Identification of non-charged 7.44 analogs interacting with the NHR2 domain of RUNX1-ETO and exhibiting an improved, selective antiproliferative effect in RUNX-ETO positive cells

The RUNX1/ETO fusion protein is a chimeric transcription factor in acute myeloid leukemia (AML) created by chromosomal translocation t(8;21)(q22;q22). t(8;21) abnormality is associated with 12% of de novo AML cases and up to 40% in the AML subtype M2. Previously, we identified the small-molecule inhibitor 7.44, which specifically interferes with NHR2 domain tetramerization of RUNX1/ETO, restores gene expression down-regulated by RUNX1/ETO, inhibits proliferation, and reduces RUNX1/ETO-related tumor growth in a mouse model. However, despite generally favorable physicochemical, pharmacokinetic, and toxicological properties, 7.44 is negatively charged at physiological pH and was predicted to have low to medium membrane permeability. Here, we identified M23, M27, and M10 as non-charged analogs of 7.44 using ligand-based virtual screening, in vivo hit identification, biophysical and in vivo hit validation, and integrative modeling and ADMET predictions. All three compounds interact with the NHR2 domain and show KD,app values of 39-114 {micro}M in Microscale Thermophoresis experiments as well as IC50 values of 33-77 M as to cell viability in RUNX1/ETO-positive KASUMI cells, i.e., are [~]5 to 10-fold more potent than 7.44. M23 is [~]10-fold more potent than 7.44 in inhibiting cell proliferation of RUNX1/ETO-positive cells. M23 and M27 are negligibly protonated or in a [~]1:1 ratio at physiological pH, while M10 has no (de-)protonatable group. The non-protonated species are predicted to be highly membrane-permeable, along with other favorable pharmacokinetic and toxicological properties. These compounds might serve as lead structures for the optimization of binding affinity, bioavailability, and anti-leukemic effects of compounds inhibiting RUNX1/ETO oncogenic function in t(8;21) AML.