Preclinical Testing Of a PI3K Inhibitor In T Lineage Leukemia: Target Validation and Notch1/Myc Down-Regulation In Drug Resistant Clones

Mutations that deregulate Notch1 and Ras/PI3K/Akt signaling are common in T-ALL and often coexist. Thus, inhibiting these pathways alone and in combination has been advocated as a rational therapeutic strategy. GDC-0941 is a pan-PI3K inhibitor that is advancing in clinical development that we evaluated for efficacy and mechanisms of drug resistance in T-ALL cell lines and primary murine T-ALLs. Cell lines are uniformly sensitive to GDC-0941; however gradual dose escalation results in resistant lines. By contrast, primary leukemia cells display variable in vivo sensitivity to treatment with GDC-0941 +/- a MEK inhibitor. Importantly, mice invariably relapse with drug resistant clones. A majority of GDC-0941-resistant T-ALL cell lines and primary leukemias unexpectedly showed reduced levels of activated Notch1 protein, down-regulate many Notch1 target genes including Myc, and display cross-resistance to gamma secretase inhibitors (GSIs). In multiple cases, Notch1 mutations that were present in drug-sensitive parental leukemias were absent at relapse. Importantly, resistant clones that emerge both in vitro and in vivo up-regulate PI3K signaling indicating an “on pathway” mechanism of resistance. Consistent with these data, inhibition of Notch1 signaling promotes GDC-0941 resistance and enhances PI3K signaling, whereas expression of activated Notch1 increases the GDC-0941 sensitivity of mouse and human T-ALL cell lines. Thus, oncogenic Notch1 mutations that promote clonal outgrowth during malignant transformation unexpectedly undergo negative selection during treatment with GDC-0941. Together, these in vivo studies: (1) validate PI3K as an important therapeutic target in T-ALL; (2) demonstrate that active Notch1 and elevated Myc expression are dispensable for T-ALL growth in vivo; and, (3) indicate that therapeutic PI3K inhibition selects for the outgrowth of leukemia cells with reduced Notch1 signaling. Importantly, our data also raise the possibility that simultaneously administering Notch1 and PI3K inhibitors will accelerate drug resistance in T-ALL, and support alternative approaches for deploying combination regimens against these dominant oncogenic “driver” pathways.