Coordinated DNA methyltransferase 3A and methyltransferase-like 7A activity reprograms the tumor microenvironment through discoidin domain receptor 1 signaling

High-throughput transcriptome sequencing

Sequencing libraries were generated and sequenced by CapitalBio Technology (Beijing, China) using 5 μg of total RNA per sample. Briefly, ribosomal RNA (rRNA) was removed from the total RNA using the Ribo-Zero Magnetic Kit (MRZG12324; Epicentre, Madison, WI, USA). Sequencing libraries were constructed with the NEBNext Ultra RNA Library Prep Kit for Illumina (E7760S; New England Biolabs, Ipswich, MA, USA). RNA was fragmented into ~300 bp fragments in the NEBNext First Strand Synthesis Reaction Buffer (5×). First-strand cDNA was synthesized using reverse transcriptase with random primers, followed by second-strand cDNA synthesis in the presence of dUTP Mix (10×) in the Second Strand Synthesis Reaction Buffer. The cDNA fragments were subjected to end repair, poly(A) tail addition, and Illumina adapter ligation. The USER Enzyme (M5508; NEB) was used to digest the second cDNA strand to construct strand-specific libraries after adapter ligation. Library DNA was amplified, purified, and enriched by PCR. The libraries were qualified using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and quantified with the KAPA Library Quantification Kit (kk3605; Merck, Darmstadt, Germany). RNA-seq experiments were performed in three biological replicates per group on the Illumina NovaSeq platform (Illumina, San Diego, CA, USA) using paired-end 150 bp sequencing and generating an average of 50 million reads per sample³¹.

Transcriptome sequencing data analysis

Quality control of raw paired-end sequencing reads was performed using FastQC (v0.11.8). Raw data was processed with Cutadapt (v1.18) to remove Illumina sequencing adapters and poly(A) tail sequences. Reads containing > 5% ambiguous bases (N) were filtered out using a Perl script. The FASTX Toolkit (v0.0.13) was used to retain reads with ≥ 70% of bases that had a Phred quality score > 20. Paired-end reads were error-corrected with BBMap and high-quality reads were aligned to the human reference genome using HISAT2 (v0.7.12).

Differential expression analysis of mRNA read counts was performed using limma (|log2FC| > 3; adj. P < 0.05). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed with ClusterProfiler and the results were visualized as bubble plots. Feature selection was carried out using glmnet (LASSO, ridge, ElasticNet) and caret (SVM-RFE). Overlapping genes were identified using VennDiagram.

5-methylcytosine (5mC) and m⁶A gene expression analysis

A literature-curated list of 20 5mC- and 26 m6A-related genes was analyzed. Expression profiles were extracted from transcriptome data and visualized as heatmaps generated with pheatmap (v1.0.12), comparing normal mammary epithelial and breast cancer cells. Pearson correlation analysis (stats package) was used to identify 5mC and m6A genes most strongly associated with DDR1 expression.

m⁶A modification site prediction

Potential m⁶A modification sites on DDR1 mRNA were predicted using the SRAMP web server. The DDR1 nucleotide sequence was retrieved from NCBI in FASTA format and submitted to the online tool for analysis.

Cell transduction and experimental groups

Lentiviral particles were produced in HEK293T cells using Lipofectamine 2000 (#11668030; Thermo Fisher, Waltham, MA, USA). pHAGE-puro (#118692; Addgene, Watertown, MA, USA) was co-transfected with pSPAX2 (#12260; Addgene, Watertown, MA, USA) and pMD2.G (#12259; Addgene, Watertown, MA, USA) for overexpression. pSuper-retro-puro was packaged with gag/pol (#14887; Addgene, Watertown, MA, USA) and VSVG (#8454; Addgene, Watertown, MA, USA) for knockdown. Viral supernatants were collected at 48 and 72 h, filtered (0.45 μm), concentrated, pooled, and titrated. Lentiviral transduction of MCF-7 cells was carried out using packaging services provided by Sangon Biotech (Shanghai, China).

MCF-7 cells in logarithmic growth were seeded at 1 × 10⁵ cells/well in 6-well plates. Cells at ~75% confluence were transduced with lentivirus (MOI = 10; ~5 × 10⁶ TU/mL) and 5 μg/mL of polybrene (TR-1003; Merck) for 4 h, followed by medium dilution and replacement after 24 h. Cells were cultured with a stepwise increase in puromycin (2–10 μg/mL, E607054; Sangon Biotech) until no further cell death occurred. Knockdown efficiency was verified by western blot and RT-qPCR (shRNA sequences listed in Table S1).

The experimental groups were as follows: control group (MCF-7 cells transduced with empty vector lentivirus); oe-DNMT3A group (MCF-7 cells overexpressing DNMT3A); oe-METTL7A group (MCF-7 cells overexpressing METTL7A); oe-DNMT3A + oe-METTL7A group (MCF-7 cells co-overexpressing DNMT3A and METTL7A); and oe-DNMT3A + oe-METTL7A + sh-DDR1 group (MCF-7 cells co-overexpressing DNMT3A and METTL7A with DDR1 knockdown). Each group was tested in triplicate³².

Site-specific m⁶A methylated RNA immunoprecipitation (MeRIP)-qPCR assay

The DDR1 mRNA m⁶A modification levels were evaluated using the MeRIP m⁶A kit (17-10499-2; Merck). Co-precipitated samples and input controls were digested with proteinase K and RNA extraction. DDR1 mRNA enrichment levels were quantified by RT-qPCR with experiments performed in triplicate. The levels of expression were normalized to input controls. The DDR1 (human) primer sequences were as follows: forward, 5′-GATCTCGACTCCGCTTCAAGGA-3′; and reverse, 5′-CAAAGGGTGTCCCTTACGCACA-3′.

Clonal bisulfite sequencing

DNA methylation analysis at cg13329862 (DDR1) was performed using clonal bisulfite sequencing (CBS). Genomic DNA was extracted with the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) and quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher). Bisulfite conversion of 500 ng of genomic DNA was performed with the EZ DNA Methylation Kit (Zymo Research, Irvine, CA, USA). Converted DNA was resuspended in 40 μL of TE buffer and stored at −20°C.

PCR amplification was carried out in a 20-μL reaction containing 2 μL of bisulfite-treated DNA, 2 μL of each forward (5′-TGATTTTGGGGTTGTTTGTTAGTA-3′) and reverse (5′-AATACTTTTTCCCCACTCAACACTA-3′) primer (2 μM each), 1× PCR buffer, 10 mM dNTPs (TR1122; Thermo Fisher), 1 mM MgCl₂ (AM9530G; Thermo Fisher), and 1 U ZymoTaq™ DNA polymerase (E2002; Zymo Research). The cycling conditions were as follows: 95°C for 10 min; 35 cycles at 95°C for 30 s, 59°C for 40 s, and 72°C for 1 min; and a final extension at 72°C for 10 min. PCR products were resolved on a 1.5% agarose gel stained with purified GelRed (K310001; Thermo Fisher), and cloned into pCR^®4-TOPO^® (K4575-02; Thermo Fisher, Waltham, MA, US).

The cloning products were transformed into E. coli TOP10 competent cells (Thermo Fisher) and plated on LB/ampicillin (50 μg/mL; Beyotime, China) agar plates. Positive clones were PCR-verified (T3/T7 primers) and sequenced (Genewiz) after an overnight incubation at 37°C. Bisulfite sequencing DNA methylation analysis (BISMA) was used to analyze methylation, excluding redundant clones and sequences with < 95% conversion efficiency. The methylation frequency at target CpGs was calculated statistically.

Co-immunoprecipitation assay

Protein-protein interactions (PPIs) between DDR1 and STAT3 were analyzed using the Pierce Co-immunoprecipitation (Co-IP) Kit (26149; Thermo Fisher). Cell lysates were incubated with DDR1 antibody (ab288675; Abcam, UK)/STAT3 (ab68153; Abcam) or IgG control antibody (ab172730; Abcam), followed by the addition of Protein A/G magnetic beads (ab286842; Abcam) to capture the antigen-antibody complexes. The presence of STAT3 in immunoprecipitates was detected after elution by western blot analysis.

Cell Counting Kit-8 cell viability assay

Cell viability was assessed using the cell counting kit-8 (CCK-8) assay. Cells from each experimental group were seeded in 96-well plates (CLS3922; Corning, USA) at a density of 5 × 10³ cells/well (n = 5 replicates per group). After 24, 48, and 72 h of culture, 10 μL of CCK-8 reagent (CK04; Dojindo, Japan) was added to each well and incubated at 37°C for 2 h. Absorbance at 450 nm (OD450) was monitored utilizing a microplate reader (168–1130; Bio-Rad, USA) to evaluate cell viability. The experiment was performed in triplicate.

EdU cell proliferation assay

Cells (1 × 10⁵/well) in 24-well plates (Corning) were incubated with 50 μM EdU (RiboBio, Guangzhou, China) for 2 h at 37°C, fixed with 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA), and stained using Apollo and DAPI (Thermo Fisher). Images were captured using a Leica microscope (Germany) from five randomly selected fields per sample and the number of EdU⁺ cells was quantified. All experiments were performed in triplicate.

Colony formation assay

Cells (500/well) were seeded in 6-well plates (TMO140644; Corning) and cultured for 14 d at 37°C with medium replaced every 3 d. The colonies were fixed with 4% paraformaldehyde (158127; Sigma-Aldrich), stained with 0.1% crystal violet (C3886; Sigma-Aldrich), and counted after washing. Experiments were performed in triplicate.

Wound healing assay

Cells were pretreated with 10 μg/mL of mitomycin C (M5353; Sigma-Aldrich) for 2 h. Subsequently, 5 × 10⁵ cells per well were seeded in 6-well plates (3516; Corning) and cultured until > 90% confluence. Uniform wounds were created in the monolayer using a sterile 10-μL pipette tip, followed by 2 gentle washes to remove detached cells. Serum-free medium was added and the wound areas were photographed at 0 and 36 h using an inverted microscope (Olympus, Japan) at identical positions. Wound widths were measured using ImageJ 6.0 software (Media Cybernetics, USA), and cell migration ability was calculated as follows: wound closure (%) = [(width at 0 h − width at 36 h)/width at 0 h] × 100%.

Transwell invasion and migration assays

Cell invasion was assessed using Transwell chambers (8-μm pores; Corning) pre-coated with Matrigel (BD Biosciences), diluted 1:8, and incubated at 37°C for 30 min. After a 12-h serum starvation, 1 × 10⁵ cells/mL in serum-free medium were seeded in upper chambers, while lower chambers contained 600 μL of medium with 10% FBS as a chemoattractant. Following a 24-h incubation, non-invaded cells were removed and the invaded cells were fixed with methanol, stained with 1% crystal violet (Beyotime), then quantified by counting five random fields per membrane. Migration assays were performed using uncoated Transwell chambers (8-μm pore size, 3422; Corning) following the same procedure.

Treg differentiation assay

Human CD4⁺ T cells (JY-J1251; Shanghai Jinyuan Biotechnology, Shanghai, China) were cultured in RPMI-1640 (11875093; Gibco) supplemented with 10% FBS, 1% penicillin-streptomycin, 10 mM HEPES (15630080; Gibco), 2 mM L-glutamine (G8540; Sigma-Aldrich), 1 mM sodium pyruvate (11360070; Gibco), MEM non-essential amino acids (11140050; Gibco), and 55 μM β-mercaptoethanol (31350010; Gibco). The cells (1 × 10⁶/mL) were stimulated with anti-CD3 (1 μg/mL, 317326; BioLegend) and anti-CD28 (1 μg/mL, 302934; BioLegend) and co-cultured with breast cancer cells (1:1 ratio) from different treatment groups for 72 h. Flow cytometry (BD LSRFortessa, USA) analyzed CD4+CD25+Foxp3+ Tregs via fixation and staining with anti-CD4-FITC (11-0049-42; Invitrogen), anti-CD25-PE (12-0259-42; Invitrogen), and anti-Foxp3-Alexa Fluor 647 (560045; BD Biosciences), per manufacturers’ protocols, followed by data analysis using FlowJo (v10.7.1) was accessed as a commercial software³².

Chemotaxis assay

Treg migration was evaluated using Transwell chambers (8 μm-pore; Corning) with 600 μL of 10% FBS medium (1:1 mixed with conditioned medium) in the lower chamber and 5 × 10⁵ CD4⁺ T cells (serum-free) in the upper chamber. Non-migrated cells were removed after 12 h, while migrated cells were fixed in 4% PFA for 30 min, stained with 0.1% crystal violet for 20 min, dried, and counted in five random fields.

Animal model establishment and grouping

Female nude mice (BALB/c, 4–6 weeks old; Charles River, China) were maintained under specific pathogen-free (SPF) conditions at 22 ± 2°C with 50 ± 10% humidity and 12 h light/dark cycles. The mice had access to standard chow and sterile water ad libitum. Breast cancer xenografts were established by subcutaneously injecting 1 × 10⁶ MCF-7 cells in 100 μL of PBS with different genetic modifications into the right flank under sterile conditions. All animal procedures were approved by the Ethics Committee of the General Hospital of Ningxia Medical University (Approval No.: KYLL-2025-1285) and performed in compliance with ethical guidelines to minimize suffering.

The experimental groups (n = 6 per group) were as follows: control (empty vector-transduced MCF-7 cells); oe-DNMT3A (DNMT3A-overexpressing cells); oe-METTL7A (METTL7A-overexpressing cells); oe-DNMT3A + oe-METTL7A (dual-overexpressing cells); and oe-DNMT3A + oe-METTL7A + sh-DDR1 (dual-overexpressing with DDR1 knockdown). Tumor growth was measured every 3 d using calipers (Mitutoyo, Japan) and the volume (V) was calculated as 0.5 × length (L) × width (W)². At day 21 post-inoculation, the tumors were surgically excised, weighed (Mettler Toledo, Switzerland), and recurrence rates and times were monitored for 15 d post-excision³³.

Flow cytometric analysis of tumor-infiltrating Tregs

Single-cell suspensions were prepared from tumor tissues using a Tumor Dissociation Kit (#130-096-730; Miltenyi Biotec). Cells were stained with anti-CD4-FITC (1:200, #100405; BioLegend), anti-CD25-PE (1:100, #113703; BioLegend), and anti-Foxp3-Alexa Fluor 647 (1:100, #126407; BioLegend) for 30 min after RBC lysis (#R7757; Sigma-Aldrich). Intracellular fixation/permeabilization was performed using eBioscience buffers™ (#88-8824-00 and #00-5523-00; Thermo Fisher). Flow cytometry and FlowJo analysis quantified CD4+CD25+Foxp3+ Tregs among tumor-infiltrating lymphocytes.

Statistical analysis

Data are presented as the mean ± SD from ≥ 3 independent experiments. Two-group comparisons were performed using an independent samples t-test, multiple-group comparisons were performed using one-way analysis of variance (ANOVA), and data from different time points were analyzed using two-way ANOVA. Analyses were performed in GraphPad GraphPad Prism 9.5.0 (GraphPad Software, San Diego, CA, USA) and R 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria) with *P* < 0.05 considered significant.

Bioinformatics reveals that discoidin domain receptor 1 is epigenetically regulated to promote regulatory T cell infiltration in breast cancer

RNA-seq analysis of MCF-10A (normal mammary epithelial) and MCF-7 (breast cancer) cells (Figure 1A) identified 1454 DEGs (|log2FC| > 3; adj. P < 0.05), including 586 upregulated and 868 downregulated genes (Figure 1B). GO analysis revealed enrichment in epidermal development (biological process [BP]), collagen-containing extracellular matrix (cellular component [CC]), and receptor–ligand activity (molecular function [MF]; Figure 1C). KEGG analysis indicated enrichment in the PI3K–Akt and JAK–STAT signaling pathways (Figure 1D). We next focused on the JAK/STAT3 pathway because recent evidence suggested that DDR1 not only promotes breast cancer cell adhesion, migration, and invasion but also mediates immune-related signaling that regulates immune cell recruitment²⁰. Specifically, DDR1 activates downstream STAT3 signaling, inducing CXCL5 expression, a chemokine essential for neutrophil and Treg migration²¹.

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Identifying key DEGs in breast cancer via RNA-seq and multiple machine learning algorithms. (A) Workflow of differential gene expression analysis between MCF-10A and MCF-7 cells based on RNA-seq; (B) Volcano plot of DEGs analyzed using the limma package; green indicates downregulated genes, purple indicates upregulated genes (|log2FC| > 3, adj. P Value < 0.05); (C) GO enrichment analysis of significantly enriched BP, MF, and CC; (D) KEGG pathway enrichment analysis of DEGs; (E–G) Screening of DEGs using ElasticNet, LASSO, and Ridge algorithms, respectively; (H) Screening results using the SVM-RFE algorithm; (I) Intersection analysis of key genes identified by the four machine learning algorithms (ElasticNet, Ridge, SVM-RFE, and LASSO). Only DDR1 was consistently selected by all four methods; (J) DDR1 expression in MCF-7 and MCF-10A cells, derived from RNA-seq data. Data are presented as the mean ± standard deviation (mean ± SD). **** indicates P < 0.0001.

Four machine learning algorithms were used to screen the DEGs to identify key genes involved in the onset and progression of breast cancer, which resulted in the identification of 649 genes using the ElasticNet algorithm (Figure 1E), 13 genes using the LASSO algorithm (Figure 1F), 9 genes using the Ridge algorithm (Figure 1G), and 5 genes using the SVM-RFE algorithm (Figure 1H). The intersection of these results yielded a single common gene, DDR1, highlighting its critical role in breast cancer progression. Thus, DDR1 was selected as the core candidate gene (Figure 1I), which was significantly upregulated in breast cancer tissues (Figure 1J). These findings suggested that DDR1 has a crucial role in regulating the development and progression of breast cancer.

Methylation of DNA and RNA, especially 5mC and m⁶A, has a pivotal role in various BPs. Large-scale analyses of 11,080 samples from 33 human cancer types demonstrated that most regulators of 5mC and m⁶A modifications exhibit similar expression patterns and extensive regulatory interactions^34,35, indicating cooperative involvement in cancer development and potential reciprocal regulation.

Twenty well-recognized 5mC- and 25 m⁶A-related genes were identified through a literature review to further explore the epigenetic genes influencing DDR1 expression. Expression data for these genes were extracted from the RNA-seq dataset and visualized as heatmaps (Figure 2A, B). Co-expression analysis showed that DDR1 exhibited the strongest positive correlation with the 5mC regulator, DNMT3A (Figure 2C), and the m⁶A regulator, METTL7A (Figure 2D). Furthermore, multiple potential m⁶A modification sites on DDR1 mRNA were predicted using the SRAMP database (Figure 2E, F).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Co-expression analysis of DDR1 with genes related to 5mC and m⁶A modifications. (A) Heatmap of 20 5mC-related genes in the RNA-seq dataset; (B) Heatmap of 25 m⁶A-related genes in the RNA-seq dataset; (C) Co-expression correlation between DDR1 and 5mC-related genes; (D) Co-expression correlation between DDR1 and m⁶A-related genes; (E) Predicted potential m⁶A modification sites on DDR1 mRNA using the SRAMP database; (F) Distribution of predicted m⁶A sites on DDR1 mRNA based on SRAMP analysis.

The pivotal role of DDR1 in breast cancer was established through these analyses, revealing likely regulation by the epigenetic modifiers, DNMT3A and METTL7A.

DNA methyltransferase 3A and methyltransferase-like 7A collaborate to regulate discoidin domain receptor 1 and activate breast cancer signaling pathways

Overexpressing DNMT3A or METTL7A cell lines were constructed in MCF-7 cells via lentiviral transduction to determine how DNMT3A and METTL7A regulate DDR1 expression through epigenetic modifications (Figure S1A).

RT-qPCR and western blot analyses of transduction efficiency revealed significant upregulation of DNMT3A in the oe-DNMT3A group and METTL7A in the oe-METTL7A group with both DNMT3A and METTL7A showing higher expression in the combined oe-DNMT3A+oe-METTL7A group compared to the single overexpression group (Figure S1B, C).

Site-specific MeRIP-qPCR revealed significantly increased m⁶A modification of DDR1 mRNA in the oe-METTL7A and oe-DNMT3A+oe-METTL7A groups relative to controls, whereas no change was observed in the oe-DNMT3A group (Figure S1D). CBS analysis of DNA methylation in the promoter region of DDR1 revealed significant increases in methylation levels in both the oe-DNMT3A and oe-DNMT3A+oe-METTL7A groups with no significant change in the oe-METTL7A group (Figure S1E). These results indicated that DNMT3A primarily regulates DNA methylation, while METTL7A influences m⁶A modification, contributing to the epigenetic regulation of DDR1.

RNA-seq analysis revealed significant enrichment of DEGs in the JAK-STAT pathway. The effect of DDR1 on STAT3/CXCL5 signaling was examined given reports that DDR1 activates STAT3 and promotes CXCL5-mediated neutrophil infiltration and T-cell activation^36–38. RT-qPCR and western blot showed upregulated DDR1, STAT3, and CXCL5 mRNA/protein levels in the oe-DNMT3A and oe-METTL7A groups compared to controls with the highest levels in the combined overexpression group (Figure S1F, G). Co-IP confirmed a DDR1-STAT3 interaction (Figure S1H).

These findings demonstrated that DNMT3A and MET-TL7A collaborate to regulate DDR1 expression and activate the DDR1/STAT3/CXCL5 axis, potentially influencing the breast cancer microenvironment.

Silencing discoidin domain receptor 1 reverses activation of the breast cancer signaling axis by DNA methyltransferase 3A and methyltransferase-like 7A

Rescue experiments with DDR1 silencing confirmed that DNMT3A and METTL7A epigenetically regulate DDR1 expression. First, the silencing efficiency of shRNA targeting DDR1 was evaluated. RT-qPCR and western blot results demonstrated that sh-DDR1 significantly reduced DDR1 RNA and protein expression compared to sh-NC and this sequence was selected for subsequent assays (Figure S2A, B). DDR1 expression was significantly elevated in the oe-DNMT3A+oe-METTL7A group, whereas co-transfection with sh-DDR1 restored mRNA and protein levels to near-control values (Figure S3A, B). These findings demonstrated that DDR1 knockdown reverses the effects of DNMT3A and METTL7A, supporting the epigenetic co-regulation of DDR1.

Further analyses of STAT3 and CXCL5 mRNA and protein levels by RT-qPCR and western blot showed significant upregulation in the oe-DNMT3A+oe-METTL7A group compared to the control group, while significant downregulation was observed in the oe-DNMT3A+oe-METTL7A+sh-DDR1 group with no significant difference from the control group (Figure S3C, D). In the absence of DNMT3A or METTL7A overexpression DDR1 knockdown alone significantly decreased STAT3 and CXCL5 expression at both the mRNA and protein levels compared to the sh-NC group (Figure S2C, D). These results demonstrated that DDR1 silencing effectively blocked activation of the DDR1/STAT3/CXCL5 axis by DNMT3A and METTL7A, further supporting the central role of DDR1 in activating this signaling pathway.

Collectively, the findings demonstrated that DNMT3A and METTL7A epigenetically upregulate DDR1, thereby activating the DDR1/STAT3/CXCL5 axis, which may facilitate breast cancer progression and modulate the TME.

DNA methyltransferase 3A and methyltransferase-like 7A collaborate to enhance breast cancer cell proliferation and invasion capabilities

Overexpression models were established via lentiviral transduction to investigate the roles of DNMT3A and METTL7A in breast cancer progression. CCK-8 assays revealed that oe-DNMT3A and oe-METTL7A significantly enhanced cell viability with a further increase in the oe-DNMT3A+oe-METTL7A group (Figure 3A). EdU staining confirmed higher proliferation rates in the overexpression groups, especially in the combined group (Figure 3B). Similarly, clone formation assays showed an increase in the number of colonies in the oe-DNMT3A and oe-METTL7A groups, which was further elevated in the dual-overexpression group (Figure 3C). Western blot analysis revealed upregulated proliferation markers (Ki67 and PCNA) in all overexpression groups with the highest expression occurring in the combined group (Figure 3D).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

DNMT3A and METTL7A synergistically promote breast cancer cell proliferation and migration. (A) CCK-8 assay for cell viability; (B) EdU staining assay to assess cell proliferation; (C) Colony formation assay to evaluate clonogenic capacity; (D) Western blot analysis of proliferation markers (Ki67 and PCNA); (E) Scratch wound assay to assess cell migration; (F) Transwell and invasion assay to evaluate invasive capacity; (G) Western blot analysis of migration and invasion markers (E-cadherin and MMP-9). All experiments were performed in triplicate. Multiple-group comparisons were performed using one-way ANOVA, and data from different time points were analyzed using two-way ANOVA. Data are expressed as the mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Scratch and Transwell assays demonstrated that both oe-DNMT3A and oe-METTL7A enhanced cell migration and invasion compared to controls with further increases in the combined oe-DNMT3A+oe-METTL7A group (Figure 3E, F). Western blot showed corresponding E-cadherin downregulation and MMP-9 upregulation in the oe-DNMT3A and oe-METTL7A groups with more pronounced changes in the dual-overexpression group (Figure 3G).

These findings indicated that DNMT3A and METTL7A markedly promote breast cancer cell proliferation, migration, and invasion with synergistic effects when co-expressed, likely through coordinated regulation of key signaling pathways.

Silencing discoidin domain receptor 1 significantly suppresses the proliferative and invasive effects of DNA methyltransferase 3A and methyltransferase-like 7A in breast cancer cells

A rescue experiment was performed by silencing DDR1 to assess the pivotal role of DDR1 in mediating the proliferative and migratory effects of DNMT3A and METTL7A.

CCK-8 assays showed higher cell viability in the oe-DNMT3A+oe-METTL7A group versus controls, which was reversed by DDR1 knockdown (oe-DNMT3A+oe-METTL7A+sh-DDR1; Figure 4A). EdU staining revealed a higher proportion of EdU⁺ cells in the oe-DNMT3A+oe-METTL7A group that was reduced upon DDR1 silencing (Figure 4B). Colony formation assays showed similar trends with an elevated number of colonies in the dual-overexpression group and a significant reduction after DDR1 knockdown (Figure 4C). Western blot analysis confirmed upregulation of Ki67 and PCNA in the oe-DNMT3A+oe-METTL7A group, which was markedly decreased by DDR1 silencing (Figure 4D).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Silencing DDR1 suppresses DNMT3A- and METTL7A-induced proliferation and migration of breast cancer cells. (A) CCK-8 assay for cell viability; (B) EdU staining assay to assess cell proliferation; (C) Colony formation assay to evaluate clonogenic capacity; (D) Western blot analysis of proliferation markers (Ki67 and PCNA); (E) Scratch wound assay to assess cell migration; (F) Transwell and invasion assay to evaluate invasive capacity; (G) Western blot analysis of migration and invasion markers (E-cadherin and MMP-9). All experiments were performed in triplicate. Multiple-group comparisons were performed using one-way ANOVA and data from different time points were analyzed using two-way ANOVA. Data are expressed as the mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The scratch assay demonstrated a significantly greater migration distance in the oe-DNMT3A+oe-METTL7A group compared to controls, which was markedly reduced by DDR1 knockdown (Figure 4E). Transwell migration and invasion assays further showed that the number of cells was substantially higher in the dual-overexpression group, whereas DDR1 silencing suppressed both migration and invasion (Figure 4F). Western blot analysis of migration and invasion markers revealed that E-cadherin was significantly downregulated and MMP-9 significantly upregulated in the oe-DNMT3A+oe-METTL7A group. However, silencing DDR1 reversed these effects, leading to upregulation of E-cadherin and downregulation of MMP-9 (Figure 4G).

These findings highlight the essential role of DDR1 in mediating the cooperative effects of DNMT3A and METTL7A on breast cancer cell proliferation, migration, and invasion with DDR1 silencing effectively abolishing the synergistic activity and underscoring the potential as a therapeutic target.

The discoidin domain receptor 1 / signal transducer and activator of transcription 3 / C-X-C motif chemokine ligand 5 axis promotes regulatory T cell differentiation and migration induced by breast cancer cells

DDR1 and CXCL5 have been shown to induce NET formation, thereby promoting Treg immune infiltration and driving tumor growth and metastasis³². Breast cancer cell lines overexpressing DNMT3A and METTL7A were established via lentiviral transduction to investigate the role of the DDR1/STAT3/CXCL5 axis in regulating Treg differentiation and migration. These cells were then co-cultured with human CD4⁺ T cells, as illustrated in the experimental workflow (Figure 5A).

• Download figure
• Open in new tab
• Download powerpoint

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

The DDR1/STAT3/CXCL5 axis promotes Treg differentiation and migration induced by breast cancer cells. (A) Schematic of the experimental design; (B) Flow cytometry analysis of Treg proportions (CD4⁺CD25⁺Foxp3⁺); (C) Immunofluorescence staining of the Treg marker, Foxp3; (D) Transwell assay assessing the migration capacity of Tregs. All experiments were performed in triplicate. Data are expressed as the mean ± SD. Multiple-group comparisons were performed using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Flow cytometry analysis of Treg cell proportions (CD4⁺CD25⁺Foxp3⁺) revealed that both the oe-DNMT3A and oe-METTL7A groups exhibited significantly increased Treg ratios compared to the control group with the highest proportion in the oe-DNMT3A+oe-METTL7A group (Figure 5B). Immunofluorescence staining showed a marked upregulation of the Treg marker, Foxp3, in both single overexpression groups with even higher expression in the combined overexpression group (Figure 5C). These results suggested that the DDR1/STAT3/CXCL5 axis may synergistically promote Treg differentiation in the context of breast cancer.

To assess the ability of breast cancer cells to induce Treg migration, a Transwell assay was performed using conditioned media from the different groups. Both oe-DNMT3A and oe-METTL7A groups exhibited significantly enhanced Treg migration relative to the control with further enhancement observed in the oe-DNMT3A+oe-METTL7A group (Figure 5D). These results indicate that the DDR1/STAT3/CXCL5 axis promotes Treg migration by stimulating CXCL5 secretion.

Collectively, the DDR1/STAT3/CXCL5 signaling axis markedly promotes Treg differentiation and migration with these effects amplified by the cooperative activity of DNMT3A and METTL7A. These findings provided critical evidence for understanding the regulatory mechanisms of the breast cancer immune microenvironment.

Silencing discoidin domain receptor 1 significantly suppresses the ability of breast cancer cells to induce regulatory T cell differentiation and migration

Rescue experiments with DDR1 knockdown demonstrated the essential role in DNMT3A/METTL7A-driven Treg induction. Flow cytometry analysis revealed a marked increase in CD4⁺CD25⁺Foxp3⁺ Tregs in the oe-DNMT3A+oe-METTL7A group, which was significantly reduced following DDR1 silencing (oe-DNMT3A+oe-METTL7A+sh-DDR1; Figure 6A). Immunofluorescence staining confirmed elevated Foxp3 expression in the combined overexpression group with substantial attenuation upon DDR1 knockdown (Figure 6B), indicating DDR1 is essential for Treg differentiation in this axis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Silencing DDR1 suppresses the ability of breast cancer cells to induce Treg differentiation and migration. (A) Flow cytometry analysis of Treg proportions (CD4⁺CD25⁺Foxp3⁺); (B) Immunofluorescence staining of the Treg marker, Foxp3; (C) Transwell assay assessing Treg migration. All experiments were performed in triplicate. Data are expressed as the mean ± SD. Multiple-group comparisons were performed using one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

A Transwell chemotaxis assay was performed using conditioned media to assess the effect of DDR1 on Treg migration. The results demonstrated that Treg migratory capacity was enhanced in the oe-DNMT3A+oe-METTL7A group relative to controls, whereas DDR1 knockdown substantially reduced this effect (Figure 6C). These results suggested that CXCL5, acting through activation of the DDR1/STAT3 axis, markedly enhances Treg chemotaxis, an effect that can be effectively reversed by DDR1 silencing.

In conclusion, DDR1 facilitated Treg differentiation and migration in the breast cancer microenvironment. DDR1 silencing effectively inhibited the immunoregulatory effects of DNMT3A and METTL7A, providing important mechanistic insights into the epigenetic regulation of tumor immune evasion in breast cancer.

DNA methyltransferase 3A and methyltransferase-like 7A synergistically promote breast cancer growth, recurrence, and immune microenvironment modulation

The in vivo roles of DNMT3A and METTL7A in breast cancer recurrence and modulation of the immune microenvironment were examined using a nude mouse xenograft model, followed by molecular and histologic analyses, as outlined in the experimental workflow (Figure S4A).

Tumor volumes increased significantly in the oe-DNMT3A and oe-METTL7A groups relative to controls with the greatest increase observed in the oe-DNMT3A+oe-METTL7A group (Figure S4B). The excised tumors in these groups exhibited higher weights after 28 d, especially in the combined overexpression group (Figure S4C, D). Additionally, the oe-DNMT3A+oe-METTL7A group displayed accelerated recurrence and higher recurrence rates (Figure S4E), indicating that DNMT3A and METTL7A synergistically promote breast cancer growth and recurrence.

Immunohistochemistry (IHC) analyses revealed significantly increased expression of DDR1, STAT3, and CXCL5 in tumor tissues from the oe-DNMT3A and oe-METTL7A groups compared to controls with further upregulation in the combined overexpression group (Figure S4F). Flow cytometry demonstrated a significantly higher proportion of Tregs in tumor tissues from the oe-DNMT3A and oe-METTL7A groups with the highest proportion in the combined group (Figure S4G). Immunofluorescence staining for Cit-H3 and DNA further showed that NET formation was markedly enhanced in the oe-DNMT3A and oe-METTL7A groups and significantly elevated in the combined group (Figure S4H).

Collectively, these findings indicated that DNMT3A and METTL7A promote breast cancer growth, recurrence, and immune microenvironment remodeling through activation of the DDR1/STAT3/CXCL5 axis, thereby enhancing Treg infiltration and NETs formation. This study provided new evidence supporting the molecular mechanisms underlying breast cancer progression.

Silencing DDR1 attenuates DNMT3A- and METTL7A-mediated breast cancer recurrence and immune microenvironment modulation

A rescue experiment involving DDR1 silencing was performed to verify the critical role of DDR1 in DNMT3A- and METTL7A-mediated breast cancer recurrence and immune microenvironment regulation. Tumor growth, recurrence, and related molecular markers were subsequently analyzed, as outlined in the experimental design (Figure S5A).

Tumor volume and weight were significantly increased in the oe-DNMT3A+oe-METTL7A group compared to controls, whereas DDR1 knockdown (sh-DDR1) reversed these effects (Figure S5B–D). Recurrence rates were elevated and recurrence times shortened in the oe-DNMT3A+oe-METTL7A group, but both parameters were suppressed by DDR1 silencing (Figure S5E). These results indicated that DDR1 is critical for DNMT3A/METTL7A-mediated tumor growth and recurrence.

IHC staining showed elevated DDR1, STAT3, and CXCL5 expression in the oe-DNMT3A+oe-METTL7A group relative to controls, which was reduced following DDR1 knockdown (Figure S5F). Flow cytometry revealed increased Treg infiltration in the oe-DNMT3A+oe-METTL7A group, an effect reversed by DDR1 silencing (Figure S5G). Immunofluorescence demonstrated enhanced NET formation (Cit-H3/DNA co-localization) in the oe-DNMT3A+oe-METTL7A group, while DDR1 silencing suppressed NETs (Figure S5H).

Taken together, these results demonstrated that DDR1 silencing significantly impairs the ability of DNMT3A and METTL7A to promote breast cancer recurrence and modulate the immune microenvironment, supporting DDR1 as a promising therapeutic target in breast cancer.