Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Genome-wide DNA methylation and transcriptome sequencing analyses of lens tissue in an age-related mouse cataract model

  • Yuzhu Hu ,

    Contributed equally to this work with: Yuzhu Hu, Dongmei Su, Yue Zhang

    Roles Data curation, Formal analysis, Investigation, Software, Visualization, Writing – original draft

    Affiliation Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Dongmei Su ,

    Contributed equally to this work with: Yuzhu Hu, Dongmei Su, Yue Zhang

    Roles Conceptualization, Methodology, Project administration, Supervision, Writing – review & editing

    Affiliations Department of Genetics, Health Department, National Research Institute for Family Planning, Beijing, China, Graduate School, Peking Union Medical College, Beijing, China

  • Yue Zhang ,

    Contributed equally to this work with: Yuzhu Hu, Dongmei Su, Yue Zhang

    Roles Data curation, Investigation, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Yanjiang Fu,

    Roles Resources

    Affiliation Daqing Ophthalmology Hospital, Daqing, Heilongjiang, China

  • Sijia Li,

    Roles Resources, Software, Validation

    Affiliation Hongqi Hospital of Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Xiaoya Chen,

    Roles Data curation, Formal analysis, Investigation, Visualization

    Affiliation Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Xiao Zhang,

    Roles Data curation, Investigation

    Affiliation Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Shunfei Zheng,

    Roles Data curation, Investigation

    Affiliation Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

  • Xu Ma ,

    Roles Writing – review & editing

    hushanshan@mdjmu.edu.cn (SH); jswkysgc@126.com (XM)

    Affiliations Department of Genetics, Health Department, National Research Institute for Family Planning, Beijing, China, Graduate School, Peking Union Medical College, Beijing, China

  • Shanshan Hu

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing

    hushanshan@mdjmu.edu.cn (SH); jswkysgc@126.com (XM)

    Affiliation Hongqi Hospital of Mudanjiang Medical University, Mudanjiang, Heilongjiang, China

Abstract

DNA methylation is known to be associated with cataracts. In this study, we used a mouse model and performed DNA methylation and transcriptome sequencing analyses to find epigenetic indicators for age-related cataracts (ARC). Anterior lens capsule membrane tissues from young and aged mice were analyzed by MethylRAD-seq to detect the genome-wide methylation of extracted DNA. The young and aged mice had 76,524 and 15,608 differentially methylated CCGG and CCWGG sites, respectively. The Pearson correlation analysis detected 109 and 33 differentially expressed genes (DEGs) with negative methylation at CCGG and CCWGG sites, respectively, in their promoter regions. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses showed that DEGs with abnormal methylation at CCGG sites were primarily associated with protein kinase C signaling (Akap12, Capzb), protein threonine kinase activity (Dmpk, Mapkapk3), and calcium signaling pathway (Slc25a4, Cacna1f), whereas DEGs with abnormal methylation at CCWGG sites were associated with ribosomal protein S6 kinase activity (Rps6ka3). These genes were validated by pyrosequencing methylation analysis. The results showed that the ARC group (aged mice) had lower Dmpk and Slc25a4 methylation levels and a higher Rps6ka3 methylation than the control group (young mice), which is consistent with the results of the joint analysis of differentially methylated and differentially expressed genes. In conclusion, we confirmed the genome-wide DNA methylation pattern and gene expression profile of ARC based on the mouse cataract model with aged mice. The identified methylation molecular markers have great potential for application in the future diagnosis and treatment of ARC.

Introduction

A cataract is partial or complete opacity of the lens of the eye. Cataracts are an acquired conditions among adults, and the most prevalent type is age-related cataract (ARC) [1]. The lens is formed from ectodermal tissue and contains lens epithelial cells (LECs) that produce lens fibers throughout life, and becomes more compact and thicker with age [2]. Because LECs retain intact nuclei and organelles, they help to maintain lens metabolic homeostasis [3]. LECs produce cysteine and glutathione to maintain the lens environment in the presence of harmful external stimuli [4]. Cataract surgery has improved, and most patients recover vision quickly. Despite this, cataracts remain a serious public health issue and will become more important as the global population and life expectancy increase [5].

Epigenetics provides a new explanation for ARC [6]. Epigenetics pertains to heritable alterations in genomic structure that occur without any changes in the DNA sequence. Investigations of epigenetic mechanisms encompass DNA methylation, histone modifications, and microRNAs. DNA methylation is a major mechanism for epigenetic regulation of genes [7]. Changes in the expression of DNA repair genes have been linked to the pathogenesis of ARC [8]. Human lens development requires the expression of multiple important genes, and DNA methylation regulates the expression of many development-related genes. Environmental exposures also alter DNA methylation patterns. Therefore, DNA methylation modifications may be associated with the pathogenesis of cataracts that are influenced by age and environmental factors, as well as lens development [7]. Abnormal methylation is an important characteristic of epigenetics and a crucial factor that can lead to cataracts and other ocular disorders [6]. Therefore, a comprehensive exploration of the particular mechanism of DNA methylation in the progression of ARC is of great significance.

In this study, the experimental procedure and design analysis is shown in Fig 1. Specifically, we first identified methylation across the entire genome in the anterior lens capsule membranes of young and aged mice by MethylRAD-seq. Then, we used the previously published transcriptome sequencing data for the same disorder in the same model [9] to identify differentially expressed genes (DEGs) with negative methylation in their promoter regions. Next, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses on the DEGs with negative methylation. We verified gene methylation levels by pyrophosphate sequencing to identify novel therapeutic targets using bioinformatics tools.

thumbnail
Fig 1. Experimental procedure and design analysis.

DNA samples were collected from the anterior lens capsule membrane tissues of three young mice (8 months old) and three aged mice (20 months old) for genome-wide methylation analysis. Subsequently, GO and KEGG analyses were conducted on the DMGs at CCGG and CCWGG sites. Next, the distribution of DNA methylation sites in different functional regions was explored. Then, a joint analysis of methylation and previous transcriptome data was carried out. After that, GO and KEGG enrichment analyses were performed on the DEGs with differential methylation at CCGG and CCWGG sites in the promoter region. Finally, pyrosequencing was used to validate the candidate genes with five samples from the young group (normal group,8 months old) and five samples from the natural aging group (ARC group,20 months old), thereby confirming the accuracy of the methylation sequencing data.

https://doi.org/10.1371/journal.pone.0316766.g001

Materials and methods

1. Mice and isolation of the lens anterior capsule membrane

We raised 40 wild-type young mice (8 months old) at 22 ± 2°C, 55 ± 5% humidity, and 12-h/12-h light/dark cycle. The lenses of the young mice were observed and found to be transparent. Ten of the young mice were chosen randomly as the normal control group. To simulate the occurrence of cataracts in the natural aging state of individuals, we raised 40 mice for 20 months under the same conditions. The lenses of the aged mice were observed. We found that the cataracts gradually became turbid with increasing age using a slit lamp. We selected 10 mice with cataracts that had Grade 3 turbidity as the ARC group [9].

The experimental mice were anesthetized with 2.0% isoflurane inhalation and samples were taken. Then, the mice were euthanized by CO2 asphyxiation. We collected DNA samples from the anterior lens capsule membrane tissues (only contain LECs) of three young mice (8 months old) and three aged mice (20 months old) for genome-wide methylation analysis as described previously [1013]. For young mice, the number of anterior lens capsule samples is three, that is, n = 3. They are in the healthy group and are named SQ-1, SQ-2, and SQ-3 respectively. For naturally aging mice, the number of anterior lens capsule samples is also three. n = 3 is the disease group, that is, suffering from cataracts. They are named SL-1, SL-2, and SL-3 respectively.All the animal experiments were approved by the Animal Protection and Use Committee of Mudanjiang Medical University (No. 20220228–41) and were in line with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research. The experimental procedure also complied with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.

2. DNA extraction and MethylRAD experimental process

Genomic DNA from the anterior lens capsule membrane tissue was extracted using the TIANamp Micro DNA Kit according to the manufacturer’s instructions. Libraries were constructed following the guidelines established by Wang et al. [14]. Four units of the methylation-modifying endonuclease FspEI (New England BioLabs, cat. no. R0662L) were used to break down 200 ng of DNA at 37°C for 4 h. Then, the DNA was joined to an aptamer. To initiate the ligation procedure, 5 μl of fragmented DNA was added to 10 μl of ligation master mixture (800 units of T4 DNA ligase (New England Biolabs), 1 mM of adenosine triphosphate (ATP), and 0.2 μM of each of the two aptamers). Subsequently, the amplified PCR product was purified by 8% polyacrylamide gel electrophoresis. The approximately 100-base pair band was extracted, and the DNA was spread at 4°C for 12 h, then rinsed off the gel using nucleic acid-free water. The products were amplified again by PRC. The PCR products were purified using a QIAquick PCR Purification Kit (Qiagen cat. No.28104.), eluted with 15 μl of purified water, and measured by Qubit (Qubit 2.0 fluorometer (Invitrogen)). to quantify the product. The purified PCR product were sequenced on an Illumina Nova PE150 platform (Shanghai OE Biotechl Co.,Ltd.,Shanghai,China). FspEI has different preferences for sequences on each side of methylation-modified cytosines, and therefore the types of sites analyzed by FspEI are different. For a mCG site, FspEI analyzes only the CCGG site. For a mCHG site, where H is either A, C, or T, when H is base C, the data processing recognizes the site as CCGG. Therefore, FspEI only provides results for CCAGG and CCTGG for the mCHG site, denoted by CWG, where W is either A or T.

3. Alignment and quality control

First, adapter trimming and quality filtering were conducted on the raw reads. A custom Perl script was used to remove reads with >8% ambiguous bases (N), low-quality reads (15% of the bases had a Phred quality score <30), or reads that lacked enzyme restriction sites. Subsequently, Bowtie 2 (v.2.3.4.1) [15] was used to align the reads that contain enzyme restriction sites (enzyme reads) to the mouse genome (GRCm39) (https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/001/635/GCF_000001635.27_GRCm39/GCF_000001635.27_GRCm39_genomic.fna.gz) [16] using the default settings.The default settings are as follows:the parameter is set as:—no-unal, that is, records that do not meet the alignment requirements are not output. Since the reference sequence has repeat regions, reads in the repeat regions can be aligned to multiple positions, making it difficult to determine the actual position of the methylation site. After random alignment, reads with multiple alignments are deleted. Obtain Clean Reads that are aligned to a unique position on the reference sequence. Perform electronic digestion on the genome and count the number of various enzyme digestion sites on each chromosome (or scaffold). Next,The distributions of cytosine methylation sites across the mouse chromosomes and other functional genomic regions were analyzed. The normalized read depth metric RPM (reads per million) was used to obtain the relative expression level of each methylation site (CpG).The sequencing depth of methylation tags can reflect the methylation level of sites. Since the FspEI enzyme can only recognize methylated sites and cut them, theoretically, the higher the methylation level of this site, the more tags are cut off. Therefore, the higher the tag depth, the higher the methylation level. However, the depth needs to be normalized by the amount of sequencing data of the sample. Therefore, there is the RPM method, that is RPM = [(read coverage at each site/number of high-quality reads in the library) × 1,000,000]. The R package DESeq [17] was used to assess changes in the methylation level in accordance with the sequencing depth of each site within the relative quantitative outcomes of methylation.The sequencing data were used to determine the fold change (log2FC >1), and P-value <0.05 among diverse sites was considered to be statistically significant. (GO) and KEGG functional enrichment analyses were carried out on the differentially methylated genes (DMGs).

4. GO and KEGG functional enrichment analyses

The functional enrichment analyses were performed using Python scripts to identify methylation-related GO(http://geneontology.org/) terms and KEGG (http://www.genome.jp/kegg/) pathways associated with the DMGs. The hypergeometric distribution was used to determine the significance of the annotations of the DMGs to each GO term and each KEGG pathway, against the complete list of protein coding genes in the mouse genome as the background.

5. DNA methylation combined with transcriptome sequencing

To refine the methodology, transcriptome sequencing was performed using a NEBNext® Ultra™ II Directional RNA Library Prep Kit for low-input samples. Library quality was assessed with a 2100 Bioanalyzer, and sequencing was performed on an Illumina HiSeq X Ten platform to obtain paired-end reads. HISAT2 was used to align the reads to the reference mouse genome, and FPKM values were used to normalize gene expression. The R package DESeq identified DEGs with P <0.05(the P value here is corrected by FDR) and fold change >2. To explore the relationship between gene expression and DNA methylation, we performed a Pearson correlation analysis between DEGs with abnormal methylation at CCGG and CCWGG sites in their promoters and their mRNAs. Subsequently, using a P-value threshold of <0.05(the P value here is corrected by FDR) and a correlation coefficient of < -0.8, we integrated the analyses of DMGs and DEGs within promoter regions.

6. Pyrosequencing verification

DNA was extracted using a TIANamp Micro DNA Kit. Methylation was carried out using a Qiagen EpiTect Bisulfite Kit (Qiagen59104), ensuring a comprehensive and standardized approach. Primer design was performed using PyroMark Assay Design 2.0. After PCR amplification, pyrosequencing was conducted on an advanced PyroMark Q48 Autoprep platform (Qiagen). The methylation profiles of individual sites were systematically analyzed using the proprietary PyroMethylation Analyzer software, thereby facilitating an automated and efficient assessment. The primer sequences used in this study are listed in Table 1.

thumbnail
Table 1. Primer sequences and product lengths in each assay.

https://doi.org/10.1371/journal.pone.0316766.t001

7. Statistical analysis

Data are presented as the mean ± standard deviation for at least three experiments. GraphPad Prism 8 software was used to process the data. To evaluate the differences between the control and ARC groups we used a t-test. For multiple comparisons, one-way analysis of variance (ANOVA) was used to detect significant differences. Differences were considered statistically meaningful for P-values <0.05.And the P-values are all P-values after FDR correction.

Results

1. MethylRAD-seq quality control

We performed methylation sequencing by MethylRAD-seq, which specifically recognizes alterations in cytosine C5 methylation on DNA using the FspEI enzyme. MethylRAD-seq systematically performs digestion at predetermined intervals, and selectively gathers DNA fragments of 13–17 base pairs that are methylated across the entire genome. This process facilitates both qualitative and relative quantitative assessments of methylation sites. After re-filtering the raw data, paired-end sequencing reads are mapped following the criteria outlined in the Materials and Methods section, to obtain diverse read types, including enzymatic and clean reads for each individual sample as shown in Table 2. Each sample yielded an average of approximately 78.02 million reads. The FspEI treatment resulted in an average of 32.25 million enzymatic reads and 27.94 million cleaned reads (Table 2). The depth of sequencing coverage for MethylRAD-seq sites, specifically CCGG and CCWGG motifs, for each sample is visualized by the box plots in Fig 2. The samples contained an average of 830,000 CCGG and 380,000 CCWGG motifs. The mean coverage depth was 21 for CCGG and 7.97 for CCWGG (Table 3). The enzymatic reads for each sample were aligned against the established reference sequences. To ensure the reliability of the methylation status, only sites with a sequencing depth of three-fold or more were considered validly methylated [18,19]. The sequencing reads that satisfied these prerequisites were used for subsequent analyses.

thumbnail
Fig 2. Depth of coverage statistics for methylation sites.

Green color indicates a box plot representing the sequencing depth of CCGG methylation sites in each sample. Orange color stands for the boxplot of the sequencing depth of CCWGG methylation sites in every sample.

https://doi.org/10.1371/journal.pone.0316766.g002

thumbnail
Table 3. Statistics on the depth of coverage for methylation sites.

https://doi.org/10.1371/journal.pone.0316766.t003

2. Genome-wide methylation analysis by MethylRAD-seq

To identify differentially methylated sites(DMSs), depth information (base mean estimated expression) was normalized using DESeq software, differential multiplicity was determined, and the number of reads was checked for significance using the negative binomial distribution test. Credible methylation sites with P <0.05 and multiplicity of difference >2 are shown in Fig 3. The tissue samples from young and aged mice had 76,524 and 15,608 DMSs at CCGG and CCWGG sites (S1 and S2 Tables), respectively; 27,337 were hypermethylated and 49,187 were hypomethylated at the CCGG sites (Fig 3A and 3C), and 6,327 were hypermethylated and 9,281 were hypomethylated at the CCWGG sites (Fig 3B and 3D).

thumbnail
Fig 3. Genome-wide methylation analysis of the anterior lens capsule membranes of young and old mice by methyl-rad-seq.

(a) Histogram of DMSs at CCGG sites. Red: Total number of DMSs; yellow: Hypermethylated; purple: Hypomethylated. (b) Histogram of DMSs at CCWGG sites. Red: Total number of DMSs; yellow: Hypermethylated; purple: Hypomethylated.(c) Volcano plot of DMSs at CCGG sites. Gray: Non-DMS; red: Hypermethylated; green: Hypomethylated. (d) Volcano plot of DMSs at CCWGG sites. Gray: Non-DMS; red: Hypermethylated; green: Hypomethylated.All DMSs at CCGG and CCWGG sites can be found in S1 and S2 Tables.

https://doi.org/10.1371/journal.pone.0316766.g003

3. GO and KEGG analyses of the differentially methylated genes at CCGG and CCWGG sites

We carried out a comprehensive GO functional enrichment analysis for the three main categories of GO, namely biological process, cellular component, and molecular function.For the entries that are significantly enriched (P < 0.05) in each category, they are all ranked according to their P-values. It should be noted that the larger the -log10P value, the higher the significance of the enrichment of this functional category in the gene set. During the GO enrichment analysis process, we screened out the top 30 GO entries. The specific operation is to first select the GO entries with the number of genes related to the differentially methylated sites greater than 2 in these three functional classifications, and then sort them from large to small according to the -log10 P value corresponding to each entry, and finally select 10 entries from each classification. DMGs at CCGG sites were annotated with 106,297 GO terms; 60,246 under biological process, 36,911 under cellular component, and 9,140 under molecular function. DMGs at CCWGG sites were annotated with 46,895 GO terms; 27,346 under biological process, 36,911 under cellular component, and 9,140 under molecular function. The terms assigned to the top 10 DMGs in each category were selected for further analysis. For the CCGG-based DMGs, the biological process terms were primarily linked to regulation of protein localization to the plasma membrane (e.g., Vamp8, Cdkl5, Pik3r2), the cellular component terms were mainly associated with actin filament (e.g., Tek, Capzb, Rac2), and the molecular function terms was mainly associated with store-operated calcium channel activity (e.g., Trpc7, Trpc1, Trpm8) as shown in Fig 4A. For the CCWGG-based DMGs, the biological process terms were mainly associated with cellular sodium ion homeostasis (Atp1b3, Atp1b1, Slc8a3), the cellular component terms was mainly associated with the Weibel-Palade body (Col6a2, Vwf, Bmper), and the molecular function terms were mainly related to inositol 1,4,5 trisphosphate binding (Plekha7, Itpr3, Itpr2, Itpr1) as shown in Fig 4B.

thumbnail
Fig 4. GO and KEGG analyses of DMGs with differentially methylated CCGG and CCWGG sites.

(a) Histogram of GO functional classification of (DMGs where differentially methylated CCGG sites are located. In the figure, the horizontal axis represents the GO entry name, which is actually the description of the gene function; the vertical axis is the -log10P value, and the larger the value, the higher the significance of the enrichment of this functional category in the gene set. The GO enrichment analysis was used to screen out the top 30 GO terms. Among these three functional classifications, the GO terms with the number of genes related to the corresponding differentially methylated sites exceeding 2 were selected. Then, they were sorted in descending order according to the -log10 P value corresponding to each term. Finally, 10 terms were taken from each classification.(b) Histogram of GO functional classification of DMGs where differentially methylated CCWGG sites are located.The results of GO analysis show the DMGs related to molecular functions (labeled in red), cellular components (labeled in blue), and biological processes (labeled in green).(c) Bubble chart of KEGG enrichment analysis of DMGs where differentially methylated CCGG sites are located.The bubble chart of the top 20 KEGG enrichment analysis (screening the pathway entries with the number of corresponding differentially expressed genes greater than 2 and sorting them from large to small according to the -log10 P value corresponding to each entry). In the figure, the X-axis Enrichment Score is the enrichment score, the larger the bubble, the more differential protein-coding genes the entry contains, the bubble color changes from purple-blue-green-red, and the smaller the enrichment pvalue value, the greater the significance. The Y-axis is the descriptive information of the enriched pathways.(d) Bubble chart of KEGG enrichment analysis of DMGs where differentially methylated CCWGG sites are located.

https://doi.org/10.1371/journal.pone.0316766.g004

To predict the pathways in which DMGs may participate, we performed KEGG enrichment analysis (Fig 4C and 4D). Top 20 of KEGG Enrichment Analysis: Screen the pathway entries with the number of corresponding differentially methylated sites genes greater than 2 and sort them from large to small according to the -log10P value corresponding to each entry. In the figure, the Enrichment Score on the X-axis is the enrichment score. The larger the bubble, the more differential protein-coding genes the entry contains. The bubble color changes from purple to blue, green, and red. The smaller the enrichment P value, the greater the significance. The Y-axis is the descriptive information of the enriched pathways. For the CCGG-based DMGs, the enriched pathways were primarily biochemical signaling pathways, notably the mitogen-activated protein kinase (MAPK) signaling cascade (e.g., Mapk1, Mapk3, Tek, Rac2), the Ras-related protein 1(Rap1) signaling pathway (e.g., Actb, Pik3r2, Tek, Actg1), the calcium signaling pathway (e.g., Adcy1, Adora2a, Fgf9, Tek, Slc25a4), the PI3K-Akt signaling pathway (e.g., Akt1, Tek, Akt2), the Ras signaling pathway (e.g., Map2k1, Tek, Pak1), and the Hippo signaling pathway (e.g., Gsk3b, YAP1) as shown in Fig 4C. For the CCWGG-based DMGs, the enriched pathways were predominantly biochemical signaling pathways, specifically the MAPK signaling pathway (e.g., Mapk1, Mapk3), the Rap1 signaling pathway (e.g., Mapk1, Mapk3, Tiam1, Rap1a), the calcium signaling pathway (e.g., Cacna1f, Itpr1, Itpr2, Itpr3, Slc8a3), the Wnt signaling pathway (e.g., Ror1, Wnt7b), the Hippo signaling pathway (e.g., Yap1, Wnt7b), the Ras signaling pathway (e.g., Mapk1, Mapk3), the cAMP signaling pathway (e.g., Map2k1, Atp1b3, Atp1b1, Edn2), and Type II diabetes mellitus (e.g., Hk1, Mafa) as shown in Fig 4D. Next, we conducted a combined analysis of individual DMGs that were chosen based on the GO and KEGG analyses. The selected consensus CCGG-based DMGs included Rac2, Tek, and Pik3r2, and the selected consensus CCWGG-based DMGs included Atp1b3, Atp1b1, Itpr1, Itpr2, and Itpr3.

4. Distribution of DNA methylation sites in diverse functional regions

We analyzed the distribution patterns of methylated CCGG and CCWGG sites across various genomic regions, namely the 5’ untranslated region (UTR), 3’ UTR, upstream regions of genes, exons, introns, and intergenic spaces (Fig 5A and 5B). Most of the DNA methylation sites were predominantly localized in intronic regions, followed closely by exons, intergenic spaces, and upstream regions. The proportions of methylated sites were minimal in both the 5’ and 3’ UTRs. We also examined the distribution patterns of the 2-kilobase (kb) region upstream of the transcription start site (TSS), the gene body itself, and the TTS. The regions that flank the TSS are important in regulating gene expression. The DNA methylation levels at CCGG and CCWGG sites in the gene body and the TTS had comparable patterns between the control and ARC groups (Fig 5C and 5D). The DNA methylation levels were notably lower in the 2-kb region upstream of the TSS and markedly higher at the TSS. Methylation levels gradually increased throughout the gene bodies, and decreased toward the TTS. The distribution of DNA methylation at CCGG sites mirrored the TSS methylation pattern, whereas Contrastingly, the distribution of methylation at CCWGG sites had a distinct pattern. Specifically, methylation levels at CCWGG sites were initially lower in the 2-kb upstream region of the TSS, then abruptly increased proximal to the TSS (left panel, Fig 5C). Three pronounced peaks of elevated DNA methylation were detected at CCWGG sites (left panel, Fig 5D).

thumbnail
Fig 5. Distribution of DNA methylation sites in different functional regions.

(a) Histogram showing the distribution of CCGG-based DMGs on different functional regions of the genome. Y-axis, methylation site count; X-axis, genome constituents. (b) Histogram showing the distribution of CCWGG-based DMGs on different functional regions of the genome. Y-axis, methylation site count; X-axis, genome constituents. (c) Distribution of CCGG-based DMGs around the gene body. X-axis, adjacent position; Y-axis, normalized read number. (d) Distribution of CCWGG-based DMGs around the gene body.

https://doi.org/10.1371/journal.pone.0316766.g005

5. MethylRAD-seq combined with transcriptome sequencing

We identified 3,856 DEGs between the lens tissues from aged and young mice (P-value threshold <0.05, fold change >2.n compared the young group) by analyzing the previously published transcriptome sequencing data from the identical disease model. Among the identified DEGs, 917 were up-regulated and 2,939 were down-regulated, as reported previously [9]. DNA methylation has a pivotal role in regulating gene expression by modulating chromatin architecture, DNA configuration, stability, and protein–DNA interactions. Traditionally, methylation in promoter regions is considered to have a profound effect on gene expression regulation [20], prompting our focus on methylation differences specifically in these regions. Consequently, we integrated the RNA-seq transcriptome data [9] with an analysis of the DNA methylation patterns in promoter regions. To visualize the correlation between DNA methylation and gene expression levels, we constructed four-quadrant plots for CCGG and CCWGG methylation sites (Fig 6A and 6D). Red indicates the negatively correlated sites where the FC of differentially expressed RNAs is greater than 2; blue indicates the positively correlated sites where FC > 2. The first quadrant (upper right, blue dots) represents the genes with both upregulated gene expression and methylation levels, which may synergistically promote related biological processes. The second quadrant (upper left, red dots) represents the genes with downregulated gene expression but upregulated methylation levels, suggesting that there may be a mechanism where methylation inhibits expression. The third quadrant (lower left, blue dots) represents the genes with both downregulated gene expression and methylation levels, showing a consistent trend of change. The fourth quadrant (lower right, red dots) contains the genes with upregulated gene expression but downregulated methylation levels, and demethylation may promote their expression.

thumbnail
Fig 6. Results of combined analysis of DMGs at DMSs in the promoter region and DEGs.

(a) A four-quadrant diagram for differentially expressed genes and genes with differential methylation levels at CCGG sites within the promoter region. Red represents negatively correlated sites where the FC of differentially expressed RNA is greater than 2; blue denotes positively correlated sites with FC > 2.The genes in the first quadrant (upper right, blue dots) represent those with both up-regulated gene expression and methylation levels; the genes in the second quadrant (upper left, red dots) indicate those with down-regulated gene expression but up-regulated methylation levels; the genes in the third quadrant (lower left, blue dots) are those with both down-regulated gene expression and methylation levels; the genes in the fourth quadrant (lower right, red dots) contain those with up-regulated gene expression but down-regulated methylation levels. (b) Heat map for hierarchical clustering analysis of differentially expressed genes in the promoter region. On the horizontal axis, each column represents a sample; on the vertical axis, each row represents a gene. The gene expression levels are mapped to the color scale, with red representing high expression and blue representing low expression. On the left is the dendrogram of the cluster analysis of the gene expression patterns.(c) Genes with differential methylation at distinct CCGG methylation sites in the promoter region. On the horizontal axis, each column represents a sample; on the vertical axis, each row represents a gene. The gene methylation levels are mapped to the color scale, with red representing high methylation and blue representing low methylation. On the left is the dendrogram of the cluster analysis of the gene methylation patterns, which shows the similarities between genes.(d) A four-quadrant plot for differentially expressed genes and genes with differential methylation levels at CCWGG sites within the promoter region. Red represents negatively correlated sites where the FC of differentially expressed RNA is greater than 2; blue denotes positively correlated sites with FC > 2. (e) Heat map for hierarchical clustering analysis of differentially expressed genes in the promoter region. Red indicates up-regulation, and blue indicates down-regulation. (f) Genes with differential methylation at diverse CCWGG methylation sites in the promoter region. Blue indicates hypomethylation, and red signifies hypermethylation.DEGs with differential methylation at CCGG/CCWGG sites in the promoter region can be found in S3 and S4 Tables. The Protein-Protein Interaction (PPI) graph of these DEGs with differential methylation at CCGG/CCWGG sites in the promoter region can be found in S1 and S2 Figs, and the corresponding protein interaction scores can be found in S5 and S6 Tables.

https://doi.org/10.1371/journal.pone.0316766.g006

We filtered out 109 and 33 DEGs that had differential methylation at CCGG and CCWGG sites (S3 and S4 Tables), respectively, in their promoter regions by applying thresholds of P-value <0.05 and correlation coefficient <−0.8. The DEGs that have CCGG sites in their promoter regions are displayed in Fig 6B. On the abscissa, each column represents a sample; on the ordinate, each row represents a gene. The gene expression levels are compared against the color scale, with red representing high expression and blue representing low expression. On the left is the dendrogram of the cluster analysis of the gene expression patterns, which shows the similarities between genes. The shorter the branch is, the more similar the expressions of the genes are in different samples.The genes that have differentially methylated CCGG sites in their promoter regions are shown in Fig 6C.On the abscissa, each column represents a sample; on the ordinate, each row represents a gene. The gene methylation levels are compared against the color scale, with red representing high methylation and blue representing low methylation. On the left is the dendrogram of the cluster analysis of the gene methylation patterns, which shows the similarities between genes. The shorter the branch is, the more similar the methylation states of the genes are in different samples. The DEGs that have CCWGG sites in their promoter regions are shown in Fig 6E (red indicates up-regulation, blue indicates down-regulation). The genes that have differentially methylated CCWGG sites in their promoter regions are presented in Fig 6F (blue indicates hypomethylation, red indicates hypermethylation). These results indicate an inverse relationship between gene expression and DNA methylation levels, suggesting that they display opposing trends.

6. GO and KEGG enrichment analysis for DEGs with differentially methylated CCGG and CCWGG sites

An integrated the analyses of the promoter regions of the DMGs and DEGs identified 109 and 33 genes at CCGG and CCWGG sites, respectively, using a P-value threshold of <0.05 and correlation coefficient of <−0.8, then GO and KEGG functional enrichment analyses were performed on the identified genes (S7S10 Tables). Among the top 30 GO terms for CCGG-based DMGs, regulation of protein kinase C signaling was highly enriched under biological process (e.g., Akap12, Capzb), and calmodulin binding (e.g., Mapkapk3, Mknk2), calmodulin-dependent protein kinase activity (e.g., Mapkapk3, Mknk2), protein threonine kinase activity (e.g., Dmpk, Mapkapk3, Rps6ka3, Mknk2), and SH2 domain binding (Sqstm1, Syp) were highly enriched under molecular function (Fig 7A and Table 4). For CCWGG-based DMGs, canonical Wnt signaling pathway (e.g., Med12, Wnt7b) and homeostatic process (specifically Wnt7b) were highly enriched under biological process, and nuclear receptor activity (Srebf1, Esrrg) and ribosomal protein S6 kinase activity (Rps6ka3) were highly enriched under molecular function (Fig 7B and Table 5). To predict the pathways involving DEGs with differentially methylated CCGG and CCWGG sites in their promoter regions, we performed KEGG enrichment analysis (S9 and S10 Tables). The CCGG-based DMGs were primarily involved in MAPK signaling pathway (e.g., Mapkapk3, Mknk2), cAMP signaling pathway (e.g., Cacna1f, Atp2b3), and calcium signaling pathway (e.g., Slc25a4, Cacna1f) as shown in Fig 7C. The CCWGG-based DMGs were primarily involved in the mTOR signaling pathway (e.g., Rps6k, Wnt7b) and the HIF-1 signaling pathway (including Eno4) as shown in Fig 7C and Table 6. A combined analysis of the GO and KEGG terms and pathways showed that, at CCGG sites, the genes were predominantly associated with regulation of protein kinase C signaling (e.g., Akap12, Capzb), protein threonine kinase activity (e.g., Dmpk, Mapkapk3), and the calcium signaling pathway (e.g., Slc25a4, Cacna1f), whereas, at CCWGG sites, they were associated with ribosomal protein S6 kinase activity (Rps6ka3) (Fig 7D and Table 7).

thumbnail
Fig 7. Perform GO and KEGG enrichment analysis for DEGs with differentially methylated levels at CCGG and CCWGG sites.

GO annotation shows DMGs related to distinct molecular functions (MF, red), cellular components (CC, blue), and biological processes (BP, green). (a) GO enrichment analysis of DEGs with differentially methylated CCGG sites in promoter region; (b) GO enrichment analysis of DEGs with differentially methylated CCWGG sites in promoter region; (c) KEGG enrichment analysis of DEGs with differentially methylated CCGG sites in promoter region; (d) KEGG enrichment analysis of DEGs with differentially methylated CCWGG sites in promoter region.Significant results of GO enrichment analysis of DEGs with differentially methylated CCGG/CCWGG sites in promoter region (P < 0.05, correlation coefficient < -0.8) in S7 and S8 Tables. Significant results of KEGG enrichment analysis of DEGs with differentially methylated CCGG/CCWGG sites in promoter region (P < 0.05, correlation coefficient < -0.8) in S8 and S10 Tables.

https://doi.org/10.1371/journal.pone.0316766.g007

thumbnail
Table 4. The top 30 GO enrichment analysis was conducted on DEGs with differentially methylated levels at CCGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.t004

thumbnail
Table 5. The top 30 GO enrichment analyses for DEGs that have differentially methylated levels at CCWGG sites within the promoter region.

https://doi.org/10.1371/journal.pone.0316766.t005

thumbnail
Table 6. Top 20 KEGG enrichment analyses are done for DEGs with varying methylation levels at CCGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.t006

thumbnail
Table 7. Top 20 KEGG enrichment analyses for DEGs with different methylation levels at CCWGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.t007

7. Pyrophosphate sequencing validation

To evaluate the reliability and precision of the MethylRAD-seq data, we validated four genes (Capzb, Dmpk, Slc25a4, Rps6ka3) by pyrophosphate sequencing. These genes were selected based on the comprehensive analysis of methylation and transcriptome (Fig 7A). We performed pyrosequencing verification on five samples from the young group (normal group), that is, n = 5, named Young, and five samples from the naturally aging group (ARC group), named Aged. We found that the methylation status of these genes generally concurred with the sequencing outcomes obtained from the MethylRAD, confirming the reliability of the initial findings. Specifically speaking:the methylation levels of Dmpk (Fig 8B) and Slc25a4 (Fig 8C) were lower in the ARC group than they were in the control group, whereas the methylation level of Rps6ka3 (Fig 8D) was higher in the ARC group than it was in the control group (The average methylation rate of Dmpk was 59% in the young mouse group and 32% in the aged mouse group, with a p-value of less than 0.01, indicating significant statistical significance. The average methylation rate of Slc25a4 was 15% in the young mouse group and 6% in the aged mouse group, with a p-value of less than 0.05, showing statistical significance. The average methylation rate of Rps6ka3 was 13.5% in the young mouse group and 18% in the aged mouse group, with a p-value of less than 0.05, also demonstrating statistical significance). These findings align with our integrated analysis. No notable distinction was detected in methylation level of Capzb between the two groups, prompting a further analysis of the remaining three genes.

thumbnail
Fig 8. Measurement of gene methylation using pyrophosphate sequencing.

(a) The bar charts showing the statistical results of the methylation levels of Dmpk, Slc25a4 and Rps6ka3 in the two groups of samples are presented. ("ns" indicates no statistical difference;*indicates P less than 0.05; ** indicates P less than 0.01.) (b-d) The representative results of pyrophosphate sequencing for the Young and Aged groups are presented.(b) The pyrosequencing peak plots of the methylation sites where the Dmpk gene is located in the young and aged mouse groups. The left figure represents the methylation level of the methylation site where the Dmpk gene is located in the young mouse group, and the right figure represents the methylation level of the methylation site where the Dmpk gene is located in the aged mouse group. The horizontal axis shows the order of added dNTPs, and the vertical axis represents the intensity of the fluorescent signal (the detection signal of the methylated fragment).(c) The pyrosequencing peak plots of the methylation sites where the Slc25a4 gene is located in the young and aged mouse groups. The height of the peaks reflects the signal intensity of the methylation sites. The higher the intensity, the higher the degree of methylation of the site in the sample.(d) The pyrosequencing peak plots of the methylation sites where the Rps6ka3 gene is located in the young and aged mouse groups. The sequencing quality corresponding to blue is the best, followed by yellow, and red indicates poor quality. If blue appears in a certain reaction, it means that the overall quality is within an acceptable range and there are no significant problems. The specific methylation site where the gene is located is marked by the gray shaded area.

https://doi.org/10.1371/journal.pone.0316766.g008

Discussion

Approximately 94 million individuals worldwide are blind or visually impaired, and cataract is the most prevalent cause of blindness [21]. Although cataract surgery is an increasingly prevalent treatment, high costs and postoperative complications still affect many patients. Data show that if the occurrence of ARC can be postponed by a decade, the number of surgeries will decrease by half [22]. Aging has become a major risk factor for various human diseases, including diabetes, cancer, cardiovascular diseases, and neurodegenerative diseases [2325]. Among them, the onset of ARC is also closely related to age [26]. Currently, no breakthrough in drug treatment for ARC is currently available, and the unclear mechanism of lens aging is the major limiting bottleneck. In epigenetics research, numerous aspects need to be considered, including DNA methylation, histone modification, and microRNAs [27]. Previous studies have suggested that DNA methylation modifications influenced by both age and the surrounding environment, likely play a pivotal role in the pathogenesis of cataracts and the developmental processes of the lens [28]. Aberrant methylation, as a fundamental aspect of epigenetics, is one of the key factors leading to eye diseases, especially cataracts. Notably, in senile cataracts, differential epigenetic patterns in DNA surrounding specific genes are found. Studies on cataract models have detected age-dependent methylation increase in the Klotho gene promoter [29]. At the same time, lens crystallins play a crucial role in maintaining lens transparency. Among them, α-crystallin, as the main structural protein component, accounts for 35% of all crystallins in the lens and acts as a molecular chaperone to prevent the aggregation of other crystallins [30,31]. In age-related nuclear cataracts, the level of α-crystallin is reduced, which is related to the hypermethylation of CpG islands in the CRYAA gene promoter. Treatment with DNA methylation inhibitors can restore CRYAA gene expression [32]. In addition, there is epigenetic silencing of nuclear factor erythroid 2-related factor 2 (Nrf2). Nrf2 is a transcriptional activator that can protect the lens but is negatively correlated with the Keap1 protein. The expression of Keap1 increases with age and can stimulate proteasome-mediated degradation of Nrf2 and inhibit the Nrf2-dependent antioxidant protection of the lens. Analysis of age-related cataract lenses shows significant demethylation of Keap1 and a decrease in Nrf2 [33]. These studies provide strong evidence for the importance of epigenetics in cataracts and bring new hope for the development of targeted treatment methods for cataracts.However, previous studies on the methylation of pathogenic genes for ARC focus on one site, multiple sites, or a single gene, while the research on genome-wide methylation is relatively lacking. Therefore, this article has significant advantages in conducting genome-wide methylation research on cataracts. Genome-wide methylation research can comprehensively and systematically discover epigenetic changes related to the occurrence and development of cataracts, which helps to understand its pathogenic mechanism and provides potential targets for precise treatment.

For ophthalmic diseases, DMSs also have diagnostic value. A comprehensive study identified significant variations in mitochondrial DNA methylation in the peripheral blood of patients with diabetes, differentiating patients with retinopathy from those without [3436]. It has also been shown that the hypomethylated state of the promoters of NLRP3, TGFβ1, MCP-1, and TNFSF2 may increase the risk of developing diabetic retinopathy [37]. This observation result shows that the methylation patterns of these genes have the potential to be biomarkers for early diagnosis and are of great value. At the same time, they can also provide key insights for targeted therapy. Drug intervention treatment for abnormal methylation sites can open up broad prospects for non-surgical treatment of ARC.

Genome-wide methylation studies can comprehensively and systematically discover epigenetic alterations associated with the occurrence and progression of cataracts, which assists in understanding the pathogenic mechanism and offers potential targets for precise treatment. Hence, in-depth exploration of the mechanism by which DNA methylation affects the pathogenesis of ARC is of great significance. Changes in methylation patterns are expected to indicate potential therapeutic targets. Regulating methylation levels can improve gene expression and open up new avenues for drug treatment. In addition, DMG sites have early diagnostic significance and can improve diagnostic accuracy and identify high-risk populations. In this study, we used MethylRAD-seq technology to detect genome-wide methylation and conducted a combined genome-wide analysis of methylation with transcriptomic analysis to expand the research scope of ARC, contribute to new perspectives and achievements to this field, and promote new breakthroughs in cataract research and treatment.

Numerous animal model studies have discovered 180 cataract-related genes in mice [38,39]. Ultraviolet induction or selenite induction methods have been used to construct mouse cataract models. Although models can be obtained in a short time, it is difficult to truly simulate the natural aging process of the lens [4043]. Therefore, in this study, we constructed a naturally aging mouse model that simulates the disease process of human ARC more accurately and truly than previous models. Our model reflects the dynamic changes of the natural aging of the lens with time, thereby providing a more reliable reference for studying the pathogenesis of human cataracts. This mouse model was successfully verified in our previous research to be capable of simulating the progression of human ARC. Importantly, this model fills the deficiency in the study of DNA methylation sequencing in animal models of cataract and offers a new tool for a profound exploration of the genetic and epigenetic mechanisms of cataracts. By conducting DNA methylation analysis of the anterior lens capsule of naturally aging mice, alterations in the regulation of ARC-related gene expression were discovered. This provides a foundation for the early diagnosis and intervention of ARC.

In this study, we conducted MethylRAD-seq analysis on the anterior lens capsules of young and aged mouse cohorts and identified 76,524 and 15,608 DMSs at CCGG and CCWGG sites, respectively (S1 and S2 Tables). The comprehensive GO and KEGG functional enrichment analyses indicated the functional significance of the DMGs located at these sites. The top 30 GO terms for overall DMGs at CCGG sites were mainly related to the activity of store-operated calcium channels (Trpc7, Trpc1, Trpm8) (Fig 4A). The emergence of cataracts is strongly associated with calcium pumps. Strict regulation of Ca2+ influx into LECs is crucial because the calcium concentration has profound effects on maintaining lens homeostasis and growth.When there is an intracellular overload of Ca2+, adverse events are triggered. Properly maintaining the Ca2+ level can prevent cataracts [44], which indicates the accuracy of the GO analysis of overall DMGs at CCGG sites related to calcium channel activity.

The top 30 GO terms for overall DMGs at CCWGG sites were mainly related to cellular sodium homeostasis (Atp1b3, Slc8a3, Atp1b1) (Fig 4B). Previous studies have confirmed the close relationship between cataracts and sodium pumps [45]. Oxidative stress caused by disparity between oxidizing agents and antioxidants and disturbance of redox signaling can damage tissues and cells [46]. Oxygen free radicals are involved in lens aging [47]. Antioxidant enzymes can prevent the deterioration of Na(+)-K(+)-ATPase-dependent pumps [45]. Deterioration of sodium pumps is important in cataract development, confirming the accuracy of the sequencing result.

The top 20 primary KEGG pathways identified at CCGG sites were predominantly linked to the MAPK signaling cascade and calcium signaling pathway, featuring genes such as Mapk1, Mapk3, Fgf9, and Slc25a4 (Fig 4C). MAPKs have been classified into four subfamilies, namely ERK1/2, JNK1/2/3, P38, and ERK5. MAPKs participate in cell differentiation, proliferation, and apoptosis processes. Lens epithelial cell apoptosis, a prevalent mechanism underlying cataract development, is triggered by oxidative stress. Suppression of phosphorylation of ERK1/2, JNK, and P38 shows the relationship between the MAPK pathway and cataracts [48]. Additionally, intracellular Ca2+ imbalance causes LEC death and leads to cataracts. The top 20 KEGG pathways from the sequencing analysis are consistent with these findings.

Studies of Rap1 in ophthalmic diseases have focused on lens epithelium [49], and retina [50] and cornea [51]. Currently, no studies have directly linked Rap1 and ARC. Our sequencing results show Rap1 signaling pathway is in top 20 of KEGG pathways enriched by overall DMGs at CCWGG sites (Fig 4D), suggesting a possible connection with ARC and providing new clues. The Hippo signaling pathway was one of the top 20 KEGG pathways for overall DMGs at CCWGG sites (Fig 4D), featuring genes such as Mapk1, Mapk3, Yap1. Importantly, in our previous study, we confirmed that YAP1 and MST2 in the Hippo signaling pathway regulated LEC apoptosis [52] related to cataract occurrence. This finding also reflects the accuracy of the sequencing analysis.

It is known that hypermethylation in the promoter region usually leads to transcriptional repression, while hypomethylation may result in increased gene expression [5356]. The research by Balázs Gyȍrffy et al., on the other hand, indicates that under the drive of methylation, the phenomenon of decreased SFRP1 expression can be used to predict the poor prognosis of patients with ER1/HER22 breast cancer (BCs) [57].Therefore, in this paper, we focused on the study of promoter methylation to screen genes.We combined transcriptome sequencing and MethylRAD-seq results to select common DEGs. A total of 109 DEGs with negative methylation at CCGG sites and 33 DEGs with negative methylation at CCWGG sites were identified by Pearson correlation analysis (S3 and S4 Tables). Moreover, the PPI graph of these DEGs with differential methylation at CCGG/CCWGG sites in the promoter region can be found in S1 and S2 Figs., and the corresponding protein interaction scores can be found in S5 and S6 Tables. A joint analysis of DMGs and DEGs at CCGG sites was followed by GO enrichment analysis. The primary enriched term under biological process was modulation of protein kinase C signaling, specifically involving Capzb. In this sequencing result, when comparing the elderly group of mice with the young group of mice, it was found that Capzb in the promoter region of the elderly group of mice showed a hypomethylated state and its expression level is significantly increased (Highly expressed Capzb regulates the occurrence of apoptosis by modulating Protein Kinase C signaling.) (S3 Table). This result leads to a high correlation between MAPKAPK and ARC, suggesting that it may play an important role in the formation of ARC.Under the aspect of molecular function, the enriched terms were mainly related to calmodulin binding (Mapkapk3, Mknk2) and protein serine kinase activity (Dmpk, Rps6ka3). (Fig 7A). Meanwhile, in this sequencing result, compared with young mice, Rps6ka3 in the promoter region of elderly mice shows hypermethylation and low expression; while Mapkapk3, Mknk2, and Dmpk are hypomethylated and highly expressed (S3 Table) (The highly expressed Mapkapk3 regulates the occurrence of cell apoptosis by modulating the cascade reaction of the MAPK pathway).This result shows a high correlation between Mapkapk3 and ARC. In addition,according to relevant literature reports, during the physiological activities of the lens, there are some factors that are crucial for maintaining the stability and transparency of the lens. Calmodulin and calcium ions are the key regulatory factors among them. The balance between the two is of vital importance to the health of the lens. Once this balance is disrupted, cataract [58] will be induced. This research result strongly confirms the accuracy of our sequencing results and also fully demonstrates the important status of calmodulin and calcium ions in the pathogenesis of cataract.

A joint analysis of DMGs and DEGs at CCWGG sites in the promoter region was followed by GO enrichment analysis. The primary enriched term was canonical Wnt signaling pathway (Fig 7B), featuring genes such as Med12 and Wnt7b. Meanwhile, in this sequencing result, compared with young mice, Wnt7b in the promoter region of elderly mice shows hypomethylation and high expression; while Med12 is hypermethylated and lowly expressed (S4 Table). As shown in the literature,some human eye disorders are triggered by mutations of genes in the Wnt signaling pathway, including cataracts, retinal degeneration, and ocular tumors [59]. These diseases are also associated with nuclear receptor activity (Srebf1). The transition of LECs from an epithelial to mesenchymal phenotype is the central pathological process underlying fibrotic cataracts. Lanosterol synthase can potentially impede this deleterious transformation by modulating Srebf1 [60]. Previous research was limited to the relationship between congenital cataracts and nuclear receptor activity, whereas, in this study, a connection between Srebf1 in nuclear receptor activity and ARC has been established,that is, That is, compared with young mice, Srebf1 in the promoter region of elderly mice (i.e., the ARC group) shows hypomethylation and high expression(S4 Table).

A joint analysis of DMGs and DEGs at CCGG and CCWGG sites in the promoter region was followed by KEGG pathway enrichment analysis. The enriched pathways at CCGG sites were mainly related to the MAPK signaling pathway (e.g., Mapkapk3, Mknk2) and cAMP signaling pathway (Cacna1f, Atp2b3) (Table 6). Meanwhile, in this sequencing result, compared with young mice, Mknk2 in the promoter region of elderly mice shows hypomethylation and high expression;while Cacna1f and Atp2b3 are hypermethylated and lowly expressed. (S3 Table). And the enriched pathways at CCWGG sites are mainly related to the mTOR signaling pathway (Rps6ka3, Wnt7b) and HIF-1 signaling pathway (Eno4) (Fig 7C and 7D). We selected four of the annotated genes from the enrichment analyses as candidate genes (Dmpk, Slc25a4, Rps6ka3, Capzb), and verified them by pyrosequencing methylation analysis (Fig 8A). In general, the pyrosequencing verification results were consistent with the MethylRAD-seq results. The findings indicate that the methylation levels of Dmpk and Slc25a4 in the anterior lens capsule of naturally aging mice were lower and the methylation level of Rps6ka3 was higher than their methylation levels in the young mice, which is consistent with the results from the joint analysis. However, no substantial change in the methylation level of Capzb was observed between the two groups.

Dystrophia myotonica protein kinase (Dmpk) has a kinase domain and a coiled-coil domain that is involved in polymerization. Dmpk belongs to the serine/threonine kinase family [61]. Dmpk transcripts and protein have been detected in ocular tissues of adults and fetuses. Extension of CTG repeat sequences in the Dmpk gene causes type 1 myotonic dystrophy (DM1). Cataract is the main ocular complication of patients with DM1 with unknown mechanism [62]. P53 was shown to induce Dmpk expression by up-regulating P73 expression and promoting apoptosis [63]. Apoptosis of LECs can also lead to cataract.Then, is this a potential mechanism by which Dmpk causes cataract in DM1 patients? This needs further study.

The Slc25 family, also known as the mitochondrial carrier family, is the most extensive protein transport gene family in the human solute carrier superfamily, with potential involvement in cancer progression [64]. Specifically, the mitochondrial ADP/ATP carrier, also known as adenine nucleotide translocase 1 (ANT) or Slc25a4, facilitates the import of ADP into the mitochondrial matrix and the export of ATP, which is a crucial step in oxidative phosphorylation [65]. Although Slc25a4 is known to induce cell death by up-regulating pro-apoptotic protein Bax [66], its relationship with ophthalmic diseases, especially ARC, needs further study. Previous research has demonstrated that Dmpk is associated with calcium ions, and the calcium signaling pathway is associated with cataract [67,68]. Slc25a4 is also related to the calcium signaling pathway [69]. Therefore, we verified Slc25a4 and Dmpk methylation by pyrosequencing. The results were in line with the MethylRAD-seq results. Most of the research on Rps6ka3 pertains to Coffin-Lowry syndrome [70], and currently, there are no reports of its association with cataracts. However, studies have indicated that Rps6ka3 may play a crucial role in apoptosis by activating the cAMP response element-binding protein [71]. In this study, the pyrosequencing methylation analysis detected an elevated methylation level of Rps6ka3 in the anterior lens capsule of aged mice. This finding offers a new orientation for understanding the methylation regulatory role of Rps6ka3 in ARC. In future studies, we plan to further investigate the mechanism through which these three genes act in the emergence and development of ARC. These genes could potentially serve as therapeutic targets for preventing and intervening in ARC.

To enhance the clinical significance and validate the research results, we compared the differentially expressed genes Dmpk, Slc25a4 and Rps6ka3 verified by pyrosequencing with the GSE241767 dataset related to human cataract samples in the GEO database (http://www.ncbi.nlm.nih.gov/geo/). Through the comparison, it was found that the differential methylation trends of these genes in human cataracts were consistent with those in our mouse samples.Specifically, in the GSE241767 dataset, compared with the NOR group, the Dmpk in the ARC group showed a hypomethylation state; Slc25a was also hypomethylated, while Rps6ka was in a hypermethylation state. In our mouse data, compared with the Young group, the Dmpk in the Aged group was hypomethylated; Slc25a4 was hypomethylated and Rps6ka3 was hypermethylated, and this trend was consistent with the GSE241767 dataset (S11 and S12 Tables).By comparing with the human dataset, the clinical significance of the identified genes can be determined more accurately. This measure not only bridges the gap between animal model research and human applications, but also greatly improves the clinical relevance of our research work. At the same time, it provides corresponding clinical support evidence for the effectiveness of these genes as biomarkers of ARC in the future.

Meanwhile, existing literature shows that hematological biomarkers can simplify the search for potential molecular defects in families with congenital cataracts [72]. Additionally, as ARC severity increases, connective tissue growth factor in aqueous humor tends to increase, while ascorbic acid concentration and total antioxidant capacity of aqueous humor are negatively correlated with cataract severity [73]. These results provide ideas for subsequent exploration. Detecting the differential expression of target genes (Dmpk, Slc25a4, Rps6ka3) may offer potential biomarkers for early cataract diagnosis, enabling earlier detection of cataract risk and more timely clinical intervention. Moreover, in—depth study of target genes may provide new targets for cataract treatment. We can conduct drug R & D targeting the signaling pathways of these genes to delay or prevent cataract progression.

In this study, MethylRAD-Seq is used instead of bisulfite sequencing. Although there may be some limitations, such as its detection of methylation sites is relatively incomplete and not as detailed, it is valuable in specific situations.MethylRAD-Seq has high cost-effectiveness and is particularly important in studies involving a large number of clinical samples. It can provide valuable insights into disease mechanism research and potential biomarker discovery at a relatively low cost, as well as information on DNA methylation patterns. It is a useful tool in large-scale analysis studies. In addition, it has unique advantages: simple and fast library construction, strong flexibility to control methylation tag density, good repeatability, small amount of starting DNA suitable for limited samples, no dependence on genome sequences for developing unknown sites and studying differences [7478]. However, we must also admit that it may have limitations in some specific research needs. In the future, we can consider combining other sequencing methods to more comprehensively understand the patterns and functions of DNA methylation.

Moreover, due to the extremely small anterior lens capsule of mice, only methylation analysis can be performed alone. If the same sample is split, only micro-transcriptome sequencing is possible. But the mouse anterior lens capsule is tiny, leading to insufficient sample size and not meeting the sequencing machine’s minimum requirement. Thus, only conventional transcriptomics on the complete sample is viable. Conventional transcriptomics is better than micro-transcriptomics in data quality (higher sequencing depth/coverage, detecting low-abundance transcripts, accurately quantifying gene expression and providing reliable differential expression results), experimental operation (mature/stable/simple with good repeatability and lower sample requirements), and application scope (suitable for various research needs and easy integration with other omics data for systems biology research). In many previous studies, samples from different batches were also used to carry out multi-omics research work. Xu Zhang et al. employed a two-stage design for methylome and transcriptome study. Stage 1: methylation levels of 99 ankylosing spondylitis (AS) patients and 99 healthy controls (HCs)measured. Stage 2: mRNA levels of 20 patients and 20 HCs measured. Found IL12B gene’s DNA methylation and transcriptome traits can distinguish AS patients from HCs, and hypermethylation may contribute to AS pathogenesis [79]. Yilin Zhao et al. comprehensively analyzed proteome and transcriptome of lens epithelium and fibers. At least 525 proteins in lens can generate info on key lens formation processes and are related to different cells and molecules [80]. However, interpreting methylation data alone often faces many challenges. Therefore, we use pyrosequencing to verify methylation data. This method can accurately detect the degree and site of DNA methylation [8183]. It not only confirms the correctness of the data but also provides important clues for subsequent clinical verification sites. In clinical practice, these verified methylation sites can serve as potential biomarkers for the diagnosis, prognosis evaluation, and treatment monitoring of ARC disease. Detecting these methylation sites can provide a foundation for the early diagnosis and individualized treatment of ARC. This also provides a theoretical basis for treating ARC by targeting these methylation sites.

In summary, we conducted whole-genome methylation and transcriptome analyses on young and aged mice to identify potential epigenetic biomarkers for ARC. Using this approach, we successfully identified key molecular markers for the occurrence and progression of ARC. These markers may enable early diagnosis and targeted intervention for high-risk populations, and also hold promise for the discovery of new drug targets that can help improve the well-being of patients with ARC. In future studies, in-depth exploration of the different signaling pathways, key regulatory factors, and molecular targets identified in this study can be carried out. The sequencing results can be used to develop clinical diagnostic and treatment tools. Conducting clinical trials to promote the applications of research findings, establishing a long-term monitoring system to track disease progression and treatment effects, and exploring lifestyle intervention measures for preventing ARC will contribute to the provision of personalized prevention recommendations for the public. Overall, this study provides a powerful tool for in-depth understanding of ARC and opens up new frontiers in the prevention, diagnosis, and treatment of ARC.

Conclusions

DNA methylation is a significant component of epigenetics, and is important in the pathogenesis of cataracts. In this study, we aimed to discover potential epigenetic biomarkers of ARC by integrating whole-genome methylation and transcriptome analyses of young and aged mice. MethylRAD-Seq was performed to examine the whole-genome methylation in the anterior lens capsule tissues of young and aged mice. The methylation analysis identified 76,524 and 15,608 DMSs at CCGG and CCWGG sites. Subsequently, the promoter regions of DMGs were combined with the DEGs obtained by transcriptome sequencing analysis, and 109 and 33 DEGs that were negatively correlated with methylation were identified at CCGG and CCWGG sites, respectively. The combined GO and KEGG annotations identified four ARC-related genes, namely Capzb, Dmpk, Slc25a4, and Rps6ka3. These genes were verified by pyrosequencing and the results were consistent with those of MethylRAD-seq. Our investigation provides comprehensive whole-genome DNA methylation patterns and gene expression signatures associated with ARC. The identified methylation-based molecular markers hold promise for future applications in ARC prevention, diagnosis, treatment strategies, and prognostic evaluations.

Supporting information

S1 Fig. The PPI graph of DEGs with differential methylation at CCGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s001

(TIF)

S2 Fig. The PPI graph of DEGs with differential methylation at CWGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s002

(TIF)

S3 Table. List of DEGs with differential methylation at CCGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s005

(XLS)

S4 Table. List of DEGs with differential methylation at CCWGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s006

(XLS)

S5 Table. The protein interaction scores of DEGs with differential methylation at CCGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s007

(XLS)

S6 Table. The protein interaction scores of DEGs with differential methylation at CCWGG sites in the promoter region.

https://doi.org/10.1371/journal.pone.0316766.s008

(XLS)

S7 Table. List of GO enrichment analysis for DEGs with differentially methylated CCGG sites in promoter region.

https://doi.org/10.1371/journal.pone.0316766.s009

(XLSX)

S8 Table. List of GO enrichment analysis for DEGs with differentially methylated CCWGG sites in promoter region.

https://doi.org/10.1371/journal.pone.0316766.s010

(XLSX)

S9 Table. List of KEGG enrichment analysis for DEGs with differentially methylated CCGG sites in promoter region.

https://doi.org/10.1371/journal.pone.0316766.s011

(XLSX)

S10 Table. List of KEGG enrichment analysis for DEGs with differentially methylated CCWGG sites in promoter region.

https://doi.org/10.1371/journal.pone.0316766.s012

(XLSX)

S11 Table. Differential methylation data of human cataracts in the GEO database (GSE241767).

https://doi.org/10.1371/journal.pone.0316766.s013

(XLSX)

S12 Table. The comparison between the methylation data of mice and that of human cataracts (GSE241767).

https://doi.org/10.1371/journal.pone.0316766.s014

(XLSX)

Acknowledgments

We thank Margaret Biswas, PhD, from Liwen Bianji (Edanz) (www.liwenbianji.cn/) for editing the English text of a draft of this manuscript.

References

  1. 1. Delbarre M, Froussart-Maille F. Sémiologie et formes cliniques de la cataracte chez l’adulte [Signs, symptoms, and clinical forms of cataract in adults]. J Fr Ophtalmol. 2020 Sep;43(7):653–659. pmid:32586638
  2. 2. Logue MW, Smith AK, Wolf EJ, Maniates H, Stone A, Schichman SA,et al. The correlation of methylation levels measured using Illumina 450K and EPIC BeadChips in blood samples. Epigenomics. 2017 Nov;9(11):1363–1371. pmid:28809127
  3. 3. Liu Z, Wang R, Lin H, Liu Y. Lens regeneration in humans: using regenerative potential for tissue repairing. Ann Transl Med. 2020 Nov;8(22):1544. pmid:33313289
  4. 4. Berthoud VM, Minogue PJ, Osmolak P, Snabb JI, Beyer EC. Roles and regulation of lens epithelial cell connexins. FEBS Lett. 2014 Apr 17;588(8):1297–303. pmid:24434541
  5. 5. Asbell PA, Dualan I, Mindel J, Brocks D, Ahmad M, Epstein S. Age-related cataract. Lancet. 2005 Feb 12–18;365(9459):599–609. pmid:15708105
  6. 6. Wang L, Li P, Guo X. Screening of methylation genes in age-related cataract. Int J Ophthalmol. 2018 Jul 18;11(7):1102–1107. pmid:30046524
  7. 7. Wang Y, Guan H. The Role of DNA Methylation in Lens Development and Cataract Formation. Cell Mol Neurobiol. 2017 Aug;37(6):979–984. pmid:27858287
  8. 8. Li F, Wang Y, Zhang G, Zhou J, Yang L, Guan H. Expression and methylation of DNA repair genes in lens epithelium cells of age-related cataract. Mutat Res. 2014 Aug-Sep;766–767:31–6. pmid:25847269
  9. 9. Liu S, Su D, Sun Z, Piao T, Li S, Guan L,et al. Epigallocatechin gallate delays age-related cataract development via the RASSF2/AKT pathway. Eur J Pharmacol. 2023 Dec 15;961:176204. pmid:37979829
  10. 10. Pendergrass W, Penn P, Possin D, Wolf N. Accumulation of DNA, nuclear and mitochondrial debris, and ROS at sites of age-related cortical cataract in mice. Invest Ophthalmol Vis Sci. 2005 Dec;46(12):4661–70. pmid:16303963
  11. 11. Chen M, Fu Y, Wang X, Wu R, Su D, Zhou N, et al.Metformin protects lens epithelial cells against senescence in a naturally aged mouse model. Cell Death Discov. 2022 Jan 10;8(1):8. pmid:35013152
  12. 12. Stopka W, Libby T, Lin S, Wang E, Xia CH, Gong X. Age-related changes of lens stiffness in wild-type and Cx46 knockout mice. Exp Eye Res. 2021 Nov;212:108777. pmid:34597677
  13. 13. Faranda AP, Shihan MH, Wang Y, Duncan MK. The aging mouse lens transcriptome. Exp Eye Res. 2021 Aug;209:108663. pmid:34119483
  14. 14. Wang S, Lv J, Zhang L, Dou J, Sun Y, Li X, et al. MethylRAD: a simple and scalable method for genome-wide DNA methylation profiling using methylation-dependent restriction enzymes. Open Biol. 2015 Nov;5(11):150130. pmid:26581575
  15. 15. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012 Mar 4;9(4):357–9. pmid:22388286
  16. 16. Song A, Im M, Kim MS, Noh ES, Kim C, Jang J, et al.First female Korean child with Coffin-Lowry syndrome: a novel variant in RPS6KA3 diagnosed by exome sequencing and a literature review. Ann Pediatr Endocrinol Metab. 2023 Mar;28(1):67–72. pmid:35038833
  17. 17. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11(10):R106. pmid:20979621
  18. 18. Meng R, Song J, Guan L, Li Q, Shi C, Su D,et al.Genome-wide analysis of methylation in rat fetal heart under hyperglycemia by methylation-dependent restriction site-associated DNA sequencing. PLoS One. 2022 May 11;17(5):e0268117. pmid:35544480
  19. 19. Jiang Y, Li J, Ren F, Ji C, Aniagu S, Chen T. PM2.5-induced extensive DNA methylation changes in the heart of zebrafish embryos and the protective effect of folic acid. Environ Pollut. 2019 Dec;255(Pt 3):113331. pmid:31614245
  20. 20. Li F, Gao Z, Li X, Fang L. The effect of Paecilomyces catenlannulatus on removal of U(VI) by illite. J Environ Radioact. 2014 Nov;137:31–36. pmid:24998746
  21. 21. Cicinelli MV, Buchan JC, Nicholson M, Varadaraj V, Khanna RC. Cataracts. Lancet. 2023 Feb 4;401(10374):377–389. pmid:36565712
  22. 22. Jiang H, Yin Y, Wu CR, Liu Y, Guo F, Li M, et al. Dietary vitamin and carotenoid intake and risk of age-related cataract. Am J Clin Nutr. 2019 Jan 1;109(1):43–54. pmid:30624584
  23. 23. Wei W, Li T, Chen J, Fan Z, Gao F, Yu Z, et al. SIRT3/6: an amazing challenge and opportunity in the fight against fibrosis and aging. Cell Mol Life Sci. 2024 Jan 31;81(1):69. pmid:38294557
  24. 24. Li X, Liu L, Li T, Liu M, Wang Y, Ma H, et al. SIRT6 in Senescence and Aging-Related Cardiovascular Diseases. Front Cell Dev Biol. 2021 Mar 29;9:641315. pmid:33855020
  25. 25. Li T, Mu N, Yin Y, Yu L, Ma H. Targeting AMP-Activated Protein Kinase in Aging-Related Cardiovascular Diseases. Aging Dis. 2020 Jul 23;11(4):967–977. pmid:32765957
  26. 26. Lampi KJ, Wilmarth PA, Murray MR, David LL. Lens β-crystallins: the role of deamidation and related modifications in aging and cataract. Prog Biophys Mol Biol. 2014 Jul;115(1):21–31. pmid:24613629
  27. 27. Kou R, Li T, Fu C, Jiang D, Wang Y, Meng J, et al. Exosome-shuttled FTO from BM-MSCs contributes to cancer malignancy and chemoresistance in acute myeloid leukemia by inducing m6A-demethylation: A nano-based investigation. Environ Res. 2024 Mar 1;244:117783. pmid:38048862
  28. 28. Wang Y, Guan H. The Role of DNA Methylation in Lens Development and Cataract Formation. Cell Mol Neurobiol. 2017 Aug;37(6):979–984. pmid:27858287
  29. 29. Jin SL, Zhang Y, Chen ZH, Qian DW, Qine YJ, Yongjie Q, et al. Epigenetic changes of the Klotho gene in age-related cataracts. Eur Rev Med Pharmacol Sci. 2015;19(14):2544–53. pmid:26221880
  30. 30. Thampi P, Zarina S, Abraham EC. alpha-Crystallin chaperone function in diabetic rat and human lenses. Mol Cell Biochem. 2002 Jan;229(1–2):113–8. pmid:11936835.
  31. 31. Horwitz J. Alpha-crystallin Exp Eye Res. 2003 Feb;76(2):145–53. pmid:12565801
  32. 32. Zhou P, Luo Y, Liu X, Fan L, Lu Y. Down-regulation and CpG island hypermethylation of CRYAA in age-related nuclear cataract. FASEB J. 2012 Dec;26(12):4897–902. pmid:22889833
  33. 33. Gao Y, Yan Y, Huang T. Human age‑related cataracts: epigenetic suppression of the nuclear factor erythroid 2‑related factor 2‑mediated antioxidant system. Mol Med Rep. 2015 Feb;11(2):1442–7. pmid:25370996
  34. 34. Constâncio V, Nunes SP, Henrique R, Jerónimo C. DNA Methylation-Based Testing in Liquid Biopsies as Detection and Prognostic Biomarkers for the Four Major Cancer Types. Cells. 2020 Mar 5;9(3):624. pmid:32150897
  35. 35. Kim J, Hwang J, Jeong H, Song HJ, Shin J, Hur G, et al. Promoter methylation status of VEGF receptor genes: a possible epigenetic biomarker to anticipate the efficacy of intracellular-acting VEGF-targeted drugs in cancer cells. Epigenetics. 2012 Feb;7(2):191–200. pmid:22395469
  36. 36. Duraisamy AJ, Radhakrishnan R, Seyoum B, Abrams GW, Kowluru RA. Epigenetic Modifications in Peripheral Blood as Potential Noninvasive Biomarker of Diabetic Retinopathy. Transl Vis Sci Technol. 2019 Dec 18;8(6):43. pmid:31871829
  37. 37. Chen H, Zhang X, Liao N, Ji Y, Mi L, Gan Y, et al. Identification of NLRP3 Inflammation-Related Gene Promoter Hypomethylation in Diabetic Retinopathy. Invest Ophthalmol Vis Sci. 2020 Nov 2;61(13):12. pmid:33156339
  38. 38. Wolf N, Penn P, Pendergrass W, Van Remmen H, Bartke A, Rabinovitch P, et al. Age-related cataract progression in five mouse models for anti-oxidant protection or hormonal influence. Exp Eye Res. 2005 Sep;81(3):276–85. pmid:16129095
  39. 39. Jun G, Guo H, Klein BE, Klein R, Wang JJ, Mitchell P, et al EPHA2 is associated with age-related cortical cataract in mice and humans. PLoS Genet. 2009 Jul;5(7):e1000584. pmid:19649315
  40. 40. Ogasawara S, Hashizume K, Okuno T, Imaizumi T, Inomata Y, Tezuka Y, et al. Effect of Geranylgeranylacetone on Ultraviolet Radiation Type B-Induced Cataract in Heat-Shock Transcription Factor 1 Heterozygous Mouse. Curr Eye Res. 2017 May;42(5):732–737. pmid:27732059
  41. 41. Jiang Y, Fu R, Zhao J, Wu D, Qiao G, Li R, et al. Effects of ELL-associated factor 2 on ultraviolet radiation-induced cataract formation in mice. Mol Med Rep. 2015 Nov;12(5):6605–11. pmid:26328919
  42. 42. Gross J, Wegener AR, Kronschläger M, Holz FG, Schönfeld CL, Meyer LM. Ultraviolet radiation exposure triggers neurokinin-1 receptor upregulation in ocular tissues in vivo. Exp Eye Res. 2018 Sep;174:70–79. pmid:29705099
  43. 43. Shearer TR, David LL, Anderson RS, Azuma M. Review of selenite cataract. Curr Eye Res. 1992 Apr;11(4):357–69. pmid:1526166
  44. 44. Liu L, Paterson CA, Borchman D. Regulation of sarco/endoplasmic Ca2+-ATPase expression by calcium in human lens cells. Exp Eye Res. 2002 Nov;75(5):583–90. pmid:12457870
  45. 45. Sredy J, Spector A. Phosphorylation of H2O2-treated lens Na+,K+-ATPase. Exp Eye Res. 1984 Oct;39(4):479–82. pmid:6094227
  46. 46. Gao X, Tian X, Huang Y, Fang R, Wang G, Li D, et al. Role of circular RNA in myocardial ischemia and ageing-related diseases. Cytokine Growth Factor Rev. 2022 Jun;65:1–11. pmid:35561533
  47. 47. Liu Y, Li H, Liu Y. microRNA-378a Regulates the Reactive Oxygen Species (ROS)/Phosphatidylinositol 3-Kinases (PI3K)/AKT Signaling Pathway in Human Lens Epithelial Cells and Cataract. Med Sci Monit. 2019 Jun 10;25:4314–4321. pmid:31178586
  48. 48. Moustardas P, Aberdam D, Lagali N. MAPK Pathways in Ocular Pathophysiology: Potential Therapeutic Drugs and Challenges. Cells. 2023 Feb 14;12(4):617. pmid:36831285
  49. 49. Maddala R, Nagendran T, Lang RA, Morozov A, Rao PV. Rap1 GTPase is required for mouse lens epithelial maintenance and morphogenesis. Dev Biol. 2015 Oct 1;406(1):74–91. pmid:26212757
  50. 50. Ramos CJ, Lin C, Liu X, Antonetti DA. The EPAC-Rap1 pathway prevents and reverses cytokine-induced retinal vascular permeability. J Biol Chem. 2018 Jan 12;293(2):717–730. pmid:29158262
  51. 51. Poon MW, Yan L, Jiang D, Qin P, Tse HF, Wong IY, et al. Inhibition of RAP1 enhances corneal recovery following alkali injury. Invest Ophthalmol Vis Sci. 2015 Jan 8;56(2):711–21. pmid:25574050
  52. 52. Liu S, Su D, Sun Z, Guan L, Wang Z, Zhang G, et al. High MST2 expression regulates lens epithelial cell apoptosis in age-related cataracts through YAP1 targeting GLUT1. Arch Biochem Biophys. 2022 Jul 15;723:109255. pmid:35452623
  53. 53. Lei L, Xu X, Gong C, Lin B, Li F. Integrated analysis of genome-wide gene expression and DNA methylation profiles reveals candidate genes in ovary endometriosis. Front Endocrinol (Lausanne). 2023 Mar 23;14:1093683. pmid:37033258
  54. 54. Matsuzaki H, Miyajima Y, Fukamizu A, Tanimoto K. Orientation of mouse H19 ICR affects imprinted H19 gene expression through promoter methylation-dependent and -independent mechanisms. Commun Biol. 2021 Dec 17;4(1):1410. pmid:34921234
  55. 55. Panagopoulou M, Drosouni A, Fanidis D, Karaglani M, Balgkouranidou I, Xenidis N, Aidinis V, Chatzaki E. ENPP2 Promoter Methylation Correlates with Decreased Gene Expression in Breast Cancer: Implementation as a Liquid Biopsy Biomarker. Int J Mol Sci. 2022 Mar 28;23(7):3717. pmid:35409077
  56. 56. Jeong S, Cho S, Yang SK, Oh SA, Kang YK. Parallel shift of DNA methylation and gene expression toward the mean in mouse spleen with aging. Aging (Albany NY). 2023 Jul 25;15(14):6690–6709. pmid:37494662
  57. 57. Bonder MJ, Luijk R, Zhernakova DV, Moed M, Deelen P, Vermaat M,et al. Disease variants alter transcription factor levels and methylation of their binding sites. Nat Genet. 2017 Jan;49(1):131–138. pmid:27918535
  58. 58. Hu Z, Riquelme MA, Gu S, Jiang JX. Regulation of Connexin Gap Junctions and Hemichannels by Calcium and Calcium Binding Protein Calmodulin. Int J Mol Sci. 2020 Nov 2;21(21):8194. pmid:33147690
  59. 59. de Iongh RU, Abud HE, Hime GR. WNT/Frizzled signaling in eye development and disease. Front Biosci. 2006 Sep 1;11:2442–64. pmid:16720326
  60. 60. Ma P, Huang J, Chen B, Huang M, Xiong L, Chen J, et al. Lanosterol Synthase Prevents EMT During Lens Epithelial Fibrosis Via Regulating SREBP1. Invest Ophthalmol Vis Sci. 2023 Dec 1;64(15):12. pmid:38079167
  61. 61. Elkins JM, Amos A, Niesen FH, Pike AC, Fedorov O, Knapp S. Structure of dystrophia myotonica protein kinase. Protein Sci. 2009 Apr;18(4):782–91. pmid:19309729
  62. 62. Iida Y, Hayashi T, Tokuhisa T, Mizobuchi K, Omoto S, Nakano T. Polypoidal choroidal vasculopathy in a patient with DMPK-associated myotonic dystrophy. Doc Ophthalmol. 2022 Jun;144(3):217–226. pmid:35284965
  63. 63. Itoh K, Ebata T, Hirata H, Torii T, Sugimoto W, Onodera K, et al. DMPK is a New Candidate Mediator of Tumor Suppressor p53-Dependent Cell Death. Molecules. 2019 Sep 1;24(17):3175. pmid:31480541
  64. 64. Liu AR, Liu YN, Shen SX, Yan LR, Lv Z, Ding HX, et al. Comprehensive Analysis and Validation of Solute Carrier Family 25 (SLC25) and Its Correlation with Immune Infiltration in Pan-Cancer. Biomed Res Int. 2022 Oct 8;2022:4009354. pmid:36254139
  65. 65. Cimadamore-Werthein C, Jaiquel Baron S, King MS, Springett R, Kunji ER. Human mitochondrial ADP/ATP carrier SLC25A4 operates with a ping-pong kinetic mechanism. EMBO Rep. 2023 Aug 3;24(8):e57127. pmid:37278158
  66. 66. Baines CP, Molkentin JD. Adenine nucleotide translocase-1 induces cardiomyocyte death through upregulation of the pro-apoptotic protein Bax. J Mol Cell Cardiol. 2009 Jun;46(6):969–77. pmid:19452617
  67. 67. Duncan G, Wormstone IM. Calcium cell signalling and cataract: role of the endoplasmic reticulum. Eye (Lond). 1999 Jun;13 (Pt 3b):480–3. pmid:10627828
  68. 68. Gupta PD, Johar K, Vasavada A. Causative and preventive action of calcium in cataracto-genesis. Acta Pharmacol Sin. 2004 Oct;25(10):1250–6. pmid:15456524
  69. 69. Hertweck KL, Vikramdeo KS, Galeas JN, Marbut SM, Pramanik P, Yunus F, et al. Clinicopathological significance of unraveling mitochondrial pathway alterations in non-small-cell lung cancer. FASEB J. 2023 Jul;37(7):e23018. pmid:37310411
  70. 70. Jin H, Li H, Qiang S. Coffin-Lowry Syndrome Induced by RPS6KA3 Gene Variation in China: A Case Report in Twins. Medicina (Kaunas). 2022 Jul 20;58(7):958. pmid:35888677
  71. 71. Poplin R, Chang PC, Alexander D, Schwartz S, Colthurst T, Ku A, et al. A universal SNP and small-indel variant caller using deep neural networks. Nat Biotechnol. 2018 Nov;36(10):983–987. pmid:30247488
  72. 72. Wussuki-Lior O, Abu-Horowitz A, Netzer I, Almer Z, Morad Y, Goldich Y, Yahalom V, Pras E, Pras E. Hematologic biomarkers in childhood cataracts. Mol Vis. 2011 Apr 24;17:1011–5. pmid:21541272
  73. 73. Hwang HB, Yim HB, Cho YK, Choi JA. The Association Between Aqueous Connective Tissue Growth Factor and the Severity of Age-related Cataracts as Graded by the Lens Opacities Classification System III. Curr Eye Res. 2016;41(3):350–6. pmid:26301961
  74. 74. Yang X, Han L, Ye Q, Wang H, Zhang J, Wang W, Xiao H, Wang Y, Wang L, Ding J. MethylRAD Sequencing Technology Reveals DNA Methylation Characteristics of Apostichopus japonicus of Different Ages. Animals (Basel). 2023 Nov 15;13(22):3530. pmid:38003147
  75. 75. Tang M, Li R, Chen P. Exogenous glutathione can alleviate chromium toxicity in kenaf by activating antioxidant system and regulating DNA methylation. Chemosphere. 2023 Oct;337:139305. pmid:37364644
  76. 76. Sang F. Bioinformatics Analysis of DNA Methylation Through Bisulfite Sequencing Data. Methods Mol Biol. 2021;2198:441–450. pmid:32822049
  77. 77. Wang J, Xia Y, Li L, Gong D, Yao Y, Luo H, Lu H, Yi N, Wu H, Zhang X, Tao Q, Gao F. Double restriction-enzyme digestion improves the coverage and accuracy of genome-wide CpG methylation profiling by reduced representation bisulfite sequencing. BMC Genomics. 2013 Jan 16;14:11. pmid:23324053
  78. 78. Wan M, Bennett BD, Pittman GS, Campbell MR, Reynolds LM, Porter DK, Crowl CL, Wang X, Su D, Englert NA, Thompson IJ, Liu Y, Bell DA. Identification of Smoking-Associated Differentially Methylated Regions Using Reduced Representation Bisulfite Sequencing and Cell type-Specific Enhancer Activation and Gene Expression. Environ Health Perspect. 2018 Apr 27;126(4):047015. pmid:29706059
  79. 79. Zhang X, Lu J, Pan Z, Ma Y, Liu R, Yang S, Yang S, Dong J, Shi X, Xu S, Pan F. DNA methylation and transcriptome signature of the IL12B gene in ankylosing spondylitis. Int Immunopharmacol. 2019 Jun;71:109–114. pmid:30889422
  80. 80. Zhao Y, Wilmarth PA, Cheng C, Limi S, Fowler VM, Zheng D, David LL, Cvekl A. Proteome-transcriptome analysis and proteome remodeling in mouse lens epithelium and fibers. Exp Eye Res. 2019 Feb;179:32–46. pmid:30359574
  81. 81. Jiang Z, Zhou X, Li R, Michal JJ, Zhang S, Dodson MV, Zhang Z, Harland RM. Whole transcriptome analysis with sequencing: methods, challenges and potential solutions. Cell Mol Life Sci. 2015 Sep;72(18):3425–39. pmid:26018601
  82. 82. Martorella M, Kasela S, Garcia-Flores R, Gokden A, Castel SE, Lappalainen T. Evaluation of noninvasive biospecimens for transcriptome studies. BMC Genomics. 2023 Dec 19;24(1):790. pmid:38114913
  83. 83. Hölzer M. A decade of de novo transcriptome assembly: Are we there yet? Mol Ecol Resour. 2021 Jan;21(1):11–13. pmid:33030794