Negative feedback regulation of nfkb, MAPK, and JAK-STAT drives adaptive resistance to gilteritinib in AML

Abstract Kinase activating mutations in the FMS like tyrosine kinase 3 (FLT3) gene represent the most frequent molecular lesion in acute myeloid leukemia (AML). FLT3 mutations are associated with poor treatment outcomes and reduced overall survival, even after stem cell transplantation considered to be a curative option. Despite substantial progress in targeting FLT3 by small molecule inhibitors, the emergence of resistance limits the treatment efficacy. Patients treated with highly selective inhibitors such as gilteritinib, crenolanib, and quizartinib usually relapse within eight to twelve months. In most cases, drug evading mutations from the kinase domain or activation of RAS-MAPK signaling have been reported to drive resistance. Previous studies have demonstrated that hematopoietic cytokine signaling induces a drug refractory state in leukemic cells which is a shared feature with leukemia stem cells (LSCs). These cytokine-induced persistent cells constituting minimal residual disease (MRD) comprised of LSCs serve as a reservoir for the development of therapeutic resistance; therefore, their eradication is essential for effective treatment outcomes. Using a chemical genomic approach with AML cell lines and primary cells, we identified that AML treated with tyrosine kinase inhibitors (TKI) such as gilteritinib or quizartinib, in the presence of hematopoietic cytokines, transcriptionally upregulate a distinct set of thirty-four genes that are markedly enriched in drug-resistant cells. Gene ontology analysis revealed that these genes collectively modulate negative feedback regulation for the MAPK, JAK-STAT, and NF-κB signaling pathways. Functional validations through knockdown and overexpression experiments confirmed their role in TKI resistance. Importantly, we discovered that momelotinib, which efficiently inhibits the kinase activities of JAK, IKK1 and IKK2, IRAK1, and MAPK8, relieves the feedback regulation induced by FLT3 inhibitors. Consequently, the combination of momelotinib with FLT3 inhibitors (gilteritinib, crenolanib, or quizartinib) exhibited synergistic response and effectively abrogated cytokine-induced resistance while other clinically used JAK2 inhibitors did not exhibit such synergy with FLT3 inhibitors. Furthermore, a whole genome CRISPR-Cas9 drug sensitivity screen identified BCL2 as a strong synergy with momelotinib besides genes regulating cytokine signaling, DNA damage, transcription and translation machinery. Consistently, the BCL2 inhibitor, venetoclax, exhibited enhanced synergy with momelotinib compared to gilteritinib, both in vitro and in vivo against primary human and murine AML cells. The combination of gilteritinib and momelotinib with venetoclax resulted in profound leukemia suppression of murine MAL induce d by FLT3/Tet2. Similarly, PDX mice engrafted with chemoresistant AML cells exhibited effective leukemia clearance and prolonged survival when treated with momelotinib, gilteritinib, and venetoclax. This triple combination exhibited superior efficacy compared to each of the dual-drug combinations. Subsequent transcriptomic and chromatin accessibility analyses revealed that the combination of momelotinib and gilteritinib modulates the crosstalk among AP1, STAT5, and NF-κB. Under gilteritinib treatment, these transcription factors induce the expression of their own negative feedback regulators, facilitating cellular persistence, while momelotinib suppresses these transcriptional activities thereby restoring the TKI sensitivity. These findings explain how elevated AP1 and NF-κB/JAK-STAT activities transcriptionally induce negative regulators to favor cellular survival, as unchecked MAPK signaling would otherwise trigger apoptosis. Collectively, these data strongly support the rationale for evaluating the combination of gilteritinib and venetoclax with momelotinib as a therapeutic strategy to achieve effective and durable responses in AML.