Oncofusion

start / graph / mechanism-clec12a-dependence-nup98-nsd1

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q48-surface-antigens

Clec12a is REQUIRED for NUP98::NSD1 AML pathogenesis - a marker turning out to be a dependency, in a different fusion family, and untested in KMT2A-r

also called Clec12a knockout NUP98-NSD1 · CLEC12A cell-intrinsic dependency · CD371 dependency

CLEC12A was thought of only as a flag on the surface of leukaemia cells - something to aim a drug at, not something the cell needs. In a different kind of leukaemia (driven by the NUP98::NSD1 fusion), deleting it made the cells die and made the disease much less able to take hold in mice, which means it may be a genuine vulnerability and not just a label.

The same thing, for a scientist

Mohanty and colleagues (*Blood Advances*, 2025) report strong CLEC12A overexpression in both NUP98::NSD1 patients and murine AML cells transformed with the NUP98::NSD1 fusion oncogene; CRISPR/Cas9 depletion of Clec12a in NUP98::NSD1+NRAS-G12D-immortalised cells produced **higher apoptosis and lower colony numbers in vitro**, and deletion of Clec12a **significantly reduced leukaemic engraftment and prolonged survival** of the NUP98::NSD1+NRAS-G12D murine model [E1]. **Why this is on the KMT2A beat's agenda.** The PI's standing priority is analogous biology in other fusion families, and this is a clean instance. It also changes the *kind* of claim CLEC12A supports. Everywhere else in this graph CLEC12A is a **marker** - something to be bound by a CAR. Here it is a **dependency**: the cell needs it. A dependency and a marker have completely different consequences. A marker's worst failure mode is antigen-negative escape, which is exactly what defeated KITE-222 [E2]; **a dependency cannot be escaped by losing the antigen, because losing the antigen is the killing mechanism.** If CLEC12A were a dependency in KMT2A-rearranged AML, antigen-loss escape would be self-defeating rather than the expected outcome. **Whether it generalises to KMT2A-r is a bounded, dated, controlled absence: nobody has looked.** See [E3].

Go deeper
What the NUP98::NSD1 experiment does and does not show

Shows. In one genetic background - NUP98::NSD1 plus NRAS-G12D - CRISPR knockout of Clec12a raises apoptosis, lowers colony formation, reduces engraftment and prolongs survival 1.

Does not show:

  • That the requirement is fusion-specific. The model carries a cooperating NRAS-G12D. This graph's own cooperating-lesion beat established that RAS is a dissemination modifier; a phenotype measured in a RAS-mutant background may be a RAS-context phenotype. The abstract gives no Clec12a-knockout arm without NRAS-G12D.
  • A mechanism. Why the cell needs an inhibitory ITIM-bearing lectin receptor is not addressed at abstract level. CLEC12A's known physiological job is to restrain neutrophil activation and NET formation [see gene-clec12a E11]; no obvious route from that to leukaemic colony formation is stated.
  • That it is human. The knockout is murine Clec12a. Human CLEC12A overexpression in NUP98::NSD1 patients is expression data, not a dependency test.
  • Anything about KMT2A. NUP98::NSD1 and KMT2A-r are different fusion families with overlapping but distinct HOXA-driven programmes.
The counted absence, bounded

Claim: as of 2026-09-20, no published study tests whether CLEC12A is required for KMT2A-rearranged leukaemia.

Tool: NCBI PubMed E-utilities esearch (db=pubmed) and Europe PMC REST search.

Queries and counts, run 2026-09-20:

querytoolcount
CLEC12APubMed156 (control)
KMT2A OR MLLPubMed9,229 (control)
CLEC12A AND leukemiaPubMed54 (control)
CLEC12AEurope PMC1,361 (control)
CLEC12A AND (KMT2A OR MLL)PubMed2
CLEC12A AND (KMT2A OR MLL)Europe PMC (full text)142
CLEC12A AND (KMT2A OR MLL) AND (knockout OR knockdown OR CRISPR OR dependency)PubMed0
(CLEC12A OR CLL-1 OR CD371 OR MICL) AND (KMT2A OR MLL) AND (knockout OR knockdown OR CRISPR OR dependency OR essential)PubMed1
CLEC12A AND (KMT2A OR MLL) AND (knockout OR CRISPR OR dependency)Europe PMC (full text)74

Rule 67 controls passed: all four control queries return large non-zero counts through the identical code path, so the instrument returns ones.

Title-checking, because a count is not a result:

  • The single PubMed alias-broadened hit is PMID 36814396, Phenotypic characterization of disease-initiating stem cells in JAK2- or CALR-mutated myeloproliferative neoplasms - myeloproliferative neoplasms, not KMT2A, not a functional test. Excluded.
  • Of the 74 Europe PMC full-text co-mentions, exactly 2 have titles naming CLEC12A or an alias: PMID 40334081 (this entry's own NUP98::NSD1 paper) and PMID 41941266 (the KITE-222 phase 1). Neither is a KMT2A-r CLEC12A functional test.

What this absence is NOT. It is not evidence that CLEC12A is dispensable in KMT2A-r disease. It is a measurement that the experiment has not been reported under these strings on this date. The residual risk is a result inside a body text with a title that names neither CLEC12A nor an alias - the seventy-two unread Europe PMC records are exactly that residual, and they were title-checked, not read. 3

Facts and figures
mechanism classcell-intrinsic genetic dependency (established in NUP98::NSD1, untested in KMT2A-r)
fusion familyNUP98::NSD1
modelNUP98::NSD1 + NRAS-G12D immortalised murine cells; murine transplant
perturbationCRISPR/Cas9 Clec12a knockout
in vitro readoutincreased apoptosis, reduced colony number
in vivo readoutreduced leukaemic engraftment, prolonged survival
absence queries
{
  "date": "2026-09-20",
  "tools": [
    "NCBI PubMed E-utilities esearch db=pubmed",
    "Europe PMC REST search"
  ],
  "controls": {
    "PubMed CLEC12A": 156,
    "PubMed KMT2A OR MLL": 9229,
    "PubMed CLEC12A AND leukemia": 54,
    "EuropePMC CLEC12A": 1361
  },
  "tests": {
    "PubMed CLEC12A AND (KMT2A OR MLL) AND (knockout OR knockdown OR CRISPR OR dependency)": 0,
    "PubMed (CLEC12A OR CLL-1 OR CD371 OR MICL) AND (KMT2A OR MLL) AND (knockout OR knockdown OR CRISPR OR dependency OR essential)": 1,
    "EuropePMC CLEC12A AND (KMT2A OR MLL) AND (knockout OR CRISPR OR dependency)": 74
  },
  "title_checked_exclusions": [
    "36814396 - MPN stem cell phenotyping, not KMT2A, not functional",
    "72 of 74 EuropePMC records have titles naming neither CLEC12A nor an alias"
  ]
}
How this could bite us back
  • **If CLEC12A is a dependency, killing the antigen-positive cells is the wrong experiment.** A dependency argues for small-molecule or degrader disruption of CLEC12A function, not for a CAR that selects for antigen-negative escape - and the two strategies are in tension, because CAR pressure would enrich exactly the CLEC12A-low cells a dependency argument predicts are unfit.
  • **The NRAS confound goes the wrong way for a KMT2A extrapolation.** This graph's cooperating-lesion work places RAS as a dissemination modifier in KMT2A-r disease too, so importing a phenotype measured in an NRAS-G12D background into a KMT2A context imports the confound with it.
  • **Murine Clec12a is not human CLEC12A.** Species differences in C-type lectin receptor ligand repertoires are substantial, and the human dependency has not been tested at all.
  • **Publication-bias floor.** A negative Clec12a knockout in a KMT2A model would be unlikely to be published, so the counted absence [E3] is compatible with the experiment having been done and shelved.
Still open
  1. Is Clec12a required for KMT2A-rearranged leukaemia? The identical CRISPR experiment in a KMT2A::MLLT3 or KMT2A::AFF1 model is the single highest-value follow-up on this beat and, as of 2026-09-20, nobody has published it.
  2. Is the NUP98::NSD1 requirement fusion-specific or NRAS-context-dependent? A Clec12a knockout in a NUP98::NSD1-only background, and in an NRAS-G12D-only background, separates them.
  3. What does CLEC12A actually DO inside a leukaemic cell? An ITIM-bearing inhibitory receptor whose loss increases apoptosis is counter-intuitive and no mechanism is offered.
  4. Does the requirement depend on the receptor's ligand - extracellular NET DNA - and therefore on the marrow microenvironment rather than on the cell alone? That would predict the knockout phenotype disappears in a NET-free culture and is a cheap test.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    CLEC12A is strongly overexpressed in NUP98::NSD1 patients and in murine AML cells transformed with the NUP98::NSD1 fusion oncogene; CRISPR/Cas9 depletion of Clec12a in NUP98::NSD1+NRAS-G12D-immortalised cells gave higher apoptosis and lower colony numbers in vitro than Clec12a-wild-type cells, and deletion of Clec12a significantly reduced leukaemic engraftment and prolonged survival in the NUP98::NSD1+NRAS-G12D murine model.

    Clec12a is required for the pathogenesis of NUP98::NSD1 AML. · Mohanty S, Charles Cano F, Gabdoulline R, Lai CK, Othman B, Sudarsanam H, et al. · Blood Adv 2025
    PMID 40334081DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. **The model carries a cooperating NRAS-G12D**, and the abstract reports no Clec12a-knockout arm in a NUP98::NSD1-only background, so fusion-specificity is not established. Murine Clec12a, not human CLEC12A. Effect sizes, cohort sizes and the number of independent knockout clones are not available at abstract level. The authors' own conclusion is deliberately modest - 'suggest to further explore CLEC12A as a potential target' - and this entry should not be read as stronger than that.

    Evidence that argues against this (1)
    • CLEC12A's established physiological function is as an inhibitory ITIM receptor restraining neutrophil activation and NET formation, which does not obviously predict a cell-intrinsic requirement for colony formation. Either the leukaemic requirement is a distinct, unexplained function, or the knockout phenotype is partly a consequence of the CRISPR/immortalisation context rather than of CLEC12A signalling. The abstract does not distinguish these. Malamud M et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the KITE-222 phase 1 study, CLL-1-positive marrow blasts were nearly completely depleted in expanding patients while CLL-1-negative blasts persisted and total blasts did not fall, and the authors conclude that future studies must address CLL-1 heterogeneity.

    Phase 1 Study of KITE-222, an Autologous CLL-1-Directed CAR T-cell Therapy in Patients with Relapsed/Refractory Acute Myeloid Leukemia. · Daver N, Blachly JS, Ghobadi A, Advani A, Muffly L, Garciaz S, et al. · Clin Cancer Res 2026
    PMID 41941266DOI
    What this evidence cannot show

    Abstract read in full from raw NCBI efetch XML on 2026-09-20; full text not read. n=12 adults with R/R AML. Cited here only to establish the contrast between a marker (escapable by antigen loss) and a dependency (not escapable by antigen loss); the trial itself says nothing about CLEC12A dependency.

    Evidence that argues against this (1)
    • The persistence of CLL-1-negative blasts in R/R AML is itself weak evidence AGAINST CLEC12A being a pan-AML dependency: a cell that survives and proliferates without detectable surface CLL-1 is a cell that does not obviously require it. The NUP98::NSD1 result would then be a fusion-family-specific dependency rather than a general one - which is exactly the interesting version. Xiu Y et al. 2018
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    As of 2026-09-20, no published study tests whether CLEC12A is required for KMT2A-rearranged leukaemia: PubMed returns 0 records for `CLEC12A AND (KMT2A OR MLL) AND (knockout OR knockdown OR CRISPR OR dependency)` and 1 for the alias-broadened form, which title-checks to an unrelated myeloproliferative-neoplasm phenotyping paper; of 74 Europe PMC full-text co-mentions only 2 have titles naming CLEC12A or an alias, and neither is a KMT2A functional test.

    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

    **Bounded counted absence.** Date 2026-09-20; tools NCBI PubMed E-utilities esearch and Europe PMC REST search; exact query strings recorded in the section above and in `data.absence_queries`. Rule 67 positive controls run through the identical code path returned 156, 9,229, 54 and 1,361, so the instrument demonstrably returns ones. `querytranslation` was read back for every PubMed query and no query contained a parenthesised `der(n)` token or an unquoted multi-word phrase. **Residual risk, stated plainly:** 72 Europe PMC full-text co-mentions were title-checked and not read, so a KMT2A-r CLEC12A functional result buried in a body text whose title names neither CLEC12A nor an alias would be missed. The alias set searched was CLEC12A, CLL-1, CD371, MICL; `CLL1` unhyphenated was deliberately excluded because it collides with COLEC10.

    Evidence that argues against this (1)
    • The graph's own bibliography was not exhaustively grepped for this claim, which rule 64 requires. A repo regex sweep on 2026-09-20 found CLEC12A in 4 node files and CD371 in 7, and none of those describes a KMT2A-r knockout - but those eleven files were checked by string match, not read end to end. Q48-surface-antigens 2026
Connected to