PI3K-δ inhibitor resistance could be overcome by targeting IGF1R. Dose-response curves of IGF1R inhibition in A20 cells transduced with empty vector (A) or Igf1r cDNA (B). The IC₅₀ values of empty vector vs Igf1r cDNA–transduced A20 cells were 100.5 nM vs > 20 µM for iPI3K-δ treatment, 16.3 µM vs >20 µM for linsitinib treatment, and 28.8 nM vs 142.5 nM for combination treatment with linsitinib and iPI3K-δ. (C) Western blot analysis for signaling changes in A20 cells transduced with empty vector or Igf1r cDNA upon treatment with DMSO, GS-649443, or linsitinib in the presence or absence of anti–immunoglobulin G stimulation. (D) RT-qPCR analysis of IGF1R expression after 24 hours of treatment with 1 µM idelalisib in CLL patient samples with high and low IGF1R expression (n = 12). The sample from the single in vivo idelalisib–treated patient is shown in red. The expression levels were normalized to a commonly included representative patient sample with low IGF1R (sample 1). (E) Western blot analysis of IGF1R, p-ERK, and ERK levels in CLL patient samples with low IGF1R (1-6) or high IGF1R (8-10) and 1 sample from an idelalisib-treated patient (11). Sample 7 showed high IGF1R expression for mRNA but not protein. One sample with high IGF1R measured by RT-qPCR was not included in the western blotting because of limitations in sample availability. (F) Dose-response curves from an idelalisib-treated patient sample, treated in vitro with idelalisib, linsitinib, or the combination for 4 days. Viability was assessed using DiOC6/propidium iodide/CD19 staining and fluorescence-activated cell sorting. Kaplan-Meyer survival curves for recipient mice that were transferred with PI3K-δ–sensitive tumor cells (G) or PI3K-δ–resistant cells (H) from the third transfer, followed by treatment with vehicle, linsitinib, GS-649443, or a combination of linsitinib and GS-649443. The P values were calculated using the log-rank (Mantel-Cox) test. **P ≤ .01; ***P ≤ .001.