Two leukemia drug targets, PRMT5 and LSD1, team up to push AML cells to differentiate and die

AML is an aggressive blood cancer in which immature blasts cannot mature and so accumulate. PRMT5, an enzyme that places symmetric dimethyl marks on arginine, supports leukemia stem cell survival and is a promising drug target. Full inhibition is toxic to normal cells, so researchers are looking for partner drugs that allow lower, safer PRMT5 doses.

The team screened a targeted panel of small molecules in mouse MLL-AF9 AML cells with either full or half PRMT5 levels. They looked for compounds that killed the half-PRMT5 cells preferentially. Hits included known partners (type I PRMT and SF3B1 inhibitors), chemotherapy, and new candidates such as LSD1, CK2, CDK4/6, and ERK1/2 inhibitors. In human AML lines MV4-11, K562, and U937, the PRMT5 inhibitor PRT220 and the LSD1 inhibitor GSK-LSD1 were synergistic across all five synergy models tested. In MV4-11 xenografts, the combination cut tumor volume by 68%, versus 31% for the PRMT5 inhibitor and 45% for the LSD1 inhibitor alone, with no apparent body-weight loss.

RNA sequencing showed that the two drugs affect overlapping gene sets and that the combination altered more genes than either drug alone. Antigen presentation and myeloid differentiation programs rose, while cell division genes fell. Flow cytometry confirmed higher CD11b, CD86, and other maturation markers. PRMT5 inhibition mainly changed splicing, including skipping of MDM4 exon 6, which lowered MDM4 protein and raised p53. LSD1 inhibition mainly changed transcript levels. Knocking out TP53 in MV4-11 cells reduced both the synergy in cell viability and the caspase-driven apoptosis, so the effect depends on p53.

The researchers then designed six hybrid molecules that join a tranylcypromine-based LSD1 warhead to a GSK591-like PRMT5 warhead. All six inhibited LSD1 more potently than tranylcypromine, and all inhibited PRMT5 in the nanomolar range with selectivity over PRMT1 and PRMT7. The lead compounds were MC4455 (PRMT5 IC50 14 nM) and MC4491 (43 nM). A crystal structure showed MC4455 bound covalently to the FAD cofactor of LSD1, and docking suggested it occupies the same PRMT5 pocket as GSK591. In MV4-11 cells, both compounds reproduced the combination's transcriptional and splicing changes, and MC4455 increased CD11b-positive cells.

Caveats are important. The in vivo work used a small cohort of subcutaneous xenografts in one cell line. The dependence on p53 may limit use in TP53-mutant AML, which is common in high-risk disease. Dual-compound data come from cell-based and biochemical assays only, and PRMT5 binding was modeled rather than crystallized. Safety, pharmacokinetics, and efficacy of the hybrids in animals remain untested.