Supplementary Data, Supplementary Tables 1, 3, and 4, Supplementary Figures 1-12 from EZH2 Loss Drives Resistance to Carboplatin and Paclitaxel in Serous Ovarian Cancers Expressing ATM

Table S1: Clinical attributes of primary ovarian cancer patients used in this study. Table S3: Proteins and protein modifications detected in differential amounts in responsive and non-responsive ovarian cancers. Wilcoxon-Mann-Whitney Test; W = rank sum; RESP = responsive (DFS > 12 months); NRESP = non-responsive (DFS {less than or equal to} 12 months). Table S4: Gene variants (Variant), amino acid change (AA change) and allele frequencies of sequenced tumors from 17 patients (SXXX). Tumor content of tissues is listed at the end of the table. Figure S1: Screening of eight non-responsive and nine responsive tumor samples from ovarian cancer patients using DigiWest protein array analysis with 279 antibodies against proteins and their modifications (Wilcoxon-Mann-Whitney, p{less than or equal to}0.05). Figure S2: ATM/EZH2 protein ratios in 17 tumor samples with non-responders and responders quantified in the DigiWest analysis. Figure S3: Controls of immunohistochemistry staining. Figure S4: Correlation analysis of mean immunoreactive scores from stainings of EZH2 and ATM (left or phospho-Rb (Ser807/811) (right) in 34 ovarian cancer patients. n=2 per tumor sample Figure S5: A: Representative image of PLA analysis on tissue microarray spots. DAPI: staining of cell nuclei. FITC: Autofluorescence of tissue in FITC channel and determination of area for analysis (830 µm diameter). Cy5: Channel for PLA measurement after subtracting background signals of FITC channels. Analyzed: Automated counted spots of interactions. Bars represent 200 µm. B: Representative control (rabbit isotype) for PLA analysis on tissue microarray spots. C: Quantification of PLA interactions in an ovarian cancer tissue. Positive control: original PLA protocol using EZH2 and phospho-ATM (S1981) antibodies. Negative control: Staining procedure without primary antibodies. Isotype control rabbit: PLA protocol using rabbit isotype antibodies instead of antibodies against EZH2. Isotype control mouse: PLA protocol using mouse isotype antibodies instead of antibodies against phospho-ATM (S1981). n = 5. Figure S6: Correlation analysis of the mean IRS scores from stainings of ATM and mean PLA signals of EZH2 and ATM pS1981 in 34 samples on tissue microarrays; Spearman correlation; n=2 per tumor sample. Figure S7: Correlation analysis of mean IC50 values calculated in cell line cells. Values of paclitaxel mono-treatment were compared to the values of combinational treatment (CP scale). Figure S8: EZH2 protein levels in lysates of untreated cell lines analyzed with Spearman correlation against mean logIC50 values for CP (left) and TX (right); n=3. Grid line=linear regression line. Figure S9: ATM protein levels in lysates of untreated cell lines analyzed with Spearman correlation against mean logIC50 values for CP (left) and TX (right); n=3. Grid line=linear regression line. Figure S10: Hierarchical clustering of quantified total protein and protein modifications of cell cycle proteins measured with DigiWest technology. Figure S11: Proximity ligation assay on A2780 cells during treatment with CPTX using standard protocol (EZH2 + pATM) as positive control or isotype controls with mouse antibodies instead of pATM (mIgG1) or rabbit antibodies instead of EZH2 antibodies. Figure S12: Scheme suggesting a putative interaction of EZH2 and ATM during treatment with CPTX.