Fig. 2: BMP signaling potentiates RA response in a broad range of neuroblastoma cell lines.
A Pearson correlations of the expression of all genes with ATRA sensitivity in the cell lines we rescreened for 6 days. Gene expression data were from GDSC (left, P = 4.2 × 10−5 for SMAD9) and Depmap (right, P = 4.8 × 10−4 for SMAD9). P-values are from two-sided Pearson’s correlation tests. B Scatter plot of SMAD9 RNA expression from GDSC (left) and Depmap (right) plotted against ATRA 6-day IC50 values. P-values were calculated with two-sided Pearson correlation tests. The shaded area represents the Pearson correlation standard error. C Scatter plot of median RRA score of the 655 druggable genes from targeted CRISPR screens in 10 neuroblastoma cell lines. D Western blots showing BMP signaling activity (level of phosphorylated SMAD1/5/9) in neuroblastoma cell lines after FK506 and rapamycin treatment. Cells were treated for 4 h. GAPDH and β-Actin were used as a loading control, and the representative results of three independent experiments are shown. E Heatmaps of percent inhibition of cell viability in CHP-134 cells (left) conferred by drug-only treatments (values within dotted lines) and combination treatments (all other values) with ATRA and FK506. Ten-point doses were used at indicated concentrations with 1:3-fold dilutions for each drug. Each matrix represents the average of three independent experiments. Data were normalized as a function of percent cell viability using inter-plate controls. ATRA and rapamycin combination in the TGW cell line was shown on the right side. F Heatmap matrices of synergy scores derived from respective cell death values in (E). Synergy scores “ZIP δ” were calculated based on the zero interaction potency (ZIP) model, which can be interpreted similarly to BLISS synergy. G Heatmap showing the maximum ZIP δ synergy scores (color scale) of ATRA and FK506 combinations or ATRA and rapamycin combinations in 16 cell lines. Neuroblastoma cell lines are highlighted in red. H Correlation of max ZIP synergy scores of ATRA + FK506 (x-axis) and ATRA + rapamycin (y-axis) screens in the cell lines shown in (G). The P-value was calculated with two-sided Pearson correlation tests. The shaded area represents the Pearson correlation standard error. I Bar plot confirming synergy between ATRA and FK506 in CHP-134. Cells were treated for 6 days. Cell viability was measured with MTS assay and normalized to DMSO-treated samples, which were defined as 0% inhibition of cell growth. Data represent the mean ± SD of 4 independent replicates. Two-sided t-test for ATRA vs ATRA + 1 µM FK506, P = 1.29 × 10−7. J Like (I) but in the TGW cell line. Two-sided t-test for ATRA vs ATRA + 5 µM FK506, P = 4.71 × 10−6. K Barplot showing synergy between ATRA and FK506 in CHP-134 upon ACVR1 knockdown. Cells were transduced with lentiviral control shRNA or 3 independent ACVR1 shRNAs. Cells were treated with 0.05 µM ATRA and 0.1 µM FK506 for 6 days. Inhibition of cell growth was calculated as in (I). Data represent the mean ± SD of 4 independent replicates. In ATRA + FK506 group, two-sided test for control vs 0498, P = 3.22 × 10−15; control vs 2778, P = 4.7 × 10−14; control vs 3321, P = 1.24 × 10−16. Source data are provided as a Source Data file.
