Therapeutic potential of umbilical cord-derived mesenchymal stem cells in ameliorating ovarian aging: Insights from single-cell analysis

Abstract

BACKGROUND

Ovarian aging, characterized by follicle depletion and declining oocyte quality, significantly contributes to infertility in women of advanced maternal age. Mesenchymal stem cells (MSCs) offer a promising therapeutic strategy, although their precise mechanisms of action remain poorly understood.

AIM

To investigate the therapeutic effects of umbilical cord-derived MSCs (UCMSCs) on ovarian aging and to elucidate the underlying cellular and molecular mechanisms through single-cell transcriptomic analysis.

METHODS

Aged female ICR mice (10-11 months) were allocated into aged control and UCMSC transplantation (A-MSC) groups, along with young controls for comparison. A-MSC mice received tail vein injections of UCMSCs. Ovarian function was evaluated through estrous cycle monitoring, follicle counting, hormone assays, and assessment of oocyte quality (mitochondrial function, spindle morphology, aneuploidy). Single-cell RNA sequencing (scRNA-seq) was conducted on ovarian tissues to analyze transcriptional changes, cell-cell communication, and granulosa cell (GC) trajectories.

RESULTS

Transplantation of UCMSCs resulted in improved estrous cycle regularity, increased total and antral follicle counts, normalized levels of follicle-stimulating hormone, anti-Müllerian hormone, and estradiol, and enhanced yield of metaphase II oocytes. Oocyte quality improved, evidenced by higher mitochondrial membrane potential, increased mitochondrial DNA copy number, and reduced reactive oxygen species levels. scRNA-seq analysis indicated that UCMSCs primarily targeted granulosa and stromal cells, upregulating oxidative phosphorylation pathways and downregulating aging-related pathways (e.g., p53, mechanistic target of rapamycin complex 1, Wnt). Notably, the anti-aging gene Klotho was significantly upregulated. Cell communication analysis revealed a reduction in inflammatory signaling (e.g., transforming growth factor-β), and pseudotime analysis suggested that UCMSCs redirected GC fate away from a pro-inflammatory state.

CONCLUSION

UCMSCs mitigate ovarian aging by enhancing mitochondrial function, upregulating Klotho, remodeling intercellular communication, and influencing GC fate. These findings offer mechanistic insights into MSC-based therapies for age-related ovarian decline.

Key Words: Ovarian aging; Mesenchymal stem cells; Aged control; Single-cell RNA; Umbilical cord-derived mesenchymal stem cell

Core Tip: This study demonstrates that umbilical cord-derived mesenchymal stem cells (UCMSCs) effectively alleviate ovarian aging in aged mice and enhance oocyte quality. Single-cell transcriptomic analysis revealed that UCMSC treatment led to increased mitochondrial oxidative phosphorylation, elevated Klotho expression, and suppression of aging-related pathways. Additionally, UCMSCs improved intercellular communication within the ovary and influenced the developmental trajectory of granulosa cells. These findings elucidate the mechanisms underlying UCMSC-mediated ovarian rejuvenation and highlight their therapeutic potential for addressing age-related ovarian decline.

INTRODUCTION

The increasing trends of population aging and delayed childbearing indicate the clinical significance of ovarian aging disorders. The ovary exhibits functional decline earlier than other organs, contributing to systemic aging through endocrine pathways[1]. Pathologically, ovarian aging is characterized by progressive parenchymal loss and structural reorganization, marked by the irreversible depletion of the primordial follicle reserve. Human follicular loss occurs at a steady rate of approximately 1000 follicles per month, accelerating after age 37 until menopause[2]. Mouse models reflect this process, demonstrating significant follicle reduction at 6-9 months, irregular cyclicity by 9-12 months, and reproductive senescence by 15-16 months[3-5]. The competence of oocytes also declines qualitatively, directly impacting embryogenic potential and pregnancy success[6,7]. Therefore, therapeutic interventions capable of reversing this decline are urgently needed.

Mesenchymal stem cells (MSCs) are multipotent cells characterized by self-renewal and differentiation capabilities. The International Society for Cellular Therapy has established minimal criteria for MSCs, including adherence to plastic under standard culture conditions, specific surface marker expression (CD90, CD73, and CD105 positive; CD45, CD34, CD14 or CD11b, CD79α or CD19, and HLA-DR negative), and multilineage differentiation potential[8]. MSCs can be readily isolated from various fetal and adult tissues, including amniotic fluid-derived MSCs, umbilical cord-derived MSCs (UCMSCs), and placenta-derived MSCs from fetal sources; and bone marrow-derived MSCs, adipose tissue-derived MSCs, and menstrual blood-derived MSCs from adults[9]. Recent advances demonstrate the considerable therapeutic potential of MSCs and their derivatives for enhancing ovarian function[10-12]. MSCs hold promise through multiple mechanisms, including targeted homing to injury sites[13], paracrine secretion of regenerative factors[14], immunomodulation via direct cell contact and soluble mediators[14], and alleviation of oxidative stress by lowering reactive oxygen species (ROS) levels[15].

A significant challenge in understanding ovarian aging mechanisms has been the tissue’s cellular heterogeneity. Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool to deconvolute this complexity, revealing distinct aging trajectories across various cell types. In primates, aged oocytes exhibit impaired oxidative stress responses, while granulosa cells (GCs) are more susceptible to apoptosis[16]. Spatial transcriptomics in mice has identified extensive age-related gene expression alterations[17], while scRNA-seq has characterized immune infiltration and stromal dysfunction in aged ovaries[18]. A study involving highly active MSCs for treating ovarian aging in rhesus monkeys employed single-nucleus transcriptomics to illustrate that MSCs enhance ovarian structure and secretory function via multiple cellular and molecular pathways[19]. To further elucidate the mechanisms by which MSCs improve aged ovarian function, this study utilizes scRNA-seq to analyze changes in ovarian cell subpopulations and identify key signaling pathway alterations following MSC transplantation, providing new evidence to support stem cell-based therapies for ovarian aging.

MATERIALS AND METHODS

Animals and experiment grouping

Female ICR mice, aged 6-8 weeks and 8-9 months, were procured from Beijing Vital River Laboratory Animal Technology Co., Ltd. and maintained under specific pathogen-free conditions at 22-26 °C with a 12-hour light/dark cycle. Aged mice were allowed to reach 10-11 months before experimentation. All procedures received approval from the Institutional Animal Care Committee of the Chinese PLA General Hospital (No. SQ2020056). Aged mice (10-11 months) were randomly divided into two groups: Aged control [A-CON; phosphate buffered saline (PBS)] and UCMSC transplantation (A-MSC), which received tail vein injections. Young control mice (Y-CON) consisted of 6-8-week-old mice administered PBS.

UCMSCs culture and transplantation

UCMSCs were sourced from Nuwacell Biotech Co., Ltd. (Cat# RC02003, Anhui Province, China). These cells were isolated from umbilical cord tissue utilizing the explant culture method and characterized by a normal karyotype, specific surface marker expression (CD73+/CD90+/CD105+; CD14-/CD34-/CD45-/CD79α-/HLA-DR-), and trilineage differentiation potential (osteogenic, chondrogenic, and adipogenic)[20]. After thawing, the cells were seeded at a density of 6000 cells/cm² in culture dishes containing Nuwacell™ ncMission hMSC Medium (Cat# RP02010, Anhui Province, China) and incubated at 37 °C with 5% CO₂. Cells were cultured until 80%-85% confluence was reached for passaging or use. For transplantation, passage 3-5 UCMSCs were trypsinized, resuspended in PBS at a concentration of 1 × 10⁶ cells/0.2 mL, and administered via tail vein injection, with each mouse receiving two injections spaced 48 hours apart. Subsequent analyses were conducted two weeks following the second injection.

scRNA-seq and analysis

Single-cell library preparation and sequencing: Mice were euthanized using cervical dislocation, and ovarian tissues were aseptically collected following abdominal disinfection. Ovaries from three mice per group were pooled as one sample and preserved in a tissue preservation solution. The preparation of a single-cell suspension, library construction, and sequencing were performed by Beijing Capital Biotechnology Co., Ltd. Briefly, ovarian tissues were washed with PBS, minced finely, and digested in a collagenase-based tissue dissociation solution at 37 °C, with intermittent quality checks until complete dissociation was achieved (approximately 1 hour). The resultant cell suspension was filtered, centrifuged, and resuspended in PBS. Red blood cells were lysed using an appropriate lysing buffer, and the remaining cells were resuspended and adjusted to a concentration of 300-600 cells/μL. Cell viability and density were assessed via an automated cell counter, ensuring high viability for downstream applications.

Single-cell suspensions were co-incubated with barcoded gel beads and enzymes, followed by encapsulation into oil droplets to form Gel Bead-In-Emulsions (GEMs) using a microfluidic chip. Each GEM comprised a gel bead preloaded with 10× Genomics primers, a single intact cell, and reverse transcription reagents. Within the GEMs, cell lysis and mRNA reverse transcription occurred, incorporating 10× barcodes into the cDNA. After barcoding, the GEMs were disrupted, and the released cDNA was purified and amplified via polymerase chain reaction (PCR). The amplified products were quality-controlled for fragment size and yield. For library construction, cDNA was randomly fragmented to sizes of 200-300 bp, end-repaired, A-tailed, and ligated with Illumina-compatible P7 adapters containing sample-specific indexes.

Final cDNA libraries were size-selected and sequenced on the Illumina NovaSeq platform, yielding 150 Gb data output per sample. Raw sequencing data in FASTQ format underwent quality assessment using FastQC. The CellRanger software, employing the STAR algorithm, aligned reads to the mouse reference genome (mm10-2020), extracting barcodes, unique molecular identifiers, and gene sequences to generate a gene expression matrix.

QC: The gene expression matrix was processed using the Seurat R package (v5.0.2) to create Seurat objects and generate quality control metrics for each ovarian scRNA-seq sample. Cells with extreme values were filtered out based on the criteria: 500 ≤ nFeature_RNA ≤ 7500 and percent.mt ≤ 15%. Potential doublets were subsequently removed using the scDblFinder package (v1.20.2).

Data analysis pipeline: Gene expression data underwent normalization, variable feature selection, and scaling. Dimensionality reduction was achieved using PCA (20 principal components) with Harmony integration for batch correction. Cells were clustered at a resolution of 0.8 and visualized using UMAP. Cluster-specific marker genes were identified through differential expression testing.

Cell clusters were annotated based on established marker genes compiled from the literature. Differentially expressed genes (DEGs) defining each cluster were identified using the FindMarkers function in Seurat, followed by functional enrichment analysis against the MSigDB database to characterize the biological pathways and processes associated with each cell type. Detailed annotation results are provided in the results section.

Age-related transcriptional signatures were quantified by calculating senescence scores based on the GenAge database gene set using the AddModuleScore function. Pathway activity scores were derived from MSigDB gene sets using the same scoring methodology. Mitochondrial functional activity was assessed using the MitoCarta3.0 gene set, employing the AUCell algorithm from the escape package (v2.2.3). Cell-cell communication networks were inferred using CellChat (v2.1.2) to systematically analyze ligand-receptor interactions across different cell populations. Developmental trajectories were constructed with Monocle (v2.34.0) to model the pseudotemporal ordering of cellular states.

Metaphase II oocyte collection

Mice were superovulated with 10 IU pregnant mare serum gonadotropin, followed by 10 IU human chorionic gonadotropin (hCG) administered 46-48 hours later. Metaphase II (MII) oocytes were collected 13-16 hours post-hCG administration on a 37 °C heated stage. After cervical dislocation, the oviducts were dissected and transferred to pre-warmed M2 medium (Sigma, Cat# M7167, MO, United States). The ampulla was punctured to release cumulus-oocyte complexes, which were treated with hyaluronidase to remove cumulus cells. MII oocytes were washed and maintained in M2 medium for subsequent use.

Estrous cycle analysis

The estrous cycles of mice were tracked daily for 21 days via vaginal cytology. Each morning at 9:00, vaginal lavage was performed using 30 μL of saline. The collected cells were smeared onto slides, air-dried, fixed in 70% ethanol, and stained with methylene blue. After drying, slides were examined microscopically. A normal estrous cycle was characterized by the sequential transition through all stages from diestrus to estrus.

Hematoxylin and eosin staining and follicle counting

Ovaries were collected following cervical dislocation and fixed in 4% paraformaldehyde at 4 °C, after which they underwent dehydration and paraffin embedding. Serial sections were cut at a thickness of 4 μm, with every fifth section mounted for analysis. Deparaffinization of the sections was performed using xylene, followed by rehydration through a graded series of ethanols. Hematoxylin and eosin staining was conducted according to standard protocols: Sections were stained with hematoxylin for 3-5 minutes, blued in running tap water, and subsequently stained with eosin for 30-60 seconds. After dehydration through graded ethanol and xylene, the sections were coverslipped with neutral resin. For follicle counting, every fifth section was analyzed to tally primordial, primary, secondary, and antral follicles. The total number of follicles per ovary was estimated by multiplying the counted numbers by 5.

Immunohistochemical staining

After deparaffinization and rehydration, antigen retrieval was performed using heated 10 mmol/L sodium citrate buffer (pH = 6.0) at 95-100 °C for 15-20 minutes. The sections were treated with 3% H₂O₂ to block endogenous peroxidase activity, followed by blocking with 3% BSA. Primary antibody incubation was conducted overnight at 4 °C using anti-Klotho antibody (EPR6856, Abcam, Cat# ab181373, 1:300, Cambridge, United Kingdom), followed by incubation with the secondary antibody [goat anti-rabbit IgG H&L (HRP), Abcam, Cat# ab205718, 1:2000, Cambridge, United Kingdom] for 30-60 minutes at room temperature. The sections were washed three times in 0.01 M PBS (pH = 7.4) for 5 minutes each. DAB development was monitored microscopically and terminated by immersion in water once optimal staining was achieved. Nuclei were counterstained with hematoxylin for 1-2 minutes and blued in running tap water. Finally, the sections were dehydrated through a graded ethanol series, cleared in xylene, and mounted with neutral resin.

Real-time quantitative PCR

Total RNA was extracted from ovarian tissue using the TRIzol method. Briefly, tissue samples were ground in liquid nitrogen and then homogenized in TRIzol. Following chloroform separation and isopropanol precipitation, the RNA pellet was washed with 75% ethanol and dissolved in RNase-free water. RNA quality was assessed by measuring A260/A280 ratios (1.8-2.1) and by electrophoresis, with subsequent DNase I treatment performed to eliminate genomic DNA. Reverse transcription utilized the Fast All-in-One RT Kit (ESscience Biotech, Cat# RT001, Shanghai, China). Quantitative PCR analysis was conducted using the LightCycler 480 Real-Time PCR System (Roche) with 2 × Super SYBR Green qPCR Master Mix (ESscience Biotech, Cat# ES-QP002, Shanghai, China), following the manufacturer’s protocol. The thermal cycling program included an initial denaturation at 95 °C for 5 minutes, followed by 40 cycles at 95 °C for 15 seconds and 60 °C for 1 minute. Given its widespread use and stable expression in mouse ovarian tissues, HPRT1 was selected as the internal reference gene for standardized analysis[21,22]. HPRT1 served as the endogenous control for normalization of mRNA expression, with gene expression levels calculated using the 2^-ΔΔCt method. The primers are listed in Table 1.

Table 1 Primer sequences used for quantitative polymerase chain reaction analysis.

Gene	Primers (5’-3’)
Ndufb2	Forward primer	CCCCGGTACAGGGAGTTTC
	Reverse primer	GCCAAAATCGCCAAAGAATCCA
Cox5a	Forward primer	GCCGCTGTCTGTTCCATTC
	Reverse primer	GCATCAATGTCTGGCTTGTTGAA
Cox7c	Forward primer	ATGTTGGGCCAGAGTATCCG
	Reverse primer	ACCCAGATCCAAAGTACACGG
Atp5e	Forward primer	CAGGCTGGACTCAGCTACATC
	Reverse primer	GTTCGCTTTGAACTCGGTCTT
HPRT1	Forward primer	TGATAGATCCATTCCTATGACTGTAGA
	Reverse primer	CAACAAAGTCTGGCCTGTATC

Western blot

Ovarian tissues were pulverized in liquid nitrogen and homogenized in RIPA buffer containing protease inhibitors. Following centrifugation at 12000 × g for 15 minutes at 4 °C, the protein concentration in the supernatant was quantified using the BCA assay. After denaturation, proteins were separated by sodium-dodecyl sulfate gel electrophoresis and transferred to polyvinylidene fluoride membranes. The membranes were blocked and subsequently incubated with primary antibodies overnight at 4 °C: COXIV Polyclonal antibody (Proteintech, Cat# 11242-1-AP, 1:1000, Wuhan, Hubei Province, China), ATP5A1 Polyclonal antibody (Proteintech, Cat# 14676-1-AP, 1:1000, Wuhan, Hubei Province, China), and Vinculin antibody (Proteintech, Cat# 26520-1-AP, 1:5000, Wuhan, Hubei Province, China). Following this, secondary antibody incubation [HRP-conjugated Goat Anti-Rabbit IgG (H+L), Proteintech, Cat# SA00001-2, 1:2000, Wuhan, Hubei Province, China] was carried out for 1 hour at room temperature, interspersed with three washes of 10 minutes each in Tris-buffered saline with 0.1% Tween-20. Signal detection was performed using an ECL kit, and images were captured with a chemiluminescence system. Band intensities were quantified using ImageJ software.

Immunofluorescence staining of oocytes

Oocytes were fixed in 4% paraformaldehyde for either 30 minutes at room temperature or overnight at 4 °C, followed by treatment with 0.5% Triton X-100 for 20 minutes at room temperature. After blocking for 1 hour at room temperature or overnight at 4 °C with 1% BSA, samples were incubated with α-tubulin-FITC antibody (Sigma, Cat# F2168, 1:800, MO, United States) for 2 hours in the dark at room temperature. After three washes of 5 minutes each with 0.01 M PBS (pH = 7.4) in the dark, oocytes were transferred to Antifade Mounting Medium with DAPI (Beyotime, Cat# P0131, Shanghai, China) on glass slides. Coverslips were carefully applied to prevent air bubbles, and samples were observed using a confocal microscope. Prepared slides were stored in the dark at -20 °C for up to one month.

Chromosome spreads

To remove the zona pellucida, oocytes were treated with Acid Tyrode’s Solution (Bioroyee, Cat# RL2624, Beijing, China) for 3 minutes at room temperature. To ensure proper chromosome spreading, the denuded oocytes were then incubated in a hypotonic solution at 37 °C for 15-20 minutes. After three washes of 5 minutes each in M2 medium to eliminate residual acid, the oocytes were transferred into 30-40 μL of spreading solution (1% PFA, 3 mmol/L dithiothreitol, and 0.15% Triton X-100) within a hydrophobic barrier drawn on a glass slide and allowed to air-dry at room temperature under controlled humidity (40%-50%) overnight. Lastly, the samples were mounted with antifade mounting medium containing DAPI and coverslipped for microscopic observation.

Measurement of mitochondrial DNA copy number

The mitochondrial DNA (mtDNA) copy number in mouse MII oocytes was determined using absolute quantification. A fragment of the mitochondrial b6 gene was cloned into a plasmid vector to create a standard for quantification. Serial dilutions of the plasmid enabled the establishment of a standard curve, from which the mtDNA copy number in oocytes was calculated. Individual MII oocytes served as samples, amplified by PCR using the TaKaRa Taq™ Hot Start Version kit (TaKaRa, Cat# R007A, Japan) under the following conditions: Initial denaturation at 98 °C for 30 seconds; 35 cycles of 98 °C for 10 seconds, 60 °C for 30 seconds, and 72 °C for 30 seconds; followed by a final extension at 72 °C for 10 minutes. The PCR products were verified via agarose gel electrophoresis, then ligated into a T-vector and transformed into competent cells. Positive clones were selected on ampicillin-containing medium and confirmed by sequencing. Plasmids were extracted using the TIANprep Mini Plasmid kit (TIANGEN, Cat# DP103, Beijing, China) and utilized as quantification standards. The copy number was calculated using the following formula: Copy number = (plasmid concentration × 10^-9 × 6.02 × 10²³)/(plasmid length × 660). After serial dilution, these standards were subjected to real-time qPCR to generate a standard curve by plotting Ct values against lg (copy numbers). The mtDNA copy number in experimental samples was determined by comparing their Ct values to this standard curve.

Mitochondrial membrane potential assay

The mitochondrial membrane potential of oocytes was assessed using the Mitochondrial Membrane Potential Assay kit with JC-1 (Beyotime, Cat# C2006, Shanghai, China) according to the manufacturer’s protocol. Briefly, oocytes were incubated in JC-1 staining working solution at 37 °C for 20 minutes. Following incubation, the oocytes were transferred to 1X JC-1 staining buffer and washed several times. The washed oocytes were then placed in droplets of culture medium on glass slides, coverslipped, and observed immediately under a fluorescence microscope. Fluorescence intensity was quantified using ImageJ software, with the degree of mitochondrial depolarization expressed as the red/green fluorescence intensity ratio.

ROS assay

Intracellular ROS levels in oocytes were measured using the ROS Assay kit (Beyotime, Cat# S0033S, Shanghai, China) following the manufacturer’s instructions. Oocytes were incubated with 5 μM DCFH-DA diluted in PBS at 37 °C for 20 minutes. After several washes, oocytes were transferred into droplets of culture medium on glass slides, coverslipped, and examined immediately under a fluorescence microscope. Fluorescence intensity was quantified using ImageJ software.

Aneuploidy and copy number variations assay

MII oocytes were treated with Acid Tyrode’s Solution to remove the zona pellucida, followed by the extraction of the first polar body using a microinjection needle. The resulting enucleated oocytes were submitted to Peking Jabrehoo Med Tech Co., Ltd. for aneuploidy and copy number variation (CNV) analysis.

Single-cell whole-genome amplification was performed using the SurePlex DNA Amplification System (Illumina, Cat# PR-40-415101-00, CA, United States). Given the low cell count (< 3 cells), the pre-amplification cycle number was increased from 12 to 15, while the amplification cycle number was raised from 14 to 18 to enhance product yield. Amplified products were purified using 1 × XP Ampure beads. Library preparation was conducted with the TG DNA Library Prep Kit (Veriseq) (Illumina, Cat# 20029274, CA, United States), extending the transposon fragmentation time from 5 minutes to 10 minutes. Sequencing was performed on the Illumina NextSeq 550 platform using the NextSeq^® 500/550 High Output Kit v2 (75 cycles) (Illumina, Cat# FC-404-2005, CA, United States).

Raw sequencing data underwent quality control procedures, including adapter trimming and removal of low-quality reads. High-quality reads were aligned to the mouse reference genome (GRCm³8) using BWA[23] with default parameters. Duplicate reads were removed using Picard, and uniquely mapped reads with a mapping quality score (MAPQ) greater than 30 were utilized for CNV analysis. The mean sequencing depth for each single oocyte was maintained at approximately 0.1 × to 0.3 × coverage. The genome was partitioned into consecutive 600 kb windows, with read counts per window normalized by GC content, mappability, and baseline correction[24]. A Hidden Markov Model was employed to identify CNV boundaries[25]. The median corrected coverage of each segment was utilized as the copy ratio. Segments with copy ratios greater than 2.3 were classified as duplications (2.3-2.7 indicating mosaic duplication), while those with ratios less than 1.7 were classified as deletions (1.3-1.7 indicating mosaic deletion). For whole-chromosome aneuploidy, a deviation crossing these thresholds for an entire chromosome was designated as either monosomy or trisomy.

External validation dataset

To provide an independent reference for ovarian aging-associated transcriptional changes, a publicly available scRNA-seq dataset GSE232309[18] obtained from the Gene Expression Omnibus database was analyzed. This dataset contains ovarian GCs isolated from physiologically young and aged mice. The processed GC Seurat object from the original study was downloaded and analyzed in R (version 4.4.3). Cells were grouped according to metadata annotations (young vs old). Klotho expression was assessed using the normalized expression value of the Kl gene. To evaluate mitochondrial oxidative phosphorylation (OXPHOS) activity, a simplified OXPHOS-related gene signature score was calculated based on the average normalized expression of four representative OXPHOS genes (Cox5a, Cox7b, Atp5e, and Ndufb2), which were also experimentally validated in our study. Differences between groups were evaluated using the Wilcoxon rank-sum test.

RESULTS

UCMSC transplantation significantly increased the ovarian performance of aging mice

Ovarian performance was evaluated using a comprehensive set of parameters that included phenotypic, morphological, endocrine, and functional criteria. These parameters encompassed the regularity of the estrous cycle as a clinical manifestation, follicular count as a key indicator of histomorphology, serum levels of follicle-stimulating hormone (FSH), anti-Müllerian hormone (AMH), and estradiol (E2) representing the endocrine profile, and the number of oocytes retrieved reflecting overall ovarian responsiveness. The cytological features corresponding to each stage of the estrous cycle are illustrated in Figure 1A. The alterations in the estrous cycle across different experimental groups are presented in Figure 1B and C. The Y-CON group exhibited a relatively regular estrous cycle, with a total of 19 normal cycles recorded. In contrast, the A-CON group displayed prolonged diestrus phases, resulting in only 11 normal cycles. Following transplantation with UCMSCs, the A-MSC group demonstrated an improvement, recording 16 normal cycles. Additionally, the proportional distribution of time spent in each estrous stage for all groups is quantified in Figure 1B, indicating that UCMSC transplantation significantly reduced the time spent in diestrus, promoting a more regular cyclicity.

Figure 1 Umbilical cord-derived mesenchymal stem cell transplantation restores estrous cycle regularity in aged mice. A: Representative cytological images of the four estrous cycle stages. Diestrus is characterized predominantly by the presence of leukocytes. Proestrus shows mainly nucleated epithelial cells, with a small number of cornified epithelial cells and few leukocytes. Estrus is dominated by large numbers of anucleated cornified epithelial cells. Metestrus exhibits small, densely stained leukocytes mixed with a small proportion of cornified cells. Scale bar = 33.647 μm; B: Proportional distribution of each estrous cycle stage among the different groups. Note the significant prolongation of diestrus in the aged control group and its reversal in the umbilical cord-derived mesenchymal stem cell (UCMSC) transplantation group; C: Estrous cycles recorded over the 21-day observation period. UCMSC transplantation increased the total number of normal cycles in aged mice. ^bP < 0.01, ^dP < 0.0001. Y-CON: Young control; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation.

Figure 2A shows representative hematoxylin and eosin-stained images of ovarian sections depicting different follicular developmental stages. Then, quantitative analyses of follicles at each developmental stage for all groups are presented in Figure 2B and C. The A-CON group showed a significant reduction in both the total follicle count and the number of follicles at each specific developmental stage when compared to the Y-CON group. In contrast, following UCMSC transplantation, the A-MSC group demonstrated a significant increase in the total follicle count relative to the A-CON group, with notable increases across all follicle types. Among these, the rise in antral follicles reached statistical significance (P < 0.05). While the increases in primary and secondary follicles in the A-MSC group showed a clear upward trend, they did not achieve statistical significance compared to the A-CON group. Importantly, this improvement in ovarian reserve translated into enhanced functional capacity, as the A-MSC group yielded a significantly greater number of MII oocytes (5.7 ± 2.7) compared to the A-CON group (3.7 ± 3.0) (P < 0.05), although both groups remained below the yield of the Y-CON group (32.4 ± 8.2) (P < 0.00001) (Figure 2D).

Figure 2 Umbilical cord-derived mesenchymal stem cell transplantation enhances ovarian reserve and function in aged mice. A: Representative hematoxylin and eosin-stained images of ovarian sections showing different follicular developmental stages (arrows). Primordial follicle: An oocyte surrounded by a single layer of flattened granulosa cells, mainly located in the ovarian cortex. Primary follicle: An enlarging oocyte enclosed by a single layer of cuboidal granulosa cells. Secondary follicle: An oocyte surrounded by multiple layers of granulosa cells. Antral follicle: An oocyte surrounded by multiple layers of granulosa cells, with visible antral cavity formation. Scale bar = 100 μm; B: Number of follicles at each developmental stage per ovary; C: Total follicle count per ovary. Umbilical cord-derived mesenchymal stem cell transplantation (A-MSC) significantly increased the total follicle count, with a significant increase in the number of antral follicles; D: Number of metaphase II (MII) oocytes collected after superovulation. The A-MSC group yielded significantly more MII oocytes than the aged control group. Serum hormone levels measured by enzyme-linked immunosorbent assay; E: Follicle-stimulating hormone (FSH) level; F: Anti-Müllerian hormone (AMH) level; G: Estradiol (E2) level. A-MSC significantly reduced FSH levels and increased AMH and E2 levels in aged mice. Data are presented as mean ± SEM. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001, and ^dP < 0.0001. Y-CON: Young control; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation; FSH: Follicle-stimulating hormone; AMH: Anti-Müllerian hormone; E2: Estradiol.

Serum samples were obtained via orbital blood sampling from mice in each group, with FSH, AMH, and E2 levels quantified using enzyme-linked immunosorbent assay (Figure 2E-G). FSH levels were lowest in the Y-CON group (3.78 ± 0.98 mIU/mL), elevated in the A-CON group (7.03 ± 1.12 mIU/mL), and significantly reduced by UCMSC transplantation (6.14 ± 1.39 mIU/mL, P < 0.05). AMH levels were highest in the Y-CON mice (7.56 ± 1.23 ng/mL) and lowest in the A-CON mice (3.97 ± 1.06 ng/mL). Treatment with UCMSCs significantly elevated AMH levels to 5.76 ± 0.98 ng/mL (P < 0.01). Similarly, the severe reduction in E2 levels observed in the A-CON group (28.11 ± 6.75 pg/mL) compared to the Y-CON group (87.52 ± 6.4 pg/mL) was partially but significantly ameliorated by UCMSCs treatment, raising E2 levels to 36.51 ± 8.54 pg/mL (P < 0.05).

UCMSC transplantation ameliorates oocyte quality impairments in aging mice

The normal development and functionality of oocytes hinge on intact mitochondrial performance. In this context, mitochondrial membrane potential serves as a critical indicator of cellular health, with its collapse signaling cellular distress and an increased likelihood of apoptosis. The mtDNA copy number constitutes the essential genetic reservoir for OXPHOS, forming the foundation for ATP production. A decline in mtDNA copy number compromises the electron transport chain (ETC), resulting in decreased membrane potential, electron leakage, and the excessive production of ROS. Consequently, assessing ROS levels is vital, as oxidative stress can damage cellular components and precipitate a detrimental cycle of worsening mitochondrial dysfunction. The relative fluorescence intensity of ROS in oocytes was quantified across groups (Figure 3A). ROS levels were significantly elevated in the A-CON group compared to the Y-CON group (P < 0.001). Transplantation of UCMSCs resulted in a significant reduction of ROS levels in the A-MSC group (P < 0.01 vs A-CON) (Figure 3B). The reduction in mitochondrial membrane potential, indicated by a decreased JC-1 red-to-green fluorescence ratio, was most pronounced in the A-CON group (Figure 3C and D). The red-to-green ratio was significantly higher in the Y-CON group compared to the A-CON group (P < 0.0001). Transplantation of UCMSCs significantly restored mitochondrial membrane potential, resulting in a markedly increased ratio in the A-MSC group compared to the A-CON group (P < 0.01). As shown in Figure 3E, the mtDNA copy number in MII oocytes was severely depleted in the A-CON group (117960.15 ± 61982.68) relative to the Y-CON group (341984.37 ± 256102.12). UCMSC transplantation significantly elevated the mtDNA copy number to 269974.44 ± 147824.36 in the A-MSC group, representing a statistically significant difference from the A-CON group (P < 0.01).

Figure 3 Umbilical cord-derived mesenchymal stem cell transplantation ameliorates oocyte mitochondrial function in aged mice. A: Representative images of reactive oxygen species (ROS) levels in oocytes, detected by DCFH-DA fluorescence; B: Quantitative analysis of ROS levels in oocytes. ROS levels were significantly elevated in the aged control (A-CON) group and were reduced following umbilical cord-derived mesenchymal stem cell transplantation (A-MSC); C: Assessment of mitochondrial membrane potential (ΔΨm) using JC-1 staining. Representative fluorescence images are shown (red: J-aggregates, high ΔΨm; green: J-monomers, low ΔΨm); D: Quantitative analysis of the JC-1 red-to-green fluorescence intensity ratio. The reduced ratio observed in A-CON oocytes was significantly restored by A-MSC; E: Absolute quantification of mitochondrial DNA (mtDNA) copy number in individual metaphase II oocytes. The marked depletion of mtDNA copy number in the A-CON group was significantly rescued by A-MSC. Scale bar = 20 μm. Data are presented as mean ± SEM. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001, ^dP < 0.0001. Y-CON: Young control; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation; ROS: Reactive oxygen species; mtDNA: Mitochondrial DNA.

Subsequent analyses evaluated the meiotic apparatus and genetic integrity of oocytes. Spindle morphology directly reflects the structural basis for chromosome segregation, while chromosome spreading visually confirms segregation errors leading to aneuploidy, a major cause of miscarriage. Additionally, CNV detection identifies finer-scale chromosomal imbalances associated with developmental disorders. As illustrated in Figure 4A, typical normal spindles are symmetrical and fusiform, whereas abnormalities manifest as disorganized microtubules and misaligned chromosomes. Quantification of these phenotypes (Figure 4B) revealed that spindle assembly was severely compromised in the A-CON group, with only 38.89% exhibiting normal morphology compared to 84.62% in the Y-CON group. This age-related defect was significantly rescued by UCMSC transplantation, increasing the incidence of normal spindle morphology to 62.79% in the A-MSC group (P < 0.05). The proportion of MII oocytes with a normal chromosome count (20) was quantified via chromosome spreading (Figure 4C). As shown in Figure 4D, the rate was 91.43% in the Y-CON group, significantly reduced to 53.85% in the A-CON group. UCMSC transplantation resulted in a non-significant but notable increase to 70.37% in the A-MSC group. A comprehensive cytogenetic analysis was conducted on 47 MII oocytes without polar bodies, with results detailed in Table 1. The proportions of aneuploid, CNV, and cytogenetically normal oocytes across the three groups are summarized in Figure 4E. The A-CON group exhibited an elevated aneuploidy rate compared to the Y-CON group. Although UCMSC transplantation was associated with a reduction in this rate, the differences did not reach statistical significance; however, the observed trend suggests a potential beneficial effect of UCMSCs on chromosomal integrity. Further studies are warranted to determine whether UCMSCs treatment directly improves chromosome stability in aged oocytes. Additionally, no significant differences in CNV rates were observed among the groups.

Figure 4 Umbilical cord-derived mesenchymal stem cell transplantation improves the meiotic apparatus and reduces aneuploidy tendency in oocytes from aged mice. A: Representative immunofluorescence images of meiotic spindles (green, α-tubulin) and chromosomes (blue, DAPI) in metaphase II (MII) oocytes. Scale bar = 20 μm; B: Quantitative analysis of the percentage of oocytes with normal spindle morphology. Umbilical cord-derived mesenchymal stem cell transplantation (A-MSC) significantly rescued the age-related decline in normal spindle formation; C: Representative images of chromosome spreads from MII oocytes, showing a normal euploid count (20 chromosomes) and an abnormal aneuploid count (21 chromosomes). Scale bar = 20 μm; D: Quantitative analysis of the percentage of oocytes with a normal chromosome count (20); E: Bar chart showing the percentages of aneuploid oocytes and oocytes with copy number variations across the different groups. Data are presented as mean ± SEM. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. Y-CON: Young control; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation; CNV: Copy number variation.

Single-cell transcriptomics defines the cellular heterogeneity of the mouse ovary

To comprehensively delineate the cellular landscape of the murine ovary under varying conditions, scRNA-seq was performed on ovarian tissues isolated from Y-CON, A-CON, and A-MSC mice. Unsupervised clustering of transcriptomic data from 31839 high-quality cells revealed 29 distinct cell clusters (Figure 5A-C). These clusters were annotated as known GCs, theca cells (TCs), stromal cells (SCs), and immune cells (ICs) based on the expression of established marker genes (Figure 5D and E). The largest cluster was identified as GCs, characterized by the expression of genes involved in follicular hormone secretion, including Fst, Inhba, and Bex4, with Ptgfr serving as a marker for luteal cells. The second largest cluster, SCs, was characterized by the co-expression of genes related to collagen and extracellular matrix proteins, such as Col1a1, Bgn, and Dcn. TCs were identified by markers including Wnt6, implicated in embryonic development, alongside Srd5a1 and Cyp17a1, which are critical for androgen synthesis. Endothelial cells were defined by the expression of canonical vascular genes like Pecam1 and Flt1, in addition to the lymphatic endothelial marker Ccl21a. Epithelial cells expressed the adhesion molecule Epcam and the keratin gene Krt19. The IC cluster was primarily marked by Ptprc, a gene involved in antigen receptor signaling, and was further subdivided into B cells (marked by Cd79a), T cells (Cd3d), neutrophils (S100a9), and phagocytes (C1qa).

Figure 5 Single-cell transcriptomic profiling defines the ovarian cellular landscape. A: UMAP visualization of single cells from mouse ovarian tissues, colored by 29 distinct cell clusters; B: UMAP visualization of single cells from mouse ovarian tissues, colored by different cell types; C: Stacked bar plot showing the proportional composition of major ovarian cell types across the three experimental groups; D: Feature plots showing the expression of marker genes used for cell type annotation; E: Dot plots showing the expression of marker genes used for cell type annotation. GC: Granulosa cell; SC: Stromal cell; TC: Theca cell; EndoC: Endothelial cell; EpiC: Epithelial cell; IC: Immune cell.

Proportional shifts in major ovarian cell populations across groups are quantified in Figure 5C. Compared to the Y-CON group, the A-CON group exhibited a decline in the proportions of GCs (48.26% vs 50.52%) and SCs (18.86% vs 25.18%), along with increases in endothelial cells (10.15% vs 7.95%), epithelial cells (4.53% vs 0.75%), and ICs (14.19% vs 7.91%). UCMSC transplantation mitigated these aging-related changes, as the A-MSC group demonstrated an increase in GCs (56.04%) and SCs (20.66%), coupled with a decrease in endothelial (6.72%), epithelial (1.45%), and IC (12.59%) proportions. This analysis provides a foundational map for investigating the specific effects of aging and UCMSC therapy on each ovarian cell type at the transcriptional level.

UCMSC transplantation reverses aging-associated transcriptomic signatures

To elucidate the transcriptomic changes induced by UCMSC transplantation, DEGs were identified between the A-MSC and A-CON groups (log₂FC ≥ 0.5, P < 0.05). The analysis revealed a substantial number of DEGs across all major ovarian cell types (Figure 6A). In endothelial cells, 18 genes were up-regulated and 44 down-regulated; in epithelial cells, 12 up-regulated and 13 down-regulated; in ICs, 47 up-regulated and 85 down-regulated; and in TCs, 41 up-regulated and 16 down-regulated. Notably, GCs and SCs experienced the most profound transcriptomic alterations, with GCs exhibiting 808 up-regulated and 397 down-regulated genes, while SCs showed 226 up-regulated and 30 down-regulated genes. This pattern indicates that these two somatic cell types underwent the most significant gene expression reprogramming in response to UCMSCs treatment.

Figure 6 Umbilical cord-derived mesenchymal stem cell transplantation induces cell-type-specific transcriptomic reprogramming and upregulates the anti-aging gene Klotho. A: Volcano plots showing the numbers of upregulated and downregulated differentially expressed genes (DEGs) in major ovarian cell types between the umbilical cord-derived mesenchymal stem cell (UCMSC) transplantation and aged control groups. Granulosa cells (GCs) and stromal cells (SCs) exhibited the most pronounced changes; B: Functional enrichment analysis of upregulated DEGs; C: Functional enrichment analysis of downregulated DEGs; D: Heatmap showing the differential expression levels of key anti-aging genes from the GenAge database across cell types. Klotho was significantly upregulated in granulosa cells; E: Senescence scores calculated for each cell type based on the GenAge database gene set. UCMSC transplantation significantly reduced senescence scores in granulosa cells and epithelial cells; F: Klotho expression by cell type and sample; G: Proportion of Klotho-positive cells by cell type and sample; H: Representative immunohistochemical images of KLOTHO protein in ovarian sections; I: Quantification of KLOTHO protein expression in ovarian sections. UCMSC transplantation markedly enhanced KLOTHO expression. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001, ^dP < 0.0001. GC: Granulosa cell; SC: Stromal cell; TC: Theca cell; EndoC: Endothelial cell; EpiC: Epithelial cell; IC: Immune cell; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation.

Functional enrichment analysis on DEGs from various cell types was conducted to elucidate the biological processes affected by UCMSC transplantation. This bioinformatic approach identifies statistically overrepresented functional categories within a gene set, revealing potential regulatory mechanisms. As presented in Figure 6B and C, the “Ribosome” pathway was significantly enriched among up-regulated genes in all cell types except immune and epithelial cells. Additionally, “OXPHOS” was up-regulated across all lineages except in immune and endothelial cells. In GCs, up-regulated DEGs were linked to processes such as protein synthesis, reproductive structure development, and lipid transport, while down-regulated DEGs were associated with aging-related pathways, including p53, mechanistic target of rapamycin complex 1 (mTORC1), and Wnt signaling, as well as apoptosis and autophagy. In contrast, ICs exhibited an opposing trend: Up-regulated genes were enriched for immune regulation and cytokine responses, while down-regulated genes were involved in ribosome biogenesis and IC activation. This transcriptomic reprogramming suggests that UCMSC transplantation enhances metabolic activity and anabolic functions in functionally critical cells like GCs, while concurrently suppressing pro-inflammatory states in the aging ovary.

To elucidate the mechanism by which UCMSCs ameliorate ovarian aging, this study focused on aging-associated genes derived from the intersection with the GenAge database. Transcriptomic analysis identified GCs and SCs as the primary targets of this rejuvenation, with the anti-aging gene Klotho showing the most substantial upregulation (Figure 6D). Consistently, a senescence scoring analysis revealed a significant attenuation of pro-aging transcriptional programs in granulosa and epithelial cells following MSC transplantation (Figure 6E). Klotho was prioritized for experimental validation, with results confirming that UCMSC transplantation robustly enhanced the expression intensity of membrane-localized KLOTHO protein in GCs and oocytes (Figure 6F-I), designating it as a pivotal mediator of MSC-induced ovarian rejuvenation.

Activity scores of key signaling pathways - OXPHOS, p53, transforming growth factor (TGF)-β, and Wnt - are summarized in Figure 7A-D. Following UCMSC transplantation, OXPHOS activity was significantly elevated across all cell types, except for ICs. Concurrently, the p53 pathway was suppressed in endothelial, granulosa, stromal, and TCs; TGF-β signaling was downregulated in epithelial, granulosa, and TCs; and Wnt signaling decreased in GCs but increased in ICs. The upregulation of OXPHOS was validated at the transcriptional level through qPCR analysis of associated genes (Figure 7E-H). Ct values for HPRT1 indicated no significant differences among groups, suggesting stable expression under the experimental conditions. The expression of Cox5a, Cox7b, and Atp5e significantly increased in the A-MSC group compared to the A-CON group, while Ndufb2 exhibited a non-significant upward trend. Additionally, western blot analysis corroborated this enhancement at the protein level, demonstrating significantly elevated expression of COXIV and ATP5A in the A-MSC group (Figure 7I-K).

Figure 7 Umbilical cord-derived mesenchymal stem cell transplantation enhances oxidative phosphorylation. A: Oxidative phosphorylation activity was significantly increased in most somatic cell types following umbilical cord-derived mesenchymal stem cell (UCMSC) transplantation; B: Activity of the senescence-associated p53 pathway; C: Activity of the senescence-associated transforming growth factor-β pathway; D: Activity of the senescence-associated Wnt pathway; E: The mRNA expression levels of Cox5a; F: The mRNA expression levels of Cox7b; G: The mRNA expression levels of Atp5e; H: The mRNA expression levels of Ndufb2; I: Validation of enhanced oxidative phosphorylation at the protein level by western blot analysis; J: Representative western blot images of COXIV; K: Representative western blot images of ATP5A; L: Mitochondrial pathway activity scores derived from the MitoCarta3.0 database, showing coordinated upregulation in stromal, theca, and granulosa cells following UCMSC transplantation. Data are presented as mean ± SEM. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001, ^dP < 0.0001. GC: Granulosa cell; SC: Stromal cell; TC: Theca cell; EndoC: Endothelial cell; EpiC: Epithelial cell; IC: Immune cell; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation.

Given the significant enrichment of UCMSCs-responsive DEGs in fundamental mitochondrial processes, the specific cell types undergoing metabolic reprogramming were pinpointed. Utilizing the MitoCarta3.0 database, this study assessed mitochondrial pathway activity and found that stromal, theca, and GCs were the most responsive, exhibiting a coordinated upregulation of mitochondrial gene sets after transplantation (Figure 7L). This cell-type-specific pattern suggests that UCMSCs enhance ovarian function by selectively boosting the energy metabolism of critical somatic cells, thereby creating a more bioenergetically favorable microenvironment.

UCMSC transplantation remodels the intercellular communication network

To decipher the impact of UCMSC transplantation on intercellular crosstalk, CellChat was employed to infer cell-cell communication by integrating scRNA-seq data with a curated ligand-receptor interaction database. Comparative analysis between the A-MSC and A-CON groups revealed a global enhancement in intercellular communication, evidenced by an increase in both the number and strength of interactions among ovarian cell populations (Figure 8A). Analysis of the outgoing communication patterns from each cell subpopulation identified SCs as the most responsive niche, displaying the most substantial alterations in signaling (Figure 8B). Focusing on the aging-associated TGF-β pathway, the chord diagram illustrated a notable rewiring of signaling networks in the A-MSC group (Figure 8C). Specifically, ICs received significantly reduced signals from both endothelial and other ICs, accompanied by a decrease in the corresponding ligand-receptor pair interactions (Figure 8D). Additionally, signaling mediated by multiple chemokine ligand-receptor pairs and components of the collagen pathway was also attenuated in the A-MSC group (Figure 8E and F).

Figure 8 Umbilical cord-derived mesenchymal stem cell transplantation remodels the intercellular communication network in the aged ovary. A: The total number of inferred interactions (left) and the total interaction strength (right) among all cell populations are increased in the umbilical cord-derived mesenchymal stem cell transplantation (A-MSC) group compared to the aged control (A-CON) group; B: Analysis of interactions identifies stromal cells exhibiting the most significant alterations in signaling pathways following A-MSC; C: Chord diagrams of transforming growth factor (TGF)-β signaling pathway network in A-CON group (left) and A-MSC group (right); D: Alterations in ligand-receptor pairs of the TGF-β signaling pathway; E: Alterations in ligand-receptor pairs of the chemokine pathway; F: Alterations in ligand-receptor pairs of the collagen pathways. GC: Granulosa cell; SC: Stromal cell; TC: Theca cell; EndoC: Endothelial cell; EpiC: Epithelial cell; IC: Immune cell; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation; TGF: Transforming growth factor.

UCMSC transplantation influences GC fate determined by pseudotime analysis

To delineate the developmental trajectory and fate decisions of GCs, pseudotemporal ordering was performed using Monocle (Figure 9). The trajectory analysis (Figure 9A-C) revealed that GCs originate from a common starting point (state 1) and undergo bifurcation, resulting in two distinct fates: State 2 and state 3. The distribution of cells across these states varied markedly among groups (Figure 9D). In the Y-CON group, the majority of cells (85.31%) remained in state 1, with 13.59% in state 3 and only 1.10% in state 2. In contrast, the A-CON group exhibited a disrupted distribution, with only 46.65% of cells in state 1, 50.91% in state 3, and 2.44% in state 2. UCMSC transplantation significantly altered this distribution, with the A-MSC group showing 49.33% of cells in state 1, 23.88% in state 2, and 26.79% in state 3, indicating a reprogramming of GC fate.

Figure 9 Pseudotime trajectory analysis reveals umbilical cord-derived mesenchymal stem cells reprogram granulosa cell fate in the aged ovary. A: Pseudotime trajectory analysis of granulosa cells; B: The trajectory reveals three distinct cell states (state 1, 2, 3), with a bifurcation from state 1 leading to two fates (state 2 and state 3); C: Cells are colored by pseudotime and experimental group; D: Bar chart showing the distribution of granulosa cells across the three states in each experimental group. Umbilical cord-derived mesenchymal stem cells transplantation significantly altered the fate distribution, increasing the proportion of cells in state 2 and decreasing the proportion in state 3 compared to the aged control group; E: Heatmap of gene expression dynamics along the two cell fates during the pseudotime trajectory and Gene Ontology enrichment analysis of the biological processes associated with the differentially expressed genes defining cell fate 1 and cell fate 2; F: Expression patterns of the top 10 differentially expressed genes between the two cell fates. Y-CON: Young control; A-CON: Aged control; A-MSC: Umbilical cord-derived mesenchymal stem cell transplantation; GO: Gene Ontology; TORC1: Target of rapamycin complex 1; TOR: Target of rapamycin.

Gene expression patterns defining the two cell fates at the branch point were next analyzed (Figure 9E). Genes that were significantly differentially expressed (q-value < 10^-8) between cell fate 1 (state 2) and cell fate 2 (state 3) underwent Gene Ontology enrichment analysis. Results indicated that genes highly expressed in the terminal phase of cell fate 1 (state 2) were primarily associated with inflammatory responses, chemotaxis, endothelial cell proliferation, and angiogenesis. Conversely, genes upregulated during the mid-to-late phase of cell fate 2 (state 3) were enriched for processes such as autophagy, vesicle formation, immune responses, and phosphorylation. The top 10 DEGs are displayed in Figure 9F.

External validation reveals aging-associated mitochondrial transcriptional remodeling

To independently assess age-related transcriptional changes in ovarian GCs, the publicly available scRNA-seq dataset GSE232309 was analyzed, encompassing 2542 GCs from young mice and 1222 GCs from aged mice. The UMAP plot presented in Figure 10A indicates significant overlap between young and aged GCs, suggesting that cellular transcriptional states are broadly conserved during physiological aging. Next, Klotho expression was evaluated at the single-cell level, revealing no significant differences between young and aged GCs, indicating the absence of marked transcriptional changes in Klotho during physiological ovarian aging (Figure 10B). To investigate mitochondrial metabolic activity, a gene signature score related to OXPHOS was calculated based on representative OXPHOS genes. Aged GCs exhibited significantly higher transcriptional activity related to OXPHOS compared to the Y-CON group (P < 0.05, Figure 10C), suggesting significant remodeling of mitochondrial transcription during physiological ovarian aging. These findings imply that while mitochondrial metabolic reprogramming occurs during ovarian aging, Klotho expression remains relatively stable. This independent dataset thus serves as an external reference framework for interpreting the enhanced mitochondrial activity and Klotho expression induced by UCMSCs observed in this study.

Figure 10  External validation of ovarian aging-associated transcriptional alterations using GSE232309. A: UMAP of granulosa cells from young and aged mice showing substantial overlap between age groups; B: Klotho expression in young and aged granulosa cells; C: Oxidative phosphorylation-related gene signature score calculated from representative oxidative phosphorylation genes. OXPHOS: Oxidative phosphorylation.

DISCUSSION

Ovarian aging is primarily characterized by the depletion of the follicular pool and a decline in oocyte quality. In recent years, the application of MSCs has garnered increasing attention as a therapeutic strategy to address ovarian dysfunction. MSC-based therapy is considered a highly promising approach for managing ovarian hypofunction and has demonstrated beneficial effects in various animal models of ovarian decline[26]. Common routes of MSC administration in these models include tail vein injection, intra-ovarian transplantation, and intraperitoneal injection. Several studies have evaluated the efficacy of these different delivery methods[27-30]. Generally, tail vein and intra-ovarian transplantation yield better outcomes than intraperitoneal injection. Tail vein injection is less invasive, allows for repeated administration, offers greater flexibility in treatment planning, and may promote multi-organ repair; however, the number and timing of stem cells reaching the target organ can be suboptimal. In contrast, intra-ovarian transplantation ensures higher local cell retention but is more invasive and carries a greater risk of ovarian hemorrhage. Considering the systemic nature of aging and the clinical preference for minimally invasive procedures, this study adopted the tail vein injection approach. A notable limitation is that this study did not track the dynamic process of cellular homing in vivo and thus failed to quantify the exact number and duration of UCMSC colonization in ovarian tissue. Nevertheless, the therapeutic effects of stem cell-mediated treatments do not necessarily depend on the formation of persistent and dense local colonies.

Accumulating evidence suggests that systemically administered stem cells primarily exert their effects through the activation of paracrine signaling and immunomodulatory actions, rather than through direct cell replacement[31,32]. Through tail vein injection, although a small fraction of UCMSCs may be drawn to the aging ovary by injury-associated chemotactic gradients, the majority of cells can exert significant systemic effects through the secretion of a wide range of bioactive factors[33,34]. Concurrently, our scRNA-seq data support this systemic paracrine effect. In the absence of dense, long-term UCMSC implantation, key survival pathways in GCs were activated, and mitochondrial function was restored, indicating that systemic administration primarily reactivates the aged ovarian microenvironment through indirect, paracrine-driven molecular cascades. To address this limitation, future studies will employ long-term in vivo imaging and lineage tracing techniques to further elucidate the spatial and temporal distribution of transplanted cells.

In this study, UCMSC transplantation exhibited comprehensive efficacy in mitigating ovarian aging at the functional level. Key outcomes included the restoration of regular estrous cyclicity, an increase in both total and antral follicle counts, improved serum hormone profiles (reduced FSH, elevated AMH and E2), and enhanced oocyte yield. Single-cell transcriptomics revealed that the improvements in ovarian reserve and function were not isolated phenomena; rather, they were rooted in altered gene expression within key cell types, particularly GCs and SCs. Differential gene expression and enrichment analyses indicated that these cells are central targets of UCMSCs action. In these populations, aging-related pathways such as p53 and mTORC1 were downregulated, while energy and anabolic processes, including OXPHOS and ribosome biogenesis, were upregulated. Concurrently, a majority of mitochondrial-related pathways were enhanced in the A-MSC group. Mitochondria play a pivotal role in various cellular functions: In addition to generating the majority of cellular ATP[35], they regulate Ca²⁺ levels, release cytochrome c and proteins[36], and produce ROS, NAD+/NADH, NADP+/NADPH, and various metabolites[37]. The integrity and activity of mitochondria are fundamental to signaling, intracellular redox homeostasis, and apoptosis regulation, thereby positioning them as key mediators of cellular and organismal aging[38]. Alterations in mitochondrial function within GCs significantly affect ovarian aging and fertility. GCs and oocytes engage in close interactions through bidirectional communication and gap junctions: Oocytes modulate GC activity via specific factors, while GCs are essential for constructing the follicular microenvironment necessary for oocyte maturation[39]. Under physiological conditions, GCs primarily rely on OXPHOS for energy conversion from glucose, which is critical for regulating mitochondrial function and maintaining OXPHOS capacity[40]. Proteins encoded by mtDNA that are integral to the OXPHOS system directly impact mitochondrial energy metabolism and are essential for oocyte development and maturation[41].

Despite the increase in OXPHOS pathway activity following MSC treatment, ROS levels decreased, presenting a seemingly paradoxical outcome that uncovers the intricate relationship between mitochondrial dysfunction and oxidative stress. During ovarian aging, while OXPHOS activity declines, electron flow within the ETC persists. Impaired complex function increases the likelihood of electron leakage, resulting in the formation of ROS when electrons react with oxygen. Additionally, aged or damaged mitochondria, though inefficient in ATP production, can still generate substantial amounts of ROS. Thus, ovarian aging is characterized by both reduced OXPHOS activity and elevated ROS levels. Following MSC treatment, OXPHOS activity improved, ETC function was restored, and ROS production decreased. Alberico and Woods[42] reported mtDNA mutations in GCs from older women and in animal models of ovarian aging, indicating that ultrastructural changes in aged GCs may correlate with ROS accumulation in follicular fluid. As age advances, the levels of the ROS scavenger superoxide dismutase 1 decline in GCs, hampering mtDNA repair and steroid hormone production. The activation of OXPHOS accounts for the observed increases in mitochondrial membrane potential and mtDNA copy number, along with the reduction in ROS levels in oocytes and somatic cells. This enhancement in energy metabolism facilitates the ATP required for folliculogenesis, steroidogenesis, and oocyte maturation, thereby supporting the phenotypic recovery of follicle numbers, hormone levels, and oocyte yield.

Notably, although UCMSC treatment significantly improved spindle morphology and various indicators of oocyte quality, no statistically significant reduction in oocyte aneuploidy rates or chromosomal instability burden was observed. This finding suggests that restoring meiotic spindle organization does not directly translate into improved accuracy in chromosome segregation. The assembly and morphology of the spindle apparatus constitute only one aspect of the complex mechanisms governing chromosome stability during meiosis. Despite enhancements in cellular metabolism and cytoskeletal organization, age-related defects involving cadherin maintenance, centromere attachment, DNA repair capacity, and chromosome structure may persist. Hence, structural restoration of spindle integrity may occur prior to a measurable correction of chromosomal abnormalities. Additionally, the lack of statistical significance could stem from the limited number of oocytes available for chromosomal analysis, thus reducing statistical power to detect modest therapeutic effects. Consequently, while findings suggest that UCMSCs positively impact oocyte quality and spindle organization, the potential for these cells to directly restore chromosomal stability in aged oocytes necessitates further research. Subsequent studies with larger sample sizes and more comprehensive chromosomal analyses will be crucial in elucidating the effects of UCMSCs on chromosomal integrity during meiosis.

The Klotho gene encodes an anti-aging protein with diverse biological roles, including the regulation of calcium and vitamin D homeostasis, prevention of chronic fibrosis, antioxidant and anti-inflammatory actions, protection against cardiovascular and neurodegenerative diseases, and modulation of apoptosis and senescence[43]. Klotho-knockout mice exhibit a shortened lifespan, growth arrest, and aging-related phenotypes such as vascular calcification, cognitive decline, and multi-organ atrophy and injury[44]. Klotho exerts its anti-aging effects by suppressing four major aging-associated pathways: TGF-β, insulin-like growth factor-1, Wnt, and nuclear factor kappa B (NF-κB)[45]. Recent studies have highlighted Klotho’s significance in the reproductive system. Liu et al[46] reported that Klotho is involved in cyclophosphamide (CTX)-induced ovarian failure in mice, where CTX-induced microRNAs target Klotho mRNA, leading to reduced KLOTHO protein expression. This reduction impairs autophagy of damaged organelles, increases ROS levels, causes cellular damage, and ultimately results in ovarian failure characterized by elevated FSH and decreased E2[46]. Similarly, Xu et al[47] observed significantly lower KLOTHO expression in GCs and serum from women with diminished ovarian reserve, with levels further declining with age. In the present study, Klotho expression was increased in GCs, while the TGF-β and Wnt pathways were downregulated, aligning with previous findings. However, the mechanisms through which UCMSCs upregulate Klotho and the specific pathways via which Klotho exerts its effects require further investigation.

To elucidate these findings, an independent public scRNA-seq dataset of mouse ovarian GCs undergoing physiological aging was analyzed. The results indicated that, despite significant transcriptional changes related to OXPHOS in aging GCs, Klotho expression remained largely unchanged. These observations suggest that mitochondrial metabolic remodeling may represent a fundamental aspect of physiological ovarian aging, while Klotho is not necessarily a direct marker of age-related decline[48]. Instead, Klotho may function as a treatment-responsive anti-aging mediator activated during ovarian regeneration. Thus, the significant upregulation of Klotho observed following UCMSC transplantation may reflect a regenerative response distinct from the transcriptional changes associated with normal aging. Several studies have indicated that increased expression of mitochondrial respiratory genes in aging tissues may signify a compensatory response to declining mitochondrial efficiency rather than an improvement in bioenergetic performance[49,50]. In this context, the elevated transcriptional activity linked to OXPHOS observed in aged GCs in external datasets may reflect an adaptive effort by cells to maintain energy balance amid age-related mitochondrial stress. In contrast, UCMSC transplantation not only enhanced the expression of genes associated with OXPHOS but also improved mitochondrial membrane potential, reduced ROS accumulation, and restored mtDNA copy number, indicating a more comprehensive recovery of mitochondrial function.

Aging at both organ and organismal levels is frequently accompanied by inflammatory activation. Inflammation-related molecular patterns further drive cellular senescence, creating a vicious cycle that accelerates aging[51], leading to the proposal of inflammation as a hallmark of aging[52]. In particular, inflammatory activation is a critical feature of ovarian aging, contributing to oxidative stress, apoptosis, and accelerated cellular senescence. Ovarian aging is characterized by a chronically inflamed and dysfunctional microenvironment, evidenced by increased IC infiltration and pathway activation. In this study, transplantation of UCMSCs reduced the proportion of ICs, downregulated the inflammatory TGF-β pathway in granulosa, theca, and epithelial cells, and suppressed pathways associated with biosynthesis and IC activation in ICs. By remodeling the cell-cell communication network, UCMSCs significantly attenuated TGF-β-mediated activation signals received by ICs. Consistent with these findings, Deng et al[53] reported that human UCMSC transplantation elevated anti-inflammatory cytokines [interleukin-10 (IL-10), tumor necrosis factor-stimulated gene-6, vascular endothelial growth factor] while reducing pro-inflammatory cytokines (IL-6, IL-1β), thereby suppressing inflammation and improving the GC microenvironment. Another study demonstrated that exosomes derived from human UCMSCs inhibited the NF-κB pathway, enhanced IL-10 expression, and suppressed tumor necrosis factor-α and interferon-γ, alleviating GC inflammation[54]. However, our cell communication analysis revealed an upregulation of ligand-receptor pairs for several chemokines following UCMSC transplantation, suggesting that UCMSC therapy does not merely suppress immunity but dynamically modulates the immune response. The upregulation of CCL chemokines may recruit ICs and induce localized inflammation, as observed with dendritic cells, naïve T cells, and natural killer T cells in this study. This localized inflammation could facilitate the clearance of senescent cells and pave the way for tissue repair and angiogenesis. Furthermore, pseudotime trajectory analysis indicated that UCMSCs treatment altered the developmental trajectory of a subset of GCs, leading to a new state enriched in genes linked to inflammation, chemotaxis, endothelial proliferation, and angiogenesis, ultimately enhancing GC function and ovarian activity.

This study has several limitations. First, the scRNA-seq analysis relied on pooled ovarian samples from each experimental group rather than independent biological replicates. While this pooling approach minimizes technical variability and facilitates a comprehensive characterization of ovarian cell heterogeneity, the absence of biological replicates may hinder the assessment of inter-individual differences and reduce the reliability of cell proportion and differential expression analyses. Future studies that incorporate independent biological replicates and external validation datasets would enhance the reproducibility of the observed transcriptional changes. Second, UCMSCs were administered via tail vein injection; however, no direct in vivo tracking of the transplanted cells was performed. As a result, the extent of ovarian homing, cell retention within the ovarian tissue, and persistence remain unclear. Previous studies provide extensive evidence suggesting that the therapeutic effects of MSCs primarily arise from paracrine signaling, extracellular vesicle secretion, and systemic immunomodulatory mechanisms. Future research should employ cell labeling and lineage tracing methods to clarify the relative contributions of direct ovarian implantation and indirect systemic regulation. Third, the study was conducted exclusively in a mouse model of ovarian aging. Although the identified molecular pathways show broad conservation across species, the translational relevance of these findings to human ovarian aging requires validation. Additional studies utilizing human primordial GCs, ovarian tissue samples, and clinical cohorts are essential to confirm the applicability of these mechanisms in humans. Finally, despite the significant improvements in mitochondrial function, oxidative stress status, and spindle morphology following UCMSC transplantation, no statistically significant reduction in oocyte aneuploidy or CNV burden was observed. These results imply that enhancements in cellular and structural indicators of oocyte quality may not directly translate into measurable restoration of chromosomal stability. Consequently, larger-scale cytogenetic studies are warranted to determine whether UCMSCs directly influence the accuracy of chromosomal segregation during oocyte aging.

CONCLUSION

In conclusion, this study demonstrates that UCMSC transplantation ameliorates ovarian aging by restoring ovarian function and improving oocyte quality through coordinated mechanisms at single-cell resolution. Significant enhancement of mitochondrial OXPHOS and upregulation of the anti-aging gene Klotho in GCs were observed. Furthermore, UCMSCs remodeled intercellular communication networks and redirected GC fate trajectories. These multi-level effects collectively reverse key aging phenotypes, providing mechanistic insights into UCMSC-based regenerative therapy for age-related ovarian decline.