Fig 2: SMAD3 K53/K333 methylation predicts poor survival of patients with cancer.(A) Kaplan-Meier survival analysis comparing high and low levels of SMAD3 K53/K333 trimethylation using microarray results from a breast cancer specimen. P = 0.0004. (B) Cox regression analysis showing the significance of the relationship between SMAD3 K53/K333 trimethylation expression and prognosis for patients with breast cancer in the presence of other clinical variables. (C) Scatter plot of SMAD3 K53/K333 trimethylation IHC staining score comparing different N stages using microarray results from a breast cancer specimen (n = 80). P = 0.0051, by ?2 test. (D) Scatter plot of the number of cancer-positive lymph nodes comparing high and low levels of SMAD3 K53/K333 trimethylation using microarray results from a breast cancer specimen (n = 80). P = 0.0039, by ?2 test. (E) Working model of EZH2-mediated SMAD3 K53/K333 methylation crosstalk with TGFB-mediated SMAD3 phosphorylation, membrane localization, and activation.

Fig 3: Targeting SMAD3 K53/K333 methylation inhibits cancer metastasis.(A) The aa sequence of different TAT peptides. (B and C) HEK293T cells were treated with different TAT peptides (Pep-1, Pep-2) and TGFB1 (5 ng/mL), and WCEs were collected for IP with anti-SMAD3 antibody, followed by IB analysis. (D) MDA-MB-231 were silenced with TAT peptides and treated with TGFB1 (5 ng/mL), and WCEs were collected for IB analysis. (E) Quantitative analysis of Transwell cell migration and invasion assays using MDA-MB-231 cells treated with different TAT peptides. (F and G) Representative lung image (F) and H&E-stained lung sections (G). Scale bars: 10 mm. (H and I) Scatter plots show the number of lung metastatic nodes (I) and lung weights (H). All immunoblotting was performed 3 times, independently, with similar results. Data indicate the mean ± SD. **P < 0.05, by 2-tailed Student’s t test.

Fig 4: EZH2 promotes cell migration and invasion dependent on methylation of SMAD3 at K53 and K333.(A) Quantitative analysis of Transwell assay in the indicated MDA-MB-231 cells treated with TGFB1 (5 ng/mL). (B) MDA-MB-231SMAD3–/– cells were stably transfected with WT SMAD3 or SMAD3 K53/333R plasmids and silenced with control or EZH2 shRNA (nos. 1 and 2). WCEs were collected for IB analysis. (C) WT Flag-EZH2 or a Flag-EZH2 H689A plasmid was transfected into MDA-MB-231SMAD3–/– cells ectopically expressing WT SMAD3 or SMAD3 K53/333R, and WCEs were collected for IB analysis. (D and E) A Transwell cell invasion assay was performed using MDA-MB-231SMAD3–/– cells stably transfected with WT SMAD3 or SMAD3 K53/333R plasmids and transfected with a vector, EZH2WT, or EZH2H689A. Representative images (D) and quantitative analysis (E). Original magnification, ×200. (F–H). Representative lung image (F),H&E-stained lung sections (G), and scatter plot showing lung weights (H). Scale bars: 5 mm. All immunoblotting was performed 3 times, independently, with similar results. Data indicate the mean ± SD. **P < 0.05, by 2-tailed Student’s t test (A, E, and H).

Fig 5: Deletion of SMAD3 K53/K333 methylation inhibits the SMAD3 oncogene in vitro and in vivo.(A) MDA-MB-231SMAD3–/– cells were stably transfected with WT SMAD3 or SMAD3 K53/333R plasmids and treated with TGFB1 (5 ng/mL). Quantitative RT-PCR analysis of TGFB/SMAD3 signaling pathway downstream genes, including CTGF, PAI1, PDGFB, and SMAD7, in the indicated cells with or without TGFB1 (5 ng/mL) treatment. (B) Quantitative analysis of Transwell assay in the indicated cells. (C) IF and (D) IB analysis of EMT markers in the indicated cells. Scale bar: 50 um. (E and F) Representative lung image (E) and H&E-stained lung sections (F). Scale bars: 5 mm. (G and H) Scatter plots showing lung metastatic nodules (G) and lung weights (H). All immunoblots were performed 3 times, independently, with similar results. Data indicate the mean ± SD. **P < 0.05, by 2-tailed Student’s t test (A, B, G, and H).