Increased DNMT1 acetylation leads to global DNA methylation suppression in follicular granulosa cells during reproductive aging in mammals

Abstract

With increasing age, the reproductive performance of women and female animals declines. However, the molecular mechanisms underlying ovarian aging and age-related fertility decline remain unclear. Granulosa cells (GCs) are suspected to play an important role in reproductive aging, and their proliferation, apoptosis, and steroid hormone secretion are used to determine the fate of follicles and ovarian function. First, we found that the proliferative ability of GCs from the old mouse group (10-month-old) decreased compared with that from the young mouse group (6-week-old), and cell cycle arrest occurred in old mice. To investigate changes in protein modification, we compared the levels of protein acetylation in GCs from young and old mice. We found that the K1118, K1120, K1122, and K1124 sites of DNA methyltransferase 1 (DNMT1) were increasingly acetylated with age, resulting in a decrease in DNMT1 protein expression. Therefore, we performed whole-genome methylation sequencing of GCs in the two groups and found that the CG methylation levels in the old group were lower than those in the young group. Furthermore, the inhibition of DNMT1 expression in GCs resulted in cell cycle arrest. This study revealed the dynamics and importance of protein acetylation and DNA methylation in GCs during reproductive aging. The findings provide a theoretical basis for studying the mechanism of reproductive aging in mammals.

Introduction

With increasing age of female mammals, the reproductive system and fertility declines. This is mainly caused by a complex process that occurs when the ovary undergoes senescence. Currently, most studies suggest that ovarian aging is caused by accumulation of oxidative stress, alterations in epigenetic modifications, and diminished levels of sex hormones, which leads to a progressive decline in the number and quality of oocytes [1,2,3]. Granulosa cells are the most abundant cells in the follicles surrounding oocytes. The proliferation, apoptosis, autophagy, and hormone secretion by GCs directly determine the fate of the follicle, ovarian function, and oocyte development [4]. Previous studies have proven that the apoptosis of GCs can lead to follicular atresia, resulting in a decline in follicular reserve [5]. However, the mechanisms underlying ovarian aging remain unknown.

DNA methylation, the covalent addition of the methyl group to the fifth carbon of cytosines (5mC) within CpG dinucleotides, is the most well-characterized type of the epigenetic control of gene expression and developmental process, which has essential roles in aging [6]. Studies showed that age affects DNA methylation at almost one-third (29%) of the sites, of which 60.5% become hypomethylated and 39.5% hypermethylated with increasing age [7]. Koch et al. have introduced an Epigenetic-Aging-Signature consisting of five CpG sites. Four of them were hypermethylated CpG sites, and one was a hypomethylated CpG site [8]. DNMT1 is the main enzyme that uses the information on DNA methylation patterns in the parent strand and methylates the daughter strand in freshly replicated hemimethylated DNA [9]. Loss or reduced activity of DNMT1 could potentially result in cell cycle arrest [10]. Yang et al. found that DNMT1 inhibition altered the methylation status of CDKN1A, leading to cell cycle arrest [11]. The expression of DNMT1 is affected by its acetylation modification. It has been shown that acetylation of KG linker lysine residues impairs DNMT1–USP7 interaction and promotes DNMT1 degradation [12]. Although there have been many studies on DNMT1 and cell cycle, there is no influence on the cell cycle caused by DNA methylation after aging.

In this study, the results showed decreased cell proliferation and cell cycle arrest. The acetylation level of DNMT1 increased, and the expression of genes and proteins decreased in the old group. In the inhibition experiment of DNMT1, DNMT1 could affect the cell cycle. Genome-wide methylation analysis of GCs revealed a decreasing trend in DNA methylation in old mice. The methylation of protein tmod1 and nrp2 related to the cell cycle decreased, and the expression increased. We revealed the relationship between age-related methylation changes and cell cycle changes.

Materials and methods

Animals

All experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals, prepared by the Institutional Animal Care and Use Committee of Hebei Agricultural University, China. All mouse strains were maintained on a Kunming (KM) background and were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). We selected 6-week-old and 10-month-old female KM mice as the young and old groups, respectively, with average weights of 32 g and 50 g, respectively; healthy male mice were selected to mate with the female mice. There are 24 mice in young group and old group. Among the animals, there are 10 mice per group designated for birth count analysis, while 14 mice per group are allocated for cycle detection and DNA methylation analysis. Mice were housed at a density of 2–5 adults/cage or 1–3 adults with one litter/cage in a specific pathogen-free animal facility, with a light/dark cycle of 12:12 h, a constant temperature (20 ± 2 ℃), and free access to food and water. The mice were administered pentobarbital sodium anesthesia and subsequently humanely euthanized through careful cervical dislocation.

Approximately 16 pig ovaries were procured from a local slaughterhouse for the purpose of collection. we were transported at 37 °C in saline (containing 100 IU/mL penicillin and 100 µg/mL streptomycin) to our laboratory within 1 h.

Cell culture and treatment

We collected GCs from 3 to 6 mm healthy follicles on pig ovaries using a 7-gauge syringe needle (Jiangxi Hongda Medical Equipment Group Co., Ltd., China). Mouse ovaries were collected by gently puncturing them using a 4.5 syringe needle. The average number of ovarian aspiration follicles per pig or mice is presented in Supplementary Table 1. GCs were seeded into six-well plates containing 15% fetal bovine serum (Sigma, USA) in DMEM/F12 (HyClone, Logan, UT, USA) with 100 IU/mL penicillin and 100 µg/mL streptomycin and incubated at 37 °C with 5% CO₂. We cultured porcine GCs for 12 h after inoculation, and when they just adhered to the wall, they were treated with thioguanine (Cat #S1774; Selleck Chem) at concentrations of 0, 1, 5, and 10 µmol/L for 48 h.

Litter size in mice

Six-week-old male mice were mated with female mice in the estrus phase. The number of births was recorded approximately 21 d after administering a vaginal suppository.

Cell proliferation assay

The Cell Counting Kit-8 (CCK-8) (TransGen, Beijing, China) was used to detect the viability of mouse follicular GCs, and a blood cell counting plate was used to determine and maintain the density of GCs at approximately 5 × 10⁴ cells. GCs were inoculated into a 96-well plate and cultured for 12, 24, 36, or 48 h. The CCK-8 solution (10 µL) was added, and the samples were incubated for 2 h at 37 ℃ and 5% CO₂. Absorbance was measured at 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

Cell cycle analysis

Mouse or pig GCs treated with thioguanine (Cat #S1774; Selleck Chem) were washed with precooled PBS, incubated with 400 µL of a propidium iodide solution for 30 min, and the fluorescence was measured at 488 nm using flow cytometry.

Total RNA extraction and qRT-PCR

Total RNA was extracted from cells and tissues using TRNzol Universal (Tiangen, Beijing, China). Reverse transcription was performed to synthesize cDNA using the Prime Script RT reagent kit (TaKaRa Bio, Inc., Otsu, Shiga, Japan). qRT-PCR was performed using the THUNDERBIRD SYBR qPCR mix (TOYOBO, Tokyo, Japan), and the number of cycles was monitored using an ABI QuantStudio 6 real-time fluorescence quantitative PCR system (ABI, Grand Island, NY, USA). Each treatment group included at least three biological replicates. All primer sequences are listed in Table 1.

Table 1 Primer sequence for RT-qPCR

Genomic DNA extraction

Mouse GCs were collected, and genomic DNA was extracted according to the instructions of a DNA extraction kit (DP305; TIANGEN, Beijing, China). Finally, DNA was dissolved in a Tris-EDTA(Ethylenediaminetetraacetic acid) buffer solution. EDTA is the ethylenediaminetetraacetic acid.

Western blotting

Protein lysates were obtained from approximately 10⁶ cells. Proteins were separated using a double-plate vertical electrophoresis device (DYCZ-24DH; Beijing Liuyi Biotechnology Co., Ltd., Beijing, China) and transferred onto polyvinylidene fluoride membranes (Millipore, St. Louis, MO, USA). Antibodies against succinyllysine (PTM-419), crotonyllysine (PTM-502), 2-hydroxyisobutyryllysine (PTM-802), malonyllysine (PTM-902), lactyllysine (PTM-1401), acetyllysine (PTM-102), and CDK1 (PTM-6109) were purchased from PTM BioLab (China). The antibody against DNMT1 (A16729) was obtained from ABclonal Technology (Wuhan, China). The antibody against PTX3 (13797-1-ap) was purchased from Wuhan Sanying (Wuhan, China). Antibodies against HAS2 (ab140671), TGFB2 (ab238249), and BUB1 (ab195268) were purchased from Abcam, and antibodies against CDKN1A (2947) and E2F1 (3742) were purchased from Cell Signaling Technology (Danvers, MA, USA). Antibodies against Tmod1 (DF12333) and Nrp1 (DF7032) were purchased from Affinity Biosciences (OH, USA). The antibody against Tnfaip6 (MAB2104) was purchased from R&D Systems (Minneapolis, MN, US), and the antibody against actin (GB15001) was purchased from Servicebio (Wuhan, China). Secondary antibodies used were horseradish peroxidase (HRP)-labeled goat anti-mouse and goat anti-rabbit IgG (Servicebio). A Tanon 5200 multi-imaging system (Tanon, Shanghai, China) was used to detect signals generated by the ECL HRP substrate (WBKLS0500; Millipore).

RNA sequencing

Total RNA from porcine follicles (at least three biological replicates) was sequenced using an Illumina NovaSeq 6000 by Gene Denovo Biotechnology Co. (Guangzhou, China). Differential expression analysis of RNA was performed using the DESeq2 software (Love et al.,2014) [13] between two different groups, and edgeR (Robinson et al.,2010) [14] was used between two samples. Genes and transcripts with a false discovery rate below 0.05 and an absolute fold change ≥ 2 were considered differentially expressed. Hierarchical clustering analysis was performed on all transcriptome samples using the pairwise similarity of each pair of samples determined by Spearman’s correlation coefficient.

Whole-genome methylation sequencing (WGMS)

After genomic DNA was extracted from samples, the DNA concentration and integrity were determined using a NanoPhotometer^® spectrophotometer (Implen, CA, USA) and agarose gel electrophoresis, respectively. Then, DNA libraries for bisulfite sequencing were prepared. DNA fragments (at least three biological replicates) were PCR amplified and sequenced using an Illumina HiSeq™ 2500 by Gene Denovo Biotechnology Co. (Guangzhou, China). The clean reads obtained were mapped to the species reference genome using the BSMAP software [15] (version: 2.90) by default. BSMAP is a fast and accurate algorithm to map bisulfite reads to the reference genome. Differential DNA methylation between two samples at each locus was determined using Pearson’s chi-square test (χ²) in methylKit [16] (version: 1.7.10). To identify differentially methylated cytosines, the minimum read coverage to determine the methylation status of a base was set to four. Differentially methylated regions (DMRs) for each sequence context (CG, CHG, and CHH, H represents A, C, or T) were determined according to the following criteria: for CG and CHG, the number of GCs in each window ≥ 5, absolute value of the difference in the methylation ratio ≥ 0.25, and q ≤ 0.05; for CHH, the number of GCs in a window ≥ 15, absolute value of the difference in the methylation ratio ≥ 0.15, and q ≤ 0.05.

Correlation of DNA methylation and gene expression in samples

To determine whether gene expression influences DNA methylation in a sample, genes were categorized into four classes based on their expression levels, including a no expression group (reads per kilobase per million reads mapped [RPKM] ≤ 1), a low expression group (1 < RPKM ≤ 10), a moderate expression group (10 < RPKM ≤ 100), and a high expression group (RPKM > 100).

To analyze whether DNA methylation influences gene expression in a sample, genes were categorized into four classes according to their methylation level: no methylation, low methylation, moderate methylation, and high methylation (nonmethylated genes were excluded, and the rest were divided into three groups on average).

Spearman’s correlation analysis was performed to statistically discern the relationships between DNA methylation and gene expression within the ± 2-kb flanking regions and the gene body region (rho > 0 indicates positive correlation; rho < 0 indicates negative correlation).

Correlation of DNA methylation and gene expression between groups

To analyze whether DEGs influence DNA methylation between groups, DEGs were categorized into four classes based on their different expression patterns, including a special-down group (genes specifically expressed in the control group), a special-up group (genes specifically expressed in the treatment group), an other-down group (genes downregulated in the treatment group), and an other-up group (genes upregulated in the treatment group).

To determine whether DNA methylation in DMRs influences gene expression between groups, genes were classified according to the genomic location, including the ± 2-kb flanking regions and the gene body region.

Statistical analyses

All results are expressed as the mean ± SEM. An unpaired Student’s t-test was used to determine statistical significance, and P < 0.05 indicated a significant difference.

Results

Reduced reproductive performance in old mice

The mice in the old group showed a significant decrease in the number of litters compared with that in the young group (7.08 ± 4.92 vs. 14.36 ± 3.64, Fig. 1A). The proliferative ability of GCs was significantly reduced (P = 0.0001) in 10-month-old mice after 24 h of culture (Fig. 1B). Because the cell cycle is closely related to the proliferative capacity of cells, we examined the cell cycle of GCs in the two groups and found that the cell cycle in the old group tended to stagnate in the G2/M phase (13.35% vs. 11.28%, Fig. 1C–E). These results indicate that with increasing age, the reproductive capacity of mice becomes defective, the ability of follicular GCs to proliferate is reduced, and the cell cycle appears to be stalled. The expression levels of the cell cycle-related gene CDK1 and GCs proliferation-related genes HAS2, TNFAIP6, and PTX3 were significantly decreased in the old group (Fig. 1F–H, P < 0.05).

Fig. 1

Reduced reproductive performance in old mice. A) Comparison of litter size in young and old mice. B) Decreased proliferation of granulosa cells in older mice. C–E) Cell cycle analysis in young and old mice. F–H) The mRNA (F) and protein (G-H) expression levels of CDK1, HAS2, TNFAIP6 and PTX3 in GCs. (n = 3). *P < 0.05, **P < 0.01

Protein acetylation modification

The realization of cell function is determined not only by protein expression levels but also by posttranslational modifications of proteins. Thus, to better understand the potential mechanisms leading to the cycle arrest of the GCs in ovarian aging, we detected six types of posttranslational protein modifications in the young and old mouse groups, including protein acetylation, crotonylation, succinylation, malonylation, lactylation, and 2-hydroxyisobutyrylation (Fig. 2A–F). The level of protein acetylation was found to change, and we performed an in-depth omics analysis of the protein acetylation modification. The difference in the Pearson correlation between the groups was greater than that among samples from the same group (Fig. 2G). A total of 606 sites with increased and 516 sites with decreased protein acetylation were detected in GCs from the old mice compared with those from the young mice (Fig. 2H). The KEGG analyses showed that these 730 differentially acetylated modification proteins were significantly engaged in butyrate metabolism, tryptophan metabolism, reproductive metabolism, and the citric acid cycle (Fig. 2I).

Fig. 2

Protein modification analysis in the old and young groups of mice. A–F) Protein modification levels in the young and old groups; from left to right: acetyllysine, succinyllysine, crotonyllysine, 2-hydroxyisobutyryllysine, malonyllysine, and lactyllysine. G) Heatmap of the Pearson correlations between GCs in the old and young groups. H) Volcano map showing differentially acetylated modification sites. I) KEGG pathway analysis of differentially acetylated modification proteins

Acetylation modification level change of DNMT1 in old mice

Analysis of the protein acetylation modification showed that the acetylation levels of DNMT1 increased 3.716-fold at the K1118, K1120, and K1122 sites and 3.204-fold at the K1124 site in 10-month-old mouse GCs compared with those in 6-week-old mouse GCs. The expression of DNMT1 protein was observed to decline to 0.6 times the level observed in the young group. The structure of dnmt1 is shown in Fig. 3A. In this study, sites with elevated acetylation levels were located within the KG linker. Therefore, we examined the expression levels of DNMT1 in GCs from both groups of mice, and both mRNA and protein level results showed that DNMT1 expression decreased in 10-month-old mice compared with that in 6-week-old mice (Fig. 3B–D).

Fig. 3

Decreased DNMT1 expression in old mice. A) DNMT1 structure. B–D) mRNA and protein level of DNMT1. *P < 0.05, **P < 0.01

Decreasing trend of DNA methylation in old mice

DNMT1 levels were significantly decreased in GCs of 10-month-old mice, suggesting potential changes in genome-wide methylation modifications. Therefore, we performed WGMS, as well as transcriptome sequencing, for both groups of GCs. Figure 4A shows the degrees of CG (CpG sites), CHG (CHG sites), and CHH (CHH sites) (H represents A, C, or T) methylation and the ratio of the gene density in the chromosomes of the young and old groups. Subsequently, methylated C-base sites were counted, and the data showed that the old group had a decreased CG site content and increased CHG and CHH site contents. In the old group, CG methylation (76.95%) was higher than CHG (4.73%) and CHH (18.32%) methylation. In the young group, CG methylation (87.52%) was higher than CHG (2.82%) and CHH (9.67%) methylation (Fig. 4B–C). The distribution of methylation levels showed that the rate of CG methylation decreased in the old group compared with that in the young group (Fig. 5D–E). Additionally, CG methylation in the old group was decreased in the 2 kb of gene body flanking region, whereas CHH and CHG methylation were increased (Fig. 4F–H). To further elucidate the dynamics of changes in DNA methylation in aging mice, we grouped the entire mice genome into intergenic and genic regions; each genic region was further divided into exon, intron, five UTR, three UTR, upstream 2k, and downstream 2k. Finally, analysis of the incidence of CG methylation in gene regions revealed a decreased trend in each gene region in the aging group (Fig. 4I). These results suggest that CG methylation in the aging group has a decreasing trend, similar to DNMT1 expression.

Fig. 4

Genomic landscape of DNA methylation in old and young mice. A) Circos plot showing the gene density, TE density, and ratios of CG, CHG, and CHH methylation in old and young mice using 1-Mb sliding 200-kb windows. The outer track represents the mouse chromosomes. B–C) Ratios of CG, CHG, and CHH methylation in old and young mice. D–E) CG, CHG, and CHH DNA methylation distribution in different methylation levels of old and young groups. F–H) Average CG methylation levels in the gene body and upstream and downstream of all genes. I) Methylation rates of CG in different gene functional elements

DNA methylation and gene expression

To explore how DNA methylation affects gene expression, a combined whole-genome methylomic and transcriptomic analysis was performed. First, we identified 5284 DMRs and 3355 DEGs. DNA methylation in upstream 2 kb was negatively correlated with gene expression (Fig. 5A). Additionally, Spearman’s correlation analysis of DNA methylation and gene expression levels in the upstream and downstream 2 kb regions of the encoded genes showed the same results in the two groups (Fig. 5B–C). Then, all genes were divided into four groups according to their expression levels: none, low expression, middle expression, and high expression genes. The methylation level of these four groups in the gene, upstream, and downstream 2 kb regions were evaluated. The results showed that in the CG category, the DNA methylation level of non-expressed genes was the highest, whereas that of the highly expressed genes was the lowest. The lower the expression of the gene, the higher the level of gene methylation. However, in the CHG and CHH types, gene expression was not strongly associated with methylation (Fig. 5D–F). Thereafter, we compared the differentially methylated genes with the DEGs (Fig. 5G–I). Detailed information on the methylate in each region of the Tmod1 and Nrp2 genomes is shown in. The methylation level of both TMOD1 and NRP2 is notably high, whereas their corresponding mRNA levels are conspicuously low (Fig. 5J–K). They could inhibit the cell cycle and were increasingly expressed while their methylation was decreased (Fig. 5L–M).

Fig. 5

DNA methylation and expression in old and young mice. A) Expression of genes with down- and upregulated DMRs in different gene regions. B–C) Correlation analysis between methylation rates and gene expression levels. D–F) Methylation level of genes at different expression levels in gene regions of CG (D), CHG €, and CHH (F). G–I) Venn diagram of differentially methylated genes vs. differentially expressed genes of CG (G), CHG (H), and CHH (I). J-K) Methylation level and MRNA level of TMOD1 and NRP2 gene. L-M) The protein expression levels of TMOD1 and NRP2 in GCs. *P < 0.05, **P < 0.01

Inhibition of DNMT1 affects the cell cycle

As described above, the GCs cycle was arrested (Fig. 1), and the overall level of DNA methylation modifications was reduced (Fig. 4) in the old group of mice. To explore whether the reduction in the DNMT1 level is a key factor affecting the cell cycle, we treated porcine GCs with the DNMT1-specific inhibitor thioguanine at the optimal concentration (5 µM), which was selected at a cell proliferation rate greater than 80% (Fig. 6A–C). The KEGG analyses showed that these 708 DEG were significantly engaged in the cell cycle pathway (Fig. 6D). Cell cycle detection results also showed that the GCs were arrested in the G2/M phase after being treated with thioguanine (Fig. 6E–G). Finally, we examined the expression levels of the DEGs CDKN1A, TGFB2, BUB1, and E2F1, which are associated with the cell cycle. qRT-PCR and western blotting showed a significant increase (P < 0.05) of the cell cycle inhibitor CDKN1A and cell growth-related gene TGFB2 after treatment. The transcript and protein levels of the cell cycle maintenance genes E2F1 and BUB1 were significantly reduced (P < 0.05) (Fig. 6H–J).

Fig. 6

Inhibition of DNMT1 caused cell cycle arrest. A) Proliferation of porcine granulosa cells after treatment with thioguanine. B-C) DNMT1 protein level after treatment of porcine granulosa cells with thioguanine. D) KEGG enrichment pathway analysis of differentially expressed genes in the control and inhibitor groups. E–G) Cell cycle of porcine granulosa cells after treatment with thioguanine. H–J) The mRNA (H) and protein (I-J) expression levels of CDKN1A, TGFB2, E2F1 and BUB1 in GCs. (n = 3). * P < 0.05, **P < 0.01

Discussion

In the present study, we procured ovaries from mice of 6-week-old (young) and 10-month-old (old), all synchronized in their diestrus phase. Additionally, we sourced pig ovaries from slaughterhouses. These pigs have not yet experienced estrus. From these pig ovaries, we selectively extracted follicles with diameters ranging from 3 to 6 mm, these follicles exhibit a vibrant yellow and transparent appearance, while blood vessels are distinctly visible, displaying a clear and vivid red color. Based on our prior research, 3–6 mm follicles were identified as non-atretic and healthy follicles [17].

We find that aged mouse GCs were cycle-arrested, and a reduction in DNMT1 expression was an important cause of cell cycle arrest. The reduction in protein acetylation led to the expression of DNMT1, which is a crucial methyltransferase regulating intracellular DNA methylation. Consequently, as the expression of DNMT1 diminished in aged mouse GCs, there was a corresponding tendency for DNA methylation to decrease. In subsequent DNMT1 inhibition experiments, we verified that the reduced DNMT1 level was one of the main causes of cell cycle arrest.

The main cause of reproductive senescence is ovarian senescence, which is characterized by a decrease in the number of available follicles and the number and quality of oocytes. The causes of ovarian senescence are complex and involve oxidative stress and epigenetic factors [18, 19]. Furthermore, recent developments in anti-aging drugs have focused on oxidative stress, DNA damage, and mitochondrial and protein dysfunction [20]. The availability of new technologies, such as single-cell sequencing, has helped unravel the mysteries of aging.

The primary manifestation of ovarian aging is the deterioration in oocyte quality [21]. GCs play a crucial role in nourishing oocytes by supplying them with vital nutrients, hormones, and essential small molecules that are indispensable for their growth and development [22]. GCs apoptosis are the most direct factor for follicular atresia. Blockage of the cell cycle also prevents the follicle from developing into a dominant follicle and thus, from ovulating. Therefore, GCs were used in the present study. However, oocyte quality is more important; therefore, we may later focus on whether posttranslational modifications of proteins and DNA methylation during reproductive senescence are also present in oocytes, similar to GCs.

First, we found that GCs in senescent mice had a reduced proliferative capacity and stalled cell cycles, similar to what is traditionally thought to occur after cellular senescence. Subsequent assays of posttranslational modifications of GC proteins in both old and young groups of mice revealed that all six modifications were subject to changes during aging; however, acetylation changes were more pronounced at 15, 30, 55, 70, and 130 kDa for almost every size band and at 15 kDa, mainly for histone modifications. Therefore, acetylation modifications were selected for subsequent studies.

Among the differentially acetylated proteins in senescent GCs, the DNMT1 protein showed reduced acetylation and reduced gene expression. DNMT1 is a DNA methyltransferase that maintains DNA methylation, whereas DNMT3a and DNMT3b are de novo synthetases. Recently, numerous studies have revealed a close association between DNA methylation and aging, during which genome-wide methylation decreases [23]. Therefore, the findings of the present study are, to some extent, in line with those of previous studies. Herein, we focused on the level of CG methylation because it is more relevant to mammals. Our study found higher methylation levels in the CG subtype than in the CHG and CHH subtypes, with the highest methylation levels in the gene region, followed by the upstream and downstream regions. Moreover, methylation levels in the upstream and gene regions were inversely correlated with gene expression. Subsequently, we found that CG methylation decreased, but CHG and CHH methylation increased, which may be due to nongenetic factors, such as environmental factors.

Some studies have found that the main regulatory mechanisms leading to cell cycle arrest in senescent cells are the P53/P21^WAF1/CIP1 pathway, p16^INK4A/pRB pathway, and DREAM complex [24,25,26]. Our results identified DNMT1 as a gene that potentially regulated cell cycle arrest in senescent cells; as DNMT1 expression decreases, the cell cycle appears to be prolonged. The relationship between DNMT1 and cell cycle has been previously reported. dnmt1 has several regulatory subdomains related to cell cycle regulation, among which the PBD and TS structural domains are closely associated with the cell cycle and cell division [27]. DNMT1 knockdown in human endometrial cancer cells resulted in decreased proliferation and cell cycle arrest [28]. Some researchers have found that lnc-MAP3K13-7:1 regulated DNMT1 expression in KGN cells by affecting the ubiquitination of DNMT1. Subsequently, DNMT1 increased CDKN1A transcription by hypomethylating the CDKN1A promoter, which led to reduced proliferation and cell cycle arrest in KGN cells [29]. This is similar to our findings, which showed increased CDKN1A expression in DNMT1 inhibition experiments. However, our findings suggest that the DNMT1 protein is not solely affected by altered ubiquitination but is also controlled by acetylation. Most studies on DNMT1 in relation to the cell cycle do not mention senescence; therefore, it is of interest to investigate the relationship between DNMT1 in senescent cells and cell cycle arrest.

Conclusions

In conclusion, this study suggests that cell cycle arrest accompanying cellular senescence is associated with DNMT1 and reveals changes in DNA methylation in mouse senescent GCs, suggesting targets for intervention to address the problem of mammalian senescence. The acetylation modification of DNMT1 in aged GCs increased, causing the degradation of protein molecules. Subsequently, owing to reduced methylation of cell cycle-related genes, such as Tomd1 and Nrp2, transcription is promoted, which leads to a change in the expression and, finally, to cell cycle arrest (Fig. 7).

Fig. 7

The mechanistic scheme of the regulation of GC proliferation and the cell cycle by acetylated DNMT1 by changing the methylation of key factors