KDM3B inhibitors disrupt the oncogenic activity of PAX3-FOXO1 in fusion-positive rhabdomyosarcoma

Abstract

Fusion-positive rhabdomyosarcoma (FP-RMS) is an aggressive pediatric sarcoma driven primarily by the PAX3-FOXO1 fusion oncogene, for which therapies targeting PAX3-FOXO1 are lacking. Here, we screen 62,643 compounds using an engineered cell line that monitors PAX3-FOXO1 transcriptional activity identifying a hitherto uncharacterized compound, P3FI-63. RNA-seq, ATAC-seq, and docking analyses implicate histone lysine demethylases (KDMs) as its targets. Enzymatic assays confirm the inhibition of multiple KDMs with the highest selectivity for KDM3B. Structural similarity search of P3FI-63 identifies P3FI-90 with improved solubility and potency. Biophysical binding of P3FI-90 to KDM3B is demonstrated using NMR and SPR. P3FI-90 suppresses the growth of FP-RMS in vitro and in vivo through downregulating PAX3-FOXO1 activity, and combined knockdown of KDM3B and KDM1A phenocopies P3FI-90 effects. Thus, we report KDM inhibitors P3FI-63 and P3FI-90 with the highest specificity for KDM3B. Their potent suppression of PAX3-FOXO1 activity indicates a possible therapeutic approach for FP-RMS and other transcriptionally addicted cancers.

Fig. 1. Identification and characterization of P3FI-63.

A Small molecule screen strategy to identify compounds that selectively inhibit PAX3-FOXO1-driven transcription B Top 64 compounds and their inhibitory characteristics of ALK super-enhancer-driven luciferase vs. CMV-driven luciferase (20 µM). C RNA-seq GSEA analysis of RH4 cells treated with compound P3FI-63. Normalized Enrichment Score (NES), False Discovery Rate (FDR). D Western validation of RNA-seq showing upregulation of MYOG and PARP cleavage (n = 1). E EC₅₀ calculations of P3FI-63 on PAX3-FOXO1 positive RMS cell lines RH4, RH30, and SCMC. Data presented as mean values ± SEM. n = 3 biological replicates, error bar = Standard Error. F RNA-seq GSEA analysis of P3FI-63-treated RH4 cells reveal KDMs as possible targets. G Direct enzymatic inhibition assays of P3FI-63 against KDMs, HDACs, and PRMT5. Single experiment. H Western validation of KDM inhibition by P3FI-63 showing increases in methylation of H3K9me3, H3K4me3, and H3K27me3 (n = 1). Source data are provided as a Source Data file.

Fig. 2. Characterization of P3FI-90.

A P3FI-90 structure compared to P3FI-63. B EC₅₀ calculations for P3FI-90 on PAX3-FOXO1 positive RMS cell lines RH4, RH30, and SCMC. Data presented as mean values ± SEM. n = 3 biological replicates, error bar = Standard Error. C Direct enzymatic inhibition assays of P3FI-90 against KDMs, HDACs, and PRMT5 at 8 concentrations. Single Experiment. D Western validation of KDM inhibition showing increases in H3K9 and H3K4 methylation. One representative experiment from n = 2 is shown. E Left panel: Binding site of P3FI-90 (green) in KDM3B (PDB: 4c8d) overlaps with the binding site of the cofactor, 2-oxoglutarate (magenta). Middle panel: P3FI-90 docks well into the active site of KDM1A/LSD1 (PBD: 2Z3Y) forming hydrogen bonds with the main chain NH-groups of A331 and V-333. Pyridine moieties fit into hydrophobic pockets formed by the site chains of V-333, T335, A806, Tyr 761, L659, and Trp 751. Right panel: P3FI-90 alternative binding site on KDM3B that overlaps with N-phenyl-N’-pyridin3-yl-urea on KDM3B structure PDB:5RAZ and that is structurally similar to P3FI-90. The pocket is on the opposite side of the active site. In the alternative binding site, P3FI-90 forms hydrogen bonds with Ser1418, His 1394, and Glu1608. F RNA-seq GSEA analysis of RH4 cells treated with P3FI-90 showing downregulation of PAX3-FOXO1 target gene set, and upregulation of Apoptosis gene set. G Western validation of RNA-seq showing upregulation of MYOG and PARP cleavage. One representative experiment from n = 2 is shown. H Surface plasmon resonance show binding of P3FI-90 to KDM3B. KD of 7.68 × 10^-6 M, R_max of 130.5 RU, and Chi² of 4.78. I Ligand-observed nuclear magnetic resonance (NMR) technique WaterLOGSY (WL) with positive peaks showing P3FI-90 binding to KDM3B when KDM3B is present. N-Methyl-L-Valine (NMV) as non-binding control. J Ligand-observed NMR technique Carr-Purcell Meiboom-Gill (CPMG) showing attenuation in P3FI-90 peaks in the presence of KDM3B indicative of binding. Source data are provided as a Source Data file.

Fig. 3. CRISPR-based characterization of KDMs.

A CRISPR knockout with sgRNAs tiled across the entire KDM3B coding region. Analysis of fold depletion of sgRNAs as walking average of 4 sgRNAs. B RNA-seq TPM of KDMs after CRISPRi knockdown of KDMs. C Western validation of CRISPRi knockdown of KDMs. One representative experiment from n = 2 is shown. D GSEA analysis of RNA-seq data by P3FI-90 and KDM knockdowns. Refer to Supplementary Data 3, Tables 3–10 for Normalized Enrichment Scores (NES), p-values, and False Discovery Rates (FDR). E GSEA using P3FI90_UP and P3FI90_DOWN gene sets created using top 50 differentially expressed transcripts and F GSEA plots (Gene List on Supplementary Data 7, Table 3). Source data are provided as a Source Data file.

Fig. 4. scRNA-seq analysis of P3FI-90.

A UMAP of clusters after treatment with DMSO vs P3FI-90 for 16 h. B Cells in clusters as percent of total and as Log 2-fold change for P3FI-90 compared to DMSO. C Clusters ordered by average expression of Myoblast genes, Myocyte genes, and PAX3-FOXO1 target genes. n = 1 independent experiment; box plots of median and quartiles, whiskers showing 1.5 × interquartile ranges. Statics from GSEA analysis and adjusted for multiple testing by GSEA. D UMAP of clusters by cell cycle. E percentage quantitation of cell cycle overall, and G2M across clusters. F Flow cytometry of cell cycle in SCMC cell line after 24 h treatment with P3FI-90 at 1 µM. DMSO n = 3 biological replicates, P3FI-90 n = 2 biological replicates. Source data are provided as a Source Data file.

Fig. 5. Histone methylation profile of P3FI-90.

A ChromHMM analysis of read counts at chromatin states for H3K4me3, H3K9me2, H3K27me3, H3K27ac, and PAX3-FOXO1. n = 1. B ChIP-seq of H3K4me3, H3K9me2, H3K27me3, H3K27ac, and PAX3-FOXO1 genome-wide analysis at transcription start sites (TSS) and PAX3-FOXO1 (P3F) binding sites. Showing one representative profile from n = 3 except H3K27me3 which is n = 1. C ChIP enrichment analysis (ChEA) of H3K9me2 differential peaks in RH4 cell line treated for 24 h with P3FI-90 (1 µM) vs DMSO, n = 3. *p = values can be found in supplementary Data 5, Table 2. ChEA analysis uses Fisher’s exact test. D Gene ontology analysis of H3K4me3 differential peaks (top 20%) 24 h after treatment with P3FI-90 (1 µM) vs DMSO, n = 3. Gene ontology uses Fisher’s exact test. Source data are provided as a Source Data file.

Fig. 6. RNA Pol2 analysis of P3FI-90.

A ChIP-seq of Total RNA Pol2, Pol2 Ser5, and Pol2 Ser2 from TSS to transcription end site (TES) genome-wide. n = 1. B ChIP-seq of Total RNA Pol2, Pol2 Ser5, and Pol2 Ser2 at enhancer sites. n = 1. C IGV track view of MYOD1 gene. Orange highlight for enhancer regions and green highlight for gene body. D GSEA of ranked differential peaks of RNA Pol2 after treatment with P3FI-90 vs DMSO showing negative enrichment of PAX3-FOXO1 core regulatory circuitry.

Fig. 7. Efficacy of P3FI-90 in metastatic and orthotopic models FP-RMS and pictorial depiction of P3FI-90 action in FP-RMS.

A In vivo xenograft model of metastatic FP-RMS showing delay in tumor progression when treated with P3FI-90 vs DMSO. Data presented as mean values ± SEM. n = 5 for DMSO and n = 4 for P3FI-90. *p = 0.0016 (2-way ANOVA). Error bar = Standard error. B In vivo xenograft model of orthotopic intramuscular FP-RMS. n = 4 and *p = 0.0046 (2-way ANOVA). Error bar = Standard error. C RNA-seq GSEA of a tumor isolated from in vivo subcutaneous xenograft model of FP-RMS after treatment with P3FI-90 vs DMSO. D Model of P3FI-90 action in FP-RMS. Left panel show active transcription of PAX3-FOXO1 target genes which is downregulated by inhibition of KDM3B by P3FI-90 resulting in increased methylation at H3K9. Right panel show disrupted transcription of apoptosis and myogenesis genes which becomes active due to inhibition of KDM1A by P3FI-90 resulting in increased methylation at H3K4. Source data are provided as a Source Data file.