RNAi screening for genetic mediators of response to dasatinib in vivo using a transplantable BCR-ABL1⁺BCP-ALL model. (A) Schematic representation of the longitudinal screening strategy. BCR-ABL1⁺ BCP-ALL cells were infected with a pool of 25 206 shRNAs tagged with GFP at MOI <1. GFP⁺ cells were isolated 48 hours postinfection by FACS, and cultured for 3 days before plating in 3 replicate cultures for in vitro screen and 4 days before injection into 8 nonirradiated syngeneic mice for in vivo screen. Input samples were taken at beginning of screens; pretreatment samples were taken after disease developed from cultures or blood of recipient mice, then mice/cultures were treated with dasatinib, allowed to relapse, and then posttreatment samples were taken from cultures or blood of recipient mice. Dasatinib was used in vitro at 0.6 nM (∼LD30) for 3 days, and in vivo at 10 mg/kg for 3 days once per day (q.d.) starting at 11 days postinjection. Hairpin representation was determined by high-throughput sequencing. (B) Bar graph showing the number of unique hairpins (1 or more reads) detected in each in vivo and in vitro screen sample both before (pre-tx) and after treatment (post-tx), as well as at transplant/plating (input). (C-D) Dendrograms generated by hierarchical clustering of normalized hairpin representation in vivo (C) and in vitro (D). (E) Principal component analysis of log₂ fold changes before (pre/input) and after treatment (post/pre) in vivo and in vitro. (F) Waterfall plots representing the log₂ fold changes after dasatinib therapy of all shRNAs in the library in vivo (blue) and in vitro (red), with shRNAs arranged in rank ascending order based on their log₂ fold change in vivo.