The Oncogenic Function of the Intrinsically Disordered Transcriptional Regulator MN1 Depends on the Chaperone DNAJB6b

Meningioma 1 (MN1) overexpression (OE) is associated with the progression from myelodysplastic syndrome to acute myeloid leukemia (AML) and correlates with poor survival in AML patients. In murine models, MN1-OE in common myeloid progenitors has been shown to drive AML transformation. Molecularly, MN1-induced inhibition of myeloid differentiation is linked to the over-stabilization of the BAF complex at binding sites of transcription factors involved in hematopoietic differentiation. However, the exact mechanisms through which MN1 promotes AML development remain unclear. We performed single-cell RNA-sequencing on 21 AML samples (17 diagnosis and 4 paired relapses – 190387 cells). Among the diagnosis samples, 8/17 were classified as adverse risk according to the ELN2022 risk stratification. We identified that MN1 expression was mainly confined to the leukemic stem cell populations (LSC) in each patient. Within these LSC populations, differential gene expression analysis identified 267 upregulated genes in MN1^high LSCs, which were associated with TNFa and IFNa inflammatory signaling and B- and T-cell differentiation. Conversely, 196 downregulated genes were linked to MYC targets and oxidative phosphorylation (OXPHOS). Single cell GSEA indicated that MN1^high LSCs exhibit a more quiescent and multilineage profile, as indicated by the LSC-UP and LMPP-UP programs. RNA-seq on NB4 MN1-OE cells mirrored the findings observed in MN1^high LSCs, with up-regulation of 1670 genes associated with LSC-UP and B- and T-cell programs, while genes related to the TP53 signaling pathway and OXPHOS were downregulated. Molecular profiling of 823 AML patients revealed that those with MN1^high often carry mutations in genes linked to ELN2022-adverse risk. These patients are also enriched in quiescent LSC cells and molecular programs associated with immature B-cell differentiation and LMPP-UP, suggesting that high MN1 expression is associated with an immature LSC-like multipotent phenotype. Lentiviral MN1-OE in primary APL samples (n=5 patients) allowed efficient engraftment in vivo using PDX mouse models with colonization of secondary organs and splenomegaly, which was absent in controls. MN1 is an intrinsically disordered protein with a long polyQ-stretch, required for transformation. We found that the expression of the chaperone DNAJB6 isoform 2 (DNAJB6b) was increased in CD34⁺ AML cells with high MN1 levels. We hypothesized that the disordered nature of MN1 creates a reliance on chaperones like DNAJB6b to prevent inappropriate cytoplasmic phase separation, making MN1 dependent on this chaperone for maintaining its oncogenic functions. shRNA-induced knockdown (KD) of DNAJB6b in AML cell models led to cytosolic retention of MN1, reduced cell survival and enhanced drug-induced apoptosis (with venetoclax, VEN and cytarabine) and myeloid differentiation (with ATRA). We validated these findings in an MN1-OE murine AML model, where Dnajb6b-KD strongly reduced cell proliferation and survival, with increased sensitivity to AML-related drugs such as VEN and ATRA. To unravel mechanisms via which MN1 controls oncogenic transformation, we performed RNA-seq and ChIP-seq on DNAJB6-KD cells to identify direct MN1 targets and examine the epigenetic marks H3K4me3 and H3K27ac. DNJAB6b-KD led to a decrease in MN1 binding and downregulation of MN1-bound loci (241), including known targets such as FLT3, GATA2, and MEIS1, and genes linked to B cell programs (CD79A, TYMS, and IKZF1). Gene set enrichment analyses in DNAJB6b-KD cells revealed an association with terms like MYELOID DIFFERENTIATION and HSC_DOWN, in line with loss of MN1 oncogenic properties upon DNAJB6b-KD. Additionally, we observed reduced H3K4me3 and H3K27ac marks at the MEIS1, HOXA gene family, and FLT3 loci upon DNAJB6b-KD. To identify potential therapeutic agents for targeting MN1^high AMLs, we analyzed the ex vivo screening of 30 AML samples using a 159-drug panel (Lee et al., Nat Comm 2015), which revealed bortezomib and tamoxifen as highly effective in MN1^high AMLs. We validated these results in our MN1-OE models, indicating that these AMLs, due to their biphenotypic features, may be particularly responsive to therapies employed for lymphoid malignancies. Our findings show that MN1 oncogenic functions in AML hinge on the chaperone DNAJB6b and that targeting such chaperones presents a promising therapeutic strategy to treat MN1^high AMLs.