Genetic Modeling and Therapeutic Targeting of ETV6-NTRK3 with Loxo-101in Acute Lymphoblastic Leukemia

Introduction: Philadelphia chromosome-like acute lymphoblastic leukemia (Ph-like ALL) is a high-risk subtype characterized by kinase-activating alterations. One recurrent alteration is the ETV6-NTRK3 fusion, which results in constitutive activation of NTRK3, a member of the neurotrophic receptor kinase family. ETV6-NTRK3 has been identified in a range of malignancies, including breast cancer, pediatric glioma and infantile fibrosarcoma. The oncogenic role of ETV6-NTRK3 in B-cell ALL has not been investigated. The goals of this study were to assess the development of leukemia in genetically engineered models of ETV6-NTRK3, and to investigate efficacy of the specific TRK A, B and C inhibitor, LOXO-101, currently in clinical trials for the treatment of solid tumor patients who harbor NTRK fusions.

Methods: For in vitro studies, kinase fusions were expressed in IL3 dependent Ba/F3 cells. To generate a genetically engineered mouse model, we used a previously reported conditional knockin model of Etv6-NTRK3 (Cancer Cell 2007;12:542-558), whereby the human portion of NTRK3 cDNA encoding the tyrosine kinase domain was inserted into exon 6 of the mouse Etv6 locus, downstream of a floxed transcriptional terminator sequence. Expression of the Etv6-NTRK3 protein was accomplished using Cre-recombinase driven by the B-lineage promoter CD19. A patient derived xenograft (PDX) model of ETV6-NTRK3 was established by engrafting primary human ALL cells expressing luciferase into NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice. Phosphoflow cytometry analysis and sensitivity to LOXO-101 was assessed in vitro and in vivo.

Results: Etv6-NTRK3/+, CD19-Cre mice developed aggressive disease with 100% penetrance and a median latency of 38 days (n=27). The average body weight of Etv6-NTRK3/+, CD19-Cre mice was significantly reduced compared to age-matched Etv6-NTRK3/+ controls (13.9 vs 20.2g, p<0.001). We observed increased spleen weight in Etv6-NTRK3/+, CD19-Cre mice compared to controls (142 vs 71mg, p=0.02), but no difference in peripheral white blood counts (9.7 vs 13.4 x 10⁹/L, p=0.3). Presence of the Etv6-NTRK3 fusion was confirmed in bone marrow samples by RT-PCR. Immunophenotyping of bone marrow indicated arrest at the pre-B stage (Hardy stage C: B220+, CD19+, CD43+, BP1+, IgM-), recapitulating human ALL. Pathological analysis using hematoxylin and eosin and B220 staining showed infiltration of leukemic cells into the bone marrow, spleen, liver and lung. Interestingly, we observed extensive infiltration of leukemic cells into the central nervous system, specifically ventral to the thoracic and lumbar vertebrae, and the meninges within the brain. Copy number alteration and sequence mutation analysis is currently being performed to determine additional genetic lesions. Leukemia cells from the bone marrow displayed constitutive activation of the MAPK pathway via pERK1/2.

We next assessed the in vitro efficacy of the TRK inhibitors crizotinib, which also inhibits ALK, and a more specific inhibitor, LOXO-101. Compared to crizotinib (IC₅₀ 205 nM), LOXO-101 was 10 times more potent against BaF3-ETV6-NTRK3 cells (IC₅₀17 nM), and had no effect on other kinase fusions (ABL1, ABL2, CSF1R, FLT3, JAK2) up to 10µM. In addition, LOXO-101 was remarkably selective for TRK A, B and C in a cytotoxicity screen of 77 human cancer cell lines as compared to crizotinib. Using a PDX model of ETV6-NTRK3, we demonstrate that treatment with LOXO-101 (200mg/kg/day p.o for six weeks) reduced leukemic infiltration to undetectable levels in the bone marrow (0 vs 75.8% human CD45/CD19 bone marrow blasts, n=5 each group) and spleen compared to vehicle-treated mice (splenic weight 316 vs 20mg, p<0.001). Notably, treatment with dexamethasone had a modest effect against this tumor (average 55.3% bone marrow blasts and spleen weight 134mg, n=5). Mice treated with LOXO-101 were still alive and leukemia-free four weeks after the cessation of treatment, as determined by Xenogen imaging.

Conclusion: We have described the first genetically engineered mouse model of Ph-like ALL with an ETV6-NTRK3 fusion, and reported remarkable efficacy of LOXO-101 against the NTRK3 fusion, with complete suppression of leukemic cell proliferation when administered as a monotherapy. These findings warrant screening for ETV6-NTRK3 in newly diagnosed ALL patients, and testing the efficacy of LOXO-101 in combination with chemotherapy regimens.