EWS::FLI1 expression in human embryonic mesenchymal stem cells leads to transcriptional reprograming, defective DNA damage repair and Ewing sarcoma

Authors

Inmaculada Hernández-Muñoz

Irene Cuervas

Estela Prada

Julian Pulecio

Ramón Gimeno

Evelyn Andrades

Soledad Gómez-González

Pau Berenguer-Molins

Ariadna Acedo-Terrades

Júlia Perera-Bel

Marta Bódalo-Torruella

Lara Nonell

Elena Pérez

Daniela Grases

Caterina Mata

José Yélamos

Yvonne Richaud-Patin

Enrique Vidal

Yasmina Cuartero

François Le Dily

Mariona Suñol

Alejandro Manzanares

Angel Raya

Jaume Mora

Doi

10.1038/s41467-025-64475-y

PMID: 25186949 · DOI: 10.1038/s41467-025-64475-y · Journal: Nature Communications (2025)

TL;DR

Hernández-Muñoz et al. isolated human embryonic mesenchymal stem cells (heMSCs) from heSC-derived experimental teratomas and showed that lentiviral expression of the EWSR1::FLI1 fusion oncogene alone — without cooperating mutations — is sufficient to (i) impose a Ewing sarcoma-like transcriptional program enriched for neural and endothelial markers, (ii) bind preferentially to intronic CA-dinucleotide microsatellites (not GGAA promoter repeats as in transformed A673 cells), (iii) directly upregulate BRCA1 while paradoxically impairing the DNA damage response via defective ATM/ATR phosphorylation, and (iv) generate metastatic Ewing-like tumors in NOD/SCID mice. The authors propose early embryonic mesenchymal progenitors as the cell of origin for Ewing sarcoma and a two-hit model with pubertal hormones as the second hit.

Cohort & data

  • Cell models: 3 independent heMSC lines (isolated from heSC-derived teratomas, ES4 line), 4 hpMSC lines (pediatric bone marrow aspirates from HSJD), and the Ewing sarcoma cell line A673. Additional Ewing lines CHLA25, COG-E-352 (EWS::ERG), and TC205 (EWS::FEV) used as fusion-variant comparators.
  • In vivo: 21-day-old NOD/SCID mice injected intramuscularly (gastrocnemius) with 1×10^6 EWS::FLI1-transduced heMSCs in Matrigel; ~5-month tumor latency; 40% tumor penetrance in EF-heMSC arm vs. 0% in controls (n=7 tumors analyzed).
  • Cancer type: [[ES|Ewing sarcoma]] (OncoTree ES).
  • Assays: RNA-seq on Illumina HiSeq 2500 (GEO GSE272957); ChIP-seq with Flag antibody (GEO GSE272959); 10x Genomics Visium CytAssist spatial transcriptomics on FFPE tumors (GEO GSE272958); ChIP-qPCR; RT-qPCR; western blot; alkaline comet assay; immunohistochemistry; AmpliSeq Childhood Cancer Panel (Illumina, 130-gene targeted DNA panel).
  • Cross-referenced public data: DepMap (Ewing vs. other cell-line expression); MSigDB Ewing signatures (MIYAGAWA, ZHANG, RIGGI, STAEGE); GEO datasets GSE6460, GSE7637, GSE7896, GSE8884, GSE9451, GSE9440, GSE9510, GSE9520, GSE9593, GSE10315, GSE13604, GSE13828, GSE17679, GSE34620, GSE37371, GSE31215 (n=236 sarcoma profiles).

Key findings

  • heMSCs are bona fide early mesenchymal progenitors: CD73+/CD105+/CD90+/HLA-I+, CD45−/CD34−/CD31−, trilineage differentiation competent. Mitochondria are punctate/perinuclear (like heSCs and A673), with elevated UCP2 and UCP3 expression and increased TCA-cycle metabolites compared with hpMSCs — features of an earlier developmental stage than pediatric MSCs.
  • EWS::FLI1 induces an ES-like transcriptome in heMSCs: RNA-seq identified 3,836 DEGs in EF-heMSCs (2,204 up, 1,632 down; p<0.05). GSEA showed significant enrichment for the Ewing signatures of Miyagawa, Zhang, Riggi, and Staege, but not for osteosarcoma or rhabdomyosarcoma signatures. The transcriptome reflects an aberrant endothelial–neural hybrid differentiation program rather than a pure differentiation block.
  • EWS::FLI1 induces known and novel targets: Established targets EZH2, IGF1, NGFR, PADI2, DKK2, NR0B1 and the kinase PRKCB (described by Surdez et al. as essential for Ewing tumor cell survival) are induced; ENC1 and LOX are repressed. Most strongly induced genes are previously undescribed EWS::FLI1 targets specific to Ewing in DepMap.
  • EWS::FLI1 ChIP-seq in heMSCs binds intronic CA microsatellites, not GGAA promoters: 3,086 peaks (2,836 annotated). Mapping: 1,065 peaks in introns, 1,186 in intergenic regions; only ~4% in exons; <0.1% in promoters. De novo motif analysis shows intergenic peaks at canonical GGAA repeats but intronic/promoter peaks at tandem CA dinucleotide repeats (≥10 CA). 85% of bound chromatin is in quiescent (low histone mark) regions.
  • Binding is qualitatively distinct from transformed A673 Ewing cells: Direct peak overlap with A673 Ewing-cell ChIP-seq is virtually nonexistent, yet 37% of heMSC-bound genes are also bound in A673 — with binding shifting from intronic/intergenic sites in heMSCs to 1-kb upstream promoter regions in A673 (22% of shared genes; 16% specifically intron→promoter). Suggests EWS::FLI1 “primes” intronic CA microsatellites in the cell of origin before transitioning to enhancer/promoter occupancy in transformed cells.
  • Only ~5% of DEGs in heMSCs correspond to direct EWS::FLI1-bound genes, indicating most transcriptional reprogramming is indirect and likely mediated by chromatin remodeling, R-loop accumulation, or altered splicing.
  • EWS::FLI1 directly upregulates BRCA1 but impairs the DNA damage response: Strong EWS::FLI1 ChIP-seq peaks land in BRCA1 exons 11 and 15 (confirmed by ChIP-qPCR). EF-heMSCs show elevated BRCA1 mRNA and protein; siRNA against the EWS::FLI1 breakpoint abolishes BRCA1 induction. Despite higher BRCA1 levels, EF-heMSCs harbor significantly increased basal DNA damage (alkaline comet assay) and defective phospho-ATM and phospho-ATR induction after etoposide treatment, with DNA-PKc (PRKDC) activation preserved. IC50 of etoposide is >2-fold lower in EF-heMSCs than in controls.
  • In vivo tumorigenesis: 40% of mice (mean latency ~5 months) developed soft whitish spinal masses, hemorrhagic abdominal/mesothelial masses, and lung/liver/kidney metastases with classic Ewing histology (small round blue cells, “salt and pepper” chromatin, hemorrhage, mitotic figures). No tumors in control-heMSC mice. Tumors stain positive for human HLA, human Ku80, Flag (EWS::FLI1), PRKCB, BRCA1, Ki67, and cyclin A.
  • Spatial transcriptomics confirms ES identity of experimental tumors: Visium CytAssist on primary abdominal and pulmonary metastatic tumors recovered ~70% shared transcriptome between sites, enriched in chromatin organization and mRNA splicing genes. Identifies BCL11B and ITM2A as Ewing-specific markers (overexpressed in Ewing cell lines per DepMap; immunostained in experimental tumors but not in normal thymus/spleen/liver). RING1B and UCP2 (markers of the proposed cell of origin) are retained in tumors.
  • p53–p21–RB1 axis is intact: EWS::FLI1 induces p53, p21, and [[RB1|RB]] in heMSCs (consistent with growth arrest in primary cells), but DNA content remains diploid. NGS with the AmpliSeq Childhood Cancer Panel detected no pathogenic variants in any heMSC line, confirming EWS::FLI1 is the sole genetic event.

Genes & alterations

  • EWSR1::FLI1 — Ewing sarcoma–defining fusion oncogene; expressed via Flag-tagged lentivirus at levels well below A673. Sole genetic event required to induce ES-like transcriptome, BRCA1 dysregulation, DDR defect, and metastatic tumor formation when expressed in heMSCs. Binding mode in cell-of-origin context: intronic CA microsatellites (≥10 dinucleotides), not the GGAA microsatellite enhancers used in established Ewing cells.
  • BRCA1 — Direct EWS::FLI1 transcriptional target via binding to exons 11 and 15. Upregulated in EF-heMSCs and primary Ewing tumors but functionally compromised: reduced phospho-BRCA1, dispersed BRCA1 foci, defective DDR by comet assay. Authors support the “BRCA1-loss-of-function despite elevated expression” model of Ewing chemosensitivity.
  • ATM / ATR — Both kinases are basally hyperphosphorylated in EF-heMSCs but fail to further phosphorylate after etoposide. Authors hypothesize EWS::FLI1-driven phosphatase activity. [[PRKDC|DNA-PKc]] activation is preserved.
  • TP53 / CDKN1A / RB1 — Pathway intact in EF-heMSCs; p53–p21–RB1 induction limits long-term in vitro culture of transduced cells. heMSCs harbor no pathogenic variants by AmpliSeq Childhood Cancer Panel.
  • PRKCB — Strongly induced novel direct target. EWS::FLI1 binds intron 7 in heMSCs and the TSS/intron 2 in A673. PRKCB knockdown via EWS::FLI1 siRNA abolishes induction. Confirms Surdez et al. designation as a Ewing dependency.
  • CD99 — Constitutive Ewing marker; expression unaffected by EWS::FLI1 introduction in heMSCs.
  • EZH2, IGF1, NGFR, PADI2, DKK2, NR0B1 — Established EWS::FLI1 targets, induced upon oncogene expression and validated by RT-qPCR.
  • BCL11B, ITM2A — Newly nominated Ewing-specific markers from spatial transcriptomics of experimental tumors; previously linked to neural/endothelial features of Ewing.
  • [[RNF2|RING1B]] — Identified in prior work by the group as a Ewing cell-of-origin trait; retained in experimental tumors per spatial transcriptomics.
  • UCP2 — Mitochondrial uncoupling protein; high expression marks the stem-cell phenotype of heMSCs and is retained in experimental Ewing tumors.

Clinical implications

  • Therapeutic vulnerability to DNA-damaging agents: EWS::FLI1–induced DDR defect in the cell of origin provides a mechanistic basis for the well-known clinical chemo- and radiosensitivity of Ewing sarcoma. EF-heMSCs are >2-fold more sensitive to etoposide (a frontline Ewing chemotherapeutic) than control heMSCs, supporting continued exploitation of DDR vulnerability in Ewing.
  • Biomarker candidates: BCL11B and ITM2A proposed as Ewing-discriminating markers (DepMap overexpression; positive IHC in experimental Ewing tumors; negative in normal thymus, spleen, liver). PRKCB reaffirmed as a Ewing-restricted dependency consistent with Surdez et al.
  • Disease-origin model: Authors propose a two-hit model — (1) in-utero acquisition of EWS::FLI1 in early heMSCs as the chromatin-priming initiation event, (2) postnatal/pubertal hormone-driven escalation of oncogene activity as the transforming hit. This is consistent with the clinical observation of EWSR1 rearrangements in pediatric hemangiomas that later progress to Ewing during puberty (Mora et al. 2015).
  • No direct treatment/prognosis claim from this paper beyond DDR-targeting; results are mechanistic.

Limitations & open questions

  • EF-heMSCs cannot be maintained in long-term culture (rapid p53–p21–RB1 induction); only acute (48–72 h post-infection) phenotypes and in vivo outgrowth could be assayed.
  • Tumor penetrance was 40% with ~5-month latency, suggesting additional stochastic events still rate-limit tumorigenesis even from the proposed cell of origin. The proposed pubertal/hormonal “second hit” remains experimentally unvalidated.
  • The chromatin binding pattern of EWS::FLI1 in heMSCs (intronic CA microsatellites) is qualitatively different from the GGAA-enhancer pattern in established Ewing lines like A673; mechanism of the intron→promoter transition during transformation is unresolved.
  • Only one heSC line (ES4) was used as the source of teratoma-derived heMSCs; generalizability across genetic backgrounds untested.
  • AmpliSeq Childhood Cancer Panel (130 genes) excludes the bulk of the exome — cryptic cooperating variants outside the panel cannot be excluded.
  • Authors note conflicting prior literature on whether Ewing is “BRCA1-deficient via R-loops” (Gorthi et al. 2018, [PMID:29643511]) or “BRCA1-dependent with functional ATM deficiency” (Menon et al. preprint). Their data support the latter interpretation but the conflict is not adjudicated.
  • The two Ewing genomic-landscape companion papers from 2014 (PMID:24336570, Tirode et al.; and Crompton et al.) showed STAG2 and TP53 as the only recurrent co-mutations beyond EWS-FLI1; this study reinforces that no cooperating mutation is strictly required for transformation when EWS::FLI1 is expressed in the right developmental context.

Citations from this paper used in the wiki

  • “EWS::FLI1 expression in heMSCs … resulting in the acquisition of an ES transcriptome, with the oncogene not preferentially binding to gene promoters, but to intronic and intergenic microsatellites.” (Abstract)
  • “In heMSCs, EWS::FLI1 directly regulates the expression of the DNA repair protein BRCA1, although cells expressing EWS::FLI1 show DNA damage.” (Abstract)
  • “Xenografting of EWS::FLI1-transduced heMSCs results in the formation of tumors expressing characteristic ES markers.” (Abstract)
  • “the preferential binding sequence in intron and promoter regions is the tandem iteration of 10 or more CA dinucleotides” (Results, Fig. 4D)
  • “only a few EWS::FLI1-bound genes displayed significant transcriptional changes in heMSCs, accounting for about 5% of DEGs” (Results)
  • “the IC50 of etoposide is over half-fold lower in EF-heMSCs compared to control heMSCs” (Results, Fig. 5F)
  • “40% of the mice inoculated with EF-heMSCs showed health problems … hemorrhagic masses in the abdominal cavity affecting the mesothelium, the liver, and other organs; and nodules in the lungs” (Results)
  • “we propose early heMSCs as the cell of origin of ES” (Discussion, conclusion)
  • “based on H. Kovar’s study and our own results, we propose a two-step model of ES tumorigenesis, alike to the currently accepted model for acute leukemia” (Discussion)

This page was processed by crosslinker on 2026-05-11.