Background: A novel class of small-molecule inhibitors of the Menin-MLL1 interaction are efficient in pre-clinical models and show remarkable efficacy in clinical trials in relapsed and therapy refractory MLL-rearranged and NPM1 mutant AML by disrupting the oncogenic Menin-MLL1-complex on chromatin. Targeting Menin leads to profound epigenetic reprogramming of leukemia cells resulting in terminal myeloid differentiation and eventually disease eradication. Nevertheless, the phenomenon of acquired drug resistance is limiting outcomes under long-term Menin-inhibitor treatment.
Aims: We aimed to characterize the phenomenon of acquired, non-genetic Menin-inhibitor resistance and identify an epigenetic switch to reprogram resistant cells towards a sensitive state.
Methods: Primary patient material, patient-derived xenografts (PDX) and AML cell lines were used for targeted DNA sequencing, RNAseq, ChIPseq, ATACseq and single-cell-ATACseq. A genome-wide CRISPR-Cas9 screen allowed an unbiased view on differential gene dependencies and co-vulnerabilities of resistant cells and CRISPR-based single gene deletions and highly selective small molecule inhibitors were used to validate the findings.
Results: We have previously identified point mutations in the MEN1 gene that are acquired during drug treatment and mediate resistance to Menin-inhibition. Although, these mutations appear to be a major determinant of therapy failure, our studies of patient material revealed that non-genetic adaptation processes may also play a role in mediating treatment escape.
In cohorts of PDX we discovered individuals in which drug resistance occurred without newly acquired somatic mutations. In contrast to MEN1-mutant leukemias, those cases of non-genetic resistance showed profound reprogramming of gene expression and cell identity and failed to re-express key MLL-target genes including MEIS1 and HOXA genes. In contrast to resistant leukemias with MEN1 mutations, the inhibitor retained its ability to displace Menin from chromatin.
To generate a cell line system that would allow to study the underlying mechanistic principles with more granularity, we transplanted OCI-AML2 cells into immunodeficient mice and treated them with the Menin-inhibitor VTP-50469 followed by harvest of resistant cells from relapsed animals. In accordance with our observations from PDX and primary patient material, those resistant cells did not acquire additional somatic mutations and showed profound transcriptional reprogramming, including irreversible loss of MEIS1. Importantly, CRISPR-Cas9 deletion experiments revealed that those cells lost their dependency on Menin but remained addicted to the MLL-fusion protein. Using a genome-wide CRISPR-Cas9 screen and subsequent validation experiments, we discovered that inactivation of the histone acetyltransferase KAT6A was able to reverse the resistance phenotype and restore sensitivity to the Menin-inhibitor. Co-immunoprecipitation, ChIPseq and RNAseq experiments demonstrated that KAT6A binds to the MLL-fusion complex and associates with Menin and MLL1 on chromatin. Finally, combined perturbation of Menin-MLL binding and KAT6A activity cooperated to silence MLL-target gene expression and increased therapeutic efficacy in both Menin-inhibitor resistant and sensitive leukemias.
Summary/Conclusion: We were able to establish the phenomenon of non-genetic cellular adaptation as a mechanism of resistance towards Menin-MLL1 inhibition in MLL-rearranged leukemia. Importantly, we provided preclinical proof-of-principle that such a reprogrammed cell state is reversible by pharmacologic targeting of a critical epigenetic switch.
Keywords: Resistance, MLL, Epigenetic, AML