EWSR1::ZNF384 - the same EWSR1 exon 7 fragment that drives Ewing sarcoma, fused instead to a zinc-finger protein, in B-ALL

FULL TEXT READ from an Internet Archive snapshot dated 2009-09-29 of the original HighWire page; aacrjournals.org is Cloudflare-blocked and NO DOI exists for this pre-DOI AACR article. n = 2 cases with t(12;22) in this series. Exon-level breakpoints only, no residue coordinates. NOTE an error in the paper's own Introduction, which reads 'the EWSR1 or TAF15 genes on chromosome 17 and 22, respectively' - inverted relative to its own Results and to the gene loci; the Results assignment used here (t(12;22) = EWSR1, t(12;17) = TAF15) is the correct one. expected_absent_terms names ZNF384 because the gene is called CIZ/NMP4 throughout.

• The same paper's functional results argue against the obvious interpretation of this architecture. EWSR1-CIZ transformed NIH3T3 cells but 'did not show enhanced (or decreased) transactivation' of a CIZ-site reporter and induced none of mmp3, eat2 or manic fringe, while EWSR1-FLI1 controls in the same experiment induced strongly. So the one paper that has compared an EWSR1::ZNF384 fusion and an EWSR1::FLI1 fusion side by side found them behaving DIFFERENTLY on transcriptional output, with the same EWSR1 module - which is a partial answer to the comparison this entry says nobody has run, and it points away from the portable-activation-module model. Martini A et al. 2002

FULL TEXT READ via Europe PMC fullTextXML for PMC8550960. This is a retrospective collection from 16 consortia spanning 1992-2018 with no total B-ALL denominator, so the partner distribution reflects detection methods as much as biology - rearrangements were detected 'by FISH, RT-PCR, and/or NGS as per individual study group criteria', and a FISH-only centre cannot name a partner at all. Partner information was missing for 25 of the 218 patients. No outcome data are reported separately for the EWSR1 cases.

• The 'other' category in this table is heterogeneous and partly an artefact of the detection method. Note in particular that it contains a DUX4 case - DUX4 rearrangement defines its own major B-ALL subtype in every modern classification, so a DUX4::ZNF384 entry raises the question of how the case was assigned, and the paper does not discuss it. A long tail of one-case partners assembled across 16 consortia over 26 years should not be read as a stable partner spectrum. Li JF et al. 2018

support_location is 'this-project' because the 264 figure is an arithmetic result computed inside this graph, not a number printed in any paper. It was computed for EWSR1::FLI1 type 1, and this entry APPLIES it to EWSR1::ZNF384 on the basis that both use EWSR1 exon 7. That application assumes the same transcript and the same exon numbering in Martini 2002 as in the Ensembl model used, which this beat did not verify. The source entry itself records that '264 versus 265 is a convention' because the junction codon is chimeric, and that the EAD boundary at 1-285 is a compositional convention with no structural support - InterPro covers no model over residues 1-225.

• The equivalence claim has a specific weak joint and it should be named: Martini 2002 gives exon numbers for a 2002 transcript model, and this graph's 264 comes from a current Ensembl model. Exon numbering for EWSR1 has been stable in the Ewing literature, but this beat did not check that Martini's exon 7 is the same exon as Ensembl's exon 7 on ENST00000397938. If the numbering differs by one, the fragment lengths differ by roughly the length of an exon and the 'identical module' claim fails. The check is cheap - align Martini's Figure 1B junction sequence to the current transcript - and it was not done here. Martini A et al. 2002

ABSTRACT ONLY, and this is an incidental finding in a paper about BCR-ABL1-like ALL, not a ZNF384 study. 1.1% of 190 is 2 patients, so the Wilson 95% interval runs roughly 0.3-3.7% and this is an order-of-magnitude figure. Detection used anchored multiplex PCR enrichment over 81 ALL-related genes, so the assay could see ZNF384 fusions, but the paper reports no other ZNF384 partner in the cohort - which is surprising given that EP300 and TCF3 are commoner everywhere else, and may indicate incomplete reporting of the ZNF384 group rather than a real partner distribution.

• The most informative thing about this figure is what is missing beside it. This cohort reports EWSR1-ZNF384 and no EP300-ZNF384 or TCF3-ZNF384 at all, while every dedicated ZNF384 series finds those two partners accounting for about three quarters of cases. Either this single-centre Korean cohort has a genuinely different partner distribution, or ZNF384 fusions other than the incidentally-noted one were not reported in a paper whose subject was BCR-ABL1-like ALL. The second is much more likely, and it means this 1.1% is a lower bound on ZNF384 rearrangement in the cohort and says nothing reliable about EWSR1's share. Hirabayashi S et al. 2021

A REVIEW, read in full text via Europe PMC, and deliberately graded WEAK. THIS BEAT COULD NOT TRACE THE 'EXON 7 OF ZNF384' CLAIM TO ANY PRIMARY SOURCE and does not endorse it. It is recorded solely because it is the only statement found that would break ZNF384-side invariance for a canonical in-frame fusion, and because an untraced review sentence is exactly the kind of claim that propagates. Martini 2002, the primary source for this fusion, reports only CIZ exon 2 and CIZ exon 3.

• Every primary source this beat read puts the ZNF384 junction at exon 2 or exon 3: Martini reports exon 2 and exon 3 for EWSR1; Nyquist found the ZNF384 breakpoint in the intron between exons 2 and 3 in both TAF15 cases; Gocho reports exon 3, the first coding exon, for EP300 and states it matches the earlier TAF15 and E2A fusions; McClure found exon 3 in 8 of 9 EP300 cases and exon 2 in the ninth. Against four primary sources converging on exons 2-3, an untraced 'or 7' in a review should be treated as unsupported until somebody finds the case. Gocho Y et al. 2015