Discovery of JAK2V617F mutant specific allosteric inhibitors for the treatment of myeloproliferative neoplasms

Abstract The V617F mutation in Janus Kinase 2 (JAK2V617F) is a recurrent driver of myeloproliferative neoplasms (MPNs), including myelofibrosis (MF), polycythemia vera (PV), and essential thrombocythemia (ET). Current therapies targeting the JAK-STAT pathway such as pan-JAK inhibitors (JAKi) like Ruxolitinib have demonstrated clinical benefit but are limited by a lack of disease-modifying activity. These agents bind the orthosteric site of the JAK2 JH1 domain and fail to selectively inhibit JAK2V617F over wild-type JAK2, resulting in insufficient suppression of the mutant clone at tolerable doses. Consequently, they do not significantly reduce mutant allele burden, lack disease-modifying effects, and rarely achieve clinical remission. Additionally, inhibition of wild-type JAK2 disrupts normal hematopoiesis, contributing to treatment discontinuation due to adverse events. These limitations underscore the need for mutant-selective, disease-modifying therapies. Allosteric inhibition of JAK2 represents a promising strategy to selectively target JAK2V617F through novel mechanisms of action. However, identifying suitable allosteric sites and ligands remains a significant challenge. To address this, Atavistik Bio has developed the AMPS™ (Atavistik Metabolite Proprietary Screening) platform, a scalable and systematic discovery engine for identifying unique ligand-macromolecule interactions. Leveraging this platform, we identified JAK2V617F selective inhibitors that bind the allosteric pseudokinase JH2 domain, which harbors the V617F mutation. Through structure-based drug design, we optimized initial hits from the AMPS™ screen to develop potent, picomolar-binding compounds with high specificity for the JAK2V617F JH2 domain. These compounds demonstrated strong binding to JAK2V617F JH2 and no measurable activity against JAK2 JH1 or the JH1 domains of other JAK family members at concentrations up to 50 µM. They also exhibited >100-fold selectivity over TYK2 and JAK1 JH2 domains. Compounds with potent and selective binding to the JAK2V617F JH2 domain also showed strong inhibition of JAK2V617F dependent signaling and proliferation in multiple human hematopoietic cancer cell lines harboring the mutation. In contrast, these compounds displayed significantly weaker potency against cytokine-stimulated JAK2 signaling and growth in cell lines expressing wild-type JAK2. To further assess selectivity, both copies of the JAK2V617F allele in UKE-1 cells were CRISPR-edited to wild-type JAK2. Using this isogenic cell pair, we confirmed that multiple compounds consistently exhibited >10-fold selective inhibition of cytokine-independent JAK2V617F signaling compared to cytokine-dependent wild-type JAK2 signaling. Additional studies in primary MPN patient-derived hematopoietic stem cells to assess mutant selectivity are ongoing. Select inhibitors were evaluated in mice bearing subcutaneous JAK2V617F expressing SET2 tumors. Pharmacokinetic and pharmacodynamic analyses in this model demonstrated robust target engagement, as evidenced by decreased phospho-STAT5 levels within xenograft tumors. Compounds with favorable PK profiles, tolerability, and minimal off-target activity were selected for further advancement. Additional studies in preclinical models of JAK2V617F driven splenomegaly are currently underway. In summary, we have discovered novel, potent, and selective allosteric inhibitors of JAK2V617F that effectively suppress pathogenic, cytokine-independent JAK2V617F driven signaling and cell proliferation while sparing normal cytokine-dependent signaling. These findings support the development of highly mutant selective inhibitors for the treatment of JAK2V617F driven MPNs.