Small-Molecule U2 Auxiliary Factor Homology Motif (UHM) Domain Inhibitors Cause Splicing Pattern Changes in U2AF1 Mutant Leukemia Cells and Induce Sub-G1 Cell Cycle Arrest

Mutations in factors involved in the spliceosomal pathway have emerged to be an important contributor to the development of myeloid neoplasms including MDS and secondary AML (sAML). A set of frequently mutated genes encoding RNA splicing factors including SF3B1, U2AF1, SRSF2, and ZRSR2 or haploinsufficiency of LUC7L2 have beenreported, but pharmacological intervention targeting these genes remains limited and under-developed. Combination of mutations of these splicing factors are rarely found in patients indicating that multiple defects in a spliceosome pathway may be deleterious. Heterozygous mutation in U2AF1 found in MDS leads to the acquisition of the neomorphic mutant U2AF1 (U2AF1 ^wt) that relies on wild-type U2AF1 (U2AF1 ^wt) to survive and proliferate.

To support our hypothesis, we undertook a discovery campaign to identify small-molecule compounds targeting U2AF1. By recognizing the challenge of discovering compounds selectively targeting the mutant zinc-finger domains in U2AF1 found clinically, we devised a tractable intervention strategy to inhibit the UHM domain of U2AF1 (U2AF1-UHM). U2AF1-UHM binds with U2AF2 to form a protein complex (U2AF1/U2AF2) that recognizes the 3' splice site in mRNA processing. U2AF1-UHM inhibitors may decrease the binding between U2AF1 and U2AF2 and abrogate U2AF1/U2AF2 functions to change expression patterns of protein isoforms. To identify hit compounds for development, we performed screening of ~3000 fragment molecules against U2AF1-UHM using the thermal shift assay. Two of the top hits were derivatives of the same chemical scaffold. We verified that both hits inhibited the binding of U2AF1-UHM and U2AF2-ULM (U2AF ligand motifs) in our Homogenous Time-Resolved Fluorescence (HTRF) assay. We next evaluated the top hit, SF-1-8, in our K562-U2AF1 ^S34F mutant cell line and obtained an IC50 at the micromolar range. In contrast, SF-1-8 had no activity in K562-U2AF1 ^wt cell line and bone marrow cells obtained from healthy individuals. Based on SF-1-8, we performed chemical modifications to develop structure-activity relationship (SAR) of SF-1-8 and obtained SF-1-50 that was two-fold more potent than SF-1-8. To assess the selectivity of SF-1-8 and analogs to other UHM containing proteins, we further determined the IC50 values of our SF-1-8 analogs to RBM39-UHM, SPF45-UHM, and PUF60-UHM to construct the selectivity profiles of our compounds.

The selectivity profiles of SF-1-8 and SF-1-50 showed they were selectively more effective to U2AF1-UHM than RBM39-UHM, SPF45-UHM, and PUF60-UHM. We then studied pathways impacted by these inhibitors in K562-U2AF1 ^S34F cells by performing RNA-seq in K562-U2AF1 ^S34F cells treated with SF-1-8 at 5 uM and control. We found that 36 and 63 genes were significantly (p <0.002) up- and down-regulated respectively by SF-1-8 in K562-U2AF1 ^S34F cells. Downregulated genes included CBL, CBLL1, and a subset of collagen genes ( COL1, COL3, COL5) and upregulated genes included ATF3 and BCL2. When analyzing the transcript changes affected by SF-1-8, we identified substantial protein isoform changes in genes involved with proteosome, endocytosis, apoptosis, extracellular matrix/cell adhesion, histones, and stress response. Primary effects from changes of protein isoform patterns ameliorated the restoration of the trans-Golgi network ( ERGIC3, COPB2), secretory pathway ( RUSC1, AP4E1) and impairment of clathrin mediated endocytosis and endosome-lysosome transport ( CBL, PICALM, VAMP7, ASAP1). NKM-1 cell line was previously characterized to carry U2AF1 mutation. We also found SF-1-50 caused accumulation of sub-G1 cells in NKM-1 cells in a dose-dependent manner.

In summary, SF-1-8 caused disruption of extra- and intracellular protein transport in K562-U2AF1 ^S34F cells and represented a new class of small-molecule inhibitors to target U2AF1-UHM. Further optimization of this class of compounds will allow us to develop effective chemical probes for study in the U2af1 murine models and assess the potential of U2AF1 as a therapeutic target in MDS and sAML.