Activating Mutations of Insulin Receptor Substrate 2 (IRS2) in Patients with Tyrosine Kinase Inhibitor-Refractory Chronic Myeloid Leukemia

The success of the ABL1 tyrosine kinase inhibitor imatinib in the treatment of chronic myeloid leukemia (CML) set a precedent for molecularly targeted cancer therapies. Although the majority of chronic phase CML patients treated with imatinib achieve durable remission, approximately 15-20% of patients develop resistance to therapy. The best-characterized mechanism of imatinib resistance is via acquired point mutations in the BCR-ABL1 kinase domain that interfere with drug binding and subsequently reactivate BCR-ABL1 kinase activity. These mutations are generally well inhibited by newer, more potent ABL1 inhibitors such as nilotinib, dasatinib, and ponatinib. However, ~40% of patients with imatinib resistance demonstrate sustained inhibition of BCR-ABL1 kinase activity and the absence of a BCR-ABL1 point mutation, implying persistent activation of additional pathways for cell growth and survival. To identify and characterize these novel pathways, we have integrated whole-exome sequencing and high-throughput ex vivo kinase inhibitor panel screening of primary patient samples from patients with CML with BCR-ABL1 kinase-independent resistance. From our initial cohort of 22 patient specimens, we noticed common inhibitor sensitivity patterns in the PI3-K/AKT, JAK/STAT, and MAPK pathways. Deep sequencing revealed that a subset of these patient specimens (3/22; 13.6%) harbored a mutation in the insulin receptor substrate 2 (IRS2) gene. IRS2 belongs to the IRS family of cytoplasmic adaptor proteins that function as essential signaling intermediates for intracellular signaling pathways to modulate normal growth, cell survival, metabolism and differentiation (Mardilovich et al, Cell Commun Signal. 2009). IRS proteins are amplified and/or mutated in several other malignancies including breast, colorectal, prostate, pancreatic, and hepatocellular carcinomas, with IRS1 and IRS2 generally playing roles in proliferation and metastasis, respectively (Day et al., Int J Exp Patho 2013; Esposito et al., Oncol Rep 2013).  In contrast, while additional whole exome sequencing analysis performed by our laboratory identified IRS2 mutations in a small percentage of Philadelphia chromosome-negative myeloproliferative neoplasm patients (3/56; 5.4%), IRS2 variants were not detected in patients with acute myeloid leukemia (N=145) or acute lymphoblastic leukemia (N=83). Expression in each of the two of the IRS2 mutations identified in refractory CML patients (S594W and H1328R) in Ba/F3 cells (which are dependent on IL-3 for survival) demonstrated transformation capacity in the absence of IL-3. When co-expressed in Ba/F3 cells with BCR-ABL1, these IRS2 mutants conferred varying degrees of reduced sensitivity to imatinib in vitro and showed persistent AKT phosphorylation despite inhibition of BCR-ABL1 kinase activity by imatinib. Further evaluation of IRS2 mutants in NIH3T3 cells revealed sensitivity to PI3K/AKT/MTOR inhibitors. Together, our data suggest that targeting kinases in the PI3K/AKT/MTOR pathway in combination with imatinib in patients with BCR-ABL1 kinase-independent resistance and activating IRS2 mutations may be an effective alternative therapy.