Unlocking Terminal Differentiation in AML: Targeting ISWI-Mediated Transcription Repression for Selective Therapy

Acute myeloid leukemia (AML) cells exhibit abnormal expression levels of master transcription factors, similar to or even exceeding those in normal granulocyte-monocyte progenitors (GMP), yet they fail to terminally differentiate into granulo-monocytic lineages. Despite the presence of proliferation and myeloid-commitment genes, gene expression analyses reveal a suppression of monocyte and granulocyte terminal-differentiation genes targeted by PU.1.

To address this repression, liquid chromatography-tandem mass spectrometry (LCMS/MS) was utilized to analyze the protein composition of the PU.1/RUNX1/CEBPA transcription factor complex in AML cells. The study found an enrichment of corepressors over coactivators, a known imbalance caused by leukemia oncoproteins. Notably, ATP-dependent chromatin remodelers SMARCA5 and CHD4 were identified as enriched corepressors, playing a critical role in epigenetic repression by obstructing gene transcription start sites.

A high throughput screen led to the discovery of a novel compound series, ED2-AD101, which inhibits the activity of SMARCA5/CHD4. In silico modeling, structure-activity relationship studies, and various assays demonstrated that ED2-AD101 allosterically inhibits ISWI family ATP-ase activity by binding to the HELICc-DExx domain. ED2-AD101 potently suppressed AML cell growth in vitro via terminal differentiation without inducing early apoptosis, while sparing the growth of normal CD34+ hematopoietic stem and progenitor cells.

The compound also induced terminal granulo-monocytic differentiation in THP-1 cells, as indicated by CD11b/CD14 markers. Importantly, ED2-AD101 induced differentiation-based cell cycle arrest even in chemo-resistant p53-null AML cells and other cancer cells, demonstrating its potential in overcoming chemoresistance. The lead compound showed pharmaceutical range efficacy with a GI50 of approximately 0.9 μM and a therapeutic index consistent with previous observations using decitabine or 5-azacytidine to inhibit DNMT1 in AML cells.

Drug pharmacology and pre-clinical in vivo experiments using patient-derived xenotransplant models of leukemia are ongoing. The findings suggest that targeting druggable corepressors could provide a novel approach to non-cytotoxic, normal hematopoiesis-sparing, differentiation-based oncotherapy for AML and other cancers.