Deregulation of DUX4 and ERG in acute lymphoblastic leukemia

Authors

Jinghui Zhang

Charles G. Mullighan

Doi

10.1038/ng.3691

PMID: 27776115 · DOI: 10.1038/ng.3691 · Journal: Nature Genetics (2016)

TL;DR

Zhang, Mullighan and colleagues characterized a subtype of B-progenitor acute lymphoblastic leukemia (BLL, ~7% of B-ALL) defined by IGH–DUX4 rearrangement and concurrent transcriptional deregulation of ERG. In a cohort of 1,913 children, adolescents and young adults profiled by gene-expression arrays, SNP arrays, and whole-genome/exome/transcriptome sequencing, all cases of this subtype harbored DUX4 insertion adjacent to the IGH enhancer, ~56% had focal RAG-mediated ERG deletions, and the majority expressed ERGalt — a novel ERG isoform initiated from a non-canonical first exon (exon 6 alt) bound directly by DUX4. ERGalt acts as a dominant-negative inhibitor of wild-type ERG and is transforming in mouse hematopoietic transplant models, and the subtype is associated with favorable outcome despite frequent IKZF1 alterations. (PMID:27776115)

Cohort & data

Key findings

  • A distinct gene-expression cluster identified 141 of 1,913 (7.6%) B-ALL cases, comprising 5.2% of childhood standard-risk, 9.4% of childhood high-risk, 10.2% of adolescent and 5.4% of adult B-ALL. (PMID:27776115)
  • All sequenced cases of this subtype carried IGH–DUX4 rearrangement, placing DUX4 under the IGH enhancer with C-terminal truncation (e.g., DUX4 E415*, Q334*) and in-frame read-through into the IGH locus. IGH–DUX4 was confirmed in all 6 cases selected for RT-PCR/genomic PCR validation; DUX4 overexpression was confirmed by immunoblotting. (PMID:27776115)
  • Focal ERG deletions were present in 85/153 (55.6%) of subtype cases at 21q22.3, most commonly involving exons 3–7 (N=27) or 3–9 (N=22) of ERG NM_182918.3; breakpoints contained conserved heptamer recombination signal sequences and intervening non-consensus nucleotides indicating RAG-mediated origin. ERG deletions were not seen in other B-ALL or T-ALL cases. (PMID:27776115)
  • Mean 17.5 non-silent sequence mutations per case (range 2–42), with a paucity of structural alterations beyond the defining lesions. (PMID:27776115)
  • Lymphoid transcription-factor lesions in 46.5% of cases (IKZF1 36.7%, PAX5 11.3%); recurrent transcriptional-regulator mutations (21%) included MYC, MYCBP2, MGA, and ZEB2. Activating Ras-pathway mutations in 35.2%, cell-cycle regulators 22.5%, and epigenetic modifiers in 56.3% — most commonly KMT2D, SETD2, ARID2 and NCOR1. (PMID:27776115)
  • A 28 kDa C-terminal ERG protein (“ERGalt”) was detected by immunoblotting in 50/79 (63.2%) tested DUX4/ERG cases, including those without an ERG deletion. ERGalt is initiated from a non-canonical first exon (exon 6 alt) located 197 nt proximal to ERG exon 7 and splices in-frame into exons 7–10 of ERG isoform 1. (PMID:27776115)
  • DUX4 ChIP-seq in NALM-6 and Reh confirmed direct DUX4 binding at ERG exon 6 alt; lentiviral expression of patient-derived truncated DUX4 alleles (E415*, Q334*) in CD34+ cord-blood cells and Reh cells induced expression of ERGalt at transcript and protein levels. Of 45 transcripts using non-canonical first exons that were significantly deregulated in DUX4/ERG ALL, three were directly DUX4-bound: ERG, NSD1, and RNGTT. (PMID:27776115)
  • ERGalt lacks the N-terminal pointed and central regulatory domains but retains the ETS DNA-binding and transactivation domains; in a gpIX reporter assay, ERGalt acted as a competitive inhibitor of wild-type ERG. (PMID:27776115)
  • In vivo transformation: Arf−/− mouse hematopoietic cells transduced with ERGalt + NRASG12D serially replated as lymphoid colonies; transplanted ERGalt cells generated lymphoid-precursor, biphenotypic, or pre-B leukemias (longer latency than wild-type ERG, which generates erythromegakaryoblastic leukemia). (PMID:27776115)
  • Subclonal-to-clonal ERG-deletion evolution: of 46/54 transcriptome-sequenced cases expressing ERGalt, 27 had a clonal ERG deletion; 7 ERGalt-positive cases without a clonal deletion had a subclonal deletion detectable by genomic PCR, and clonal ERG deletions emerged in xenografts and at relapse. (PMID:27776115)

Genes & alterations

  • DUX4 — IGH–DUX4 rearrangement universal in this subtype (100% of sequenced cases); insertion adjacent to the IGH enhancer with C-terminal truncation alleles (e.g. E415*, Q334*) and read-through into IGH. DUX4 directly binds the ERG exon 6 alt promoter region (two binding motifs within a 372 nt region). (PMID:27776115)
  • ERG — focal RAG-mediated 21q22.3 deletions in 55.6% of subtype cases (most commonly exons 3–7 or 3–9 of NM_182918.3); expression of a novel dominant-negative isoform ERGalt initiated from non-canonical exon 6 alt; intron-6 retention; expression of an antisense long non-coding RNA (“ALE”) proximal to the first exon of ERG, restricted to this subtype. (PMID:27776115)
  • IKZF1 — alterations in 36.7% of subtype cases; in this subtype, IKZF1 alterations were not associated with poor outcome (in contrast to other childhood B-ALL subtypes). (PMID:27776115)
  • PAX5 — alterations in 11.3% of subtype cases. (PMID:27776115)
  • MYC, MYCBP2, MGA, ZEB2 — transcriptional-regulator mutations enriched in this subtype relative to 209 other B-ALL and 16 T-ALL cases sequenced as comparators. (PMID:27776115)
  • NRAS (and other Ras-pathway genes) — activating mutations in 35.2% of cases; NRASG12D cooperated with ERGalt to sustain lymphoid colony replating in Arf−/− mouse cells. (PMID:27776115)
  • KMT2D, SETD2, ARID2, NCOR1 — most frequently mutated epigenetic modifiers (epigenetic-modifier mutations in 56.3% of cases overall). (PMID:27776115)
  • NSD1, RNGTT — additional loci where DUX4 binding induces transcription from non-canonical first exons in DUX4/ERG ALL. (PMID:27776115)

Clinical implications

  • Favorable prognosis despite IKZF1 loss: the authors state DUX4/ERG ALL is associated with favorable outcome irrespective of co-occurring IKZF1 deletion (otherwise an adverse marker in B-ALL), which has direct implications for risk stratification. (PMID:27776115)
  • Diagnostic detection: DUX4 rearrangement is invisible to conventional karyotyping and FISH because of the repetitive subtelomeric DUX4 locus. The authors argue that ERG deletion alone is an unreliable surrogate (only 56% of subtype cases) and that transcriptome and/or genome sequencing are required to identify this subtype reliably. Quantitation of DUX4 expression may also flag candidate cases. (PMID:27776115)
  • The authors do not propose any direct targeted-therapy intervention; the clinical implication is risk re-classification rather than treatment selection. (PMID:27776115)

Limitations & open questions

  • Mouse model gap: ERG exon 6 alt is incompletely conserved in mouse, so primary mouse hematopoietic cells could not be used to model induction of ERGalt by DUX4 from the endogenous locus; the authors call out the need for engineered mouse models that account for this. (PMID:27776115)
  • Outcome inference: the favorable-prognosis claim derives from a heterogeneous multi-cohort dataset spanning St. Jude, COG, Alliance/CALGB, ECOG and MD Anderson on different treatment protocols; direct treatment-controlled survival analysis is not presented in the main figures. (PMID:27776115)
  • Cell-of-origin / lineage drift: ERGalt-transplanted mice developed lymphoid-precursor, biphenotypic and pre-B leukemias, leaving the precise B-lineage cell-of-origin unresolved. (PMID:27776115)
  • Ordering of events: the model that DUX4 rearrangement precedes RAG-mediated ERG deletion is supported by subclonal-to-clonal evolution in xenografts and at relapse but not by single-cell lineage tracing. (PMID:27776115)
  • Assay sensitivity for IGH–DUX4: identification of DUX4 rearrangement from genome/RNA sequencing is technically challenging owing to repetitive breakpoints and DUX paralogy; how robustly this detection scales to standard clinical pipelines remains open. (PMID:27776115)

Citations from this paper used in the wiki

  • “Microarray and transcriptome sequencing data identified 141 (7.6%) ALL cases with a distinct gene expression profile … 5.2% of childhood standard risk, 9.4% of childhood high risk, 10.2% of adolescent and 5.4% of adult ALL cases.” (p. 2)
  • “Eighty-five (55.6%) of these cases had focal deletions of ERG at chromosome 21q22.3 … most commonly involved exons 3–7 (N=27) or 3–9 (N=22) of ERG transcript variant 1 (NM_182918.3).” (p. 2)
  • “All cases sequenced exhibited rearrangement of DUX4 to IGH, placing DUX4 under the control of the immunoglobulin heavy chain enhancer.” (p. 3)
  • “We identified a mean of 17.5 non-silent sequence mutations per case (range 2–42) … Alterations of lymphoid transcription factor genes were present in 46.5% of cases (IKZF1 36.7% and PAX5 11.3%).” (p. 3)
  • “ERGalt … retains the DNA-binding ETS and transactivation domains … displayed diminished transactivating activity and acted as a competitive inhibitor of wild type ERG.” (p. 5–6)
  • “DUX4/ERG ALL is associated with favorable outcome, irrespective of the presence of concomitant genetic alterations otherwise associated with poor outcome in other contexts, such as deletion of IKZF1.” (p. 7–8)

This page was processed by crosslinker on 2026-05-14.