Trabectedin Inhibits EWS-FLI1 and Evicts SWI/SNF from Chromatin in a Schedule-dependent Manner

Abstract

Purpose: The successful clinical translation of compounds that target specific oncogenic transcription factors will require an understanding of the mechanism of target suppression to optimize the dose and schedule of administration. We have previously shown trabectedin reverses the gene signature of the EWS-FLI1 transcription factor. In this report, we establish the mechanism of suppression and use it to justify the reevaluation of this drug in the clinic in patients with Ewing sarcoma.Experimental Design: We demonstrate a novel epigenetic mechanism of trabectedin using biochemical fractionation and chromatin immunoprecipitation sequencing. We link the effect to drug schedule and EWS-FLI1 downstream target expression using confocal microscopy, qPCR, Western blot analysis, and cell viability assays. Finally, we quantitate target suppression within the three-dimensional architecture of the tumor in vivo using ¹⁸F-FLT imaging.

Results: Trabectedin evicts the SWI/SNF chromatin-remodeling complex from chromatin and redistributes EWS-FLI1 in the nucleus leading to a marked increase in H3K27me3 and H3K9me3 at EWS-FLI1 target genes. These effects only occur at high concentrations of trabectedin leading to suppression of EWS-FLI1 target genes and a loss of cell viability. In vivo, low-dose irinotecan is required to improve the magnitude, penetrance, and duration of target suppression in the three-dimensional architecture of the tumor leading to differentiation of the Ewing sarcoma xenograft into benign mesenchymal tissue.

Conclusions: These data provide the justification to evaluate trabectedin in the clinic on a short infusion schedule in combination with low-dose irinotecan with ¹⁸F-FLT PET imaging in patients with Ewing sarcoma.

Figure 1:. The suppression of EWS-FLI1 by trabectedin is concentration dependent.

(A) Direct comparison of identical exposures of trabectedin for the indicated time followed by replacement with drug-free medium in TC32 cells (Exposure = Concentration*Time). Greater suppression of cell viability (percent confluence) occurs above a 10 nM threshold (10 nM, 60 min) relative to solvent control (solvent). (B) Sustained suppression of EWS-FLI1 activity as measured by NR0B1-Luc (black bars) in comparison to CMV-driven (gray bars) reporter. Cells exposed to drug for 1 hour followed by a 17 hour incubation in drug-free medium. (C) Sustained suppression of EWS-FLI1 target genes (EZH2, WRN, NR0B1) favors high concentration (Cmax) exposure to drug. Data is direct comparison of identical exposure of 24nM trabectedin for 1 hour followed by 23 hours in drug free medium (Cmax) or 1nM trabectedin for 24 hours (AUC) exposure as measured by qPCR fold change relative to GAPDH (2^ddCt) ****, p-value<0.0001. (D)(E) Western blot in 5 Ewing sarcoma cell lines comparing the effect of Solvent (S) to Cmax or AUC exposure on the expression of the EWS-FLI1 downstream targets NR0B1, EZH2, WRN relative to the GAPDH loading control. (F) Dose response curves of cell number as a function of exposure (concentration*time = logAUC) in TC32 Ewing sarcoma cells. Trabectedin was incubated at 10 concentrations for the indicated time and then replaced with normal medium for a total of 48 hours. Concentrations tested were 25, 20, 15, 12.5, 10, 5, 2.5, 1.25, 0.625, and 0.3125 nM. Above a threshold concentration, 6 minutes of drug exposure leads to sustained effects on viability 48 hours after drug is removed as indicated by the red curve.

Figure 2:. Trabectedin redistributes EWS-FLI1 within the nucleus in a schedule-dependent manner.

Redistribution of EWS-FLI1 within the nucleus in TC32 Ewing sarcoma cells with (A) high dose exposure (Cmax, 24 nM for 1 hour), drug removal and incubation for the indicated time but not with (B) low dose continuous exposure (AUC, 1 nM for 24 hours). (C) Similar redistribution of EWS-FLI1 only with high Cmax exposure (24 nM for 1 hour) in TP53 mutant A673 cells. Confocal microscopy stained for nucleolin (NCL), EWS-FLI1.

Figure 3:. Trabectedin evicts SWI/SNF from chromatin in a schedule-dependent manner.

(A) Trabectedin evicts SMARCC1 and EWS-FLI1 from chromatin with high dose (Cmax, 24 nM for 1 hour) followed by incubation in drug-free medium but not (B) continuous low dose (AUC, 1nM continuous) exposure in TC32 Ewing sarcoma cells. Western blot analysis showing total lysate (Total), chromatin fraction (chromatin) with H3 histone control (H3) and soluble fraction (soluble) with GAPDH control. Lysates collected at 1, 9 and 16 hours. (C) Chromatin immunoprecipitation of IgG or SMARCC1 at known EWS-FLI1 and SWI/SNF target genes (MYT1, SOX2, CCND1, NR0B1) in comparison to (D) GAPDH locus control following 24 nM trabectedin treatment for 1 hour (1h Trab.) followed by collection immediately or after 8 more hours in drug free medium (9h Trab.) in TC32 cells. Data is represented as percent input quantitated against a standard curve.

Figure 4:. Trabectedin treatment reverses the pioneering activity of EWS-FLI1.

(A) Venn diagram of the total number of H3K27me3 (left) and H3K9me3 (right) peaks as measured by chromatin immunoprecipitation and sequencing (ChIP-seq) following treatment with DMSO solvent, 24nM trabectedin for 1 hour (1 Hour Trab.), or 24 nM trabectedin for 1 hour followed by an 8 hour recovery in drug-free media (9 Hour Trab.) in TC32 Ewing sarcoma cells (B) Heatmap displaying the genome-wide distribution of (B) H3K27me3 or (C) H3K9me3 peaks relative to transcriptional start sites (TSS) following 24 nM trabectedin for 1 hour (Hour 1), or 24nM trabectedin for 1 hour followed by 8 hours in drug-free media (Hour 9) (D) Genome-wide distribution of reads of H3K9me3 peaks relative to TSS following 24 nM trabectedin for 1 hour (Hour 1), or 24nM trabectedin for 1 hour followed by 8 hours in drug-free media (Hour 9) (E) Total number of GGAA microsatellites marked (+/− 50KB) with H3K9me3 (15,400, dark green), H3K27me3 (2767, light green), both (105, darkest green) or neither (8443, grey) after treatment with 24 nM trabectedin for 1 hour. (F) Number of EWS-FLI1 target genes containing GGAA microsatellite sequences within 50kb of TSS. (G) Total number of GGAA microsatellites associated with EWS-FLI1 target genes marked (+/− 50KB) with H3K9me3 (30, blue), H3K27me3 (6, light blue), both (40, dark blue) or neither (7, grey) after treatment with 24 nM trabectedin for 1 hour (H) Genome browser tracks of H3K9me3 at TSS following indicated solvent or trabectedin treatments at the NR0B1 gene.

Figure 5:. Trabectedin suppresses EWS-FLI1 activity as measured by ¹⁸F-FLT imaging.

(A) Mice bearing TC32 Ewing sarcoma xenografts in right gastrocnemius show suppression of ¹⁸F-FLT signal 6 to 54 hours after treatment with trabectedin but not vehicle control. The bladder shows high ¹⁸F-FLT signal across all samples due to excretion of tracer. (B) High PET avidity of two mice 24 hours after treatment with vehicle (day 2). Data is a 3-Dimensional reconstruction of the tumor (tumor) followed by cross-sections in the X, Y, and Z axes. (C) Suppression of ¹⁸F-FLT PET avidity in two mice 24 hours after treatment with 0.18 mg/kg of trabectedin (day 2). Data is a 3-Dimensional reconstruction of the tumor (tumor) followed by cross-sections in the X, Y, and Z axes. Scale indicates signal intensity. (D) 3-D reconstruction and single cross-section of tumors at multiple time points following treatment with vehicle (1-hour), trabectedin (1-hour), irinotecan (24 and 48 hour), or the combination of trabectedin and irinotecan. Rows indicate time and treatments, columns represent 3-dimensional reconstruction and single cross-section for each of the treatment groups. The intensity scale is the same as (B), and (C).

Figure 6:. Combination treatment of trabectedin and irinotecan induces differentiation of TC32 Ewing sarcoma cells in vivo.

(A) (left to right) 4× and 20× magnification of H&E staining of TC32 IM xenograft tumor 3 days after treatment with vehicle (control) or trabectedin and irinotecan (treated). 60× magnification of MTCO2 human mitochondrial stain and 60× SP7 Osterix osteoblast stain showing human cells expressing SP7 in treated but not control. (B) (left to right) 4× and 20× magnification of H&E staining of TC32 IM xenograft. 60× magnification of MTCO2 human mitochondrial stain and 60× PicroSirius Red stain indicating specific human collagen cells 5 days after treatment with vehicle (control) or trabectedin and irinotecan (treated). (C) (left to right) 4× and 20× magnification of H&E staining of TC32 IM xenograft. 20× and 60× magnification of MTCO2 human mitochondrial stain showing human adipocyte. 5 days after treatment with vehicle (control) trabectedin and irinotecan (treated).