DYRK1A interacts with EZH2 to regulate transcriptionally active chromatin in myeloid leukemia associated with down syndrome

Abstract Children with Down syndrome are predisposed to hematological malignancies due to the presence of an extra copy of chromosome 21. Trisomy 21 (T21) alone has been shown to disrupt fetal hematopoiesis. Further examination identified that the Down syndrome critical region (DSCR) on chromosome 21 plays a key role in T21-driven leukemogenesis due to gene dosage imbalances. DSCR harbors several genes essential for hematopoietic development and leukemogenesis, such as DYRK1A, one of the dosage sensitive genes. As a dual-specificity kinase, DYRK1A plays diverse roles in cell cycle progression, stem cell differentiation, and DNA damage response. Interestingly, DYRK1A has been reported to interact with RNA polymerase II at active promoters, which shows a potential role in transcriptional regulation. DYRK1A overexpression leads to disrupted hematopoiesis, yet its distinct role in ML-DS remains unclear. In this study we showed DYRK1A as a novel interactor of EZH2 in CMK, a ML-DS cell line, and 3 distinct relapsed ML-DS PDX samples by co-immunoprecipitation. Given that both proteins are known to function in cytoplasmic and nuclear compartments, we performed co-immunoprecipitation in these subcellular fractions and found that the interaction only occurs in the nuclear fraction in CMK and ML-DS PDX cells. Considering the canonical role of EZH2 as a methyltransferase and of DYRK1A as a dual-specificity kinase, we examined if their canonical roles are necessary for this interaction. Therefore, we treated cells with EZH2 inhibitor (GSK126) or DYRK1A inhibitor (EHT1610) and performed co-immunoprecipitation. The data demonstrated that the interaction between DYRK1A and EZH2 was maintained even in the presence of inhibitors which suggest that the interaction between these two proteins is not dependent on either methyltransferase or kinase activity. Since DYRK1A has been implicated in transcriptional regulation and we showed that it interacts with EZH2 in the nucleus only, we hypothesized that DYRK1A may collaborate with EZH2 to shape the chromatin landscape in ML-DS. To explore this possibility, we performed CUT&RUN chromatin profiling of EZH2, DYRK1A, H3K27me3, and H3K4me3 in CMK cells. CUT&RUN using anti-DYRK1A antibody identified 12,348 peaks (p <0.001). These peaks partially overlapped with the H3K4me3 peaks, but they were not enriched by H3K27me3 peaks. Overlapping DYRK1A peaks with EZH2 peaks revealed 186 regions occupied by both, and these regions are highly enriched by H3K4me3, suggesting active gene expression. Pathway enrichment analysis of DYRK1A and EZH2 co-occupied regions indicated enrichment in key signal transduction pathways, including MAPK, ERK, RAS, and neurotrophin (NTRK) signaling (p < 0.05). To further investigate the function of EZH2 and DYRK1A in ML-DS, we depleted the expression of EZH2 or DYRK1A in CMK cells by CRISPR-cas9 mediated knockout. While depletion of EZH2 resulted in decreased protein level of DYRK1A in the nucleus, it had no effect on cytoplasmic DYRK1A. Interestingly, the downregulation of DYRK1A protein did not affect the level of EZH2 in both fractions. These data indicate that DYRK1A requires EZH2 for its nuclear localization and likely regulates expression of specific genes required for leukemogenesis. Together, these findings suggest that DYRK1A and EZH2 may functionally cooperate to regulate gene expression at transcriptionally active regions in ML-DS cells, especially within key signal transduction networks. This interaction may represent a previously unrecognized mechanism contributing to leukemogenesis in T21, with potential implications for targeting non-canonical EZH2 functions in ML-DS.