Oncofusion

start / graph / target-cd65s-sialylated-vim2-ganglioside

Targetstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q48-surface-antigens

CD65s (VIM-2, sialylated ceramide dodecasaccharide) - the myeloid antigen that, with CD15, makes NG2 a perfect predictor of KMT2A rearrangement

also called CD65s · VIM-2 · VIM2 · sialylated CD65 · X3-NeuAc-VII3Fuc-nLc10Cer · VIM-2 dodecasaccharide · VIM-2 decasaccharide · sialyl-CDw65

CD65s is a sugar chain on a fat molecule in the cell membrane, carrying a sialic acid on the end, and no gene makes it. When a leukaemia cell carries it together with CD15, a third marker called NG2 identifies a KMT2A rearrangement with perfect accuracy in the one large adult study that measured it.

The same thing, for a scientist

CD65s is the **sialylated** member of the CDw65 pair: the ganglioside X3-NeuAc-VII3Fuc-nLc10Cer (VIM-2 dodecasaccharide) and its decasaccharide relative, recognised by antibody **VIM-2**, whose binding is abolished by neuraminidase - the opposite response to the neutral CD65 antigen bound by VIM-8 and VIM-11 [E1]. **No gene encodes it.** **Why it is the most decision-relevant non-gene antigen on this beat.** In 313 adult ALL patients, NG2/CSPG4 expression was significantly associated with a CD10(-)/CD34(-)/CD24(-)/**CD65s(+)**/CD15(+)/CD13(-)/CD33(-) phenotype and predicted MLL rearrangement with sensitivity 0.89, specificity 0.89 and PPV 0.93 - and **NG2 predicted MLL rearrangement with 100% accuracy among patients with a CD65s-positive and/or CD15-positive immunophenotype** [E2]. CD65s is not decoration on that result; it is the conditioning variable that turns a good predictor into a perfect one within a subgroup. **And it is a maturation switch, not a lineage flag.** Paietta and colleagues, studying 711 AML patients across seven ECOG trials, open with the observation that **CD65s appears when the progenitor antigen CD34 disappears**, marking a turning point in normal myeloid differentiation; 198 of 711 (28%) were CD65s-low, that group was enriched in FAB M0/M1 and in CD34, CD117 and **terminal transferase**, and CD15 and CD11b were rarely detected in it [E3]. **The KMT2A-rearranged phenotype inverts that relationship and that is the interesting part.** The defining KMT2A-r B-ALL phenotype is CD65s-**positive** *and* CD34-**negative** [E2], which matches the maturation logic - but those same blasts are frequently **TdT-negative** [see `gene-dntt`], where the AML maturation logic pairs TdT with CD34 and with CD65s-*low*. **A KMT2A-rearranged blast wears a combination that normal myeloid maturation does not produce**, and this graph has no entry explaining that. Measured 2026-09-20: CD65s appears in **8** node files with no entry [E4].

Go deeper
The phenotype, stated in full and in the source's own words

Schwartz 2003 reports NG2 expression significantly associated with CD10(-)/CD34(-)/CD24(-)/CD65s(+)/CD15(+)/CD13(-)/CD33(-) in B-precursor ALL 2.

Three things get dropped when this is quoted, and all three matter:

  1. CD34(-). Routinely omitted. It is what makes the CD65s pairing anomalous, because CD65s normally appears as CD34 disappears 3 - so here the two are consistent, while TdT-negativity is not.
  2. CD13(-)/CD33(-). The phenotype is myeloid-antigen-positive for the glycan antigens and myeloid-antigen-negative for the two classic protein myeloid antigens. CD15 and CD65s carry the signal; CD13 and CD33 do not. A panel built on CD13/CD33 alone would miss this phenotype entirely.
  3. The discordances. NG2 was positive in 3 patients with no detectable MLL rearrangement and negative in 8 patients with MLL-AF4 transcripts 2. The 100% accuracy is conditional and the false-negative rate is real.
Why this is a non-gene antigen and what that costs

CD65s is a ganglioside - a sialylated glycan on ceramide. There is no HGNC record for CD65 or CD65s 5. Its biosynthesis requires the same alpha-1,3-fucosylation chemistry as CD15 6 plus a sialylation step that this beat did not resolve to a specific sialyltransferase.

That unresolved step is the honest gap. For CD15 this beat could name FUT4 and FUT9 and hand the graph a gene-keyed bridge. For CD65s it cannot: the ST3Gal family member responsible for the VIM-2 ganglioside in myeloid cells was not established from any source read here, and the candidate enzymes (ST3GAL4, Q11206, 333 aa; ST3GAL6, Q9Y274, 331 aa) were looked up and deliberately not asserted, because no source read on this beat attributes the VIM-2 epitope to either. Recording that as unresolved is the correct outcome, and it is a concrete, bounded task for a future cycle.

Facts and figures
antigen classsialylated glycosphingolipid (ganglioside) epitope
chemical structureX3-NeuAc-VII3Fuc-nLc10Cer (VIM-2 dodecasaccharide) and VIII3-NeuAc-V3Fuc-nLc8Cer (VIM-2 decasaccharide)
defining antibodyVIM-2
neuraminidase effectbinding ABOLISHED by Vibrio cholerae neuraminidase
encoded by a genefalse
why not a genea ganglioside: a sialylated fucosylated lacto-series glycan on ceramide, assembled by glycosyltransferases
kmt2a relevanceCD65s(+) is a component of the phenotype in which NG2 predicts MLL rearrangement with 100% accuracy in adult B-precursor ALL
maturation relationshipappears as CD34 disappears in normal myeloid differentiation (PMID 12886241)
biosynthesis resolved
{
  "fucosylation": "alpha-1,3-fucosyltransferase, FUT4 implicated by correlation (PMID 9376686)",
  "sialylation": "UNRESOLVED - no source read on this beat attributes the VIM-2 epitope to a specific sialyltransferase; ST3GAL4 (Q11206, 333 aa) and ST3GAL6 (Q9Y274, 331 aa) were looked up and deliberately NOT asserted"
}
graph mentions 2026 09 20
{
  "CD65s": 8
}
How this could bite us back
  • **This is a diagnostic marker, not a target.** CD65s is abundant on normal neutrophils and CML granulocytes - more so than on acute leukaemia cells - so a CD65s-directed agent would have inverse selectivity.
  • **The 100% accuracy is conditional and the denominator is unstated.** It applies within a CD65s+/CD15+ subgroup of an adult series, and the same series has 8 NG2-negative MLL-AF4 cases.
  • **Sample handling moves this antigen.** Neuraminidase abolishes VIM-2 binding while enhancing the neutral CD65 antigen's [E1], so a processing difference between laboratories can invert a CD65/CD65s call.
  • **A maturation marker cannot be assumed stable under a differentiating agent.** If menin inhibition drives myeloid differentiation, CD65s and CD15 should move - which makes them unreliable for MRD monitoring during menin-inhibitor therapy, an inference this graph has not tested.
Still open
  1. Which sialyltransferase makes the VIM-2/CD65s epitope in myeloid cells? Naming it would give CD65s the gene-keyed bridge that FUT4 gives CD15, and this beat could not resolve it from any source it read.
  2. Does the CD65s+/CD15+/NG2+ combination retain its accuracy in INFANT KMT2A-rearranged ALL? Every number in this entry comes from a 313-patient adult series, and the index population of this project is infants.
  3. How can a blast be CD65s-positive (a post-CD34 maturation antigen) and TdT-negative and CD34-negative and CD33-negative simultaneously? That combination is not produced by normal myeloid maturation and no entry in this graph explains it.
  4. Do CD65s and CD15 add independent information, or is the '100% accuracy' driven by one of them? The source reports them as 'CD65s-positive and/or CD15-positive', which deliberately does not separate them.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    VIM-2 binding is abolished by Vibrio cholerae neuraminidase and VIM-2 strongly bound the sialylated VIM-2 decasaccharide VIII3-NeuAc-V3Fuc-nLc8Cer, while the asialo form V3FucnLc8Cer was not recognised by VIM-2 but was strongly bound by VIM-8 and VIM-11; the CDw65 cluster was originally assigned to antibodies binding the ganglioside X3-NeuAc-VII3Fuc-nLc10Cer (VIM-2 dodecasaccharide), and VIM-2 and the other CDw65 antibodies represent two different antigen specificities.

    The CDw65 monoclonal antibodies VIM-8 and VIM-11 bind to the neutral glycolipid V3FucnLc8Cer. · Kniep B, Peter-Katalinic J, Muthing J, Majdic O, Pickl WF, Knapp W · J Biochem 1996
    PMID 8830039DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. Three antibodies from a larger workshop cluster; modern anti-CD65/CD65s clones were not tested and their specificity must be checked per reagent. This is a binding study on purified glycolipids, which is the right assay for the question but says nothing about density or accessibility on a live blast.

    Evidence that argues against this (1)
    • If a diagnostic laboratory's 'CD65' reagent is in fact VIM-2-like, then its CD65 result is a CD65s result, and the distinction this entry rests on becomes a labelling artefact in that lab's data rather than a difference in what was measured. Xiu Y et al. 2018
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 313 adult ALL patients, with MLL rearrangement determined by RT-PCR and FISH in 120 B-precursor cases, NG2 expression was significantly associated with a CD10(-)/CD34(-)/CD24(-)/CD65s(+)/CD15(+)/CD13(-)/CD33(-) phenotype with sensitivity 0.89, specificity 0.89 and PPV 0.93 for MLL rearrangement; NG2 predicted MLL rearrangement with 100% accuracy among patients with a CD65s-positive and/or CD15-positive immunophenotype; NG2 was positive in 57% of pro-B ALL, 2% of common ALL and 20% of pre-B ALL and absent in T-ALL and mature B-ALL; and NG2 was positive in 3 patients without detectable MLL rearrangement and negative in 8 patients with MLL-AF4 transcripts.

    PMID 12886247DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. **Adults, n=313 screened with MLL status in 120** - not children and not infants, which is the index population of this project. The 100% accuracy is conditioned on the CD65s+/CD15+ subgroup whose size the abstract does not give; a perfect accuracy in an unstated denominator is the statistic most likely not to replicate. The 8 NG2-negative MLL-AF4 cases bound the claim in the other direction.

    Evidence that argues against this (1)
    • Eight patients with MLL-AF4 transcripts were NG2-negative in the same series, so the marker combination has a substantive false-negative rate. Any diagnostic algorithm that treats a CD65s+/CD15+/NG2+ result as necessary rather than sufficient for KMT2A rearrangement will miss those cases. Xiu Y et al. 2018
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    CD65s appears when the progenitor antigen CD34 disappears, marking a turning point in normal myeloid differentiation; of 711 AML patients on seven ECOG trials, 198 (28%) were CD65s-low, a group more common in FAB M0/M1, with more frequent CD34, CD117 and terminal transferase, less myeloperoxidase, and in which the more mature myeloid antigens CD15 and CD11b were rarely detected (all P < 0.0001); CD65s-low incidence rose with age from 20% under 50 to 67% over 80, and among patients over 55 the CD65s-low group had a lower complete remission rate (33% vs 44%, P = 0.055).

    PMID 12886241DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. Adult AML on trials enrolling 1986-1999; **not paediatric, not B-lineage, and explicitly not KMT2A-stratified** - the abstract states the CD65s-low and CD65s-high groups did not differ in the distribution of cytogenetic prognostic groups. The CR-rate difference in the over-55 group is P = 0.055, i.e. not significant, and the authors describe it as 'intriguing' rather than established. Cited here for the maturation relationship, not for prognosis.

    Evidence that argues against this (1)
    • Because CD65s-low and CD65s-high AML did not differ by cytogenetic group, this paper is evidence that CD65s tracks differentiation stage rather than genotype. That argues CD65s positivity in KMT2A-rearranged B-ALL reflects the blast's maturation state, not a fusion-driven programme - the weaker and more parsimonious reading, and this entry does not claim otherwise. Xiu Y et al. 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    As of 2026-09-20, CD65s appears in 8 of this graph's 1,465 node files and CD65 in 13, with no entry for either; PubMed returns 15 records for the bare token `CD65s` and 9 for `CD65s AND leukemia`.

    Q48 surface-antigen beat: bounded counted-absence sweep, PubMed E-utilities and Europe PMC REST · Q48-surface-antigens · Oncofusion knowledge graph (internal measurement, 2026-09-20) 2026
    What this evidence cannot show

    In-graph counts by Python regex with word boundaries on 2026-09-20; `\bCD65\b` and `\bCD65s\b` are disjoint patterns so the counts do not overlap. The dispatch brief quoted 7 for CD65s; this run gives 8 and the difference is unreconciled - treat as approximate. PubMed counts from E-utilities esearch with `querytranslation` read back; the bare `CD65s` token gets no MeSH expansion, so papers writing 'sialylated CD65', 'VIM-2' or the glycolipid name are missed and the true literature is larger.

    Evidence that argues against this (1)
    • A 15-record PubMed literature for a marker with a 100%-accuracy result attached to it is itself suspicious and suggests the marker is under-searched rather than under-studied: the VIM-2 literature is substantially larger (28 records for `VIM-2 AND (CD65 OR myeloid)`) and is reachable only under the antibody name. Q48-surface-antigens 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    HGNC returns no record for the string CD65 as approved symbol, alias symbol or free-text search, in contrast to CD15 (which returns FUT4) and CD371 (which returns CLEC12A) through the identical calls on the same day.

    HGNC record HGNC:31713 (CLEC12A) · HUGO Gene Nomenclature Committee · HGNC 2026
    What this evidence cannot show

    Measured 2026-09-20 against the live HGNC REST API. **This is a lookup negative with a positive control**: the same two endpoints returned FUT4 for CD15 and CLEC12A for CD371 in the same run, so the empty result is a property of the string and not of a broken call. It bounds today's HGNC indexing, nothing more. The citation attached is the CLEC12A record because it is the positive control; no HGNC record exists for CD65 to cite.

    Evidence that argues against this (1)
    • The honest blank is better than CD15's misleading resolution to FUT4, but both are failures of the same assumption. An instrument that returns nothing for CD65 and something wrong for CD15 is not more trustworthy on CD65 - it is merely failing more visibly. HUGO Gene Nomenclature Committee 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Fuc-TIV (FUT4) mRNA upregulation at the HL-60 promyelocyte stage correlated with upregulation of Lewis-x, VIM-2 and Lewis-y and with transitory downregulation of sialyl-Lewis-x, so the same alpha-1,3-fucosylation chemistry contributes to CD15 and to the VIM-2/CD65s epitope.

    PMID 9376686DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. Myeloid cell lines as maturation proxies, none KMT2A-rearranged; correlation between mRNA and surface antigen, no knockdown. **This establishes the fucosylation step only.** The sialylation step that distinguishes CD65s from CD65 is not addressed by this paper and was not resolved to a specific sialyltransferase by this beat.

    Evidence that argues against this (1)
    • The same FUT4 peak raises VIM-2 while lowering sialyl-Lewis-x, which are both sialylated fucosylated structures. That the two move in opposite directions shows the epitopes are not interchangeable and that reasoning from 'more fucosyltransferase' to 'more CD65s' is not safe. Kadirvelraj R et al. 2023
Connected to