KMT2A-Rearranged ALL Requires DYRK1A for Regulation of ERK Signaling and Cell Proliferation

Background: B-acute lymphoblastic leukemia (ALL) is the most common childhood cancer and is driven primarily by oncogenic chromosomal rearrangements and gene fusions. While overall survival rates are nearly 90% in children and about 40% in adults with good- and intermediate-risk ALL subtypes, clinical outcomes of patients with high-risk ALL subtypes, including KMT2A-rearranged (KMT2A-R), are significantly lower, posing an urgent need to develop novel therapeutic strategies for these patient populations.

Results: To identify targetable kinases and potential precision medicine approaches in KMT2A-R ALL, we conducted domain-specific kinome-wide CRISPR screens and found the serine/threonine kinase DYRK1A essential for KMT2A-R ALL proliferation. The Cancer Dependency MAP suggests DYRK1A is non-essential for normal tissues, indicating a potential therapeutic window. RT-PCR and Western blot analyses showed increased DYRK1A expression levels in KMT2A-R ALL cell lines and patient-derived xenograft (PDX) samples compared to other leukemia subtypes, suggesting direct regulation by the KMT2A-fusion oncogene. Meta-analysis of ChIP-seq data confirmed direct binding of the KMT2A-fusion protein to the DYRK1A promoter. Inhibition of menin, a tumor suppressor protein that forms a complex with the KMT2A-fusion protein, reduced DYRK1A expression at RNA and protein levels and prevented the binding of the KMT2A-fusion protein to the DYRK1A promoter, validating the direct regulation of DYRK1A by the KMT2A-fusion oncogene. To test the consequence of direct pharmacologic inhibition of DYRK1A using EHT1610 and harmine, we treated multiple KMT2A-R ALL cell lines and PDX models, validating potent inhibition of leukemia cell growth without inducing apoptosis. Furthermore, we detected potent upregulation of MYC and hyperphosphorylation of ERK. Premature ERK signaling activation in developing B cells is known to induce cell cycle arrest as a protective mechanism to prevent the development of autoimmune diseases. We hypothesized that DYRK1A inhibitor-induced ERK hyperactivation leads to cell cycle arrest in KMT2A-R ALL via a similar mechanism. Supporting our hypothesis, combining DYRK1A inhibitors with the MEK inhibitor trametinib rescued KMT2A-R ALL cell proliferation, validating that ERK hyperactivation is the main driver of DYRK1A inhibitor-mediated cell cycle arrest. Interestingly, although DYRK1A inhibition mainly affected cell proliferation, we detected a potent increase in the pro-apoptotic molecule BIM. BIM is inactivated by the anti-apoptotic molecule BCL2, which was also slightly increased after DYRK1A inhibition, indicating that DYRK1A inhibitors could also sensitize KMT2A-R ALL cells to BCL2 inhibition (e.g., venetoclax). In support of our hypothesis, combined treatment with DYRK1A inhibitors and venetoclax synergistically killed KMT2A-R ALL cells in vitro. Given the poor pharmacodynamics and severe toxicity of EHT1610 and harmine, we synthesized a new DYRK1A inhibitor, GNF2133, and tested its activity in vitro against KMT2A-R ALL cell lines and in vivo +/- venetoclax in KMT2A-R ALL PDX models. Importantly, both GNF2133 and venetoclax were well tolerated in xenograft mice, and combination inhibitor therapy resulted in superior inhibition of in vivo leukemia proliferation and long-term animal survival.

Conclusion: Our results identify DYRK1A as a critical regulator of ERK signaling in KMT2A-R ALL. DYRK1A inhibition leads to BIM accumulation, sensitizing cells to BCL2 inhibition by venetoclax in vitro and in vivo and resulting in therapeutic benefit in preclinical PDX models that is highly translatable to the clinic.