Epigenetic Resistance to Menin-MLL1 Inhibition Is Driven By Loss of the Non-Canonical Polycomb Repressive Complex 1.1 in NUP98-Rearranged AML

Chromosomal translocations producing oncogenic fusion proteins are common drivers of many types of cancer, including leukemia. Translocations involving the Nucleoporin 98 (NUP98) gene produce NUP98-fusion proteins and are associated with a dire prognosis in acute myeloid leukemia (AML). NUP98-fusion proteins interact with the mixed-lineage leukemia (MLL1) chromatin modifying enzyme, and MLL1 is a molecular dependency in NUP98-rearranged (NUP98-r) AML. Our group recently showed that targeting the interaction between MLL1 and its binding partner Menin is therapeutically effective in NUP98-r AML models. Menin-MLL1 inhibition using a small molecule VTP50469 simultaneously represses pro-leukemogenic genes and upregulates markers of myeloid differentiation.

Mechanistically, we showed that disruption of the Menin-MLL1 interaction using VTP50469 leads to eviction of NUP98-fusion proteins from the transcriptional start site (TSS) of pro-leukemogenic genes. However, it remains unclear whether loss of the NUP98-fusion-Menin-MLL1 complex is sufficient to silence pro-leukemogenic gene expression and induce differentiation. We hypothesized that differentiation must be enforced through epigenetic silencing of pro-leukemogenic genes by a repressive chromatin complex.

We performed a bespoke CRISPR screen targeting 616 epigenetic regulators in NUP98-r leukemia cells treated with the Menin-MLL1 inhibitor VTP50469. Our results revealed that the non-canonical polycomb repressive complex 1.1 (PRC1.1)-a dynamic, repressive chromatin regulatory complex with an established role in lineage specification and cancer-are required for Menin-MLL1 inhibitors to inhibit cell proliferation. Loss of each PRC1.1 complex component resulted in a competitive growth advantage in our CRISPR screen, and this effect was enhanced in cells treated with VTP50469, suggesting that loss of PRC1.1 may mediate resistance to Menin-MLL1 inhibition.

To validate our screen and further define the role of PRC1.1 in NUP98-r leukemia, we used CRISPR/Cas9 to genetically inactivate each component of the PRC1.1 complex. We found that loss of PRC1.1 components resulted in a competitive growth advantage of NUP98-r leukemia cells, and fitness of PRC1.1-deficient leukemia cells was enhanced by treatment with VTP50469.

Next, we determined how PRC1.1 localization and activity changes upon eviction of NUP98-fusion proteins from chromatin using Menin-MLL1 inhibitors. We found that PRC1.1 components colocalize with the NUP98-Menin-MLL1 complex at the TSS of pro-leukemogenic genes whose expression is significantly downregulated upon treatment with VTP50469 such as MEIS1, EYA1, and PBX3. Menin-MLL inhibition resulted in increased PRC1.1 activity at these loci, as measured by the repressive histone mark H2aK119Ub. These results suggest that the NUP98-fusion-Menin-MLL1 complex and PRC1.1 dynamically compete for transcriptional control of pro-leukemogenic gene expression in NUP98-r leukemia.

To confirm these findings in a different model, we engineered a system in which NUP98-fusion proteins can be rapidly and inducibly degraded using heterobifunctional small molecules such as dTAG-13. We tagged NUP98-fusion proteins with a modified version of FKBP and used these constructs to transform mouse lineage negative, Sca-1 +, cKit+ (LSK) cells. We found that degradation of NUP98-fusion proteins resulted in rapid changes in transcription at key pro-leukemogenic loci upon dTAG-13 treatment. Genes which lost active transcription were bound by PRC1.1, and degradation of NUP98-fusion proteins led to accumulation of this repressive complex, increased H2aK119Ub, and loss of activating histone marks such as H3K27Ac.

In summary, these results demonstrate that the NUP98-fusion-Menin-MLL complex and PRC1.1 dynamically compete for transcriptional control of developmentally regulated, stem-cell associated genes. Loss of PRC1.1 may promote NUP98-fusion protein-driven leukemogenesis and mediates resistance to Menin-MLL1 inhibition by failing to epigenetically silence genes that are essential for maintaining an undifferentiated, stem cell-like state.

Armstrong:Imago Biosciences: Consultancy, Other: Shareholder; Janssen: Research Funding; Neomorph Inc: Consultancy; C4 Therapeutics: Consultancy, Other: SHareholder; Mana Therapeutics: Consultancy, Other: Shareholder; Twenty eight-seven Therapeutics: Consultancy, Other: Shareholder; Cyteir Therapeutics: Consultancy, Other: Shareholder; Accent Therapeutics: Consultancy, Other: Shareholder; Novartis: Research Funding; Syndax: Research Funding; -: Patents & Royalties: MENIN inhibition WO/2017/132398A1.