miR-378a-3p participates in the decline of ovarian reserve function by targeting ZFP36L2 to regulate granulosa cell mitochondrial function and apoptosis.

The research protocol was approved by the Ethics Committee of Reproductive Medicine at the Affiliated Hospital of Shandong University of Traditional Chinese Medicine (TCM) and registered with the Chinese Clinical Trial Registry (Registration No. ChiCTR2400079975;  https://www.chictr.org.cn/ ); all participants had provided informed consent. Selecting patients undergoing ART-assisted conception who were recruited from the Department of Reproductive Medicine at the Affiliated Hospital of Shandong University of TCM between May and October 2025. The criteria for inclusion in the DOR group were based on the Bologna criteria, which required a diagnosis supported by at least two of the following indicators: an AMH value < 1.1 ng/mL; an AFC of less than 5–7 on both sides; and baseline FSH levels measured in two consecutive menstrual cycles ≥ 10 U/L. Another 15 NOR individuals underwent intracytoplasmic sperm injection (ICSI) attributed to male factors comprising the control group. Participants had to satisfy two criteria: (1) ages 20–45; (2) infertility lasting more than one year. Participants with endometriosis, adenomyosis, congenital gonadal dysgenesis, severe metabolic disorders, or any other systemic diseases that might affect fertility or the effects of ovulation induction were excluded. All participants underwent the GnRH antagonist regimen, with transvaginal ultrasound evaluations and baseline serum reproductive hormone measurements carried out during menstrual cycle days 2–3. Based on the hormone levels and ultrasound findings, combined with their age, early ovarian response and body mass index (BMI), appropriate doses of recombinant human follicle-stimulating hormone (r-FSH, Puregon®, Merck Sharp & Dohme, New Jersey, USA) were administered to stimulate follicular development. Once a dominant follicle developed with an average diameter of 12 mm or more, GnRH antagonist (GnRH-ant, Orgalutran®, Merck Sharp & Dohme, New Jersey, USA) was administered until the trigger day. Throughout the treatment period, medication adjustments were made in a timely manner according to the patient's sex hormone levels and follicle development status. When at least 1–2 dominant follicles reached a diameter of 18–20 mm, or there was a batch (usually more than 3) of follicles with diameters over 16–18 mm, and the serum estradiol(E 2 ) level matched the count of mature follicles, while the progesterone(P) level had not risen prematurely, recombinant human chorionic gonadotropin (r-hCG, Ovidrel®, Merck Serono, Geneva, Switzerland) was administered to stimulate oocyte maturation, which was the trigger. A transvaginal ovarian puncture was performed under B-ultrasound guidance to retrieve oocytes about 34 to 36 h after triggering. Follicular fluid from non-bloody follicles exceeding 17 mm in diameter was collected by trained personnel for storage in the specimen bank. All follicular fluid samples were centrifuged at 3000 rpm for 10 min, and the supernatant was carefully aspirated. The sediment was transferred into a glass tube containing 4 mL of phosphate-buffered saline (PBS) to prepare a cell suspension. This cell suspension was then gently added dropwise into a tube containing 2.5 mL of human lymphocyte separation medium, with careful handling to maintain a clear interface. The tube was centrifuged at 2000 rpm for an additional 10 min. The floating cell layer at the interface was collected into a 1.5 mL centrifuge tube. washed with PBS, and resuspended by gentle pipetting. The suspension was centrifuged at 2500 rpm for an additional 10 min. Upon completion, the supernatant was removed by aspiration, and the GC pellet was stored at −80 °C for subsequent analysis. The TargetScan ( http://www.targetscan.org ), miRDB ( http://www.mirdb.org/miRDB/ ), and Starbase ( http://rnasysu.com/encori/ ) databases were used to predict and analyze the target genes of miR-378a-3p. Subsequently, the overlapping target genes were extracted utilizing a Venn diagram. The human embryonic kidney (HEK) 293 cell line and its derivatives have become an ideal tool for preliminary validation of transcriptional regulatory mechanisms due to their high transfection efficiency and strong exogenous gene expression capability [ 16 ]. HEK-293 T cells were cultured in DMEM medium (Gibco, California, USA) with 10% fetal bovine serum (FBS). Then, HEK-293 T cells in the logarithmic growth phase were seeded into 6-well culture plates at 1 × 10 5 cells/well. On the day of transfection, the cells were approximately 50–60% confluent. Subsequently, each transfection condition—comprising wild-type or mutant reporter plasmids co-transfected with miR-378a-3p mimics or negative controls (NCs)—was separately diluted in OPTI-MEM low-serum medium and combined with 15 µL Lipofectamine 2000 to form transfection complexes. Table 1 presents the different plasmid combinations. These complexes were incubated at 20–25 °C, and then added to the respective cell culture wells. After 4–6 h of transfection, the medium was replaced with fresh DMEM containing 10% FBS, and the cells were cultured for an additional 24 h in an incubator at 37 °C with 5% CO 2 . They were then washed twice with PBS by gentle pipetting. Each well received 100 µl of Passive Lysis Buffer (PLB), and the plates were mildly agitated for 15 min at 20–25 °C to obtain the cell lysates. A volume of 20 µL from these lysates was transferred to luminescence plates, where background readings were taken for two seconds using a GloMax luminometer. Subsequently, 100 µL LAR II substrate followed by Stop & Glo® reagent was added sequentially to assess the luminescent activities of luciferases. Finally, the regulatory effect of miR-378a-3p on ZFP36L2 was assessed by calculating the ratio of the luminescence intensities. Table 1 Plasmid combinations Group Plasmid combinations 1 miRNA negative control + ZFP36L2 3 ‘UTR-WT 2 miR-378a-3p mimics + ZFP36L2 3 ‘UTR-WT 3 miRNA negative control + ZFP36L2 3 ‘UTR-MT 4 miR-378a-3p mimics + ZFP36L2 3 ‘UTR-MT Plasmid combinations After digestion with trypsin, KGN cells (Pricella, CL-0603, Wuhan, China) were centrifuged to collect the pellets, resuspended, and counted. The cells were seeded into 6-well plates at a density of 3 × 10 5 cells per well in 2 mL of DMEM/F12 basal medium supplemented with 10% FBS and 1% penicillin–streptomycin. Cells were cultured in a 37 °C incubator with 5% CO₂until they reached approximately 80% confluence for transfection. For transfection, the culture medium was removed, the cells were washed once with PBS, and 1.5 mL of Opti-MEM was added to each well. Transfection mixtures were then prepared as follows: for the miRNA NC group, 250 µL of Opti-MEM was mixed with 5 µL of miRNA NC; for the miR-378a-3p mimics group, 250 µL of Opti-MEM was mixed with 5 µL of miR-378a-3p mimics; for the miR-378a-3p inhibitor group, 250 µL of Opti-MEM was mixed with 5 µL of miR-378a-3p inhibitor. Next, 750 µL of Opti-MEM was mixed with 15 µL of Lipo3000, incubated at room temperature for 5 min, divided into three equal portions, and each portion was combined with one of the three miRNA mixtures described above. Each resulting mixture was added to the corresponding well of the 6-well plate. After 8 h, the medium was changed. Cell function experiments were performed 48 h post-transfection. Cell function status was evaluated using a flow cytometry system (BD, C6 Plus, New Jersey, USA). Gates were set in the scatter plots of forward scattered light and side-scattered light. Debris and adherent cells were excluded, the target single cell population was delineated, and thresholds were established based on the negative or unstained controls for each fluorescence channel to distinguish positive from negative signals. Mitochondrial membrane potential (MMP) was assessed with the JC-1 probe (Beyotime, C2006, Shanghai, China), and the cells were digested by pancreatic enzyme and incubated with JC-1 working solution at 37 °C for 20 min. After centrifugation at 600 × g and 4 °C for 3 min, the cells were pelleted, the supernatant was discarded, were washed with pre-cooled JC-1 buffer and resuspended, and analyzed by red-green fluorescence ratio. Apoptosis was assessed using an Annexin V-FITC/PI Kit (Vazyme, A211-02, Jiangsu, China). Briefly, cells were collected by centrifugation at 1000 × g for 5 min and washed twice with ice-cold PBS.Subsequently, cells were resuspended in 100 μL of 1 × Bind buffer. Then, 5 µL of FITC Annexin V and 5 μL of PI were added, and the mixture was incubated at room temperature in the dark for 15 min. Finally, 300 µL of Bind buffer was added, and cell populations were distinguished using Annexin V-FITC/PI two-parameter scatter plots. ROS levels were detected by DCFH-DA probe (Beyotime, S0033S, Shanghai, China). After collection, cells were incubated with 10 µM DCFH-DA at 37 °C for 20 min in the dark. They were then washed twice with serum-free medium by centrifugation at 1000 × g for 5 min, resuspended in 200 µL of serum-free medium, and analyzed by flow cytometry within 1 h. Primers for miR-378a-3p, ZFP36L2, U6, GAPDH, and mtDNA, as detailed in Table  2 , were synthesized by Wuhan SAIWELL Biotechnology Co., Ltd. According to the manufacturer's instructions, total RNA was extracted from KGN cells transfected with miR-378a-3p mimic, inhibitor, or NC, as well as from clinical GC tissue samples, using RNA Isolater Total RNA Extraction Reagent (Vazyme, R401-01, Jiangsu, China). Reverse transcription was conducted utilizing HiScript II Q RT SuperMix for qPCR (Vazyme, R223-01, Jiangsu, China), and cDNA for miRNA detection was generated using the miRNA 1 st Strand cDNA Synthesis Kit (Vazyme, MR101, Jiangsu, China). For mtDNA analysis, mitochondria were first isolated from parallel cell samples with a Mitochondria Isolation and Protein Extraction Kit (Proteintech, PK10016, Wuhan, China), after which mtDNA was purified using silica membrane columns. Quantitative PCR reactions were executed using ChamQ SYBR qPCR Master Mix (Vazyme, Q311-02, Jiangsu, China) on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, California, USA). The cycling conditions were: an initial pre-denaturation step at 95 °C for 30 s, followed by 40 cycles consisting of 10 s at 95 °C and 30 s at 60 °C. Subsequently, a melting curve analysis was conducted following the completion of the cycles. For miR-378a-3p and ZFP36L2, expression levels were normalized to U6 and GAPDH, respectively. Regarding mtDNA copy number, mitochondrial genes were amplified with primers D41 and D56, and values were normalized against the nuclear gene GAPDH. All relative expression levels and mtDNA copy numbers were calculated using the 2 ^−ΔΔC t method. Each experiment was performed in triplicate independently. Table 2 Primers for quantitative real-time polymerase chain reaction Gene Primer Primer sequence (5’ → 3’) miR-378a-3p Primer F ACACTCCAGCTGGGACTGGACTTGGAGTCA Primer R TGGTGTCGTGGAGTCG U6 Primer F CTCGCTTCGGCAGCACA Primer R AACGCTTCACGAATTTGCGT ZFP36L2 Primer F ATCAACTCCACGCGCTACAA Primer R GGCAGAAGCCGATGGTATGA GAPDH Primer F GGAAGCTTGTCATCAATGGAAATC Primer R TGATGACCCTTTTGGCTCCC mtDNA Primer F CGA AAGGACAAGAGAAATAAGG(D41) Primer R CTGTAAAGTTTTAAGTTTTATGCG(D56) Primers for quantitative real-time polymerase chain reaction Total protein from GCs and KGN cells was extracted with RIPA buffer. Based on the molecular weight of the target proteins ZFP36L2 (55–100 kDa) and GAPDH (37 kDa), 15 mL of 12% resolving gel was prepared. After the resolving gel had solidified, 4.5 mL of 5% stacking gel was prepared. Following denaturation, the samples were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked in TBST with 5% skim milk for 1 h at a room temperature with gentle shaking. Then incubated overnight at 4 °C with primary antibodies diluted in TBST containing 0.1% Tween-20. The primary antibodies utilized included ZFP36L2 (1:1000, 30,640–1-AP, Proteintech, Wuhan, China) and the internal reference GAPDH (1:5000, 60,004–1-Ig, Proteintech, Wuhan, China). The following day, the membrane was subjected to incubation with horseradish peroxidase-labeled goat anti-rabbit IgG (H + L) secondary antibody (1:5000, 7074, CST, Massachusetts, USA) at 37 °C for 2 h. Finally, the membranes were developed using ECL chemiluminescent substrate (Vazyme, E412-02, Jiangsu, China), and grayscale quantification of the target protein bands was carried out with ImageJ software (National Institutes of Health, USA). The relative expression level of ZFP36L2 was determined by calculating the ratio of its grayscale value to that of GAPDH. Sample size estimation was performed with PASS 15.0 (NCSS, LLC., USA). Based on the preliminary experimental results of KGN cells, the apoptosis rate for the mimics-NC group was (11.11 ± 0.98)%, while that for the mimics-miR-378a-3p group was (7.04 ± 0.41)%, resulting in an intergroup difference of 4.07%. With a power (1-β) set at 0.90, a two-sided α level of 0.05, and a standard deviation (σ) of 1.0, it was assessed that a minimum of nine cases were necessary for each group. Taking into account a projected dropout rate of 10%, 15 cases were ultimately included in each group. Statistical analyses were conducted using SPSS Statistics 29.0 (IBM, USA). Continuous variables were first assessed for normality using the Shapiro–Wilk test. For normally distributed variables, descriptive statistics were presented as mean ± standard deviation, and groups were compared using independent sample t-tests. For non-normally distributed variables, data were expressed as median with interquartile range, and group comparisons were performed using the Mann–Whitney U test. Associations between variables were examined by linear regression, and P  < 0.05 was considered statistically significant.

A total of 30 human follicular fluid specimens were obtained, including 15 from NOR individuals and 15 from individuals with DOR. Participant demographics and baseline clinical data are summarized in Table  3 . Both AFC (3.27±0.88 vs. 14.67±3.11, P  < 0.001) and AMH (0.37±0.24 vs. 2.41±1.02, P  < 0.001) were significantly reduced in the DOR group compared with the NOR group, and baseline FSH levels were significantly increased. No statistically notable differences were found regarding age, duration of infertility, BMI, or three other basic reproductive endocrine hormones ( P  > 0.05). Table 3 Demographic and clinical characteristics of registered participants NOR DOR P n 15 15 Age (year) 32.40±3.04 34.33±2.41 0.064 Infertile duration (year) 2.27±0.88 2.93±1.71 0.194 BMI (kg/m2) 22.22±1.46 22.54±1.27 0.527 AFC(n) 14.67±3.11 3.27±0.88 ＜0.001 AMH (ng/ml) 2.41±1.02 0.37±0.24 < 0.001 Basic FSH (mIU/ml) 6.99±1.90 12.07±4.83 0.001 Basic LH (mIU/ml) 5.75±4.70 6.94±5.27 0.519 Basic estradiol (pg/ml) 51.04±24.60 36.51±16.92 0.071 Basic progesterone (ng/ml) 0.48±0.36 0.38±0.29 0.405 Demographic and clinical characteristics of registered participants The miR-378a-3p expression was significantly lower in the DOR group than in the NOR group ( P  < 0.01) (Fig.  1 A). Correlation analysis further showed that miR-378a-3p levels in follicular fluid GCs were positively associated with AMH (r = 0.736, P  < 0.001), AFC (r = 0.774, P  < 0.001), the count of retrieved oocytes (r = 0.680, P  < 0.001), fertilized oocyte count (r =0.633, P  < 0.001), normally fertilized oocyte count (r = 0.663, P  < 0.001), and the total embryo count (r = 0.575, P  < 0.001), and was negatively correlated with basal FSH (r = -0.432, P = 0.017). Conversely, miR-378a-3p expression was not significantly related to age at trigger, BMI, basal LH, E 2 , and P levels, or the quantity of good embryos generated (all P  > 0.05) (Figs.  1 B-N). Fig. 1 A . The expression level of miR-378a-3p in GCs of NOR group and DOR group; B - I . The association between miR-378a-3p expression levels in GCs and various clinical parameters, including age, BMI, AMH, reproductive hormones (such as basal FSH, LH, E2, and P), and AFC; J - N . The correlation between the count of retrieved oocytes, fertilized oocytes, normal fertilized oocytes, embryos and good embryos.** P  < 0.01 A . The expression level of miR-378a-3p in GCs of NOR group and DOR group; B - I . The association between miR-378a-3p expression levels in GCs and various clinical parameters, including age, BMI, AMH, reproductive hormones (such as basal FSH, LH, E2, and P), and AFC; J - N . The correlation between the count of retrieved oocytes, fertilized oocytes, normal fertilized oocytes, embryos and good embryos.** P  < 0.01 To clarify the function of miR-378a-3p in GC apoptosis, KGN cells were transfected with a miRNA NC, miR-378a-3p mimics, and an inhibitor (Fig.  2 A). Subsequently, flow cytometry was conducted to assess apoptosis rates. The results demonstrated that compared to the miRNA NC group, transfection with miR-378a-3p mimics significantly inhibited cell apoptosis ( P  < 0.05). In contrast, suppression of miR-378a-3p with its inhibitor markedly promoted cell apoptosis after transfection ( P  < 0.05) (Fig.  2 B). This evidence strongly suggests that miR-378a-3p functions as an inhibitor of KGN cell apoptosis. Moreover, transfection with miR-378a-3p mimics effectively reduced intracellular ROS levels, while suppression of miR-378a-3p expression impaired cellular ROS clearance, ultimately exacerbating oxidative stress (Fig.  2 C-D). Concurrently, transfection with the miR-378a-3p inhibitor resulted in an increased percentage of JC-1 monomers, whereas miR-378a-3p mimics suppressed this increase. These observations point to the possibility that downregulation of miR-378a-3p can result in a reduction in MMP (Fig.  2 E-F). To further explore the impact of miR-378a-3p on mitochondrial function, we took a closer look at mtDNA expression levels. The results demonstrated that transfection with miR-378a-3p mimics resulted in an elevation of mtDNA levels in KGN cells (Fig.  2 G), showing that miR-378a-3p can affect mitochondrial function and GC apoptosis. Fig. 2 A Representative flow cytometry plots of KGN cell apoptosis after transfection with miR-378a-3p mimics or inhibitors; B Bar graph showing the apoptosis rate of KGN cells; C Representative images of intracellular ROS levels in KGN cells; D Bar graph of ROS levels in KGN cells; E Representative fluorescence images assessing MMP in KGN cells using the JC-1 method; F Quantitative analysis of the red/green fluorescence ratio, reflecting MMP; G Relative mtDNA copy number in KGN cells. * P  < 0.05 A Representative flow cytometry plots of KGN cell apoptosis after transfection with miR-378a-3p mimics or inhibitors; B Bar graph showing the apoptosis rate of KGN cells; C Representative images of intracellular ROS levels in KGN cells; D Bar graph of ROS levels in KGN cells; E Representative fluorescence images assessing MMP in KGN cells using the JC-1 method; F Quantitative analysis of the red/green fluorescence ratio, reflecting MMP; G Relative mtDNA copy number in KGN cells. * P  < 0.05 Based on the sequence of miR-378a-3p, potential target genes were predicted using the TargetScan, miRDB, and StarBase databases. Taking the intersection of the predictions from the three databases, we obtained a total of 18 common candidate target genes (Table  4 ). According to a quantitative literature-based screening strategy, each candidate gene was searched in the PubMed database using the keywords “gene name” and apoptosis. ZFP36L2 and RAB10 were selected for subsequent experimental validation, as they were the two genes most strongly associated with apoptosis and best supported by the literature. Bioinformatics analysis showed that miR-378a-3p has an 8mer binding site in the 3' UTR of ZFP36L2, and a 7mer-m8 binding site in the 3' UTR of RAB10. Given that 8mer binding sites generally mediate stronger translational repression, ZFP36L2 was chosen as the primary candidate target for in-depth experimental validation. Table 4 The list of target genes predicted for miR-378a-3p Target gene Gene name SDAD1 SDA1 domain containing 1 SBDS SBDS ribosome maturation factor EIF4G3 eukaryotic translation initiation factor 4 gamma 3 CIP2A cellular inhibitor of PP2A IL6ST interleukin 6 cytokine family signal transducer RIMS4 regulating synaptic membrane exocytosis 4 ZFP36L2 ZFP36 ring finger protein like 2 RAB10 member RAS oncogene family TMEM245 transmembrane protein 245 TOB2 transducer of ERBB2, 2 MAPK1 mitogen-activated protein kinase 1 CREBRF CREB3 regulatory factor DYRK1A dual specificity tyrosine phosphorylation regulated kinase 1A KPNA6 karyopherin subunit alpha 6 DYNC1LI2 dynein cytoplasmic 1 light intermediate chain 2 SRSF10 serine and arginine rich splicing factor 10 SHANK3 SH3 and multiple ankyrin repeat domains 3 ZNF652 zinc finger protein 652 The list of target genes predicted for miR-378a-3p To investigate the interaction between miR-378a-3p and ZFP36L2, reporter plasmids containing either the wild-type or a mutated version of the ZFP36L2 3′ UTR were constructed. The findings exhibited a marked decrease in luciferase activity from the wild-type ZFP36L2 after miR-378a-3p mimic transfection, relative to the miRNA NC ( P  < 0.001), suggesting potential binding between ZFP36L2 and miR-378a-3p. Once we altered the binding site between these two molecules, however, the luciferase activity of the mutated ZFP36L2 showed no significant difference when measured against the control group after miR-378a-3p mimic introduction ( P  > 0.05). These findings strongly suggest that miR-378a-3p can directly target and bind to the 3' UTR of ZFP36L2 (Fig.  3 A-C). Furthermore, KGN cells were transfected with either mimics or inhibitors of miR-378a-3p to investigate its regulatory effects on transcription and translation of ZFP36L2. The findings revealed that expression levels of ZFP36L2 were down-regulated in the group treated with miR-378a-3p mimics while up-regulated in those treated with miR-378a-3p inhibitors (Fig.  3 D-F). These results confirm that miR-378a-3p directly interacts with ZFP36L2 mRNA in KGN cells and exerts an inhibitory effect on its transcription process. Fig. 3 A Prediction of miR-378a-3p targets based on three databases; B miR-378a-3p binding and mutational sites within ZFP36L2's 3'-UTR; C Luciferase reporter gene assays confirm direct targeting of ZFP36L2 by miR-378a-3p through its 3'-UTR region; D - F mRNA and protein expression of ZFP36L2 in KGN cells after transfection.* P  < 0.05 A Prediction of miR-378a-3p targets based on three databases; B miR-378a-3p binding and mutational sites within ZFP36L2's 3'-UTR; C Luciferase reporter gene assays confirm direct targeting of ZFP36L2 by miR-378a-3p through its 3'-UTR region; D - F mRNA and protein expression of ZFP36L2 in KGN cells after transfection.* P  < 0.05 Additional research of follicular fluid GC specimens (n = 30) showed that both ZFP36L2 mRNA and protein concentrations were significantly higher in individuals diagnosed with DOR when compared to the NOR cohort ( P  < 0.01, Fig.  4 A-C). Moreover, a compelling negative correlation emerged between the expression level of ZFP36L2 mRNA in human follicular fluid granulosa cells and miR-378a-3p expression (r = −0.874, P  < 0.01) (Fig.  4 D). Furthermore, the abundance of ZFP36L2 mRNA in human follicular fluid GCs showed a strong inverse relationship with AMH (r = -0.710, P  < 0.001), AFC (r = −0.803, P  < 0.001), the count of retrieved oocytes (r = −0.696, P  < 0.001), fertilized oocyte count (r = −0.587, P  < 0.001), normally fertilized oocyte count (r = −0.611, P  < 0.001), embryo count (r = −0.518, P  = 0.001), and good embryos count (r = -0.365, P = 0.048), and showed a positive correlation with basal FSH (r = 0.414, P = 0.023). No notable associations were detected between the expression level of ZFP36L2 mRNA and trigger day age, BMI, basal LH, E 2 , and P levels (all P  > 0.05) (Fig.  4 E-Q). Fig. 4 A - C The mRNA and protein expression of ZFP36L2 in the NOR and DOR cohort; D . The expression of miR-378a-3p is negatively correlated with that of ZFP36L2 mRNA. E - L . The association between ZFP36L2 expression levels in GCs and various clinical parameters, including age, BMI, AMH, reproductive hormones (such as basal FSH, LH, E 2 , and P) M - Q . The correlation between the expression of ZFP36L2 mRNA and the count of retrieved oocytes, fertilized oocytes, normal fertilized oocytes, embryos and good embryos.* P  < 0.05, ** P  < 0.01 A - C The mRNA and protein expression of ZFP36L2 in the NOR and DOR cohort; D . The expression of miR-378a-3p is negatively correlated with that of ZFP36L2 mRNA. E - L . The association between ZFP36L2 expression levels in GCs and various clinical parameters, including age, BMI, AMH, reproductive hormones (such as basal FSH, LH, E 2 , and P) M - Q . The correlation between the expression of ZFP36L2 mRNA and the count of retrieved oocytes, fertilized oocytes, normal fertilized oocytes, embryos and good embryos.* P  < 0.05, ** P  < 0.01

Infertility is defined as the failure of a couple to achieve pregnancy after 12 months of regular, unprotected intercourse. Statistics show that it affects about 8% to 12% of couples of reproductive age worldwide, with a decline in female fertility beginning as early as 25 to 30 years of age [ 1 ] This has led many couples to seek assisted reproductive technology (ART) to achieve pregnancy [ 2 ]. Among the various ART treatments available, the characteristics of diminished ovarian reserve (DOR) are a reduced number and/or quality of oocytes, often accompanied by a decreased antral follicle count (AFC), lower anti-Müllerian hormone (AMH) levels, and increased basal follicle-stimulating hormone (FSH) levels. Ovarian reserve serves as an indicator of a woman's fertility potential; thus, DOR directly contributes to a decline in this potential [ 3 , 4 ]. In the ovary, the layer of granulosa cells (GCs) surrounding the oocyte has a crucial impact on its development and maturation [ 5 ]. GCs are essential contributors to the reproductive process, functioning not only as supportive nutrient providers but also as endocrine cells [ 6 ]. These cells possess receptors for FSH and luteinizing hormone (LH), which are critical for follicular growth. Their presence directly influences both the development and atresia of follicles [ 7 ]. MicroRNAs are small non-coding RNA regulators influencing genetic expression through inhibiting protein synthesis and facilitating mRNA degradation [ 8 ]. Studies have found that miRNAs play a regulatory role in ovarian follicular development by modulating GC proliferation, steroidogenesis, oocyte maturation, and GC activity. Their crosstalk with oocytes influences follicular quality, female reproductive capacity, and the outcomes of ART [ 9 ]. Previous research by the project team [ 10 ] indicated that compared with young women with normal ovarian reserve (NOR), the expression of 81 miRNAs in the follicular fluid GCs of older women with diminished ovarian reserve (DOR) showed significant differences, and the expression of microRNA-378a-3p (miR-378a-3p) was significantly downregulated [ 11 ]. Recently, miR-378-3p displays abnormal regulation across numerous malignancies including gastric [ 12 ], hepatocellular carcinoma [ 13 ], and breast cancer [ 14 ]. Other research has suggested that miR-378-3p could serve as a therapeutic avenue for endometrial carcinoma [ 15 ]. Nevertheless, miR-378a-3p's role in the molecular mechanisms underlying female fertility is still not well understood, motivating us to investigate its role in human GCs.

Oocytes, recognized as one of the largest cells in the human body, play an important role in the reproductive life cycle and endocrine balance of women. GCs are critical for supporting oocyte development by supplying vital nutrients [ 17 ]. The deficiency of enzymes or transport proteins participate in lipid, protein, and carbohydrate metabolism pathways within oocytes is compensated by GCs through the supply of pertinent enzymatic metabolites. Additionally, GCs and oocytes communicate bidirectionally through multiple pathways, including gap junctions, receptor- and ligand-dependent paracrine signaling, adhesive junctions, and extracellular vesicle-mediated macromolecule transfer, which is crucial for maintaining GC viability and function [ 18 ]. Notably, when this delicate intercellular communication network is disrupted, it may compromise GC homeostasis and subsequently trigger apoptosis. This notion is supported by clinical observations: compared to women with NOR, those with DOR exhibit significantly elevated rates of early-stage, late-stage, and overall GC apoptosis [ 19 ]. This apoptotic process among GCs contributes to the onset of DOR [ 20 ]. Therefore, investigating which key molecules are abnormally expressed in DOR that disrupt GC-oocyte communication or directly trigger GC apoptotic programs is of great importance for understanding the pathological mechanisms of DOR. Existing research indicates that miRNAs are very important in fertility-related diseases [ 21 ]. Research has demonstrated that miR-6881-3p [ 22 ] and miR-484 [ 23 ] are expressed at higher levels in women diagnosed with DOR, respectively leading to the reduction of SMAD4 and YAP1, thereby increasing GC apoptosis rates. Conversely, miR-221-3p targeting FOXO1 gene [ 24 ] and miR-106a targeting ASK1 [ 25 ] show decreased expression levels in DOR patients. Among these studies, the alterations of these miRNAs cause an increase in GC apoptosis. Our previous research revealed a significant downregulation of miR-378a-3p in women with DOR. But the specific biological mechanisms linking miR-378a-3p to DOR pathogenesis have not yet been fully elucidated. The findings of this experiment indicate that the miR-378a-3p expression is closely associated with ovarian function in women, exhibiting lower levels in the DOR cohort. It is worth noting that miR-378a-3p levels in human follicular fluid GCs show a positive association with AMH and AFC. As a dependable hormonal indicator of ovarian reserve, elevated AMH levels closely reflect a greater histologically confirmed primordial follicle reserve [ 26 ]. A reduced AFC indicates ovarian aging, and compared with elevated serum FSH levels, a low AFC is an earlier objective finding and one of the most reliable indicators of ovarian reserve [ 27 , 28 ]. A robust positive association was observed among AMH levels, AFC, and the count of oocytes retrieved [ 29 ]. Furthermore, in this study, the number of oocytes obtained, fertilized oocytes, normally fertilized oocytes, and embryos were all positively correlated with miR-378a-3p levels. These results suggest that miR-378a-3p could potentially be intricately linked to ovarian function and further research is needed on its mechanism of action in DOR. Mitochondrial biogenesis is a crucial intracellular pathway involved in ovarian reserve [ 30 ]. Mitochondrial dysfunction can result in decreased activity of respiratory chain enzymes, reduced MMP, diminished ATP synthesis, disruption of intracellular calcium homeostasis, alterations in mitochondrial ion permeability, obstruction of fatty acid beta-oxidation, accumulation of intracellular fatty acids, increased oxidative stress, and oxidative damage to mtDNA. These factors collectively lead to a decline in mitochondrial biogenesis and further exacerbate mitochondrial dysfunction, ultimately culminating in cell apoptosis [ 31 ]. In female patients with DOR, elevated levels of ROS, diminished MMP, and a reduced copy number of mtDNA were observed [ 32 ]. miR-378-3p induces metabolic reprogramming, characterized by enhanced mitochondrial oxidative metabolism and elevated mitochondrial membrane potential, thereby contributing to the maintenance of mitochondrial energy homeostasis [ 33 ]. Flow cytometry data confirm miR-378a-3p regulates mitochondrial function and apoptosis in GCs, establishing its crucial role in DOR. ZFP36L2 is an RNA-binding protein that can regulate maternal mRNA decay and oocyte development [ 34 ]. However, research on this topic is relatively limited in investigating the effects of ZFP36L2 on DOR and GC apoptosis. In this study, through comprehensive bioinformatics analysis coupled with dual-luciferase reporter gene assays, we detected ZFP36L2 as a direct downstream target of miR-378a-3p. Analysis of clinical GC samples also indicated that ZFP36L2 expression, both at the mRNA and protein levels, was significantly upregulated in the DOR cohort when contrasted with the normal response group. Furthermore, our findings demonstrated a negative correlation between ZFP36L2 mRNA abundance in human follicular GCs and ovarian reserve. Therefore, silencing ZFP36L2 may represent a potential therapeutic strategy for addressing DOR. A potential new mechanism underlying the pathogenesis of DOR was identified in this study, which may inform future research on possible treatment strategies for patients with DOR. However, this study has certain limitations. First, DOR is characterized by substantial clinical and physiological heterogeneity, with patients exhibiting marked differences in age, baseline ovarian reserve parameters, etiology, and hormonal microenvironment. For clinical studies involving human samples, a relatively modest sample size may fail to fully represent the broader and more diverse DOR population, potentially introducing selection bias and limiting the generalizability of the findings. Future investigations with expanded sample sizes and stratified analyses based on distinct DOR subtypes are warranted to comprehensively evaluate the universality of this mechanism. Additionally, further rescue/interaction experiments are needed. Specifically, it is recommended to overexpress ZFP36L2 concurrently in the presence of miR-378a-3p mimics to assess whether mitochondrial function, ROS levels, JC-1 staining results, apoptosis rates, and other parameters can be partially or fully restored. Alternatively, ZFP36L2 overexpression could be employed to counteract the impacts of miR-378a-3p mimics in order to establish a causal relationship. If animal experiments cannot be conducted, at a minimum, co-culture experiments utilizing various cell lines or primary GCs in vitro should be conducted for further validation. In addition, the dynamic relationship between mitochondrial structure and function, along with the multi-dimensional regulatory mechanisms of apoptosis, continues to represent a key scientific issue warranting further investigation. Future studies may utilize transmission electron microscopy to systematically characterize alterations in mitochondrial ultrastructure. In parallel, a combination of experimental approaches—such as TUNEL assay and CCK-8 analysis—can be employed to cross-validate apoptotic events. This integrated strategy will contribute to a more comprehensive understanding of the underlying molecular mechanisms and may further expand the biological implications and translational potential of the current findings.

In conclusion, the down-regulation of miR-378a-3p may disrupt mitochondrial function and promote apoptosis in GCs by up-regulating ZFP36L2 expression, thereby influencing the pathogenesis of DOR. In addition, the low levels of miR-378a-3p and the elevated expression of ZFP36L2 observed in patients with DOR could serve as potential therapeutic targets and promising treatment options for this disease.

Supplementary Material 1. Supplementary Material 1.