Research ArticleONCOLOGY

MED12 methylation by CARM1 sensitizes human breast cancer cells to chemotherapy drugs

See allHide authors and affiliations

Science Advances  09 Oct 2015:
Vol. 1, no. 9, e1500463
DOI: 10.1126/sciadv.1500463
  • Fig. 1 Positive correlation between the expression of CARM1 and MED12 in breast cancer cell lines and human breast tumors.

    (A) Coomassie brilliant blue staining (left panel) and autoradiograph (right panel) of in vitro methylated MED12 by CARM1 in the presence of [3H]SAM. 3xFLAG-tagged MED12 protein was purified from HEK293T CARM1KO cells. (B) Western blot analyses of immunoprecipitated (IP) MED12 and input lysates from CARM1WT or CARM1KO cell lines using indicated antibodies. WT, wild type; IB, immunoblotting. (C) The Pearson correlation plot depicts the positive correlation between CARM1 and MED12 mRNA expression in 5790 human breast tumor cases collected in the bc-GenExMiner database. (D) Western blot analyses of MED12 and CARM1 proteins in human breast cancer and normal epithelial cell lines. β-Actin was used as an internal control. (E) Western blot analyses of MED12, CARM1, and ERα in patient-derived human breast tumor grafts (36). β-Actin was used as an internal control. (F) Immunohistochemical (IHC) staining of MED12 and CARM1 in human breast tumors.

  • Fig. 2 Higher levels of MED12 and CARM1 correlate with better response to chemotherapy drugs in cell line models and clinical cohorts.

    (A) KM curves stratified by MED12 mRNA levels depicting the probability of RFS in untreated, endocrine therapy–treated, or chemotherapy-treated breast cancer patients. Patient samples were divided into MED12high and MED12low groups based on the median of the expression level of MED12. Affymetrix gene ID 214275_at was used to plot the survival curves of MED12 using data sets from Gene Expression Omnibus (GEO) (Affymetrix HG-U133A and HGU-133+2 microarrays), European Genome-Phenome Archive (EGA), and The Cancer Genome Atlas (TCGA). HR, hazard ratio. (B and C) High levels of CARM1 and MED12 proteins correlate with better survival after 5-fluorouracil (5-FU) or doxorubicin treatment in breast cancer patients with 100-month follow-up (n = 254). KM estimates of DFS and OS of human patients according to the expression levels of CARM1 (n = 254) or CARM1/MED12 (n = 154). Comparison was made between groups with high or low levels of CARM1 alone or high or low levels of CARM1 and MED12. P value refers to two-sided log-rank tests. (D) Cell viability analyses of three paired CARM1WT and CARM1KO breast cancer cell lines after incubating with 1 μM 5-FU, doxorubicin, or floxuridine for 72 hours. Cell viability was determined by MTT assays. Quantitative data are presented as averages ± SD. Student’s t test was used for statistical analysis. *P < 0.05; **P < 0.01. (E) Western blotting shows the knockdown of MED12 in both CARM1WT and CARM1KO MDA-MB-231 cells. (F) Cell viability analyses of MDA-MB-231 CARM1WT shCtrl, MDA-MB-231 CARM1WT shMED12, MDA-MB-231 CARM1KO shCtrl, and MDA-MB-231 CARM1KO shMED12 cells after treatment with 5-FU or doxorubicin for 72 hours. Quantitative data are presented as averages ± SD. Student’s t test was used for statistical analysis. *P < 0.05; **P < 0.01. (G) Western blotting shows the knockdown of MED12 in both CARM1WT and CARM1KO MCF7 cells. (H) Cell viability analyses of MCF7 CARM1WT shCtrl, MCF7 CARM1WT shMED12,MCF7 CARM1KO shCtrl, and MCF7 CARM1KO shMED12 cells after treatment with 5-FU or doxorubicin for 72 hours. Quantitative data are presented as averages ± SD. Student’s t test was used for statistical analysis. *P < 0.05.

  • Fig. 3 CARM1 directly interacts with MED12 and methylates MED12 at two sites R1862 and R1912.

    (A) Schematic diagram of full-length MED12 domain structure and the truncation constructs, which are all fused to a FLAG tag cassette via the N terminus. Red arrows depict the putative arginine methylation sites predicted by the PMeS program. (B) Western blot analysis of α-FLAG immunoprecipitated MED12 fragments detected by the α-FLAG antibody (left panel) or the α-ADMA antibody (right panel) from HEK293 cell lysates transiently transfected with the corresponding MED12-expressing plasmids. (C) Mapping of CARM1-interacting domain to the PQL domain of MED12 using GST pull-down assay. FLAG-tagged MED12 fragment proteins were incubated with GST-CARM1 and then immunoprecipitated and detected with α-FLAG antibody in a Western blot (upper panel). The presence of GST-CARM1 in immunoprecipitates was detected by the anti-GST antibody in a Western blot (lower panel). (D) Western blot analysis of immunoprecipitated FLAG-tagged WT or mutant MED12 proteins using the α-FLAG or α-ADMA antibodies. (E) Western blot analysis of immunoprecipitated FLAG-tagged WT or mutant MED12 proteins using the α-FLAG or α-ADMA antibodies. (F) Coomassie brilliant blue staining (left panel) and autoradiograph (right panel) of in vitro methylated recombinant MED12WT, MED12R1862K, MED12R1912K, and MED12DM proteins by CARM1 in the presence of [3H]SAM. (G) Coomassie brilliant blue staining (left panel) and autoradiograph (right panel) of in vitro methylated GST-peptide fusion proteins by CARM1 in the presence of [3H]SAM. (H) Western blot analysis of total MED12, me-MED12, and CARM1 in control shRNA– or MED12 shRNA–expressing MDA-MB-231 CARM1WT and MDA-MB-231 CARM1KO cells. β-Actin was used as an internal control.

  • Fig. 4 Methylation of MED12 renders cells sensitive to chemotherapy drugs in vitro and in vivo.

    (A) Western blotting analysis of MED12 in MDA-MB-231 cells and shMED12-expressing MDA-MB-231 cells restored with GFP, MED12WT, or MED12DM using anti-MED12 antibody against whole-cell lysate or the α-ADMA antibody against immunoprecipitated MED12. (B) Cell survival inhibition rate plot of MDA-MB-231 shCtrl cells (x axis) or MDA-MB-231 shMED12 cells (y axis) after a 72-hour treatment with 97 FDA-approved cancer drugs (1 μM). Red dots denote 5-FU and its analog floxuridine. The inhibition rate was calculated as the difference of MTTDMSO and MTTdrug normalized to MTTDMSO. (C) Cell survival inhibition rate plot of MDA-MB-231–shMED12–MED12WT cells (x axis) or MDA-MB-231–shMED12–MED12DM cells (y axis) after a 72-hour treatment with 97 FDA-approved cancer drugs (1 μM). Red dots denote 5-FU and its analog floxuridine. (D) Cell viability curves for MDA-MB-231–shMED12–MED12WT and MDA-MB-231–shMED12–MED12DM cells after treatment with various concentrations of fluorouracil or floxuridine for 72 hours. (E) Colony yields after 2 weeks of treatment with 0.3 μM fluorouracil or floxuridine in MDA-MB-231–shMED12–MED12WT and MDA-MB-231–shMED12–MED12DM cells. (F) Cell viability assays for the indicated cell lines after 72 hours of treatment with 10 μM 5-FU, selumetinib, or crizotinib. (G) Tumor growth curve of MDA-MB-231–shMED12–MED12WT and MDA-MB-231–shMED12–MED12DM cell grafts in nude mice (n = 5). Student’s t test was used for statistical analysis. **P < 0.01; *P < 0.05. (H) Representative photographs of tumors derived from MDA-MB-231–shMED12–MED12WT and MDA-MB-231–shMED12–MED12DM cell grafts treated with PBS or 5-FU.

  • Fig. 5 MED12 methylation enhances its association to p21 gene locus and suppresses p21 transcription.

    (A) Relative mRNA levels of MED12, KRT14, KIF5a, PRSS2, CD74, and p21 in control shRNA– or shMED12-expressing MDA-MB-231 CARM1WT or MDA-MB-231 CARM1KO cells were determined by real-time qPCR. β-Actin was used as an internal control. (B) Real-time qPCR analyses of KRT14, KIF5A, PRSS2, CD74, and p21 mRNA levels in MDA-MB-231–shCtrl-GFP, MDA-MB-231–shMED12-GFP, MDA-MB-231–shMED12-MED12WT, and MDA-MB-231–shMED12-MED12DM cells. β-Actin was used as an internal control. (C) Western blot analysis of p21 protein in control shRNA– or shMED12-expressing MDA-MB-231 CARM1WT and MDA-MB-231 CARM1KO cells. (D) Western blot analysis of p21 protein in MDA-MB-231 cells overexpressing GFP or FLAG-tagged MED12WT, MED12R1862K, and MED12R1912K. (E) The raw reads of ChIP-seq tracing of MED1 (GSM560353) and MED12 (GSM560354) enrichment peaks on the mouse p21 gene locus retrieved from the GEO database (19). (F) ChIP-qPCR analysis of MED12 binding to the five genomic regions of the p21 gene in MDA-MB-231 CARM1WT or MDA-MB-231 CARM1KO cells. Normal rabbit immunoglobulin G (IgG) was used as antibody control. (G) ChIP-qPCR analysis of FLAG-tagged MED12 binding to four genomic regions of the p21 gene in HEK293 cells transiently expressing FLAG-MED12WT or FLAG-MED12DM. (H) ChIP-qPCR analysis of CARM1 binding to five genomic regions of the p21 gene in HEK293 CARM1WT or HEK293 CARM1KO cells. **P < 0.01; *P < 0.05.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/1/9/e1500463/DC1

    Fig. S1. In vitro methylation assays of MED12 by PRMT1/PRMT6 and correlation analyses of indicated genes in breast cancer specimens and cell lines.

    Fig. S2. Western blotting analysis of endogenous dimethylated MED12R1862 using a methyl-specific MED12 rabbit polyclonal antibody.

    Fig. S3. Mutation of MED12 methylation sites does not affect cell growth or EMT-associated gene expression.

    Fig. S4. The mRNA level of CDKN1A/p21, a MED12 and CARM1 co-regulated gene, correlates with 5-FU response in vitro and predicts the probability of recurrence-free survival in breast cancer patients.

    Fig. S5. Mutation of MED12 methylation sites does not affect the interaction of MED12 with other known interacting proteins.

    Fig. S6. Suppression of p21 mRNA and protein levels is retained in MED12WT- but not MED12DM-expressing HEK293 cells.

    Table S1. Primary hits from the FDA-approved oncology drug screening.

    Table S2. Differentially expressed genes regulated by CARM1 and MED12.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. In vitro methylation assays of MED12 by PRMT1/PRMT6 and correlation analyses of indicated genes in breast cancer specimens and cell lines.
    • Fig. S2. Western blotting analysis of endogenous dimethylated MED12R1862 using a methyl-specific MED12 rabbit polyclonal antibody.
    • Fig. S3. Mutation of MED12 methylation sites does not affect cell growth or EMT-associated gene expression.
    • Fig. S4. The mRNA level of CDKN1A/p21, a MED12 and CARM1 co-regulated gene, correlates with 5-FU response in vitro and predicts the probability of recurrence-free survival in breast cancer patients.
    • Fig. S5. Mutation of MED12 methylation sites does not affect the interaction of MED12 with other known interacting proteins.
    • Fig. S6. Suppression of p21 mRNA and protein levels is retained in MED12WT- but not MED12DM-expressing HEK293 cells.
    • Table S1. Primary hits from the FDA-approved oncology drug screening.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Table S2 (Microsoft Excel format). Differentially expressed genes regulated by CARM1 and MED12.

    Files in this Data Supplement:

Navigate This Article