Integration of CRISPR/dCas9-Based methylation editing with guide positioning sequencing identifies dynamic changes of mrDEGs in breast cancer progression

Abstract

Dynamic changes in DNA methylation are prevalent during the progression of breast cancer. However, critical alterations in aberrant methylation and gene expression patterns have not been thoroughly characterized. Here, we utilized guide positioning sequencing (GPS) to conduct whole-genome DNA methylation analysis in a unique human breast cancer progression model: MCF10 series of cell lines (representing benign/normal, atypical hyperplasia, and metastatic carcinoma). By integrating with mRNA-seq and matched clinical expression data from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO), six representative methylation-related differentially expressed genes (mrDEGs) were identified, including CAVIN2, ARL4D, DUSP1, TENT5B, P3H2, and MMP28. To validate our findings, we independently developed and optimized the dCas9-DNMT3L-DNMT3A system, achieving a high efficiency with a 98% increase in methylation at specific sites. DNA methylation levels significantly increased for the six genes, with CAVIN2 at 67.75 ± 1.05%, ARL4D at 53.29 ± 6.32%, DUSP1 at 57.63 ± 8.46%, TENT5B at 44.00 ± 5.09%, P3H2 at 58.50 ± 3.90%, and MMP28 at 49.60 ± 5.84%. RT-qPCR confirmed an inverse correlation between increased DNA methylation and gene expression. Most importantly, we mimicked tumor progression in vitro, demonstrating that transcriptional silencing of the TENT5B promotes cell proliferation in MCF10A cells owing to the crosstalk between hypermethylation and histone deacetylation. This study unveils the practical implications of DNA methylation dynamics of mrDEGs in reshaping epigenomic features during breast cancer malignant progression through integrated data analysis of the methylome and transcriptome. The application of the CRISPR/dCas9-based methylation editing technique elucidates the regulatory mechanisms and functional roles of individual genes within the DNA methylation signature, providing valuable insights for understanding breast cancer pathogenesis and facilitating potential therapeutic approaches in epigenome editing for patients with breast cancer.

Keywords: Breast cancer progression; CRISPR/dCas9; DNA methylation; Epigenetic modification; Guide positioning sequencing.

Fig. 1

Schematic diagram of the study. A, The overall strategy of the study. Potential mrDEGs were screened according to the data analysis derived from high-throughput sequencing (HTS) and the validation by CRISPR/dCas9-based methylation editing technology. B, The workflow of GPS library construction. The fundamental principle of GPS is based on the substitution of dCTP with 5-methyl-dCTP (5mC) at the 3’-end of DNA fragments by T4 DNA polymerase, which protects cytosines from bisulfite conversion to preserves the integrity of the base composition. This alteration enables the 3’-end to independently facilitate genetic variation profiling and guides the 5’-end, enriched with methylation information, to align more rapidly to a reference genome

Fig. 2

Hierarchical clustering of global DNA methylation profiles and screening for differentially methylated regions (DMRs). A, Hierarchical clustering using Pearson correlation across the four different stages of breast cells: MCF10A, MCF10AT, MCF10CA1h, and MCF10CA1a. B and C, Violin plots (B) and heatmaps (C) presenting dynamic methylation proportions, including the number of gains and losses of methylation in MCF10AT, MCF10CA1h, and MCF10CA1a relative to MCF10A. The substantial proportion of gain of methylation is observed in all three groups, accounting for 57.26%, 81.44%, and 72.62%, respectively. D, Venn diagram showing the overlap of three sets of DMGs: MCF10A vs. MCF10AT, MCF10A vs. MCF10CA1h, and MCF10A vs. MCF10CA1a. Based on Slow-TSB hypothesis, after comparison, we obtained a total of 2986 non-overlapping differentially methylated malignancy-specific loci, which are circled in dashed lines

Fig. 3

Screening of methylation-related differentially expressed genes (mrDEGs). A, Heatmap of the top 171 mrDEGs, 69% (118/171) of which exhibit high methylation and low expression. The y-axis shows the names of the genes. B, Scatterplot of mRNA relative expression of 118 genes shows a significantly negative correlation (r = -0.2508, **P = 0.0062) with 5mC increase. Six mrDEGs are highlighted with blue dots. C, Pyroequencing verification of DNA methylation levels of CAVIN2, ARL4D, DUSP1, TENT5B, P3H2, and MMP28 in all four cells. The y-axis shows the DNA methylation level for each CpG site. Line plots reflecting the continuity of DNA methylation changes in adjacent CpG sites. The GPS data group is represented in a light color, while the pyrosequencing data group in a dark color. D, Correlations (linear regression and 95% confidence interval) of the methylation levels of six mrDEGs detected through the pyrosequencing and derived from the GPS. Pearson correlation coefficient is shown in the upper left corner. E, RT-qPCR verification of mRNA expression of six mrDEGs normalized to H-ACTB. The y-axis represents the relative expression level, expressed as log₂(FoldChange), and the x-axis represents the name of genes. Data are the means (± SEM) of three independent experiments. Statistical significance (Student’s t-test) is indicated (**** P < 0.001). F, Heatmap summarizes the observed relationship between gene expression and 5mC increase in six mrDEGs. Data derived from RT-qPCR and pyrosequencing

Fig. 4

The efficiency validation of the CRISPR/dcas9-EpiEffector system. A, Schematic diagram of the ERRFI1 locus on chromosome 1 (36.23), with distances shown relative to the promoter and CGI. The full-length CGI is 1300 bp, containing 157 CpG sites. The first base of CGI was defined as + 1, and the first CpG site was defined as C1. The region highlighted in darkgreen represents the CGI, and regions highlighted in light red, lightgreen, light blue, blue, pink, and yellow represent PCR fragments 1–6, with PCR product lengths of 219 bp, 142 bp, 197 bp, 234 bp, 284 bp, and 294 bp, respectively. The location of gRNA is indicated by deep red vertical line near the 3’-end of PCR fragment 4. White dots on PCR fragment 4 represent the relative positions of CpG sites. B, Quantitative determination of DNA methylation levels in fragment 4 with Fseq4-1–4 − 3 using pyrosequencing. The upper part of the schematic represents the Neg group, and the bottom part is the group transfected with d9-D3A-ERRFI1_gRNA. The sequence highlighted in red indicates the position of gRNA. C, Line plots of methylation levels at the CpG sites C85–C114 around the gRNA position detected with Fseq4-3. The y-axis represents the methylation level, and the x-axis depicts the individual CpG position. Data are plotted as the mean (± SEM) value of three independent experiments. D, Relative distribution of gain of DNA methylation at individual CpG sites across the whole ERRFI1 CGI. The guide RNA binding sites are depicted in red frame. E, Boxplots comparing the increase level of DNA methylation around ERRFI1 gRNA between d9-D3L-D3A and d9-D3A, presented pairwise (Wilcoxon test, **P = 0.0048). The horizontal line within the boxplot represents the mean

Fig. 5

Targeting the TENT5B promoter with dCas9-DNMT3L-DNMT3A promotes breast cell proliferation, accompanied by histone deacetylation crosstalk. A, Pyrosequencing verification of methylation elevation of TENT5B promoter. The x-axis represents condition of different treatments, divided into Neg and d9-D3L-D3A transfection. The y-axis represents DNA methylation level. Dashed lines represent pairwise CpG sites methylation increase by paired t-test (****P < 0.0001). B, Relative expression level after artificial elevation of DNA methylation displayed 4.1-fold downregulation of TENT5B mRNA. Data are the means (± SEM) of three independent experiments. Statistical significance (Student’s t-test) is indicated (***P < 0.0007). C, Growth curve of MCF10A treated with and without d9-D3L-D3A-TENT5B_gRNA. Data are the means (± SEM). All means at each time point are statistically significant by multiple t-test with P = 0.1635 (ns), P = 0.0423 (*), P = 0.0229 (*) and P = 0.0049 (**), respectively. Experiments were performed in triplicates. D, ChIP-qPCR shows decreased H3K27ac enrichment at TENT5B gRNA adjacent regions. Data are the averages (± SEM) of three independent experiments. Statistical significance (Student’s t-test) is indicated (**P = 0.0019). E, Schematic diagram illustrating the Slow-TSB, which features hypomethylated promoters in a subset of highly efficient tumor suppressor genes, provides enhanced protection against oncogenesis. However, when subjected to aberrant hypermethylation, the integrity of the Slow-TSB is compromised, resulting in reduced effectiveness in inhibiting the growth of cancer cells, thus facilitating the development of malignant tumors