MBD2 suppresses SFRP1 expression and promotes colorectal cancer development by blocking MED19 binding to its methylated promoter

Abstract

Background

The Wnt signaling pathway antagonist SFRP1 is frequently silenced by promoter DNA hypermethylation in colorectal cancer (CRC). MBD2, a DNA methylation reader, is known to contribute to SFRP1 epigenetic silencing. Previous work showed that MBD2 critically suppresses SFRP1 expression without altering promoter methylation, though the underlying mechanism remained unclear. Elucidating how DNA methylation silences tumor suppressor genes, such as SFRP1, could reveal novel therapeutic targets with significant clinical potential.

Methods

MBD2 was inhibited in CRC models using either siRNA or a small molecule inhibitor (KCC07). The effects on SFRP1 and β-catenin expression, Wnt pathway activity, cell proliferation, and apoptosis were assessed. Tumor growth was also evaluated in vivo. Mechanistic studies investigated the role of MBD2 in mediating MED19 binding to the SFRP1 promoter and its impact on RNA polymerase II CTD-S7 phosphorylation.

Results

The IC50 of KCC07 was 23.25 μM in SW480 cells, 26.83 μM in HCT116 cells, and 39.66 μM in NCM460 cells. Inhibition of MBD2, either genetically or pharmacologically with KCC07, upregulated SFRP1 expression, downregulated β-catenin, and suppressed the Wnt pathway. KCC07 treatment also inhibited CRC cell proliferation, promoted apoptosis, and suppressed tumor growth in vivo. Mechanistically, MBD2 was found to silence SFRP1 by blocking MED19 binding to its promoter, which subsequently reduced RNA polymerase II CTD-S7 phosphorylation and impaired transcription.

Conclusions

This study reveals a novel mechanism whereby DNA methylation suppresses gene expression via MBD2, independent of changes in methylation status, by disrupting MED19 binding and subsequent transcription. Targeting MBD2 represents a promising therapeutic strategy for colorectal cancer.

Introduction

Colorectal cancer (CRC) ranks as the third most common malignancy and the second leading cause of cancer-related mortality worldwide, with over 1.9 million new cases and approximately 900,000 deaths annually.¹ Despite therapeutic advances, CRC demonstrates limited sensitivity to conventional treatments, resulting in unsatisfactory survival rates.^2-4 Therefore, identifying novel therapeutic targets remains a critical need.

DNA methylation plays a pivotal role in gene expression regulation.⁵ Neoplastic cells exhibit distinct methylation patterns, featuring genome-wide hypomethylation coupled with localized CpG island hypermethylation.^6,7 These epigenetic alterations significantly contribute to CRC pathogenesis, with tumor suppressor gene silencing via promoter hypermethylation serving as a key driver of oncogenesis.^8,9

DNA methyltransferase inhibitors (DNMTis) can reverse promoter hypermethylation and restore tumor suppressor expression.¹⁰ However, their nonspecific genome-wide hypomethylation may induce genomic instability and oncogene activation, limiting their use to hematological malignancies.^7,11,12 Thus, alternative strategies that reactivate tumor suppressor genes without altering global methylation patterns are urgently needed.

Secreted frizzled-related protein 1 (SFRP1) functions as a Wnt pathway antagonist in normal colonic epithelium,¹³ and its downregulation leads to Wnt hyperactivation in various malignancies.^14-17 We previously showed that SFRP1 promoter hypermethylation occurs early in CRC and contributes to proliferation and migration.¹⁸ While DNMTis can restore SFRP1 expression,¹⁸ the precise mechanisms underlying methylation-dependent SFRP1 silencing remain incompletely understood.

Methylation-mediated gene silencing involves two primary mechanisms: direct inhibition of transcription factor binding¹⁹ and recruitment of methyl-CpG binding domain (MBD) proteins.²⁰ Among MBD family members, MBD2 shows particular affinity for methylated CpG islands associated with silenced genes.^21,22 Notably, our prior research demonstrated that MBD2 knockdown restores SFRP1 expression without altering promoter methylation status,^23,24 challenging the conventional paradigm of transcription factor exclusion from methylated DNA. Given that many transcription factors can bind methylated sequences,²⁵ methylation-mediated silencing likely involves both direct and indirect effects, though the mechanistic details remain unclear.

How does MBD2 silence SFRP1 without altering its methylation status? MBD2 typically mediates silencing through chromatin remodeling via the MBD2-NuRD complex;^21,26 however, our previous work showed that histone deacetylase inhibition alone fails to restore SFRP expression,¹⁸ suggesting alternative MBD2-dependent mechanisms. Targeting MBD2 without altering methylation patterns could provide a safer alternative to DNMTis, underscoring the clinical importance of elucidating these mechanisms.

In the present study, we aimed to identify key factors that interact with MBD2 and are directly involved in the transcriptional regulation of SFRP1. We demonstrate that mediator complex subunit 19 (MED19) recognizes and binds methylated SFRP1 promoters. MBD2 competitively inhibits this interaction, impairing RNA polymerase II CTD serine-7 phosphorylation and thereby silencing SFRP1 expression. These findings reveal a novel mechanism of methylation-dependent gene regulation through MBD2 recruitment, offering new therapeutic targets that avoid the limitations of current DNMTis.

Materials and methods

Human tissue and follow-up data

This study involved three cases of colorectal cancer patients who were pathologically diagnosed at the Digestive Medicine Center of the Seventh Affiliated Hospital of Sun Yat-Sen University between 2022 and 2023, with complete clinical and pathological information. Tumor tissues and matched adjacent tissues were collected during the colonoscopy procedures. Additionally, normal colon mucosa samples were obtained from three individuals who underwent colonoscopy for polyp resection. All nine specimens were fixed and stained for immunohistochemical detection.

None of the included patients had received prior surgical treatment, chemotherapy, or radiotherapy. Written informed consent was obtained from all participants, and the study protocol was approved by the Ethics Committee of the Seventh Affiliated Hospital of Sun Yat-Sen University (No. KY-2025-050-01).

Cell cultures

Normal colon mucosa cell line NCM460 (ATCC, CRL-1642^TM) and the human colorectal cancer cell lines HCT116 (ATCC, CCL-247 ^TM) and SW480 (ATCC, CCL-228 ^TM) were purchased from iCell, and all had STR identification reports. All three cell lines used were tested for mycoplasma contamination using the PCR-based MycoAlert™ plus assay (Lonza, USA) prior to the experiments. All the results confirmed the absence of mycoplasma. HCT116 cell line and SW480 cell line were cultured in high-glucose medium (DMEM, Gibco; Thermo Fisher Scientific, Inc.) containing 10% fetal bovine serum (FBS, Biochannel), while NCM460 cell line was cultured in 1640 medium (1640, Biochannel.) containing 10% FBS in a humidified atmosphere at 37 °C and 5% CO². KCC07 was obtained from MedChemExpress (MCE USA, No. HY-131031). The drug was dissolved in dimethyl sulfoxide (DMSO) to prepare a stock solution and was used to treat cells at a final concentration of 25 μM.

Real-time quantitative polymerasechain reaction (RT-qPCR)

The total RNA was extracted by SteadyPure rapid RNA extraction kit (AG, AG21023) and reverse transcribed into cDNA using Evo M-MLV reverse transcription reagent premix (AG, AG11706) according to the manufacturer's instructions. The cDNA was subsequently used for RT-qPCR by using SYBR® Green Pro Taq HS Premix (AG, AG11701) on a real-time system (Bio-Rad, CFX96 Touch). The relative expression levels were analyzed via the 2^−ΔΔCt method with GAPDH as the reference gene. The primers used in this study are shown in Table S1.

Western blot

The protein was lysed by RIPA buffer and collected for centrifugation at 14,000 rpm for 10 min at 4 °C. The concentrations of the protein lysate were detected by BCA protein assay kit (KeyGEN, KGPBCA), and the same amount of protein lysate was mixed with sodium dodecyl sulfate loading buffer. The mixtures were boiled for 10 min and used for western blot detection. The Omni-EasyTM one-step PAGE gels (EpiZyme, PG212-1) were prepared, and the protein lysates were added and separated via electrophoresis. The proteins on the gels were transferred to polyvinylidene fluoride membranes (Merck Millipore, ISEQ00010). The membranes were blocked with 5% nonfat milk and incubated with primary antibodies at 4 °C overnight. After that, the membranes were incubated with secondary antibodies for 1 h at room temperature, and the protein signals were detected with Meilunbio® fg supersensitive ECL luminescence reagent (Meilunbio, MA0186-1). The images were analyzed via ImageJ, and the relative expression was calculated according to the gray values, with GAPDH or tubulin used as the endogenous reference gene. For details regarding the antibodies used in this study, including the amount, supplier, and catalog numbers, please refer to Table S2.

Cell counting kit‑8 (CCK8) proliferation experiment

For the CCK-8 proliferation experiment, SW480 and HCT116 cells were seeded into 96-well plates at a density of 3000 cells per well. GenXion Enhanced Cell Counting Kit-8 (GenXion, GxCK08-1000) was used according to the manufacturers' instructions. The optical density at 450 nm (OD450) of each well was measured at 0, 24, 48 and 72 h after seeding following the manufacturer's instructions.

Analysis of cell apoptosis

The cells were seeded in 6-well plates at 6 × 10⁵ cells/well. After transfection with siRNA or KCC07 treatment for 24 h, SW480 and HCT116 cells were harvested. An Annexin V-FITC apoptosis detection kit (Beyotime, C1062L) was used to detect apoptosis according to the manufacturer's instructions. The cell apoptosis rate was analyzed using a flow cytometer (Beckman CytoFLEX FCM), and the results were analyzed via FlowJo software.

RNA interference and lentivirus transfection

Small interfering RNAs (siRNAs) and lentiviruses were designed and constructed by IGE Biotechnology (Guangzhou, China). The siRNAs were transfected with Lipofectamine RNAiMAX (Thermo), and the lentiviruses were transfected with polybrene according to the instructions. In the “control” group, the cells were treated with transfection reagents without siRNA or lentivirus. The interference efficiency of the siRNAs was detected via RT‒qPCR 48 h later and via western blotting 72 h later. The transfection efficiency of the lentivirus was detected by RT‒qPCR 72 h later and by western blot 96 h later. The sequences of the siRNAs used in this study are listed in Table S3.

Immunofluorescence

NCM460 cells and SW480 cells were cultured in an incubator with or without KCC07 treatment. After the culture medium was removed, the cells were gently washed with PBS solution for three times (5 min each time). And then they were fixed with fresh 4% paraformaldehyde (PFA) for 15 min and then washed three times with PBS (5 min each time). PBS film breaking solution (PBST) containing 0.2% Triton X-100 was added to each well, and the film breaking was left for 20 min at room temperature. The fixed cells were then incubated with blocking solution (normal goat serum) for 1 h at room temperature and then with primary antibodies (anti-MED19) overnight at 4 °C. PBS washes were performed three times, each time for 5 min, followed by a 1-hour incubation with secondary antibody at room temperature and 1 min of 4′,6-diamidino-2-phenylindole (DAPI) staining. PBS washes were performed three times, and the pores were dried at room temperature away from light. The slides were placed under a confocal laser microscope to observe the staining of the cells and to capture the images.

HE staining and Immunohistochemistry

Tissue specimens were obtained during routine colonoscopy, and informed consent was obtained from every participant. Firstly, the sections were deparaffinized by xylene and hydrated by ethanol before staining. For H&E staining, the sections were stained with hematoxylin for 5 min and washed by running water followed by 70% alcohol containing 1% HCl. Then, the sections were stained with eosin for 3 min and sealed for observation under an optical microscope. For IHC staining, an SP Rabbit & Mouse HRP Kit (CWBIO, CW2069) was used according to the instructions. Briefly, antigen retrieval was performed on the sections with pepsase for 20 min. Then, the sections were blocked with 10% goat serum for 30 min and incubated with anti-perilipin-1 (Abcam, ab3526, 1:100) overnight at 4 °C. On the second day, the sections were incubated with secondary antibodies and DAB substrate. After that, the sections were counterstained with hematoxylin and sealed for observation under an optical microscope.

DNA pulldown

Briefly, the target DNA fragment is synthesized in vitro and constructed into the expression vector. The 5'-Biotin TEG PCR primers were designed for the preparation of biotinized target DNA fragments. The sequences of the primers used were as follows:

1F: CGTTTTTCTGATCATGACTTGG; 1R: CAGAGCCAACAGCAACAAGT.

2F: CGCGCAGTAAGTTGGCAGGA; and 2R: CATGAAAGTAACAAAATGCCCCAC.

In order to simulate the hypermethylation status of the promoter region of SFRP1 gene in CRC, synthetic DNA was treated with a methylase. The biotinized DNA was subsequently prepared via PCR amplification, and the resulting DNA fragments were subsequently purified. Nucleoprotein of SW480 cells treated with MBD2 siRNA or not were extracted, and the protein concentrations were measured by BCA reagent according to the instructions. Equal volumes of the purified DNA fragments and 400 μL of Dynabeads® C1 Streptavidin (Thermo Scientific, USA, 65001) were mixed and incubated for 0.5 h at room temperature. The mixture was added to the nucleoprotein samples at 4 °C overnight. Finally, the bound proteins were purified and detected by LC‒MS‒MS/MS.

Chromatin immunoprecipitation assay

A high-sensitivity ChIP kit (ab185913; Abcam) was used to perform ChIP assays on the cell samples according to the manufacturer's protocols. HCT116 cells (5 × 106/IP), HCT116 cells with KCC07 (5 × 106/IP), and HCT116 stably transfected cells with MBD2 knockdown (5 × 106/IP) were collected for each treatment. Next, the proteins and their bound DNA were crosslinked with 1% formaldehyde (F809702; Macklin) for 15 min at RT, followed by quenching with a final concentration of 125 mM glycine (Sigma-Aldrich, 50046) for 5 min at RT. Cross-linked cells were then centrifuged at 500 × g at 4 °C and washed twice with ice-cold PBS. The lysates were then collected and sonicated to shear the DNA into fragments of 500–600 bp in length. The MED19 antibody (PA5-44383; Thermo Fisher Scientific) and a nonimmune IgG antibody (from the High-Sensitivity ChIP Kit; Abcam) were used for immunoprecipitation. The forward primer for the SFRP1 promoter region probe was 5ʹ-GATCCCATTACAGATGCTTGGGAG-3ʹ; the reverse primer was 5ʹ-AGCAGGCTAAGCAAGCCA-3ʹ.

Luciferase reporter assay

The promoter sequences of SFRP1 were synthesized and inserted into the pMIR-Reporter luciferase plasmid vector by IGE Biotechnology (Guangzhou, China). SW480 cells were seeded, and the plasmid vectors were transfected by using Lipofectamine 3000 (Invitrogen, L3000001). The KCC07 was added, or siRNA targeting MED19 was transfected. Finally, the luciferase activities were measured by using a dual-luciferase reporter kit (Promega, E1910) in accordance with the manufacturer's directions.

Coimmunoprecipitation (Co-IP)

SW480 cells were collected, and the proteins were extracted by using RIPA lysate and centrifugation. The magnetic beads were added to the protein liquids, and the mixture was incubated at 4 °C for 4 h. Antibodies against MED19 or the corresponding IgG were subsequently added and incubated overnight. After that, the beads were retrieved and washed on a magnetic frame. The beads were resuspended in RIPA lysate and underwent a 3 min boiling step. Last, western blotting was performed to detect the abundance of MED19 and MBD2 in the supernatants and extractions.

CUT&Tag assay

The CUT&Tag assay was performed with NovoNGS® CUT&Tag 2.0 A High-Sensitivity Kit (for Illumina®) (Novoprotein Scientific, Cat# N259-YH01-01A). Briefly, SW480 cells were collected and enriched with ConA magnetic beads. The samples were incubated with MED19 antibody at 4 °C for 18 h and then with secondary antibody at room temperature for 1 h. After that, the beads were washed with Dig-Hisalt buffer, and the cells were incubated with the protein A-Tn5 transposome at 25 °C for 1 h. The cells were resuspended in 50 μL of tagmentation buffer and incubated at 37 °C for 1 h. Finally, the tagmentation was terminated with 1 μL of 10% SDS at 55 °C for 10 min, and the DNA fragments were extracted with phenol chloroform for further sequencing.

Electrophoretic mobility shift assays (EMSA)

Electrophoretic mobility shift assays (EMSA) were performed using a Chemiluminescent EMSA Kit (Beyotime, China) according to the manufacturer's instructions. Biotin-labeled DNA probes (sequences provided in Supplementary Table 4) encompassing hemi-methylated, fully methylated, and unmethylated states of the target sequence were synthesized. Recombinant MED19 protein (2 μg; Finetest) was incubated with each probe in binding reactions at 25 °C for 20 min. The reaction mixtures were then resolved on 6% native polyacrylamide gels in 0.5 × TBE buffer and subsequently transferred onto a nylon membrane. Biotin-labeled DNA was detected using the chemiluminescent substrate provided in the kit.

Animal experiments

All animal care and experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Sun Yat-Sen University (No. SYSUIACUC-2023003051) according to the standards of the National Institutes of Health. BALB/c nude mice (4~5 weeks old) were purchased from Guangdong Vital River Laboratory Animal Technology Co., Ltd. (China). The HCT116 cell line was injected subcutaneously to establish a tumor formation model in nude mice. After the formation of visible subcutaneous tumors, the mice were randomly divided into two groups (n = 6 per group), with six mice per group used as independent biological samples in a single independent experiment. KCC07 was injected intraperitoneally (100 /kg, 5 d/week) into the treatment group, while the control group received the same volume of control solution. At the 3rd week, the nude mice were euthanized using the carbon dioxide inhalation method, and the subcutaneous tumors were removed, weighed and photographed. The tumor tissues were then fixed with formalin fixation solution and embedded in paraffin. The wax blocks of the subcutaneous tumor tissue were sliced continuously, and HE staining was performed for observation under a microscope.

Statistical analyses

The data of this study were analyzed via SPSS 22.0 and are expressed as mean ± standard deviation (SD). The statistical analyses of differences between different groups were conducted by using independent-sample t tests or one-way ANOVA. p < 0.05 was regarded as significant. All repeated experiments were performed as independent biological replicates. Specifically, the number of biological replicates for the cell function experiments was six, while the number of biological replicates for the sequencing experiments and tissue staining experiments was three. The exact number of biological replicates and the p values for each experiment are provided in the figure legends.

Results

Inhibition of MBD2 upregulated the expression of SFRP1 and inhibited the Wnt signaling pathway activation

To investigate the regulatory role of MBD2 in SFRP1 expression, we performed MBD2 knockdown using four siRNAs in the colorectal cancer cell lines SW480 and HCT116. The most efficient siRNA (siRNA#4, Figure 1A, B) was selected for subsequent experiments. MBD2 depletion significantly upregulated SFRP1 expression at both the mRNA and protein levels in both cell lines (Figure 1C–E). Concurrently, we observed marked downregulation of β-catenin, a central Wnt pathway component, at both the transcriptional and translational levels. To determine whether MBD2 regulates SFRP1 through its methyl-CpG binding activity, we employed KCC07, a small-molecule inhibitor that specifically blocks the interaction of MBD2 with methylated DNA.²⁷ ChIP‒qPCR analysis confirmed that KCC07 effectively disrupted the binding of MBD2 to methylated DNA (Figure 1F). Dose‒response experiments revealed that the half-maximal inhibitory concentration (IC50) of KCC07 was 23.25 μM in SW480 cells, 26.83 μM in HCT116 cells (Figure 1G, H), and 39.66 μM in NCM460 cells (Figure S1). The significantly higher IC50 in noncancerous cells indicates a favorable therapeutic window. Consistent with the genetic knockdown results, pharmacological inhibition of MBD2 by KCC07 significantly elevated SFRP1 expression while suppressing β-catenin levels (Figure 1I–K).

Figure 1.

Inhibition of MBD2 upregulated the expression of SFRP1 and inhibited Wnt signaling pathway activation. (A and B) The interference efficiency of MBD2 siRNAs at the protein level in SW480 and HCT116 cells. (C) Silencing MBD2 increased the mRNA level of SFRP1. (D and E) Silencing MBD2 increased the protein level of SFRP1 and decreased the protein level of β-catenin. (F) KCC07 disrupted MBD2 binding to the SFRP1 promoter. (G and H) The half inhibitory concentration (IC50) of KCC07 in CRC cells. (I) KCC07 increased the mRNA level of SFRP1. (J and K) KCC07 increased the protein level of SFRP1 and decreased the protein level of β-catenin. n = 6, *indicates p < 0.05, **indicates p < 0.01.

KCC07 induced apoptosis and inhibited CRC cell proliferation both in vitro and in vivo

Our previous studies confirmed that MBD2 knockdown via siRNA suppresses proliferation and induces apoptosis in CRC cells.²³ In this study, we performed functional assays using KCC07 in CRC cells. CCK-8 assays demonstrated that KCC07 significantly inhibited proliferation in both the SW480 and HCT116 cell lines (Figure 2A). Subsequent apoptosis analysis revealed increased apoptotic rates in KCC07-treated cells compared with controls (Figure 2B). Western blot analysis of apoptosis-related proteins showed upregulation of the proapoptotic BAX and downregulation of the antiapoptotic BCL-2, along with decreased cleaved caspase 3 levels at 48 h post-treatment (Figure 2C). The decrease in cleaved caspase-3 is attributed to its transient activation peaking at earlier time points followed by degradation during late-stage apoptosis, which does not contradict the pro-apoptotic conclusion, as supported by the increased apoptotic rate and BAX/BCL-2 ratio. To evaluate therapeutic potential in vivo, we established HCT116 xenograft models in BALB/c nude mice. Animals treated with KCC07 showed significantly reduced tumor growth compared to vehicle controls (Figure 2D–F). These collective findings demonstrate that pharmacological inhibition of MBD2's methyl-DNA binding activity suppresses colorectal cancer progression through both antiproliferative and proapoptotic mechanisms both in vitro and in vivo. To verify the target specificity of KCC07 and rule out off-target effects, we performed rescue experiments in SW480 and HCT116 cells. Ectopic expression of MBD2 attenuated the efficacy of KCC07, as evidenced by reduced recovery of SFRP1 expression, diminished anti-proliferative effects, and impaired pro-apoptotic activity (Figure S2). These results confirm that KCC07 acts through an on-target mechanism.

Figure 2.

KCC07 induced apoptosis and inhibited CRC cell proliferation. (A) KCC07 inhibited the proliferation of CRC cells. (B) KCC07 increased apoptosis in CRC cells. (C) KCC07 increased the protein level of BAX and decreased the protein levels of BCL-2 and cleaved caspase 3. (D and E) KCC07 inhibited tumor growth in vivo. (F) HE staining of the tumor tissue. n = 6, ns indicates not significant, *indicates p < 0.05, **indicates p < 0.01, and ****indicates p < 0.0001.

MED19 were recruited by methylated promoter of tumor suppressor genes after inhibition of MBD2

To explore the mechanism of MBD2 regulating SFPR1 transcription, a DNA pulldown experiment was performed on the DNA sequences of the SFRP1 promoter in MBD2-downregulated SW480 cells. When compared to control cells, proteins preferentially bound to the SFRP1 promoter in MBD2-downregulated SW480 cells were enriched in the positive regulation of transcription from RNA polymerase II promoter entries according to the GO analysis (Figure 3A). This suggests that the positive transcription-regulating protein associated with RNA polymerase II may play an important role in promoting the transcription of SFRP1 gene after MBD2 knockdown. Positive transcription-regulating protein in the above entries included SUB1, MED19, TAF2, TAF5, TCF20, and CDK12 (Table S1). It's reported that MED19, TAF2, and TAF5 are associated with the transcription initiation process;^28-30 thus, we chose MED19, TAF2, and TAF5 as our research objects and found that MED19 played an important role in this process. The DNA pulldown experiment found the abundance of MED19 binding on SFRP1 promoter was significantly increased after the downregulation of MBD2 (Figure 3B). We used an MED19 antibody to conduct a CUT&Tag experiment to verify whether the inhibition of MBD2 enhanced the recruitment of MED19 to the promoter region of SFRP1. The heat map of peak distribution showed that the DNA fragments bound by MED19 mainly clustered around the transcription start site (TSS) (Figure 3C and Figure S3). In addition, the abundance of MED19 bound to the promoter region of the SFRP1 gene increased significantly in the KCC07-treated group (Figure 3D), which was consistent with the results of DNA pulldown (Figure 3B). Further KEGG analysis of the DNA fragments of the two groups showed that the MED19-bound DNA fragments in the DMSO control group of SW480 cells were mainly enriched in colorectal cancer, prostate cancer, and endometrial cancer. However, in the group of SW480 cells treated with KCC07, MED19-bound DNA fragments were not enriched in the above cancer entries but mainly enriched in the p53 signaling pathway, a recognized anticancer pathway,³¹ and other entries (Figure 3E). The results showed that genes bound to MED19 were tremendously increased, from 2451 to 16,735 in the KCC07 treatment group compared with the control group. Among the 16,735 genes, 547 genes are known tumor suppressor genes, including SFRP1, SFRP2, SFRP4, TP63, and other genes, indicating that the binding of MED19 to a large number of tumor suppressor genes can be restored after the inhibition of the methylated DNA-binding activity of MBD2 (Figure 3F). Experiment of qPCR were carried out to verify whether KCC07 could restore the expression of these tumor suppressor genes and revealed that KCC07 could restore the expression of SFRP1 (Figure 1H), TNFAIP3, ESRP1, EYA4, TP63, and other genes (Figure 3G). MBD2 is a methylation reader that tends to bind methylated DNA, and MED19 can bind DNA in large quantities after inhibiting MBD2, indicating that MED19 may tend to bind methylated DNA. In order to verify this, an EMSA to compare the binding affinity of MED19 for methylated versus unmethylated SFRP1 promoter sequences were performed. The results clearly demonstrated that MED19 indeed exhibited a stronger binding affinity for the methylated DNA probe (Figure 3H). The CUT&Tag experiments for TAF2 and TAF5 were also conducted, but the results showed that TAF2 and TAF5 did not bind to the SFRP1 promoter (not provided).

Figure 3.

DNA-pulldown screened the protein bound to SFRP1 promoter after inhibiting MBD2. (A) The top 10 enriched terms of GO analysis in the DNA pulldown experiment involving the SFRP1 promoter. (B) Downregulation of MBD2 enhanced the binding of MED19 to SFRP1 promoter. (C) The heatmap of the DNA fragments in MED19 CUT&Tag experiment. (D) KCC07 increased the abundance of MED19 bound to the SFRP1 promoter. (E) The KEGG pathway annotation of the results of MED19 CUT&Tag. (F) KCC07 increased the number of tumor suppressor genes bound to MED19 in CUT&Tag experiment. (G) KCC07 increased the mRNA levels of tumor suppressor genes. (H) MED19 bound to the methylated SFRP1 promoter in an electrophoretic mobility shift assay. n = 3 (G and H), ns indicates not significant, *indicates p < 0.05 and **indicates p < 0.01.

MED19 played a crucial role in the process of restoring SFRP1 expression during inhibition of MBD2

In order to verify the role of MED19 in SFRP1 expression, MED19 was knocked down in colorectal cancer cells via siRNAs (Figure 4A and B). The fourth siRNA was the most effective one and was used for further experiments. We used KCC07 to inhibit MBD2 while using siRNA to knock down MED19 in colorectal cancer cells and then detected the expression levels of SFRP1 and β-catenin. Our results demonstrated that the inhibition of the methylated DNA-binding activity of MBD2 led to the upregulation of SFRP1 at both the mRNA and protein levels. However, simultaneous knockdown of MED19 using siRNA prevented the recovery of SFRP1 expression (Figure 4C–F). Meanwhile, knockdown of MED19 resulted in decreased SFRP1 expression and increased β-catenin mRNA and protein levels, indicating activation of the Wnt pathway (Figure 4C–F).

Figure 4.

The effect of KCC07 on recovering SFRP1 expression requires MED19. (A and B) The interfering efficiency of MED19 siRNAs at the protein level in SW480 and HCT116 cells. (C–F) Treatment with KCC07 resulted in increased SFRP1 expression and increased β-catenin expression at both the mRNA and protein levels in CRC cells, while simultaneous knockdown of MED19 abolished these effects. n = 6, ns indicates not significant, *indicates p < 0.05, and **indicates p < 0.01.

Knockdown of MED19 weakened the inhibitory effect of KCC07 on colorectal cancer cells growth

As we described earlier, inhibition of the methylated DNA binding affinity of MBD2 by KCC07 inhibited the proliferation of colorectal cancer cells. However, the inhibitory effect of KCC07 was reversed after simultaneously down-regulating the expression of MED19 both in SW480 and HCT116 cells (Figure 5A and B). Next, we examined the role of MED19 in the ability of MBD2 in promoting the colorectal cancer cell apoptosis. We used KCC07 to inhibit MBD2 while knocking down MED19 in colorectal cancer cells and then detected cell apoptosis. The apoptotic rate of colorectal cancer cells treated with KCC07 was higher than that in the control group, and the effect was reversed by knocking down MED19 at the same time (Figure 5C and E). Apoptosis-related proteins were also detected. The results showed that after KCC07 treatment, the expression of pro-apoptotic BAX increased, while the expression of anti-apoptotic BCL-2 and cleaved caspase-3 decreased at 48 h posttreatment, which was consistent with our observations shown in Figure 2C. This pattern was reversed by simultaneous knockdown of MED19 (Figure 5D and F). Collectively, these results demonstrate that MED19 knockdown attenuates the inhibitory effect of KCC07 on colorectal cancer cell growth.

Figure 5.

Knockdown of MED19 weakened the inhibitory effect of KCC07 on CRC cell growth. (A and B) Simultaneous knockdown of MED19 abolished the ability of KCC07 on inhibiting cell proliferation in CRC cells. (C–F) Knockdown of MED19 inhibited the proapoptotic effect of KCC07. n = 6, ns indicates not significant, *indicates p < 0.05, and **indicates p < 0.01.

The expression level and intracellular distribution of MED19 had no difference in normal cells/tissues and CRC cells/tissues

We investigated the mechanism by which MBD2 regulates SFRP1 transcription by influencing the binding activity of MED19 to the SFRP1 promoter. The expression level of MED19 protein, its subcellular localization, and its binding ability to the SFRP1 promoter are the three main factors that can affect its binding abundance. We first compared the differences in MED19 protein expression and cell distribution and localization between normal colon mucosa cells (NCM460) and SW480 cells. Immunofluorescence showed that there was no significant difference in the expression of MED19 protein in NCM460 cells and SW480 cells, and it was evenly distributed in the nucleus (Figure 6A). Further tests were conducted in normal colorectal tissues, adjacent colorectal cancer tissues and colorectal cancer tumor tissues. Immunohistochemistry revealed that the MED19 protein was distributed mainly in the nucleus, and there was no significant difference in its expression between them (Figure 6B). Next, we examined the changes in MED19 expression and distribution in SW480 cells before and after using KCC07, and the results showed that the expression and distribution of MED19 did not change before and after using KCC07 (Figure 6C). This suggested that the increase of MED19 binding to the SFRP1 promoter after the inhibition of MBD2 may be due to the increased binding accessibility of MED19 to the methylated promoter.

Figure 6.

The expression level and intracellular distribution of MED19 had no difference between normal cells/tissues and CRC cells/tissues. (A) The expression and distribution of MED19 protein in NCM460 cells and SW480 cells. (B) The expression and distribution of MED19 protein in normal colorectal tissues, adjacent colorectal cancer tissues, and colorectal cancer tumor tissues. (C) KCC07 did not affect the expression and distribution of MED19 in SW480 cells. n = 3, ns indicates not significant, *indicates p < 0.05, **indicates p < 0.01, and scale bar = 100 µm.

MBD2 mediates SFRP1 expression through preventing MED19 binding to SFRP1 promoter

MBD2 protein immunoprecipitation experiment was performed to examine whether MBD2 binds directly to MED19, and the results showed that MED19 and MBD2 proteins did not bind to each other (Figure 7A). To determine whether MBD2 influences MED19's binding to the SFRP1 promoter, we performed a ChIP‒qPCR assay. The results demonstrated that MED19 binds to the SFRP1 promoter only upon MBD2 inhibition (Figure 7B). To further validate this finding, we constructed a luciferase reporter plasmid containing the SFRP1 promoter sequence. Luciferase activity increased after MBD2 inhibition with KCC07 but decreased upon simultaneous MED19 knockdown, indicating that MED19 promotes downstream gene transcription by binding to the SFRP1 promoter following MBD2 suppression in colorectal cancer cells (Figure 7C). Finally, we investigated the mechanism by which MED19 regulates SFRP1 transcription. It has been reported that MED19, MED6, and MED14 constitute a separable and regulated CDK8 kinase module that can assist CDK7 on the transcription factor TFIIH in stimulating CTD phosphorylation of RNA polymerase II.²⁸ We hypothesized that MED19 might regulate SFRP1 transcription by regulating CTD phosphorylation, an important process in transcription initiation. The phosphorylation levels of key CTD phosphorylation sites S2, S5, and S7 were detected. Results showed a significant increase in CTD-S7 phosphorylation after MBD2 knockdown and KCC07, which was reversed by simultaneous knockdown of MED19 (Figure 7D and E). To further investigate the connection between MED19 and Pol II CTD phosphorylation, coimmunoprecipitation (co-IP) experiments involving CDK7, a key kinase responsible for CTD phosphorylation, were performed. The results demonstrated that MED19 physically interacted with CDK7 (Figure 7F). These results suggest that MED19 promotes the expression of SFRP1 by promoting the phosphorylation of CTD-S7. Together, MBD2 blocked the binding of MED19 to the SFRP1 promoter instead of interacting with MED19, thus inhibiting transcriptional initiation caused by RNA polymerase II.

Figure 7.

MBD2 inhibited MED19 binding to SFRP1 promoter. (A) MED19 protein and MBD2 protein did not bind to each other directly in the IP experiment. (B) Compared with the control, the inhibition of MBD2 increased the binding ability of MED19 to the SFRP1 promoter. (C) Compared with DMSO, KCC07 increased the luciferase signal of the SFRP1 promoter, and knockdown of MED19 impaired the effect of KCC07. (D and E) MBD2 inhibition by siRNA or KCC07 increased CTD-S7 phosphorylation, which was reversed by simultaneous knockdown of MED19. (F) MED19 protein interacts with CDK7 protein in an IP experiment. n = 3, ns indicates not significant, *indicates p < 0.05, and **indicates p < 0.01.

Discussion

Dysregulation of the Wnt signaling pathway is an early event in CRC development. Silencing of SFRP1 leads to uncontrolled activation of the Wnt signaling pathway, contributing to the development of CRC. SFRP1 silencing is mediated by MBD2 recruitment to its hypermethylated promoter, which orchestrates transcriptional repression. However, the exact mechanism by which DNA methylation affects gene expression through MBD2 remains unclear. In this study, we explored the mechanism by which MBD2 regulates SFRP1 expression in CRC cell lines and first revealed that MBD2 regulates SFRP1 expression by blocking the binding of MED19 to the methylated SFRP1 promoter and affecting CTD-S7 phosphorylation of RNA polymerase II. This represents a more direct silencing mechanism than canonical chromatin remodeling pathways involving histone deacetylation or methylation. MBD2 inhibition restored the expression of SFRP1 and other tumor suppressor genes without altering promoter methylation, suggesting that blocking MBD2 binding to methylated promoters represents a promising therapeutic strategy that circumvents the genome-wide hypomethylation-related side effects associated with DNMTis.

And we found that proteins, such as MED19, TAF5, TAF3, and TAF2, can bind to methylated DNA after the removal of MBD2, thus restoring gene transcription. We hypothesize that the modification of DNA methylation is the first step in gene silencing and MBD2 blocks transcription-related protein binding to methylated DNA is the second step in gene silencing during the process of tumor development. Notably, even when promoter methylation status remains unaltered, inhibition of MBD2 allows transcription initiation factors to reoccupy the promoter, thereby rescuing transcriptional activity (Figure 8).

Figure 8.

Schematic representation of the role of MBD2 and MED19 in CRC growth. In normal colon epithelial cells, the unmethylated SFRP1 promoter allows transcription factors and the Mediator complex (including MED19) to bind, facilitating the recruitment of RNA polymerase II and the formation of the preinitiation complex. Subsequently, Mediator and general transcription factors promote the phosphorylation of the RNA polymerase II CTD, successfully initiating transcription. In contrast, in colon cancer cells, promoter hypermethylation enables MBD2 to bind and competitively inhibit the assembly of the MED19-containing preinitiation complex while concurrently recruiting the repressive NuRD complex, leading to SFRP1 silencing. However, upon MBD2 inhibition, the preinitiation complex can bind to the methylated promoter and successfully restore SFRP1 transcription. MED = mediator complex; MBD2 = methyl-CpG-binding domain 2, Pol II = RNA polymerase II, GTFs = general transcription factors, TBP = TATA-box binding protein, TAFs = TATA-box binding protein associated factors, NuRD = nucleosome remodeling and deacetylase.

MED19 is a tumor-associated gene involved in transcriptional regulation, cell proliferation and apoptosis, with reported links to multiple cancers, including lung, bladder, skin, and CRC.³² However, the role of MED19 in CRC remains controversial, with some studies identifying it as an oncogene³³ and others suggesting its tumor suppressor functions.³⁴ We conducted a detailed study on MED19 and confirmed that, upon MBD2 inhibition, MED19 plays a key role in restoring SFRP1 expression, promoting apoptosis, and suppressing proliferation in CRC cells both in vivo and in vitro. Specifically, MBD2 prevents MED19 from binding to the methylated SFRP1 promoter, thereby impairing CTD phosphorylation of RNA polymerase II and aborting transcription initiation. In our study, we identified MED19 as a tumor suppressor gene. This discrepancy may be attributed to differences in tumor stage and the extent to which MBD2 binds to target gene promoters. MED19 may mediate the conversion between cancer promotion to anticancer depending on the MBD2 binding status at distinct promoters.

Mediator (MED), act as a co-transcription factor, is involved in regulating the transcription of almost all RNA polymerase II-dependent genes. RNA polymerase II is a key enzyme for transcription in eukaryotes. Eukaryotic transcription initiation relies on the formation of a preinitiation complex (PIC) with RNA polymerase II as the core and the involvement of mediator complexes.³⁵ Human MED has 30 subunits, consisting of a core complex formed by head, middle, and tail modules and subunit kinase modules.²⁸ It has been reported that MED19 is located in the central location of the mediator complex, and a small fragment of MED19 (~133–148), with the N-terminus of MED6 and MED14, constitutes a dissociable and regulated CDK8 kinase module, which can assist CDK7 on the transcription factor TFIIH to stimulate the phosphorylation of Pol II CTD.²⁹ Agree with the literature above, our study confirmed for the first time that MED19 indeed participated in CTD phosphorylation of RNA polymerase II and thus regulated gene expression though binding to the gene promoter irrespective of promoter methylation.

We found that the number of genes bound by MED19 increased substantially from 2451 to 16,735 following MBD2 inhibition, indicating a marked reprogramming of MED19 genomic occupancy. More specifically, 547 well-established tumor suppressor genes, including SFRP1, were bound by MED19 upon MBD2 inhibition. Together, these findings suggest that MED19 is broadly involved in restoring tumor suppressor gene expression in the context of promoter methylation. It has been reported that MED19 is closely related to multiple cancers, such as lung cancer, bladder cancer, skin cancer, and CRC.^32,36 Furthermore, as a core subunit of the mediator complex, MED19 is widely involved in RNA polymerase II-dependent transcriptional regulation across multiple tissues.³⁷ Previous studies have shown that MBD2 expression is correlated with promoter hypermethylation-mediated silencing of tumor suppressor genes in gastric, breast, and lung cancers^21,26 and that MBD2 binds broadly to methylated DNA.³⁸ Taken together, we speculate that the MBD2-MED19 regulatory axis may not be limited to colorectal cancer but could extend to other solid tumors. The mechanism by which MBD2 competitively blocks MED19 binding to methylated promoters may represent a conserved epigenetic silencing paradigm across cancer types. Future validation of this regulatory axis in additional cancer types will help elucidate its universality and tumor type-specific characteristics.

Although our study identified a functional antagonistic relationship between MBD2 and MED19, the molecular details of their interaction remain to be further elucidated. We propose two possible mechanisms by which MBD2 and MED19 may interact. First, MBD2 may compete with MED19 for binding to the promoter region. Second, MBD2 binding to methylated DNA creates steric hindrance that directly prevents MED19 and its associated complexes from accessing the promoter. To distinguish between these two possibilities, future studies could employ truncation mutants or site-directed mutagenesis to identify the DNA-binding domain of MED19 and determine whether MED19 directly competes with MBD2 for methylated DNA binding or functions primarily through steric hindrance. Future studies among MBD2, MED19, and methylated DNA will help deepen our mechanistic understanding.

Several limitations of this study should be acknowledged. First, the clinical sample size was relatively modest. Larger, multicenter independent clinical cohorts are warranted to further validate the clinical relevance of the MBD2–MED19 regulatory axis. Second, while we provided preliminary insights into the molecular interaction between MBD2 and MED19, the specific mechanism involved requires further experimental confirmation. Finally, our study was primarily conducted in colorectal cancer cell lines; the universality of the MBD2-MED19 regulatory axis in other tumor types remains to be systematically validated. Future studies addressing these aspects will help advance the translational potential of our findings.

In conclusion, our study demonstrated that the inhibition of MBD2 suppresses tumor growth by restoring the expression of SFRP1 silenced by the abnormal methylation of gene promoters. Regulating MBD2 does not change the DNA methylation status, which can prevent the genome-wide hypomethylation side effects caused by DNMTis. In addition, we observed that KCC07 had a higher IC50 in normal colon epithelial NCM460 cells than in CRC cell lines, suggesting its favorable safety profile. Thus, KCC07 may be expected to be a novel antitumor drug that reverses tumor suppressor gene silencing caused by abnormal DNA methylation by inhibiting MBD2. This offers a novel therapeutic strategy that reactivates tumor suppressors by disrupting methylation-dependent silencing without altering the underlying methylation landscape.

Supplementary Material

Supplementary Material

Supplementary Material.docx

Funding Statement

This study was financially supported by the Chinese National Natural Science Foundation Projects (81972298), the Shenzhen Science and Technology Program (JCYJ20240813150427037 and JCYJ20240813150244037), and the Guangdong Basic and Applied Basic Research Foundation (2023A1515111020). The project was supported by the Sanming Project of Medicine in Shenzhen (SZSM202311017) and the Funding of Shenzhen Clinical Research Center for Gastroenterology (Gastrointestinal Surgery), Grant No. LCYSSQ20220823091203008.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethics approval and consent to participate

This study was approved by the Ethics Committee of The Seventh Affiliated Hospital, Sun Yat-sen University, Shenzhen, China, and was performed in accordance with the Declaration of Helsinki. Written informed consent about the experimental requirements and potential risks was provided by all the subjects. The animal experiments in this study have been reviewed and approved by the Institutional Animal Care and Use Committee, Sun Yat-sen University.

Data sharing

The data in this study is available from the corresponding authors on reasonable request.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/15384047.2026.2667568.