Figure 2.

ECs provide protumorigenic signals to T-ALL cells. (A) Principal component analysis (PCA) plot showing the responses of 10 individual T-ALL PDX models cultured with and without ECs connected by vectors depicting the trajectory of the EC rescue. (B) Scatter plot showing the differential specific cell death between samples cocultured with E4-ECs vs alone for the 10 models from panel A. P value was estimated by t test. (C) Heat map showing the differential responses of 10 T-ALL PDX models in the presence or absence of ECs. Color coding corresponds to the change (delta) in specific cell death (SCD) of the +EC minus the –EC conditions (here referred to as efficacy ± EC). Red indicates drugs that are rescued by ECs; blue indicates those that produce more death in the presence of ECs. (D) Pie charts indicating the drugs having greater than 20% EC-mediated rescue for the 10 T-ALL PDX models. (E) Box-and-whisker plots showing the heterogeneity score (calculated as a modified Shannon entropy) per T-ALL PDX model in the presence or absence of ECs. (F) Box plot showing specific cell death (SCD) on exposure to the 39 most active compounds from the library in either ETP (yellow bars) or non-ETP (gray bars) models (median and standard error). Middle: Drugs that are equally effective between the 2 subgroups. Left: Drugs that are more efficient in ETP T-ALL. Right: Drugs that are more efficient in non-ETP T-ALL. The Δ SCD is color-coded at the bottom of the bars. Inset: Volcano plot of the same data, depicting –log₁₀P value corrected for multiple comparison (y-axis) and Δ SCD of ETP minus non-ETP samples. Top left corner: Drugs that are significantly less effective in ETP. Top right corner: Drugs that are significantly more effective in ETP models. ns, not significant.