n = 16), Aging-con group (n = 8), and Aging-HucMSC group (n = 9), demonstrating distinct transcriptomic profiles. C Heatmap depicting the gene expression profile of GV oocytes from young, aged, and aged mice treated with HucMSC. D, E Volcano plots delineate the DEGs between young and aged oocytes, as well as between HucMSC-treated and aged oocytes, underscoring the genetic alterations associated with aging and the effects of HucMSC intervention. F Venn diagram showing the overlap between 1012 downregulated DEGs associated with aging and 1028 upregulated DEGs following HucMSC administration. G GO analysis identified 768 DEGs enriched in specific biological pathways, with those related to mitochondrial function highlighted in red. H Violin plots depicting pathway enrichment for three groups, as estimated by GSVA, for two representative pathways. I Heatmap visualizing gene expression patterns associated with mitochondrial oxidative respiratory chain function across the three mouse oocyte groups. J Representative photos of MitoSOX-positive GC within follicles in three groups. K Enumeration of MitoSOX-positive GC within similarly sized preantral follicles (n = 6–18) L Representative images of ROS levels in GV oocytes from three groups, detected by DCFH-DA staining, with scale bars at 25 μm. M Quantification of ROS fluorescence intensity in oocytes from three groups (n = 14–17). N Images of JC-1-stained GV oocytes from three groups. O Calculation of mitochondrial membrane potential (MMP) as the ratio of JC-1 red to JC-1 green signals (n = 16–28). Data presented as mean ± s.e.m. P value was determined by one-way ANOVA in three groups

HucMSCs ameliorated age-associated mitochondrial dysfunction in aged oocytes

To further investigate the functional implications of these HucMSC-responsive genes, we performed Gene Ontology (GO) enrichment analysis of the 768 downregulated genes reversed by HucMSCs in Fig. 2F, which revealed a strong association with pathways related to mitochondrial function (Fig. 2G). To validate this finding, we employed gene-set variation analysis (GSVA), a method that quantifies gene set activity at the single sample level [39]. GSVA analysis identified significant changes in mitochondrial pathways, such as “mitochondrial respiratory chain complex assembly” and “mitochondrial ATP synthesis coupled proton transport” (Fig. 2H). A heatmap further represented that the mitochondrial function-related genes were downregulated in aging oocytes, whereas the HucMSCs reversed these gene expression levels similar to young oocytes (Fig. 2I). These results from RNA-seq analysis underscored the pivotal role of mitochondria in the rejuvenating effects of HucMSCs on aged oocytes.

We next evaluate the mitochondrial function directly using MitoSOX, a specific mitochondrial superoxide indicator, to label follicles across the three groups. Aged mice exhibited a significant increase in MitoSOX-positive granulosa cells within preantral follicles of comparable size to those in the young group. Conversely, this oxidative burden was diminished in the HucMSC-treated group, suggesting that HucMSCs may alleviate the oxidative stress characteristic of the aged follicular microenvironment (Fig. 2J, K). Given the reported accumulation of ROS in aged oocytes [40], we quantified ROS levels in both GV and MII oocytes. Results demonstrated that HucMSCs significantly reduced ROS levels in aged oocytes (Fig. 2J, K and Supplementary Fig. S2B, C). Then, we utilized the JC-1 probe and Mito-tracker Red CMXRos in oocytes to assess the mitochondrial membrane potential (MMP, ΔΨm). Aged oocytes displayed significantly reduced ΔΨm; however, HucMSC treatment restored ΔΨm to near-young levels (Fig. 2N–O and Supplementary Fig. S2D, E). Oocytes from the Aging-HucMSC group exhibited a more homogeneous distribution of mitochondria and a significantly higher mitochondrial intensity compared with those in the Aging-con group (Supplementary Fig. S2F, G). These observations suggested that HucMSC treatment restored mitochondrial activity and reestablished a youthful mitochondrial distribution in oocytes.

As oxidative stress is driven by mitochondrial free radicals, antioxidants play a critical role in neutralizing these species and protecting mitochondrial function [41]. We quantified the expression of the key antioxidant proteins SIRT1 and GPX4 in MII oocytes and observed that both were significantly upregulated in the MII oocytes from the Aging-HucMSC group compared with the Aging-con group. This elevation suggested a potential role for HucMSCs in modulating cellular antioxidant responses (Supplementary Fig. S2H–K).

HucMSCs enhanced follicular development and reshaped the ovarian cell populations in aged ovary

To further elucidate the transcriptional impact of HucMSCs on the aged ovarian somatic cells, we performed scRNA-seq on ovarian samples from aged mice 4 weeks after in situ ovarian injection of either saline or HucMSCs. Each group included pooled samples from the unilateral ovaries of ten mice to ensure adequate representation for comprehensive transcriptomic analysis (Fig. 3A). A recent study [21] profiling single-cell transcriptomic dynamics across the mouse estrous cycle suggested that cycle stage has a relatively minor impact on ovarian stromal cells. This supports the validity of our approach using pooled ovarian cells from multiple mice for sequencing in the context of our conclusions. After rigorous quality control and filtering, a total of 20,019 cells were retained for subsequent analysis (Supplementary Table S1). Unbiased clustering and annotation of the cellular profiles identified six distinct cell subtypes: stromal cells (SC) marked by high Col1a2 expression, granulosa cells (GC) with elevated Hsd17b1, luteal cells (LC) characterized by Prlr, endothelial cells (EnC) marked by Cldn5, epithelial cells (EpC) with increased Krt19 expression, and immune cell (IC) identified by Ptprc (Fig. 3B, C). As depicted in Fig. 3D, the HucMSC-treated ovaries exhibited an increased proportion of GC and LC alongside a reduction in the proportion of SC compared with the saline-injected group. These changes correlate with an increased follicular count and the number of functional corpora lutea in the ovaries of aged mice (Fig. 1A).

Fig. 3.

Single-cell transcriptomic profiling revealed follicular development dynamics in aged mice following HucMSC injection. A Schematic flowchart outlining the scRNA-seq analysis workflow for ovarian samples from two comparative groups. B UMAP plots depicting the segregation of six somatic cell clusters within aged ovaries. C Violin plots showcasing the expression patterns of signature marker genes characteristic of each cell type. D Comparative analysis of the relative proportions of six cell types between Aging-con and Aging-HucMSC ovaries. E Violin plots displaying GO pathway enrichment related to “follicle development” within the clusters, as quantified by AUCell. F Heatmap illustrating the alterations in interaction intensity among six cellular populations between Aging-HucMSC and Aging-con groups. G Heatmap representing the outgoing signaling intensity in each cluster by CellChat, with a gradient from white to dark green indicating increasing expression weight values. H Cell–cell communication networks associated with fibrosis and follicle development among the six clusters, were analyzed using CellChat. Line widths correspond to the number of signaling pairs, with different colors indicating distinct signal origins

We presented the DEGs across six cell populations in Supplementary Fig. S3A. Next, we utilized the AUCell score—an indicator derived from the area under the curve (AUC)—to assess the enrichment of the gene-set associated with follicular development among the expressed genes in each cell (Supplementary Table S2). Our analysis revealed a significant upregulation of genes related to follicular development across all cell populations. Notably, SCs and GCs, two pivotal cell types within the ovary, exhibited a particularly pronounced increase in the expression of these genes (Fig. 3E).

We next conducted cell–cell communication analysis across the six cell populations. Notably, there was an increase in the relative signaling of GCs and a decrease in the signaling of SCs, indicating a substantial shift in the overall cellular communication landscape within the ovary (Fig. 3F and Supplementary Fig. S3B). To delineate the specific signaling pathways affected by HucMSC treatment, our analysis identified that both incoming (Supplementary Fig. S3C) and outgoing (Fig. 3G) signals that upregulated in GCs are predominantly associated with folliculogenesis, including pathways such as Estradiol, NECTIN, and DHEA signaling. Conversely, downregulated signaling in SC was enriched in pathways related to fibrotic processes, such as COLLAGEN and LAMININ signaling (Fig. 3G, H). Collectively, the scRNA-seq revealed that HucMSC enhanced follicular development, reshaped the ovarian cell populations, and improved ovarian microenvironment in aged ovaries.

HucMSCs enhanced transcription and glycolysis in granulosa cells

We conducted GO enrichment analysis on DEGs in GCs and found that most of the upregulated genes were associated with processes such as glycolytic processes, transcriptional regulation, steroid biosynthesis, and cell cycle progression, all of which are essential for follicle growth and differentiation (Fig. 4A and Supplementary Fig. S4A). Since follicular development is dynamic and includes multiple stages, we focused on subpopulations within GCs, aiming to characterize their transcriptomic changes and trajectories during folliculogenesis. Utilizing previously established spatial transcriptome data of the mouse ovary [42], we stratified GCs into six distinct subgroups, each representing a specific stage in the follicular development continuum (Fig. 4B, C). These subgroups include GC from preantral follicles (PAF GC), mural GC from small antral follicles (SAF mGC), mural GC from large antral follicles (LAF mGC), cumulus cells (CC), as well as proliferative mural GC (Pro. mGC) and proliferative CC (Pro. CC). This classification highlighted the heterogeneity of GCs in the orchestrated process of folliculogenesis. We employed MiloR to analyze the differential abundance of cellular subpopulations. Positive values denote enrichment in the HucMSC-treated group, while negative values indicate enrichment in the control group [43]. Consistently, this result showed a marked increase in the proportions of LAF mGC, as well as proliferative mGC and proliferative CC following HucMSC treatment (Fig. 4D). We also observed a significant increase in the proportion of LAF mGC, indicating an augmented number of large antral follicles (Fig. 4E).

Fig. 4.

The differentiation of GCs reflected the upregulated transcription and glycolysis level of follicles in the HucMSC-treated group. A GO biological process enrichment analysis of upregulated DEGs in GCs. B UMAP revealing six subpopulations of GCs. C. Dot plot showcasing the expression patterns of canonical marker genes characteristic of each cell type. D Cell differential abundance calculated by using MiloR. E Relative proportions of six subclusters of GCs between two groups. F Heatmaps depicting collectively and specifically upregulated DEGs across subpopulations, along with representative GO terms and key genes. G Bubble heatmap showing the expression of several genes in different cell subpopulations. H Representative immunofluorescence images of the expression level of RNA polymerase II (phospho S2) in follicles from two groups. Scale bar, 100 μm. I Single-cell differentiation potential was predicted using CytoTRACE, with a color gradient from light yellow representing less differentiated states to dark blue indicating highly differentiated levels. J The GC trajectory predicted by Monocle 3 and visualized by UMAP. Cells were ordered in pseudotime colored in a gradient from white to dark blue. K The change of several glycolysis-related genes in GC differentiation trajectories. L Representative immunofluorescence images of the expression level of GAPDH in follicles from two groups. Scale bar, 100 μm

To gain further insights into the transcriptome change of GCs, we conducted a detailed classification and subsequent GO analysis of the DEGs in each GC subcluster (Fig. 4F and Supplementary Fig. S4B). We discovered that all subgroups exhibited upregulation of genes associated with transcriptional regulation and glycolysis (Fig. 4F). Subsequently, we quantified the expression changes of glycolysis and transcription-related genes in GCs across different subpopulations between the two groups (Fig. 4G). Similar to our previous results of upregulated transcriptional levels in GCs after exposure to HucMSC-derived extracellular vesicles [24], transcription-related gene expression increased in almost every subpopulation, with the most pronounced changes occurring in SAF mGC and LAF mGC (Fig. 4G). We used immunofluorescence staining to confirm a significant increase in the expression of RNA polymerase II (pS2), indicative of cellular transcription levels, in granulosa cells of the HucMSCs group (Fig. 4H). As expected, genes related to cell division and the cell cycle were predominantly upregulated in Pro. mGC and Pro. CC. Genes related to the apoptotic process and negative regulation of cell proliferation were down-regulated in SAF mGC and LAF mGC, respectively (Supplementary Fig. S4B).

Furthermore, we extended our functional validation to GC subpopulations by performing glycolysis-related gene set enrichment analysis across all six GC subtypes (Supplementary Table S5). This analysis demonstrated significantly elevated glycolytic signatures in nearly all GC subclusters from HucMSC-treated aged mice compared with aged controls (Supplementary Fig. S5A, B). Following HucMSCs treatment, the expression of glycolysis-related genes (Gapdh, Ldha, Pkm, Pgk1, Eno1, and Tpi1) was most significantly elevated in proliferative GCs (pro. mGC and pro.CC) and large antral follicles (LAF mGC and CC) (Fig. 4G). The end products of glycolysis in granulosa cells, pyruvate, and lactate, can serve as primary energy sources for oocytes [44]. To capture the dynamic gene divergence of GC differentiation, we performed pseudotime trajectory analysis using Monocle3, excluding proliferative GCs. We integrated CytoTRACE to calculate the differentiation potential of individual cells, defining the state from a less differentiated (totipotent) to a more differentiated (differentiated) state [45]. The analysis identified PAF GC as possessing the highest differentiation potential, designating it as the starting point of the trajectory (Fig. 4I). The inferred direction of differentiation aligned well with the natural progression of follicular growth and development (Fig. 4J). We then focused on the process of cellular glycolysis and identified several key DEGs, such as Ldha, Pkm, Pgk1, and Gapdh, which were upregulated to varying degrees in GCs and across different stages of cell trajectories after HucMSC treatment (Fig. 4K). An increase in glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a key enzyme in glycolysis, in mGCs of the Aging-HucMSC group (Fig. 4L) was also shown. Both key glycolytic enzymes, LDHA and PKM2, showed markedly increased expression in the Aging-HucMSC group compared with controls (Supplementary Fig. S5C, D). Complementary qRT-PCR analysis of primary granulosa cells further confirmed the significant upregulation of glycolytic genes including Ldha, Pkm, and Gapdh (Supplementary Fig. S5E). In addition, we also observed a significant upregulation of the cell cycle-related genes. Among these, the Ccnd2 gene, a key cell cycle gene for GC proliferation, was significantly upregulated during the proliferative phase of mid-follicular development (Supplementary Fig. S4C, D). Collectively, these results suggested that HucMSCs promoted transcription, cell division, and glycolysis in GCs of aged ovaries.

HucMSCs redirected transcriptional evolution of ovarian stromal cells toward steroid secretion and away from fibrosis

After injecting GFP-labeled HucMSCs into the ovaries of mice, we observed that HucMSCs predominantly localized to the stromal compartment rather than integrating into follicular structures. These cells degraded into cellular debris within a few days and persisted for approximately 1 week (Fig. 5A). On the basis of this observation, we hypothesized that the therapeutic effects of HucMSCs are primarily mediated through directly changing ovarian stromal cells. Using Masson’s trichrome staining to assess ovarian fibrosis, we observed elevated levels of interstitial fibrosis in the ovaries of the aged group compared with the young group. Injection of HucMSCs significantly reduced collagen deposition around follicles (Fig. 5B). Recognizing the pivotal role of SCs in sustaining the microenvironment necessary for follicular development, we conducted an in-depth analysis to clarify the specific contributions of SCs to the ovarian milieu. On the basis of previous studies [11, 46], we conducted a meticulous clustering of SCs, categorizing them into nine distinct subgroups (Fig. 5C, D). Proportional differences in each subgroup were compared between the two groups. Changed SC subpopulation distribution was observed, including cell proportion increase in steroidogenic theca cells and WNT4-positive SCs, and cell proportion decreases in early theca cells, progenitor SCs, and fibroblasts (Fig. 5E). Additionally, CCAT scoring revealed increased differentiation potential in early theca cells in the HucMSC-treatment group (Fig. 5F), suggesting an enhanced state of folliculogenesis.

Fig. 5.

HucMSCs therapy induced profound changes in SC developmental trajectories. A Assessment of the localization and residence time of GFP-tagged HucMSC within murine ovaries post-injection. B Masson staining of ovarian sections from three distinct groups. C UMAP plot revealing nine SC subpopulations. D Dot plot highlighting the expression patterns of marker genes specific to each SC subcluster. E Comparative proportions of the nine SC subclusters between control and HucMSC-treated groups. F CCAT scores for each SC subpopulation, indicating their differentiation potential. G Trajectory analysis illustrating potential SC differentiation routes in two groups. H Heatmap displaying the scaled expression of dynamically changing genes along the SC trajectory branches. I Violin plots showing pathway enrichment in two groups, as estimated by GSVA, for fibrosis and steroidogenesis-related pathways

To elucidate the cellular dynamics induced by HucMSCs, we performed single-cell trajectory analysis using Monocle. This analysis revealed that SC differentiation follows divergent trajectories. In the Aging-con group, SCs were observed to follow a fibrotic trajectory, leading to the production of pro-fibrotic SCs, aligning with the physiological increase in ovarian fibrosis during aging [47]. By contrast, in the HucMSC-treated group, SCs exhibited a distinct differentiation trajectory (Fig. 5G). The heatmap analysis of gene expression along this alternative trajectory revealed significant enrichment of genes associated with steroidogenesis, including pathways such as ovarian follicular development, steroid biosynthesis, and response to estrogen (Fig. 5H). These findings suggested that HucMSCs redirected SC differentiation toward a follicle-supportive, steroid-secreting phenotype. Subsequently, we harnessed GSVA to compare the gene set expression differences between fibrosis- and steroidogenesis-related pathways of the SCs in two groups. Our analysis revealed a marked reduction in genes linked to fibrosis following HucMSCs treatment, coupled with an upregulation of genes linked to steroidogenesis (Fig. 5I). These findings provide strong evidence that HucMSCs attenuated fibrotic processes while promoting steroidogenic activity within SCs.

As suggested by the trajectory analysis in Fig. 5G, HucMSC treatment appears to influence stromal cell differentiation broadly toward steroidogenesis across multiple subpopulations, rather than targeting a specific subset. To substantiate this interpretation, we performed the following additional investigations. We calculated steroidogenesis-related gene set enrichment scores (Supplementary Table S6) for each of the nine stromal subclusters. This analysis revealed that nearly all subclusters exhibited significant elevation in steroidogenic signatures in the Aging-HucMSC group compared to the Aging-con group (Supplementary Fig. S6A, B). While the “steroidogenic theca” subcluster demonstrated the highest baseline expression and most substantial upregulation, the pervasive nature of this transcriptional response across diverse stromal subsets strongly supports our hypothesis of global stromal reprogramming.

Immunofluorescence analysis of CYP17A1, a key enzyme in steroidogenesis, provided spatial confirmation of this phenomenon. We observed that CYP17A1-positive steroidogenic theca cells were consistently localized around follicles, with significantly increased fluorescence intensity in the Aging-HucMSC group compared to the Aging-con group, where some theca cells even lacked detectable expression (Supplementary Fig. S6C). Higher magnification imaging (Supplementary Fig. S6D) crucially revealed that this increased steroidogenic protein expression was not confined to the canonical theca layer but showed modest elevation in extratheca stromal regions as well, despite their lower baseline expression levels. Beyond steroidogenesis-related genes, we examined INHA expression, a protein closely associated with hormonal response during folliculogenesis that was also identified as a positive marker in our differentiation trajectory (Fig. 5H). As shown in Supplementary Fig. S6E, INHA was expressed in granulosa cells across all follicle stages, with higher expression in LAFs and SAFs and weaker expression in PAFs. Importantly, higher magnification revealed INHA expression in stromal cells (Supplementary Fig. S6F), with markedly elevated levels in the SCs of Aging-HucMSC group compared with controls (Supplementary Fig. S6G), consistent with our single-cell data. These findings demonstrate a generalized increase in steroidogenic capacity across the stromal compartment rather than being restricted to a specific subpopulation, supporting our research hypothesis.

HucMSCs suppressed profibrotic signal and senescence of immune cells at the transcriptional level

Extensive research has demonstrated that the crosstalk between immune cells (ICs) and SCs is pivotal in the context of ovarian aging [21, 48]. To investigate this interaction, we systematically categorized these immune cells into 12 distinct groups on the basis of their phenotypic and functional characteristics (Supplementary Fig. S7A, B). MiloR and cell proportion analysis revealed a significant increase in macrophages and T cells following the administration of HucMSC (Supplementary Fig. S7C, D). Since macrophages are reported to be contributors to fibrosis and organ aging through interactions with SCs [7, 49], we employed CellChat to analyze signaling communication between macrophages and SCs. In the aged control group, macrophages exhibited elevated secretion of pro-fibrotic factors, such as FN1 and SPP1. Following HucMSCs treatment, however, there was an increase in signals promoting cell growth and angiogenesis, such as IGF and THBS, while fibrosis-related signals received by macrophages, including LAMININ and COLLAGEN, were significantly reduced (Supplementary Fig. S7E). Among profibrotic pathways, transforming growth factor beta (TGF-β) signaling from macrophages to SCs is recognized as a principal driver of fibrosis in multiple tissues [50]. We observed a significant reduction in TGF-β signaling from macrophages to SCs following HucMSC treatment (Supplementary Fig. S7F). Additionally, the protein expression of TGF-β in ovarian stroma was markedly decreased (Supplementary Fig. S7G, H). Furthermore, we observed a reduction in the overall expression levels of the senescence-associated secretory phenotype (SASP) gene set in ICs from the HucMSC-treated group (Supplementary Fig. S7I). These findings underscored the ability of HucMSCs to modulate immune–stromal interactions, suppress profibrotic signaling, and immune senescence, fostering a microenvironment conducive to follicular development.

Transplantation of HucMSC-treated SCs restored ovarian function in aged female mice

Beyond its fibrosis-inhibitory effects, HucMSCs treatment significantly enhanced the steroidogenic potential of SCs. This led us to investigate the causal nature of these molecular alterations—whether SCs are the initiators of follicular development or a downstream effect of stem cell-mediated follicular growth. We hypothesized that HucMSCs ameliorated follicle development of aged ovaries partially dependent on the steroidogenic function redirection of SCs. To test this hypothesis, we isolated primary stromal cells from ovarian tissue and proved its validation by expression of Col4a1 and Dcn rather than follicle-stimulating hormone receptor (Fshr), Hsd17b1, and anti-Müllerian hormone receptor (Amhr) (Supplementary Fig. S8). We extracted primary SCs from HucMSC-treated aged ovaries and transplanted them into the ovaries of other aged mice to assess their beneficial effects on ovarian function (Fig. 6A). Four weeks post-transplantation, aged mice receiving HucMSC-treated SCs (AM-SC) exhibited a significant increase in the number of antral follicles and corpus lutea compared with controls (Fig. 6B, C). Concurrently, serum estradiol levels were markedly elevated in the AM-SC group (Fig. 6D). Furthermore, protein levels of FSHR, AMH, and PCNA in the ovaries were significantly upregulated (Fig. 6E, F). Although the expression of NOBOX, MVH, and BMP15, which are markers of oocyte quality and quantity, did not achieve statistical significance owing to high intragroup heterogeneity, their mean expression levels showed a notable trend toward improvement (Fig. 6F).

Fig. 6.

Transplantation of HucMSC-treated SCs restored ovarian function in aged female mice. A Schematic representation of the primary SC injection protocol. B H&E-stained ovarian sections from the AC-SC and AM-SC groups, displaying tissue architecture. Scale bar, 200 μm. C Enumeration of ovarian follicles and corpus luteums (CLs) between the AC-SC and AM-SC groups (n = 3–4). D Determination of serum estradiol levels (n = 4–6) in two groups. E Western blot assessment of protein expression patterns in ovarian tissues. F Quantitative analysis of protein expression normalized to β-ACTIN (n = 3). G PCA of single-oocyte transcriptomes from AC-SC and AM-SC groups, highlighting distinct transcriptional signatures (n = 5). H GO analysis of biological processes enriched in upregulated DEGs in oocytes following AM-SC treatment. I, J Visualization of two key mitochondrial function-related pathways identified by GSEA. K. Fluorescent microscopy images of JC-1 stained GV oocytes from both groups, indicating MMP, with scale bars at 25 μm. L Imaging of MitoSOX-positive GCs and ROS levels in preantral follicles of similar size, with scale bars at 25 μm. M Computation of MMP as the ratio of JC-1 red to green fluorescence (n = 13–14). N Quantitative assessment of ROS fluorescence intensity in follicles from both groups (n = 9–13). Data presented as mean ± s.e.m. P value was determined by unpaired two-tailed t-test between the two groups

To further evaluate the effects of HucMSC-treated SCs on oocyte quality, we collected fully grown GV oocytes for Smart-seq to assess oocyte transcriptomes between experimental groups. Transcriptomic analysis revealed significant differences in oocyte clustering between the groups (Fig. 6G). Consistent with our results in Fig. 2F–I, GO and GSEA indicated that the upregulated DEGs in the AM-SC group were primarily associated with pathways related to mitochondrial function and ATP synthesis (Fig. 6H–J). JC-1 staining revealed increased MMP in GV oocytes from the AM-SC group (Fig. 6K, M), while ROS and MitoSOX staining indicated reduced oxidative stress within follicles (Fig. 6L, N). Collectively, our results demonstrated that transplantation of HucMSC-primed SCs significantly enhances follicular development in aged mice, yielding ovarian phenotypes comparable to those observed following direct HucMSC injection.

HucMSCs promoted steroidogenic metabolic synthesis in ovarian stromal cells

Ovarian aging is accompanied by a decline in the secretion and metabolism of ovarian sex hormones, resulting in female systemic hormonal imbalance [50, 51]. Since ovarian stromal cells participate in steroid hormone synthesis [52], a significant upregulation in the expression of pivotal genes, including Cyp17a1, Cyp11a1, Star, and Hsd3b1, in SCs was observed (Fig. 7A–C and Supplementary Fig. S9A). We supposed the HucMSC-stimulated steroidogenic cell fate redirection of SCs may play key roles in promoting follicle development. Accordingly, we orchestrated a series of experiments to substantiate our hypothesis. Initially, we assessed the in vivo response of SCs to HucMSCs. Ovaries from mice administered with either saline or HucMSCs post 4 weeks were excised for primary SCs with in vitro culture for 2 days (Fig. 7D). Untargeted metabolomic profiling of SCs from the two groups uncovered significant metabolic differences. Differentially abundant metabolites, including omega-hydroxydodecanoate, cholestan-3-one, and docosenoic acid, were associated with cholesterol and steroid hormone metabolism (Supplementary Fig. S9B, C). Targeted metabolomics of the culture supernatant revealed elevated levels of key sex hormone precursors, such as pregnenolone, progesterone, 17-hydroxyprogesterone, and androstenedione, which are indispensable for estrogen synthesis in GCs (Fig. 7E). These findings suggest an enhanced steroidogenic capacity of the SCs, with potential implications for follicular development and ovarian function.

Fig. 7.

HucMSCs enhanced the steroidogenic metabolic capacity of SCs. A Gene expression profiling of SCs derived from scRNA-seq data. B, C Immunofluorescence detection of CYP11A1 and 3β-HSD in ovarian sections, with nuclear counterstain by Hoechst 33,342. Scale bars, 100 μm. D Schematic representation for the collection of SC supernatant and SCs, and subsequent metabolomic analysis. E Targeted metabolomic assays on SC supernatant, revealing steroidogenic hormone levels between groups. F Expression analysis of Cyp19a1 in GCs from scRNA-seq data. G Flowchart summarizing the experimental design for in vitro studies involving two SC groups and their supernatants. H Graphical representation of follicle growth influenced by SC supernatant. Scale bar, 100 μm. I Statistical analysis of follicle diameters, indicative of growth rate, between two groups (n = 22–28). J Estradiol levels after primary GCs culturing with SC-supernatant from two groups (n = 3). K Relative gene expression analysis in primary GCs after exposure to SC-supernatant from two groups. Data presented as mean ± s.e.m. P value determined by unpaired two-tailed t-test between the two groups

Notably, an augmentation in Cyp19a1 expression in GCs from the HucMSC-treated group indicated that enhanced estrogen synthesis was likely driven by an abundant supply of precursors from SCs (Fig. 7F). To test this hypothesis, we incorporated SC culture medium into in vitro follicular co-culture assays (Fig. 7G, upper). When these two groups of supernatants were mixed with follicular culture medium at a 1:1 ratio for in vitro culture of preantral follicles, we observed an increase in follicular survival rates and a significant improvement in follicular growth (Fig. 7I and Supplementary Fig. S9D). Notably, we observed an increased number of Ki67-positive GCs within follicles (Supplementary Fig. S9E, F), suggesting enhanced proliferation. We also collected the supernatants from AC-SC and AM-SC and used them for primary GC culture (Fig. 7G, lower). Our results indicated that the supernatant from AM-SC significantly enhanced the ability of GCs to produce estrogen (Fig. 7J) and upregulated key genes essential for steroidogenesis, such as Cyp19a1, Hsd17b1, and Inha, as well as other phenotypes consistent with scRNA-seq data, including transcription factors Jund, glycolytic enzymes Pkm and Ldha, and cell proliferation marker Top2a (Fig. 7K). Collectively, these results demonstrate that HucMSCs modulated the secretory and synthetic functions of aged SCs, and this microenvironment transformation enhanced follicular growth and ovarian function.

HucMSCs activate conserved metabolic programs in young ovaries

To investigate whether HucMSC-mediated follicular activation in aged ovaries represents simple age reversal or a conserved therapeutic mechanism, we performed single-cell transcriptomic analysis of young mouse ovaries following HucMSC administration at 1- and 4-week timepoints (Supplementary Fig. S10A–C). Similar to observations in aged ovaries, HucMSC treatment in young mice increased granulosa cell proportions while reducing stromal cell populations (Supplementary Fig. S10D), indicating consistent follicular development outcomes. Follicle development-related gene set enrichment scores showed moderate elevation across cell types at 1 week and robust upregulation in all cellular compartments by 4 weeks (Supplementary Fig. S10E, F), demonstrating that intraovarian HucMSC injection enhances follicular development in young ovaries.

Subclustering of GC populations from all four experimental groups (Supplementary Fig. S11A–C) revealed significant expansion of Pro. mGC and CC populations following HucMSC treatment (Supplementary Fig. S11D). Glycolysis-related gene enrichment analysis (Supplementary Table S5) demonstrated substantial upregulation of glycolytic activity in bulk GC populations at both 1 and 4 weeks post-treatment (Supplementary Fig. S11E, F). This enhancement represented a global increase across all GC subtypes, with particularly pronounced effects in proliferative subpopulations exhibiting higher metabolic demands (Supplementary Fig. S11G, H). Parallel analysis of meticulously classified SC subpopulations (Supplementary Fig. S12A, B) revealed progressive upregulation of steroidogenesis-related genes (Supplementary Table S6) following HucMSC administration. While the effect was modest at 1 week, progenitor SCs already showed significant steroidogenic induction (Supplementary Fig. S12C, D), indicating early commitment toward hormonal secretion. By 4 weeks, HucMSC-induced steroidogenesis became markedly enhanced across multiple SC subtypes (Supplementary Fig. S12E, F), mirroring our observations in aged ovaries (Supplementary Fig. S9A, B).

These findings establish that HucMSC therapy activates conserved metabolic pathways in ovarian somatic cells rather than simply reversing age-associated changes. The coordinated enhancement of GC glycolysis and SC steroidogenesis represents a fundamental mechanism for ovarian reactivation that operates across different physiological states.

Discussion

Ovarian aging is a progressive process characterized by a decline in ovarian function, encompassing reductions in oocyte quantity and quality, hormonal dysregulation, and perturbations within the ovarian microenvironment. These changes harm female fertility and overall health [3]. Ovarian aging typically begins around 30 years of age, and women over 35 years are classified as being of advanced maternal age owing to the increased risk of pregnancies associated with declining oocyte quality. Currently, no therapeutic interventions are available to slow or reverse ovarian aging, and restoring reproductive potential in aged ovaries remains a significant challenge. MSCs have emerged as a promising therapeutic strategy for enhancing ovarian function [53]. Over recent years, numerous clinical studies have demonstrated the capacity of MSCs to reactivate ovarian function in individuals with POI/POF or POR, facilitating clinical pregnancies [54–56]. Despite these advances, the variability in patients’ response to MSC-induced follicular development and the persistently low live birth rates raise questions about the efficacy of MSC-based therapies. Therefore, deciphering the precise mechanisms by which MSCs act within the ovarian microenvironment is essential for optimizing therapeutic potential.

In this study, we found that transplantation of HucMSCs into aged ovaries resulted in the production of superior-quality oocytes with transcriptomic profiles resembling those of young oocytes. Notably, the upregulated genes in these oocytes were enriched in pathways related to mitochondrial function, providing direct evidence of HucMSC-mediated transcriptional enhancement and mitochondrial improvement. Through single-cell transcriptomic profiling, we revealed the cellular and molecular changes in ovarian somatic cells following HucMSC treatment. Specifically, HucMSC enhanced glycolysis-related gene expression in GCs of various subtypes and promoted steroidogenic gene expression in SCs. Using HucMSC-primed SC transplantation into aged ovaries and an in vitro co-culture system, we validated that HucMSC facilitated follicular development and improved oocyte quality, partially through increased steroid secretion by SCs. On the basis of these findings, we propose a mechanism by which HucMSCs enhance oocyte quality via an integrated axis of “steroid synthesis in SCs, glycolysis in GCs, and mitochondrial ATP production in oocytes” in aged ovaries.

Our study reveals notable alterations in SCs and GCs, two main somatic cell compartments within the aged ovary following HucMSC administration. Consistent with previous studies at the histological level, GCs exhibited enhanced cell proliferation and hormone secretion [57], along with a significant reduction in fibrotic characteristics in SCs after HucMSC injection [58]. Granulosa cells and cumulus cells surrounding oocytes play pivotal roles in protecting oocytes from harmful environments and facilitating oocyte growth and meiosis. A recent study has demonstrated that aged oocytes can be rejuvenated and acquire high developmental competence when exposed to young GCs in chimeric follicles [9], further emphasizing the critical role of GCs in determining oocyte quality. Building on our previous in vitro studies demonstrating that extracellular vesicles derived from HucMSCs can enhance transcriptional levels and cellular proliferation in aged ovaries, thereby improving aged follicle survival and growth [24], the scRNA-seq results demonstrate the similar upregulation of global transcriptional activity in HucMSC-treated GCs. GCs primarily rely on glycolysis to produce metabolites such as pyruvate, lactate, and NADPH, which are transported to oocytes to support oxidative phosphorylation within mitochondria [59]. Reduced glycolysis and oxidative phosphorylation have been observed in GCs from aged women, associated with impaired oocyte quality [60, 61]. However, the role of glycolysis in the HucMSC-treated aged ovary has not been previously reported. In this study, we demonstrated that HucMSCs significantly enhanced glycolysis across all GC subtypes, particularly in Pro. mGC and Pro. CC. This improved glycolytic activity may explain the superior oocyte quality observed after HucMSC treatment, as it provides an abundant energy supply from GCs to themselves for proliferation and adjacent oocytes for better growth. These findings further validate the key role of glycolysis metabolism in oocyte quality acquisition and highlight the potential of HucMSCs to reshape the follicular microenvironment.

Our study meticulously delineated the transcriptional profiles of oocytes, revealing a significant restoration of genes related to mitochondrial function in aged oocytes following HucMSCs injection. Mitochondrial dysfunction is closely associated with ovarian aging, contributing not only to the deterioration of oocyte quality but also to the disruption of the ovarian microenvironment, which in turn exacerbates the aging process [62]. Aged oocytes exhibit a marked decrease in mtDNA, associated with diminished fertilization rates and compromised embryonic development [4]. Numerous techniques have been employed to rejuvenate oocyte health by supplementing with heterologous or autologous mitochondrial sources to enhance energy provision, such as cytoplasm transfer from heterologous donor oocytes, and germline mitochondrial transfer from autologous DDX4-positive cells in ovarian tissue [63]. A substantial body of research indicates that MSCs may exert their tissue-reparative biological functions by transferring mitochondria to maintain mitochondrial homeostasis in target cells [64, 65]. Consistent with conclusions drawn from other published studies [32], our study provides compelling evidence that HucMSC administration attenuates oxidative stress in aged oocytes, enhances mitochondrial membrane potential, and upregulates antioxidant protein expression. These results indicate that HucMSCs improve age-related oocyte quality.

Within the ovarian architecture, stromal cells are essential for providing structural integrity and fostering a conducive microenvironment for follicle development [52]. These cells offer both physical support and biochemical signals necessary for follicle growth. Furthermore, stromal cell dysfunction contributes to ovarian aging, with fibrotic changes negatively impacting ovarian reserve [10, 66]. Previous studies have demonstrated that HucMSCs can promote the remodeling of the ovarian extracellular matrix [67]. A recent study has underscored the efficacy of MSCs derived from human embryonic stem cells (hESC) in enhancing the ovarian microenvironment of naturally aging cynomolgus monkeys, showing a reduction in stromal cell fibrosis following MSC administration [58]. Our study corroborates these findings, revealing that HucMSCs significantly mitigate fibrosis, as evidenced by a marked reduction in fibrotic markers at the transcriptional level.

While the nexus between matrix fibrosis and ovarian aging has been well-explored, with several fibrotic drugs showing promise in delaying ovarian aging [29, 68], these treatments carry the risk of systemic side effects. Our study introduces a new perspective by highlighting, in addition to reducing fibrosis, the differentiation trajectory of stromal cells post-HucMSC treatment. This process diverges into acquiring functions related to hormone secretion, a novel aspect not commonly discussed. A recent study supports this notion, indicating that ovarian stromal cells undergo two divergent differentiation pathways: one promoting steroid and lipid metabolism, and the other leading to stromal fibrosis [11]. However, the functional implications of this unique differentiation direction remain to be thoroughly explored. We also observed a reduction in the proportion of progenitor SCs, potentially due to HucMSC-induced differentiation. Consistent with other single-cell datasets from aged ovaries [11], we found a higher proportion of Wnt4-positive SCs, which significantly increased following HucMSC treatment. The potential functions of these cells were not explored in this study and will be investigated in future research.

Certain ovarian stromal cells have been identified as capable of synthesizing steroid hormones and harboring hormone receptors; for example, progesterone receptor (PGR) is detectable in mouse stromal cells, and the “synthesis of lipid” and “synthesis of steroid” pathways are downregulated in ovarian stromal cells of PGR-knockout mouse [69]. Stromal cells, as providers of hormone precursors, play a pivotal role in ovarian hormone metabolism [52], typically with theca cells being the central focus. However, our study extends this perspective by including the broader stromal cell population. Notably, reinjection of HucMSC-treated primary ovarian stromal cells into the ovaries of aged mice led to follicle development phenotypes similar to those observed with direct HucMSCs injection, accompanied by enhanced oocyte quality. This exciting finding directly implicates that HucMSCs can ameliorate the aged ovarian microenvironment by promoting the steroidogenic metabolic capacity of stromal cells. This study, for the first time, elucidates the impact of HucMSCs on the steroidogenic metabolism of ovarian stromal cells, highlighting the pivotal role of these SC’s hormonal paracrine effects in ameliorating the ovarian microenvironment and facilitating follicular development. While our subcluster analyses reveal distinct functional states within stromal cell populations, the definitive functional validation of specific subpopulations—particularly through isolation and transplantation approaches—remains technically challenging owing to the current lack of unique surface markers for their purification. Future studies employing advanced single-cell technologies, lineage tracing models, and targeted manipulation of candidate signaling pathways will be crucial to fully elucidate the mechanistic basis of HucMSC-mediated ovarian reactivation and to establish the functional hierarchy within the reprogrammed cellular ecosystem.

Chronic, low-grade inflammation, commonly termed “inflammaging”, is increasingly recognized as a hallmark of ovarian aging [70, 71]. This inflammatory milieu, characterized by the accumulation of senescent cells and their associated SASP factors—including interleukin (IL)-6, transforming growth factor (TGF)-β, and various chemokines—creates a hostile microenvironment that is detrimental to follicular development [72]. By attenuating the SASP and suppressing profibrotic signaling from immune cells to the stromal compartment, HucMSCs may help resolve the inflammatory microenvironment detrimental to follicular health. This probiotic-like immunomodulation, working synergistically with direct somatic cell reprogramming, likely contributes to reestablishing a niche conducive to folliculogenesis. However, the relatively scarce population of immune cells in the ovary and their limited representation in our single-cell data necessitate further validation through targeted approaches, such as sorted immune cell population sequencing. Future work should focus on identifying the specific HucMSC-derived factors responsible for this anti-inflammatory effect and elucidating the causal pathway linking SASP reduction to improved follicular outcomes.

Our study reveals that HucMSCs induce conserved transcriptomic shifts in both young and aged ovaries, characterized by enhanced glycolytic activity in granulosa cells and progressive steroidogenic commitment in stromal cells. However, we fully acknowledge that, while these global transcriptomic changes demonstrate consistent metabolic reprogramming across age groups, the upstream signaling pathways through which HucMSCs orchestrate this coordinated response remain largely unexplored. The precise molecular dialog between HucMSCs and resident ovarian cells—particularly the initial signals that trigger progenitor stromal cell differentiation and synchronized enhancement of glycolytic capacity—represents a compelling “black box” in our current understanding. Our future investigations are therefore focused on identifying the key paracrine factors and intracellular signaling cascades responsible for initiating these metabolic changes, knowledge that will be essential for optimizing the therapeutic potential of MSC-based interventions for ovarian aging.

Although this preclinical investigation demonstrates the efficacy of HucMSCs in reactivating ovarian function in aged mice, the clinical translation of these findings remains constrained by the limited availability of human ovarian samples. Notably, a clinical trial evaluating MSC therapy in patients with poor ovarian response (POR) revealed a marked age-dependent effect: among women under 40 years, the treatment group exhibited a significantly higher rate of spontaneous pregnancy (27.3%) compared with the control group (9.1%; P = 0.03), whereas no statistically significant benefit was observed in women over 40 years of age (16.6% versus 5.4%). A similar trend was noted in intracytoplasmic sperm injection (ICSI)/in vitro fertilization (IVF) outcomes [55]. These findings are consistent with our unpublished clinical observations in patients with POF, among whom younger individuals consistently demonstrate superior therapeutic responses. We hypothesize that this heterogeneity may reflect an age-related decline in the regenerative capacity of the ovarian microenvironment—a view that aligns with the present study’s mechanistic insight that HucMSCs act primarily by eliciting responses from ovarian somatic cells. Ovaries from older individuals or those with more advanced pathology are likely characterized by exacerbated stromal fibrosis, reduced plasticity of stromal cells, and diminished metabolic adaptability of granulosa cells.

These findings carry dual clinical implications. First, they provide a rationale for patient stratification in future MSC trials, suggesting that individuals with less severe microenvironmental deterioration may represent more suitable candidates. Second, and perhaps more importantly, our work highlights ovarian stromal cells as a promising therapeutic target in their own right. Rather than relying exclusively on cell-based therapies, future efforts could focus on developing novel pharmaceutical agents—such as small molecules or biologics—capable of mimicking the prosteroidogenic and antifibrotic effects exerted by HucMSCs on the ovarian stroma. Such an approach may offer a more accessible and scalable therapeutic strategy for a broader population of patients affected by age-related ovarian decline or premature ovarian failure. While our findings reveal new mechanisms of HucMSC-mediated ovarian reactivation, their clinical translation remains challenging. Further investigation is essential to bridge these translational gaps. Key steps include rigorous validation in physiologically relevant models such as nonhuman primates, coupled with insights from carefully curated clinical samples and standardization of treatment protocols. These lines of investigation are warranted to determine the therapeutic viability of this approach.

Although our immunofluorescence data provide partial spatial clues, this study lacks a panoramic resolution of the spatial relationships among all cell types within the ovary. Future application of spatial transcriptomics or multiplex immunofluorescence imaging could more precisely reveal the three-dimensional localization of HucMSC-induced metabolically reprogrammed cells within ovarian tissue and their proximity to follicular structures. The endpoint measurements of this study focus on local ovarian responses. Future studies should integrate systemic endocrine monitoring (e.g., gonadotropin levels) and evaluate potential effects of local ovarian cell therapy on other organs in the body. This is crucial for a comprehensive understanding of the therapeutic mechanism and for preclinical safety assessment.

It is important to acknowledge the inherent limitations of using human-derived cells in a murine model, where xenogeneic immune responses and cellular persistence kinetics may not fully recapitulate the human clinical scenario. However, the transient stromal localization of HucMSCs (less than 1 week, Fig. 5A) and the sustained functional improvements observed long after their clearance suggest a therapeutic paradigm mediated through paracrine signaling and somatic cell reprogramming rather than long-term engraftment. This mechanism is consistent with reported MSC actions across species, including recent demonstrations in nonhuman primates showing similar functional improvements in aged ovaries. While future validation using autologous systems and clinical specimens will be invaluable, the convergence of findings across experimental models strengthens the translational significance of our mechanistic insights into HucMSC-mediated ovarian reactivation.

Conclusions

We observed a pronounced shift in the fate of stromal cells, transitioning from a fibrotic phenotype to one characterized by augmented steroidogenic capabilities. This cellular transformation enhanced follicular development, promoted glycolysis and proliferation of granulosa cells, and improved oocyte quality with robust mitochondrial function. These insights are pivotal for building a theoretical framework and advancing the clinical deployment of HucMSC transplantation. Our study provides a comprehensive transcriptomic profile of both oocytes and somatic cells following HucMSC treatment, shedding new light on the cellular and molecular mechanisms underlying the rejuvenating effects of HucMSCs on ovarian aging.

Abbreviations

HucMSC

Human umbilical cord-derived mesenchymal stem cell

scRNA-seq

Single-cell RNA-sequencing

SC

Stromal cell

GC

Granulosa cell

DOR

Diminished ovarian reserve

ATP

Adenosine triphosphate

MSC

Mesenchymal stem cell

POI

Premature ovarian insufficiency

POF

Premature ovarian failure

AMH

Anti-Müllerian hormone

GV

Germinal vesicle

MII

Metaphase II

ROS

Reactive oxygen species

RT‑qPCR

Real‑time reverse transcription‑quantitative polymerase chain reaction

SIRT1

Sirtuin 1

PCNA

Proliferating cell nuclear antigen

IVO

In vivo superovulation

PCA

Principal component analysis

DEGs

Differentially expressed genes

GO

Gene Ontology

GSVA

Gene set variation analysis

MMP, ΔΨm

Mitochondrial membrane potential

LC

Luteal cells

EnC

Endothelial cells

EpC

Epithelial cells

IC

Immune cell

PAF GC

GC from preantral follicles

SAF mGC

Mural GC from small antral follicles

LAF mGC

Mural GC from large antral follicles

CC

Cumulus cells

Pro. mGC

Proliferative mural GC

Pro. CC

Proliferative CC

TGF-β

Transforming growth factor beta

SASP

Senescence-associated secretory phenotype

FSHR

Follicle-stimulating hormone receptor

PGR

Progesterone receptor

Author contributions

S-Y Z, Y-L Z, and G-J G, conceptualized the study. Y-Y Z and Y-L Z led the experimental design and wrote the manuscript. Y-Q M mainly conducted analyses of various transcriptome sequencing. Y-Y Z, W-J Y, Yan Z, Yi Z, H-J Z, and Y-Y C performed the experiments. J-M C, J-M J, and L-B S reviewed and revised the manuscript. X-M T and D H helped with the data analysis. Each author had reviewed and given final approval of the version submitted for publication.

Funding

This study was supported by grants from the Joint Funds of the National Natural Science Foundation of China (U23A20403), the “Pioneer” and “Leading Goose” R&D Program of Zhejiang (2024C03200), the Natural Science Foundation of Zhejiang Province (LD26H040001, LQN25H040003, LY23H040006), the National Key R&D Program for Young Scientists of China (2022YFC2702300), the Scientific Research Fund of Zhejiang Provincial Education Department (Y202457169), and the National Natural Science Foundation of China (82401940, 82301851).

Data availability

The raw data from SMART-seq and scRNA-seq presented in this study have been deposited in the Sequence Read Archive (SRA) database under accession numbers PRJNA1193695, PRJNA1193458, PRJNA1193466, and PRJNA1356505. The raw data from untargeted LC/MS-based metabolomics of SC has been deposited in the MetaboLights database under accession no. MTBLS11873. Additional data supporting the study’s findings will be available from the corresponding authors upon reasonable request.

Declarations

Ethics approval and consent to participate

The animal experiments were conducted in strict compliance with the ethical standards set forth in the Basel Declaration and were approved by the Laboratory Animal Welfare and Ethics Committee of Zhejiang University (approval no. ZJU20220410, 11 November 2022). The committee adheres to the guidelines of the International Council for Laboratory Animal Science (ICLAS, governing board approval 6 June 2013) to ensure the ethical conduct of all experimental procedures. Clinical Trial Number: not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.