Using an in-Vivo Degron-Based Approach to Interrogate Dependencies of Serially Acquired Mutations - Including DNMT3a-R882 and NPM1c - in Acute Myeloid Leukemia

While a number of targeted therapies have recently become available in the treatment of acute myeloid leukemia (AML), it remains to be determined whether early acquired mutations are essential for ongoing leukemia maintenance in the face of sequential acquisition of additional mutations. In some cases, our ability to answer this question is hampered by the lack of targeted therapies to a number of these mutations. Of all AML mutations, those involving DNMT3a and NPM1c are some of the most frequent - often acquired before the final transforming event. Here we report the development of an in-vivo degron-based approach to interrogate the requirement for DNMT3a-R882 and NPM1c after leukemia development.

Using an IKZF-based degron system offers us the opportunity to address the ongoing requirement of particular oncogenes for AML maintenance - by way of a conditional system that can be rapidly modulated. Thalidomide and its related analogues act by recruiting various substrates to the CRL4^CRBN E3-ubiquitin ligase complex mediating subsequent ubiquitination and proteasomal degradation. In a majority of cases - the basis of this degradation has been determined to be the result of recognition and binding to a degron motif within C2H2 zinc fingers. Using more detailed analysis with single amino-acid resolution, we have engineered 'degrons' sensitive or resistant to degradation by thalidomide and its related analogues.

In-vivo degradation using pomalidomide revealed a dependence of GMP cells on the presence of NPM1c but not DNMT3a R882. Using our cereblon (CRBN) knock-in mice engineered to mediate sensitivity to these compounds, we generated triple-mutant acute-myeloid leukemia - sequentially introducing DNMT3a-R882, NPM1c and N-RAS G12D mutations whilst altering the site of the degron tag in each case. Degradation-resistant degron constructs were used as a control in each case. Findings were subsequently correlated and validated with in-vitro treatments of triple mutant cells. RNA-seq was also performed after both in-vitro and in-vivo treatments to further delineate acute transcriptional changes in the absence of NPM1c - highlighting a number of changes including those within the Hox family of genes

Together, these findings suggest triple mutant leukemia remains dependent on the leukemia-inducing mutation of NPM1c. In contrast, whilst epigenetic alterations induced by DNMT3a mutations are essential for leukemia development - these do not appear to be essential for ongoing leukemia maintenance. Further work is currently underway to determine how modulating NPM1c impacts long-term survival from leukemia and whether the epigenetic changes associated with DNMT3a R882 mutations are indeed reversible in the context of leukemia.

This work highlights an approach to interrogate a protein of interest in-vivo using a conditional system that is reversible, rapidly acting and that uses a series of FDA-approved compounds with known kinetics. We envisage this will be generalisable as a tool to be used in other similar contexts.