Subversion of mRNA degradation pathways by EWSR1::FLI1 represents a therapeutic vulnerability in Ewing sarcoma

Abstract

Many cancers are defined by gene fusions that frequently encode oncogenic transcription factors (TFs), such as EWSR1::FLI1 in Ewing sarcoma (EwS). Here, we report that independently to its canonical roles in transcription, EWSR1::FLI1 also functions as an mRNA decay factor, reshaping mRNA stability in EwS. This function participates in EWSR1::FLI1 tumorigenicity and involves interactions of EWSR1::FLI1 with the CCR4-NOT deadenylation complex via its EWSR1-derived low-complexity domain and with the RNA-binding protein HuR/ELAVL1 via its FLI1-derived region. Strikingly, we find that EWSR1::FLI1-mediated mRNA decay antagonizes the normal mRNA protective function of HuR and renders EwS cells highly sensitive to HuR inhibition. Our findings uncover a post-transcriptional function of EWSR1::FLI1 and suggest that targeting mRNA stability mechanisms may offer therapeutic opportunities for EwS.

Fig. 1. EWSR1::FLI1 dictates an aberrant mRNA stability landscape in Ewing sarcoma.

a Schematic of the experimental design assessing transcript abundance and stability changes following EWSR1::FLI1 knockdown in shA673-1c cells. Created in BioRender, Galvan, B. (2025) BioRender.com/0001. b Half-life (HL) distribution of 7170 transcripts measured by ActD-RNA-seq in EWSR1::FLI1^high vs. EWSR1::FLI1^low shA673-1c cells. c Volcano plot showing significantly destabilized (red) and stabilized (blue) mRNAs in EWSR1::FLI1^high vs. EWSR1::FLI1^low shA673-1c cells (log2(HL ratio) > 0, p-value < 0.05, two-sided paired Student’s t-test, n = 3 independent experiments). Unaffected mRNAs are gray. Horizontal and vertical dotted lines indicate statistical cutoff (p-value = 0.05) and log2(HL ratio) = 0, respectively. Transcripts selected for validation are indicated. d GO term enrichment of destabilized transcripts from (c) (FDR < 0.2, PANTHER analysis; dot size = number of genes; color scale = p-value). e Heatmap of HL ratios for decay panel transcripts using ActD-RNA-seq, ActD-RT-qPCR or 4SU-qPCR in EWSR1::FLI1^high vs. EWSR1::FLI1^low shA673-1c cells (first 3 lanes). Validation by ActD-RT-qPCR in shSK-E17T cells and hMSCs shown in last two lanes. Results are means from 3 to 4 independent experiments; ns = not significant (one-sample t-test). f Immunofluorescence of PB markers DCP1A and EDC4 (white) and DAPI (nuclei, blue) in shA673-1c and shSK-E17T cells, -/+dox. Scale = 5 µm. g PB quantification from (f) as mean ± SD (n > 50 cells per replicate over 3 independent experiments). **p < 0.01; ***p < 0.001 (unpaired two-tailed Student’s t-test). h RIP analysis of decay panel and control mRNAs bound to EWSR1::FLI1 in shA673-1c cells (n = 5 independent experiments). Data shown as mean % of input ± SD. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs. control (unpaired two-tailed Student’s t-test). i Schematic of R-Luc-8MS2 and R-Luc-0MS2 reporter mRNAs. j Domain structure of EWSR1::FLI1. k R-Luc-8MS2 mRNA stability in HeLa cells co-transfected with MS2-CP-tagged constructs and treated with ActD for 0–4 h (n = 5 independent experiments). Mean ± SD **p < 0.01; ns = not significant compared to MS2-CP (unpaired two-tailed Student’s t-test). Source data are provided as a Source Data file. Exact p-values are provided in the Source Data file.

Fig. 2. EWSR1::FLI1 interacts with the CCR4-NOT deadenylation complex.

a Schematic of the Gaussia luciferase protein complementation assay (gPCA). GLucN1 and GLucN2 are inactive fragments of Gaussia luciferase. See Cassonnet et al. for more details. Created in BioRender. Galvan, B. (2025) BioRender.com/0002. b gPCA normalized luminescence ratios (NLR) of interactions between EWSR1::FLI1 and individual subunits of CCR4-NOT and PAN2-PAN3 complexes. Deadenylase subunits were fused to GLucN1; EWSR1::FLI1 to GLucN2. Blue bars denote interacting pairs (NLR > threshold of 3.5, dotted line); gray bars indicate non-interacting pairs. Data represent means ± SD (n = 3 independent experiments). c gPCA NLR values for interactions between CNOT2 and full-length EWSR1::FLI1, EWSR1-Nter and FLI1-Cter. EWSR1::FLI1 constructs are fused to GLucN1 (EWSR1::FLI1) or to GLucN2 (CNOT2). Data are means ± SD (n = 3 independent experiments). d Co-immunoprecipitation (IP) of FLAG-EWSR1::FLI1 with Myc-CNOT2 from HEK293T lysates ± RNAse A treatment. Input lysates and IPs were probed by anti-FLAG and anti-Myc immunoblotting. e Co-IP between endogenous CNOT2 and EWSR1::FLI1 in shA673-1c cells. f Domain architecture of CNOT2 deletion constructs. FL full-length, NAR NOT1 anchor region, CS connecting sequence. Created in BioRender. Galvan, B. (2025) BioRender.com/0003. g Co-IP of FLAG-EWSR1::FLI1 with Myc-tagged CNOT2 truncation mutants in HEK293T cells. Arrows denote expected bands; asterisk indicates non-specific signal. h Structure model of the CCR4-NOT complex. NOT module subunits (green), catalytic module subunits (red), CNOT1 scaffold (light grey), and accessory subunits (dark grey) are indicated. Alternate catalytic module configurations are shown (right box). Created in BioRender. Galvan, B. (2025) BioRender.com/0004. i, j Co-IP between endogenous CNOT1 (i) or CNOT8 (j) and EWSR1::FLI1 in shA673-1c cells. k Co-IP of FLAG-EWSR1::FLI1 with HA-tagged CCR4-NOT subunits in HEK293T cells. l Schematic of the EWSR1::FLI1-associated CCR4-NOT module. Created in BioRender. Galvan, B. (2025) BioRender.com/0005. All IPs in Fig. 2 were replicated ≥ 2 times. Source data are provided as a Source Data file.

Fig. 3. EWSR1::FLI1 mRNA decay activity depends on CNOT2 and supports Ewing sarcomagenesis.

a mRNA stability of the R-Luc-8MS2 mRNA reporter in HeLa cells transfected with control (siCTL, blue) or CNOT2-targeting (siCNOT2, grey) siRNA, and expressing either EWSR1::FLI1-MS2-CP (left) or MS2-CP (right). Data are means ± SD (n = 4 independent experiments). **p < 0.01; ns = not significant (Student’s t-test, two-tailed, unpaired). HL half-life. b Heatmap showing HL ratios for decay panel mRNAs in shA673-1c cells following CNOT2 knockdown. HL ratio = HL (siCTL) / HL (siCNOT2). Data are means from 3-5 independent experiments. c Schematic of EWSR1-Nter-deletion mutants. FETBM1: FET binding module (as described in Linden et al.). d gPCA NLR values for interactions between CNOT2 and EWSR1-Nter mutants. Constructs are fused to GLucN1 (EWSR1-Nter deletion mutants) or GLucN2 (CNOT2). Data are means ± SD (n = 3 independent experiments; n = 2 for Δ63EWSR1-Nter). Positive (> NLR cut-off of 3.5, dotted line) and negative interactions are in blue and gray, respectively. e R-Luc-8MS2 reporter stability in HeLa cells co-transfected with ∆63EWSR1-Nter-MS2-CP (blue) or MS2-CP control (grey). Data are means ± SD (n = 5 independent experiments). ns = not significant (Student’s t-test, two-tailed, unpaired). f Heatmap showing HL ratios of decay panel mRNAs upon expression of FLAG-tagged EWSR1::FLI1 or ∆63EWSR1::FLI1, or control FLAG empty vector (CTL) in hMSCs. HL ratio = HL (EWSR1::FLI1 construct) / HL (CTL). Data are means (n ≥ 2 independent experiments). g Quantifications of soft agar colony formation in A673 shEF3’.11 cells after endogenous EWSR1::FLI1 knockdown ( + dox) and rescue with FLAG empty vector (blue), FLAG-EWSR1::FLI1 (red), or FLAG-∆63EWSR1::FLI1 (yellow). Control: cells rescued with the empty vector before knockdown (-dox, gray). Data are means ± SD (n = 3 independent experiments). **p < 0.01, *p < 0.05, ns = not significant (Student’s t-test, two-tailed, paired). h Tumor volumes from A673 shEF3’.11 xenografts following endogenous EWSR1::FLI1 knockdown ( + dox) and rescue with FLAG empty vector (blue), FLAG-EWSR1::FLI1 (red), or FLAG-∆63EWSR1::FLI1 (yellow). Data are means ± SEM from 8 mice (Empty vector, ∆63EWSR1::FLI1) or 7 mice (EWSR1::FLI1). P = 0.0014 (Mann-Whitney test, one-tailed); ns = not significant. Source data are provided as a Source Data file. Exact p-values are provided in the Source Data file.

Fig. 4. Transcriptional, splicing and decay activities of EWSR1::FLI1 are uncoupled.

a Schematic of EWSR1-Nter tyrosine mutants. Degenerate hexapeptide repeats (DHRs) are indicated by vertical bars. Tyrosines outside DHRs are in black. b gPCA NLR values for CNOT2 interactions between EWSR1-Nter and tyrosine mutants. Constructs were fused to GLucN1 (EWSR1-Nter) or GLucN2 (CNOT2). Positive ( > NLR cut-off of 3.5, dotted line) and negative interactions are in blue and gray, respectively. Means ± SD (n = 3 independent experiments). c R-Luc-8MS2 mRNA stability in HeLa cells expressing SIGQQS-MS2-CP (green) or MS2-CP (grey). Data are means ± SD (n = 3 independent experiments). *p < 0.05 (Student’s t-test, unpaired, two-tailed). HL = half-life. d Overlaps of decay targets (blue) with EWSR1::FLI1-regulated (red), bound (green) or both (yellow). ns = not significant (one-sided Fisher’s Exact Test). e Schematic of EWSR1::FLI1 and mutants. f gPCA NLR for CNOT2 interactions with wild-type or transcription-deficient EWSR1::FLI1 mutants. Constructs are fused to GLucN1 (EWSR1::FLI1) or GLucN2 (CNOT2). Mean ± SD (n = 3 independent experiments). g Heatmap of decay panel mRNA HL ratios in hMSCs expressing FLAG-tagged EWSR1::FLI1 or ∆ETS-EWSR1::FLI1, or control FLAG empty vector (CTL). HL ratio = HL (EWSR1::FLI1 construct) / HL (CTL). (n ≥ 2 independent experiments). h Luciferase transactivation assay using a 12xGGAA F-Luc reporter and R-Luc normalization. Fold induction relative to empty vector. Data are means ± SD (n = 3 independent experiments). i b-isox precipitation of FLAG-tagged wild-type or mutant EWSR1::FLI1 in HeLa cells. Densitometry normalized to untreated supernatant. WCE whole cell extract. j Number and Brightness (N&B) assay in U2OS cells expressing EGFP-tagged constructs. Nuclear molecular brightness (ε) plotted relative to GR-N525. Box plots show medians and 5^th-95th percentiles (n = 2–4 independent experiments). k Overlap between EWSR1::FLI1 decay and splicing targets. ns = not significant (Fisher’s Exact Test). l SMN2 minigene assay in HeLa cells co-transfected with FLAG- or MS2-CP-tagged constructs. Data are means ± SD (n = 3 independent experiments). ∆PSI denotes change in exon 7 inclusion relative to control. ****p < 0.001; *p < 0.05; ns: not significant (one-sample or two-tailed unpaired Student’s t-test). Source data are provided as a Source Data file. Exact p-values are provided in the Source Data file.

Fig. 5. EWSR1::FLI1 co-opts HuR via its CTAD to redirect HuR-associated mRNAs for degradation.

a Enrichment of RBP motifs in 3’UTRs of EWSR1::FLI1 decay targets, using non-target 3’UTRs as control. AREBPs (dark blue), non-AREBPs (light blue) and HuR (red arrow) are shown (padj < 0.05). Adjusted P-values (Bonferroni method) were obtained using the MEME-suite. b Overlap between EWSR1::FLI1 decay targets and HuR-bound transcripts from (Mukherjee et al.) (one-sided Fisher’s Exact Test). c, d Co-Immunoprecipitation (IP) between endogenous HuR and EWSR1::FLI1 in shA673-1c cells. e Quantification of PLA signal in cells co-transfected with Myc-HuR together with a FLAG empty vector (nucleus = 284 cells, cytoplasm = 267 cells) or FLAG-EWSR1::FLI1 (nucleus = 135 cells, cytoplasm = 113 cells). ***p < 0.001 (two-way ANOVA). f, g Co-immunoprecipitation (IP: FLAG) between (f) FLAG-tagged EWSR1::FLI1, EWSR1::FLI1∆CTAD, or the empty FLAG vector and Myc-HuR or (g) between FLAG-EWSR1::FLI1 and the indicated mCherry-tagged HuR constructs. h Molecular dynamics model of the EWSR1::FLI1/HuR complex. i HL ratios of decay panel mRNAs after DHTS vs. DMSO treatments (10 µM, 1h) in shA673-1c cells. HL ratio = HL (DHTS) / HL (DMSO). Data are means (n = 4 independent experiments). j RIP analysis of decay panel mRNAs bound to EWSR1::FLI1 in shA673-1c cells after DHTS vs. DMSO treatments (10 µM, 1h). RIP enrichment is shown as fold-change relative to NDUFA12 and RPL32 control mRNAs. Data are means ± SD (n = 3 independent experiments). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 (Student’s t-test, two-tailed, unpaired). k Co-IP between endogenous CNOT2 and HuR in shA673-1c cells ± dox. l HL ratios of decay panel mRNAs after HuR knockdown (shHuR) vs. control (shCTL). HL ratio = HL (shHuR) / HL (shCTL). Data are means (n = 3-4 independent experiments). m Schematic of R-Luc-8AU and R-Luc-0AU reporter mRNAs. n HuR-reporter assay using ARE-dependent R-Luc constructs (R-Luc-8AU vs. R-Luc-0AU) and F-Luc normalization in EWSR1::FLI1^high and EWSR1::FLI1^low in shA673-1c cells transfected with mCherry-HuR or control vector. Fold induction relative to empty vector (R-Luc/F-Luc ratio normalized to control). Data are means ± SD (n = 3 independent experiments). **p < 0.01; ns = not significant (Student’s t-test, two-tailed, unpaired). Source data are provided as a Source Data file. Exact p-values are provided in the Source Data file.

Fig. 6. The decay function of EWSR1::FLI1 supports oncogenic transformation in EwS and unravels a vulnerability towards HuR inhibition.

a IC₅₀ of DHTS in EwS (red) and non-EwS (blue) cell lines. Medians are shown (solid line). n = 3 independent experiments. *p < 0.05 (Student’s t-test, unpaired, two-tailed, Welch’s correction). b, c IC₅₀ of DHTS in (b) EwS cell lines before (EWSR1::FLI1^high) and after (EWSR1::FLI1^low) EWSR1::FLI1 knockdown and (c) hMSCs +/- EWSR1::FLI1 expression. Data are means ± SD (n = 3 independent experiments). *p < 0.05 (Student’s t-test, two-tailed, unpaired). d IC₅₀ of DHTS in hMSCs expressing FLAG-tagged full-length or ∆63EWSR1::FLI1. Data are means ± SD (n = 3 independent experiments). *p < 0.05 (Student’s t-test, two-tailed, unpaired). e Soft agar colony formation in shA673-1c cells treated with DMSO or DHTS. Data are means ± SD (n = 4 independent experiments). *p < 0.05 (Student’s t-test, two-tailed, paired). Representative images are shown. f Spheroid assay images at 0, 48 and 96 h after DHTS treatment of shA673-1c cells. Scale = 200 µm. g Spheroid volume over time, normalized to t = 0 h. (n = 5–6 independent replicates). *p < 0.05, **p < 0.01, ****p < 0.0001, ns = not significant (Student’s t-test, unpaired, two-tailed). h Wound closure assay for shA673-1c cells treated with DHTS, DMSO or untreated (NT). Data are means ± SD (n = 3 independent experiments). **p < 0.01. (Student’s t-test, two-tailed, unpaired). i Spheroid assay images in shA673-1c cells at 0 and 72 h post-treatment with DHTS or DMSO. Scale = 400 µm. j Soft agar colony formation in shA673-1c cells transduced with shCTL or shHuR. Data are means ± SD (n = 4 independent experiments). ****p < 0.0001 (Student’s t-test, two-tailed, unpaired). k Spheroid volume over time in shA673-1c cells transduced with shCTL or shHuR, relative to t = 0 h. Data are mean (line) (n = 4 independent experiments). *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant (Student’s t-test, two-tailed, unpaired). l Kaplan-Meier survival of 166 EwS patients stratified by HuR expression (cut-off: best percentile, log-rank test). Source data are provided as a Source Data file. Exact p-values are provided in the Source Data file.