H2O2-induced hGC ageing, and cocultured with BMMSCs.
hGCs are the most important cells in the ovary, therefore, an hGC ageing model was established in vitro to further explore the mechanism of BMMSCs in ovarian ageing. As a small molecule oxidant, H2O2 can easily cause cell ageing by inducing oxidative stress through the biofilm system, and this technique has been widely applied to induce cell ageing[28, 29]. In our study, hGCs were exposed to H2O2 for 24 h and cocultured with BMMSCs for 48 h. First, β-galactosidase staining (blue staining with β-galactosidase activity) (Figure 1A) showed that 7.33±1.69% of the hGCs stained blue in the control group, 93.33±0.47% in the model group, and 43.66±2.05% in the coculture group (Figure 1B). Proliferation and division reflect the activation of hGCs. BrdU staining (red) showed that hGCs were proliferating and dividing (Figure 1C): 87.66±1.24% of hGCs were stained red in the control group, 16.33±1.24% in the model group, and 80.66±1.24% in the coculture group (Figure 1D) . Finally, reactive oxygen species (ROS) are an important ageing marker[30, 31], and DEH staining showed that the level of ROS was 3.77±0.12% in the control group, 8.63±0.12% in the model group, and 5.63±0.12% in the coculture group (Figures 1E,F), Next, immunohistochemical staining was performed to detect the expression of P53 protein (Figure 1G), and the results showed that 21.04±0.48% of cells expressed P53 protein in the control group, 58.20±1.21% in the model group, and 42.90±1.41% in the coculture group (Figure 1H),which suggested that the hGCs were ageing after 273 mM H2O2 treatment for 24 h, and BMMSCs recovered these ageing markers in aged hGCs.
The changes of m6A RNA methylation modificationin in aged hGCs and after BMMSCs treated
Studies have proved that the expression of demethylase FTO was down-regulated in ovarian tissue with the increase of age to promote ovarian aging[26, 27]. However, it whether BMMSCs could regulate FTO to effect methylation modification remains unclear, Interestingly, our results showed that compared with the model group, the expression of FTO was up-regulated after BMMSCs treatment in vivo (Figure 2A, B), and also significantly up-regulated after aged hGCs were co-cultured with BMMSCs in vitro (Figure 2C, D, E, F). In addition, the overall level of m6A was significantly decreased after co-cultured with BMMSCs (Figure 2G), which suggested that hGC ageing is closely related to m6A methylation modification, and BMMSCs play an key role in regulating m6A methylation modification.
Overview of the m6A methylation landscape in aged hGCs and after cocultured with BMMSCs.
m6A is the most prevalent RNA modification of mRNAs and lncRNAs and plays a key role in ageing and various ageing diseases[32, 33]. However, its specific regulatory mechanism in ovarian ageing remains unclear. In our study, after induced hGCs ageing and cocultured with BMMSCs, MeRIP-seq were performed to explore the effect of BMMSCs on m6A modification of hGC ageing. Our results showed that 7,923 transcripts displayed a total of 14,417 sites that were modified by m6A in the model group, and 6,867 transcripts displayed a total of 11,715 sites that were modified by m6A in the coculture group. Among them, 14,241 individual m6A peaks in 9,741 m6A-modified genes were detected in the model and coculture groups (Figure 3A). Notably, the coculture group had 6,109 new peaks and 9,088 missing peaks compared to the model group, revealing that the global m6A modification patterns were markedly different between the model and coculture groups (Figure 3B). As shown in Figure 3C, D, the results showed different patterns of peaks with a relative increase in the start codon region (6.4 vs. 5.7% for aged hGCs and aged hGCs cocultured with BMMSCs, respectively) and in the 3’ untranslated region (3’UTR, 39.4 vs. 37.6%) and a relative decrease in the coding sequence (CDS, 30 vs. 32.3%) and at the stop codon (20.1 vs. 21%). Figure 3E shows that the distribution of m6A signals around mRNAs and lncRNAs was comparable in the model and coculture groups. In general, m6A peaks tended to occur in CDS regions and 3’UTR, which means that m6A is likely to play a crucial role in regulating mRNA expression and affecting the stability of the mRNA, consistent with previous MeRIP-seq results[34, 35]. The m6A peak distribution analysis suggested that most mRNAs and genes had m6A peaks, and there were mostly 1 to 3 m6A modifications in the exons (Figures 3F, G, H). In addition, m6A peaks were found in all chromosomes, with the highest numbers being in chr1, chr17, and chr19 (Figures 3I).
Differentially m6A peaks in model and cocolture group
To explore the biological significance of m6A modification in BMMSCs interacting with aged hGCs, GO and KEGG pathway analyses of differentially methylated mRNAs were conducted. As shown in Figure 4A, compared to aged hGCs, after cocultured with BMMSCs had 449 significantly upregulated m6A peaks and 348 downregulated m6A peaks (fold changes≥2). Furthermore, the classic GGACU motif and the top 5 m6A motifs were observed in the model (Figure 4B) and coculture groups ( Figure 4C). GO process results showed that the altered m6a peaks were significantly enriched in chromatin modification, regulation of transcription, DNA−templated, and cell cycle ( Figure 4D). Additionally, in KEGG pathway analyses, the Spliceosome, Epstei-Barr virus infection, and Thyroid hormone signaling pathway were signifificantly correlated with genes that showed m6A peaks in aged hGCs ( Figure 4E). It can been said that m6A peaks were play a key role in ovarian ageing.
Changes in mRNA in aged hGCs and after cocultured with BMMSCs
First, we tested the transcriptome profiles of altered genes in three pairs of aged hGCs and aged hGCs after cocultured with BMMSCs using MeRIP-seq. Compared to aged hGCs, hGCs cocultured with BMMSCs had 412 significantly upregulated genes and 405 significantly downregulated genes (|log2FC| > 1, P value < 0.05) (Figure 5A), which PPI networks are presented in Figure 5B and 5C. Functional network analysis showed that the 412 genes were involved in regulation of mitotic sister chromatid segregation, positive regulation of protein localization to endosomes, and cellular senescence (Figure 5D). The KEGG pathway analysis mainly identified enrichment of the terms cytokine-cytokine receptor, pentose phosphate pathway, rheumatoid athritis, and TNF signaling pathway(Figure 5F). In addition, the 405 genes were found to be involved in purine nucleoside triphosphate biosynthetic processes, neural crest cell migration, and embryonic camera-type eye formation(Figure 5E), and were significantly enriched in the terms nonhomologous end-joining, the Hippo signalling pathway, sphingolipid metabolism, and the homologous recombination signalling pathway (Figure 5G).
Correlation between differential m6A peaks and differential expression of mRNA
Correlation analyses of altered m6A peaks with differentially expressed mRNAs (|log2FC| > 1, P value < 0.05) were performed to identify the key genes through which BMMSCs affect m6A methylation modification of hGCs during ageing (Figure 6A, B). The cumulative differential mRNA abundance is shown in Figure 6C. We identified 42 hypermethylated m6A peaks in mRNAs that were significantly upregulated (3) or downregulated (39), while 88 hypomethylated m6A peaks in mRNAs were significantly upregulated (74) or downregulated (14) (Figure 6D). Next, 130 genes that showed significant changes in both m6A modification and RNA expression levels were subjected to GO, pathway and PPI network analyses. The GO analyses of processes associated with the 130 gene sets are shown in detail in Figure 6E and identified numerous linked functional processes and pathways. Interestingly, we found that the top 3 GO terms of histone H3-K36 demethylation (KDM8 and RIOX1), regulation of apoptotic DNA fragmentation, and regulation of DNA catabolic process were enriched in the GO maps of aged hGCs cocultured with BMMSCs (Figure 6E). Interestingly, the correlation of histone H3-K36 demethylation was consistent with the results of the DEG GO analysis (Figure 5E). In addition, the top 20 KEGG pathways that are shown in Figure 6F, which were significantly enriched in metabolism, genetic information processing, and environmental information processing. Moreover, the EcCenticity analysis results showed that KDM8 ranks first in the PPI network (Figure 6G), and involved in negative regulation of osteoblast differentiation and circadian regulation of gene expression (Figures 6H). KDM8 m6A peak visualization by IGV is shown in Figure 6I. These observations indicate that KDM8 with m6A modification may play important roles in the reversal of hGC ageing induced by BMMSCs.
The expression of KDM8 in aged GCs before and after cocultured with BMMSCs
KDM8 (Lysine Demethylase 8) as the epigenetic repressive mark and important cell cycle regulator, which functioned as a transcriptional activator by inhibiting HDAC recruitment via demethylation of H3K36me2[36], and involve in osteoblast differentiation[37]. Methylation modifications of histones play critical roles in regulating gene expression, cell cycle, genome stability, and nuclear structure, therefore, we explored the regularity of KDM8 in hGCs and ovarian tissue, and the intervening effect of BMMSCs. Our results showed that compared with the elderly model group, the expression of KDM8 was down-regulated after BMMSCs treatment in macaques(Figure 7A, B), and also down-regulated after co-cultured with BMMSCs in vitro(Figure 7C,D,E), while histone of H3 was up-regulated in hGCs after cocultured with BMMSCs(Figure 7F).