Inhibition of EZH2 Enhances the Antitumor Efficacy of Metformin in Prostate Cancer

Metformin downregulates EZH2 expression by regulating miR-26a-5p. A and B, LNCaP or 22Rv1 cells were treated with DMSO, metformin (1 mmol/L), GSK126 (5 μmol/L), or both for 48 hours, followed by immunoblotting (IB). C and D, LNCaP and 22Rv1 cells were treated with metformin (1 mmol/L), GSK126 (5 μmol/L), or both for 48 hours, followed by qRT-PCR. E and F, LNCaP and 22Rv1 cells were transfected with EZH2 and pcDNA3.0, followed by 72-hour cell viability assay with indicated treatments (metformin: 1 mmol/L; GSK126: 5 μmol/L; means ± SD; n = 3). *, P ≤ 0.05; **, P ≤ 0.01. Meanwhile, some cells were harvested for Western blot to test EZH2 level after the treatments. G–I, qRT-PCR shows the expression of miR-26a-5p, miR-101-3p, let-7a-5p, let-7b-5p, and let-7c-5p of LNCaP cells treated with metformin (1 mmol/L), GSK126 (5 μmol/L), or combination for 48 hours, with all microRNA expressions being normalized to RNU6-2. J, LNCaP cells were transfected with the miR-26a-5p inhibitor or negative control miRNA inhibitor, then treated with metformin (1 mmol/L), GSK126 (5 μmol/L), or both for 48 hours, followed by IB. K, LNCaP cells were transfected with the miR-101-3p inhibitor or the negative control miRNA inhibitor, then treated with metformin (1 mmol/L) for 48 hours and harvested for IB. L, LNCaP cells were transfected with inhibitors targeting let-7a-5p, let-7b-5p, and let-7c-5p or the negative control miRNA inhibitor, then treated with metformin (1 mmol/L) for 48 hours and harvested for IB. M, qRT-PCR shows the expression of miR-26a-5p or 22Rv1 cells treated with metformin (1 mmol/L), GSK126 (5 μmol/L) or both for 48 hours, with miR-26a-5p expression being normalized to RNU6-2. N, 22Rv1 cells were transfected with miR-26a-5p inhibitor or negative control miRNA inhibitor, then treated with metformin (1 mmol/L), GSK126 (5 μmol/L), or both for 48 hours and harvested for IB. O, 22Rv1 cells were transfected with miR-26a-5p inhibitor or negative control miRNA inhibitor, treated with DMSO or the combination of metformin (1 mmol/L) and GSK126 (5 μmol/L) for 72 hours, followed by MTT assay (means ± SD; n = 3). *, P ≤ 0.05; **, P ≤ 0.01.

AR affects prostate cancer cells' response to metformin. A–D, PC3, DU145, LNCaP, and 22Rv1 cells were treated with metformin of indicated concentrations for 48 hours and harvested for immunoblotting (IB). E and F, LNCaP cells were treated with 10 nmol/L R1881 or metformin or metformin plus R1881 for 48 hours, followed by IB. Meanwhile, mRNA was extracted for the detection of levels of miR-26a-5p. G, PC3 (-AR or -Neo) cells were treated with metformin of indicated concentrations, as well as 10 nmol/L R1881 to activate AR, and subjected to IB. H, mRNA was extracted from PC3-Neo and PC3-AR cells treated with 1 mmol/L metformin for 48 hours, followed by qRT-PCR to test the levels of miR-26a-5p. I and J, 22Rv1 cells were stably transfected with sh-control, sh-AR #3, and sh-AR #4, and treated with 1 mmol/L metformin for 48 hours, followed by IB to test EZH2 protein level and qRT-PCR to detect miR-26a-5p level.

MiR-26a-5p is directly regulated by AR. A, Scheme representing the binding sequences within the miR-26a promoter relative to the designed primers. B, ChIP analysis of AR binding to the miR-26a promoter region in 22Rv1 and LNCaP cells. C and D, HEK293T cells and PC3-Neo or PC3-AR cells were cotransfected with the miR-26a promoter construct with empty vector or AR for 24 hours, treated with R1881 (10 nmol/L) for additional 24 hours, and harvested for luciferase assays. Values are means ± SD; n = 3. *, P ≤ 0.05; **, P ≤ 0.01. E and F, 22Rv1 cells were treated with DMSO, metformin (1 mmol/L), GSK126 (5 μmol/L), or both for 48 hours and harvested for anti-AR ChIP using qPCR to measure the binding of AR to the promoter of miR-26a. Values are means ± standard deviations; n = 3. *, P ≤ 0.05; **, P ≤ 0.01. G, HEK293T cells were transfected with the miR-26a promoter construct in the presence of AR, EZH2-S21D, or EZH2-S21A, and harvested for luciferase assays. Values are means ± SD; n = 3. *, P ≤ 0.05; **, P ≤ 0.01. H and I, Chromatin was precipitated with anti-AR antibody and re-precipitated with anti-AR or anti-EZH2 antibody or IgG, followed by qPCR. Values are means ± SD; n = 3. *, P ≤ 0.05; **, P ≤ 0.01. NS, not significant.

The combination of metformin and GSK126 reduced cell proliferation, increased apoptosis, and inhibited EZH2 expression in 22Rv1-derived xenograft tumors. A, Tumor growth curves of 22Rv1-derived mouse xenografts. After nude mice were inoculated with 22Rv1 cells (2.5 × 10⁵/mouse) for 2 weeks, the mice bearing tumors were treated with drugs as described in Materials and Methods. The sizes of the tumors in each group were measured every 4 days (mean ± SD; n = 9 mice for each group). *, P ≤ 0.05; **, P ≤ 0.01. B, Images of the 22Rv1-derived xenograft tumors at the end of study. C, Measurement of tumor weight upon harvest. D, Measurement of mice body weight upon tumor harvest. E, Representative images of H&E staining on formaldehyde-fixed, paraffin-embedded, 22Rv1-derived tumor sections. F, Representative images of anti-Ki67 IHC staining of tumor sections. G, Quantification of Ki67 signals as percentages of Ki67-positive cells compared with the total numbers of cells. Multiple tumor sections were calculated (means ± standard deviations; n = 4). *, P ≤ 0.05; **, P ≤ 0.01. H, Representative images of anti-cleaved caspase-3 IHC staining of tumor sections. I, Quantification of cleaved caspase-3 signals as percentages of cleaved caspase-3–positive cells compared with the total numbers of cells. Multiple tumor sections were calculated (means ± SD; n = 4). *, P ≤ 0.05; **, P ≤ 0.01. J and K, Protein lysates extracted from 22Rv1-derived tumors were subjected to Western blot for EZH2 and H3K23me3, as well as H3 and β-actin expression. L, Quantification of EZH2 protein levels in J and K.