Discovery and preclinical characterization of orally bioavailable JAK2V617F mutant selective JH2 inhibitors with disease modification potential in myeloproliferative neoplasms

Abstract

Abstract The JAK2V617F mutation is the most common genetic alteration in patients with myeloproliferative neoplasms (MPNs), present in >90% of patients with polycythemia vera (PV) and 50-60% of patients with essential thrombocythemia (ET) and myelofibrosis (MF). This mutation results in constitutive and growth factor-independent activation of the JAK-STAT signaling pathway, leading to hyperproliferation and myelofibrosis. Current therapies for MPNs, such as first-generation JAK inhibitors like ruxolitinib, target the JH1 kinase domain and inhibit both wildtype (WT) and mutant JAK2V617F (JAK2VF) with equal potency. As such, their efficacy is limited by dose-limiting cytopenias due to WT JAK2 inhibition in normal hematopoietic tissue and insufficient inhibition of mutant JAK2 needed to alter disease course. Next-generation JAK2VF selective inhibitors could achieve deeper inhibition of mutant JAK2 while sparing WT, offering the potential for improved clinical and molecular responses with disease-modifying effects. Our medicinal chemistry strategy focused on targeting the JAK2 pseudokinase (JH2) binding site wherein the V617F mutation resides. Utilizing structure-based drug design, we created de novo a series of highly potent, allosteric JAK2VF selective inhibitors that bind into the deep pocket containing the Phe-triad formed by the V617F residue along with F594 and F595. X-ray co-crystal structures of JAK2VF with multiple lead compounds confirmed our distinct “deep pocket” binding mode compared to previously reported JH2 inhibitors. These lead JAK2VF selective inhibitors exhibited >100-fold selectivity over other JAK family isoforms (JAK1, JAK3 and TYK2 JH2 domains) and were highly selective against >300 kinases, including the JAK2 JH1 domain. In cellular phosphoproteomic assays, these molecules selectively inhibited phosphorylation of downstream proteins, including STATs, MAPK and AKT, in JAK2VF cells compared to WT cells. In addition, these compounds showed potent and dose dependent anti-proliferative activity in JAK2VF cell lines and primary patient CD34+ cells in ex vivo cultures (IC50 50-100 nM) compared to WT cells (IC50>1uM). Treatment of JAK2VF patient CD34+ cells with the mutant selective inhibitors resulted in a selective reduction of JAK2VF variant allele frequency (VAF). Mechanistically, treatment with these compounds induced G1/S cell cycle arrest and apoptosis selectively in JAK2VF cell lines, with minimal effects in WT cells. In vivo, oral administration of the JAK2VF selective inhibitors resulted in potent and selective inhibition of pSTAT5 in JAK2VF tumor tissue (>90%), with minimal inhibition observed in EPO-induced pSTAT5 signaling in JAK2 WT spleen (<25%) in the SET2 xenograft model. Significant dose-dependent tumor growth inhibition (>70%), similar to ruxolitinib, was observed at well-tolerated doses in this model. In the BaF3-EPOR-JAK2VF model, treatment normalized splenomegaly (>70% decrease in spleen size) and reduced inflammatory cytokines. The JAK2 VF selective inhibitors were also evaluated in a head-to-head study with ruxolitinib in the JAK2VF bone marrow transplant mouse model (Marty et al. Blood 2010). Inhibitor treatment outperformed ruxolitinib in all clinicohematological parameters, leading to normalization of white blood cells, platelets and spleen size without causing cytopenias or other adverse effects. Importantly, there was a significant reduction in hematocrit and hemoglobin, which was not achieved with ruxolitinib and is consistent with published reports. Multiple inflammatory cytokines were also suppressed with inhibitor treatment, similar to ruxolitinib. Most importantly, analysis of stem and progenitor populations revealed a selective decrease in proliferation of JAK2VF long-term HSCs as well as megakaryocyte-erythroid progenitor populations, key contributors to MPN pathogenesis, without an impact on WT cells, in contrast to ruxolitinib which has similar effects on both JAK2VF and WT cells. These data establish our JAK2VF selective inhibitors as potential first-in-class agents with disease-modifying potential. They achieve deep molecular engagement via Phe-triad targeting, demonstrate exceptional selectivity, and show in vivo efficacy superior to ruxolitinib. This approach holds promise for transforming the treatment landscape in MPNs by directly targeting the molecular cause of disease. Prelude is now advancing a development candidate into Phase 1 clinical studies.