Additional file 1 of Elesclomol-induced increase of mitochondrial reactive oxygen species impairs glioblastoma stem-like cell survival and tumor growth

Additional file 1: Supplementary Table 1. List of drugs used for small-molecule kinase inhibitor screening (10 mM in DMSO). Supplementary Table 2. List of antibodies used for Reverse-Phase Protein microArrays (RPPA) analysis. Supplementary Table 3. Patient and GSC line characteristics. Supplementary Table 4. List of genes corresponding to significant antibodies and grouped using the Venn diagram in Fig. 4C. Supplementary Figure S1. A-D. Morphological changes of the four GSC lines used in the study (A, GSC#1; B, GSC#61; C, GSC#83; D, GSC#163) after being induced to transdifferentiate for 2 weeks. Left panel, tumorspheres in stem cell medium; right panel, net-like structures under endothelial conditions (magnification 10X). Supplementary Figure S2. (A) Fluorescent-activated cell sorting dot plots of CD34−/low and CD34high GSC#163 after two weeks of culture in endothelial conditions under hypoxia. Percentage and squares indicate the sorted subpopulations of cells with different CD34-expression levels (left, IgG1 isotype control sample; right, CD34 sample). (B-C) Immunohistochemical analysis of CD34low (B) and CD34high (C) GdEC subcutaneous tumor xenografts based on the expression of the astrocytic marker glial fibrillary acidic protein (GFAP, right panels), showing tumors with different levels of differentiation. (Left panels, haematoxylin and eosin staining; magnification 200X). Supplementary Figure S3. Concentration-response assays on U87MG and all the four glial cell lines derived from the selected GSC lines. Supplementary Figure S4. Cytofluorimetric cell-by-cell analysis of viability in four different GSC lines treated with 10, 100, or 1000 nM elesclomol in the presence or absence of the following cell death inhibitors: z-VAD, necrostatin-1, ferrostatin-1, 3-MA, NAC, and CoQ at the indicated concentrations. Results obtained from four independent experiments are expressed as percentage vs control untreated cells and reported as means ± SD. Supplementary Figure S5. Cytofluorimetric cell-by-cell analysis of cell viability (A), mitochondrial ROS production (B), mitochondrial membrane potential (C), and GSH (D) in four different GSC lines treated with 10, 100, or 1000 nM elesclomol in the presence or absence of the copper chelating agent TTM. Results obtained from four independent experiments are expressed as percentage vs control untreated cells and reported as means ± SD. Supplementary Figure S6. Cytofluorimetric cell-by-cell analysis of cell viability, mitochondrial ROS production, mitochondrial membrane potential, and GSH in HMVECs, used as a control of nontumoral endothelial cell line, treated with 10, 100, or 1000 nM Elesclomol in the presence or absence of the copper chelating agent TTM. Results obtained from four independent experiments are expressed as percentage vs control untreated cells and reported as means ± SD. Supplementary Figure S7. Illustration of the rationale suitable for the choice of rank k, a critical parameter that defines the number of metagenes used to approximate the target matrix (Gaujoux & Seoighe, 2010). A) Measurements are applied to both real data (circles) and randomized data (triangles). The rationale for choosing rank stems on diverse metrics, i) trend of the cophenetic coefficient: Brunet et al. (2004) suggest choosing the smallest value of k for which there is a decrease in the trend of the cophenetic; ii) trend of the dispersion coefficient introduced by Kim & Park. (2007); iii) explained variance by increasing rank; iv) trend of residuals; v) trend of RSS: Hutchins et al. (2008) suggest taking the first rank value for which we have an inflection point. Frigyesi et al. (2008) instead consider the first rank value for which the decrease of the RSS on real data is less than the decrease of the RSS on the random data; vi) silhouette values measured on the matrices of the base, of the coefficients and the consensus matrix; vii) trend of the sparseness introduced by Hoyer (2004). B) Multiple consensus maps corresponding to different value of k. Supplementary Figure S8. Heatmap of the most important antibodies in each of the k = 6 metagenes resulting from the model. Supplementary Figure S9. Principal Component Analysis (PCA) algorithm applied to Elesclomol data, whereby each cell line is considered as a function of the antibodies. A) Scree plot. Given the low amount of variance explained by the variables above the fifth, we considered up to 5 principal components. B) Biplots using cell lines and growth conditions as scores. Ellipses represent the 95% probability of finding sample score values. Supplementary Figure S10. Principal Component Analysis (PCA) biplots of components of the antibodies using (A) Time and (B) treatment, respectively. Ellipses represent the 95% probability of finding sample score values. Supplementary Figure S11. Concent ration-response assays on all the four selected GSC lines for setting the dose of Elesclomol most suitable for the combination with TMZ.