Oncofusion

start / graph / mechanism-flt3-is-the-b-lineage-heavier-dependency-in-kmt2a-r

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q60-cellstate2

FLT3 is the second largest KMT2A-selective dependency and it is HEAVIER in B-lineage than in myeloid cells

also called FLT3 lineage split · FLT3 in KMT2A-rearranged B-ALL

FLT3 is usually discussed as a target in myeloid leukaemia. In this project's own measurement the leukaemias that need it most among KMT2A-rearranged lines are the B-cell ones.

The same thing, for a scientist

In DepMap fetched for this beat, FLT3 has a KMT2A-rearranged-minus-all-lines difference of -0.817, second only to MEN1 (-0.901) across the twenty genes tested, and the dependency is LARGER in the B-lineage half (-1.142) than in the myeloid half (-0.859) [E1]. That inverts the default framing. The evidence class history for FLT3 is also unusually clean and worth naming: high expression was identified first by gene-expression-based classification and the requirement was then tested separately [E2], and the cytotoxicity was subsequently confirmed in PRIMARY MLL-rearranged infant ALL rather than only in lines [E3].

Go deeper
Why the same-fusion contrast must not be used here

This beat's contrast C — KMT2A::AFF1 in a myeloid cell (MV4-11) against KMT2A::AFF1 in B-lineage cells — gives -0.525 for FLT3, which would read as a large myeloid excess. It is uninterpretable for this gene, because MV4-11 carries an FLT3 internal tandem duplication (Cellosaurus CVCL_0064, from PMID 12529668) 4. The single myeloid line differs from the comparator in genotype as well as in cell type, and FLT3-ITD lines are strongly FLT3-dependent for reasons that have nothing to do with lineage.

Recording this is the point: the contrast that works for MEN1 and DOT1L fails for exactly the gene it appears most dramatic on.

Expression is not dependency, and here the order was right

FLT3 is the one gene in this worklist where a class-1 expression observation LED to a hypothesis that was then tested as class 2 and class 3 evidence, in that order 23. Elsewhere in this graph the order is often reversed or collapsed. A 2026 report supplies the mechanism for the expression itself: MBNL1 uses a structured single-stranded distal DNA element to sustain FLT3 expression in KMT2A-rearranged leukaemias 5.

And a cell-STATE absence hiding inside a large count

The bounded cell-state query for FLT3 returns 617 records, the largest in the worklist by two orders of magnitude, and essentially none of them is a cell-state mechanism. They are measurable-residual-disease monitoring of FLT3-ITD and post-transplant relapse prevention: FLT3 as a measurement of a disease state rather than a maintainer of a cell state 6. A large count is not an answer, and this is the beat's clearest example.

Facts and figures
genegene-flt3
uniprotP36888
evidence class2
cell type
{
  "expressed_in": "LMPP and early progenitors (CD135); high in KMT2A-rearranged leukaemia",
  "required_in": "KMT2A-rearranged lines of both lineages, more heavily B-lineage",
  "primary_material": "cytotoxic in primary MLL-rearranged infant ALL"
}
cell state
{
  "maintains": null,
  "bounded_absence": true,
  "note": "617 records, dominated by residual-disease monitoring"
}
depmap
{
  "pct_all_lines_depleted": 0.9,
  "mean_all_lines": -0.145,
  "mean_kmt2ar": -0.962,
  "contrast_A": -0.817,
  "myeloid": -0.859,
  "blineage": -1.142,
  "contrast_C_uninterpretable_because": "MV4-11 is FLT3-ITD"
}
How this could bite us back
  • FLT3-ITD is a resistance-generating target: kinase-domain mutations and the KIT/lactate route are both on record, and an infant B-ALL population has no established salvage after FLT3 inhibition fails.
  • FLT3 is expressed on normal LMPPs and early progenitors, so inhibition is not blast-selective and carries a myelosuppression cost.
  • Reading this beat's contrast C for FLT3 would produce a confidently wrong myeloid-selectivity claim; the genotype confound has to travel with the number.
Still open
  1. Is the B-lineage-heavier FLT3 dependency reproducible with more B-lineage KMT2A-rearranged lines? Four is not many, and KOPN-8 is missing from this release.
  2. Does FLT3 maintain any cell state, or is it purely a growth-signal requirement? Nothing in 617 records answers this.
  3. If FLT3 dependency is heavier in B-lineage, why did FLT3 inhibition disappoint clinically in infant ALL more than in FLT3-ITD AML?
Every source behind this entry (6)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    In DepMap fetched 2026-09-20, FLT3 has a KMT2A-rearranged-minus-all-lines gene-effect difference of -0.817, and within the KMT2A-rearranged panel the B-lineage half is more dependent (-1.142) than the myeloid half (-0.859).

    What this evidence cannot show

    Seven myeloid against four B-lineage lines; one line moves a mean. The myeloid half includes MV4-11, which is FLT3-ITD and therefore biases the myeloid half TOWARDS dependency, making the B-lineage excess a conservative reading. Cancer lines in culture.

    Evidence that argues against this (1)
    • THE MV4-11 CONFOUND IS THE WHOLE OBJECTION AND THIS ENTRY STATES IT IN A DIFFERENT ITEM. MV4-11 carries an FLT3 internal tandem duplication as well as KMT2A::AFF1, so a KMT2A-rearranged-minus-all-lines FLT3 difference of -0.817 is partly a measurement of ITD-driven addiction in one line rather than of a fusion-linked dependency. This project corrected a lineage analysis once already for exactly this reason - MV4-11 had to be added to the myeloid set - and the correct comparison excludes ITD-positive lines or reports the difference with and without them. Oncofusion knowledge graph (internal entry) 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    FLT3 was nominated as a therapeutic target in MLL-rearranged leukaemia by gene-expression-based classification and its inhibition was then validated functionally.

    PMID 12620411DOI
    What this evidence cannot show

    2003. The functional arm is murine and cell-line based. Abstract read in this beat. Brief rule 73: the classification cohort includes both lineages and the validation is not infant-specific.

    Evidence that argues against this (1)
    • A 2002 EXPRESSION-BASED NOMINATION IS A HYPOTHESIS, AND THE CLINICAL TEST OF IT DISAPPOINTED. Classifying MLL-rearranged leukaemia as a distinct entity and nominating FLT3 from expression data is target generation, not validation; FLT3 inhibition added to chemotherapy in KMT2A-rearranged infant ALL has since been trialled without transforming outcomes, which this graph records separately. An entry arguing FLT3 is the heavier B-lineage dependency has to carry that. Oncofusion knowledge graph (internal entry) 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    FLT3 inhibition is cytotoxic in PRIMARY MLL-gene-rearranged infant acute lymphoblastic leukaemia samples.

    PMID 15956279DOI
    What this evidence cannot show

    Primary human infant ALL material, which is the index cell type and unusually close to it for this beat. 2005; ex vivo cytotoxicity, not a clinical outcome. Abstract read in this beat.

    Evidence that argues against this (1)
    • Clinical FLT3 inhibition in infant KMT2A-rearranged ALL has not reproduced the ex vivo promise, and this graph's clinical entries carry the trial record; an ex vivo cytotoxicity is not an efficacy claim. Armstrong SA et al. 2003
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    MV4-11 carries an FLT3 internal tandem duplication in addition to KMT2A::AFF1.

    Cellosaurus cell line MV4-11 (CVCL_0064) · Cellosaurus, SIB Swiss Institute of Bioinformatics 2026
    What this evidence cannot show

    Cellosaurus annotation sourced to PMID 12529668; read back live 2026-09-20. Zygosity unspecified.

    Evidence that argues against this (1)
    • THIS ITEM IS THE CONFOUND, NOT A SUPPORTING FACT. Recording that MV4-11 carries an FLT3-ITD alongside KMT2A::AFF1 is stated here as context, but it is the single strongest reason to distrust the DepMap difference in E1. It should be read as a limitation of the entry's headline rather than as background. Oncofusion knowledge graph (internal entry) 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    MBNL1 hijacks a structured single-stranded distal DNA element to sustain FLT3 expression in KMT2A-rearranged leukaemias.

    PMID 42566532DOI
    What this evidence cannot show

    2026 report; abstract read in this beat, full text not opened. A mechanism for the expression, not for the requirement.

    Evidence that argues against this (1)
    • ONE MECHANISM PAPER, ONE GROUP, AND A STRUCTURAL CLAIM THAT IS HARD TO VERIFY INDEPENDENTLY. MBNL1 hijacking a structured single-stranded distal DNA element to sustain FLT3 is a specific and interesting mechanism; it is also unreplicated, and single-stranded-DNA-element claims depend heavily on the probing chemistry used. It does not bear on whether the dependency is B-lineage-heavier, which is this entry's claim. Oncofusion knowledge graph (internal entry) 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    A bounded PubMed query pairing FLT3 and its aliases with an eighteen-term cell-state clause and a leukaemia clause returns 617 records, dominated by measurable-residual-disease monitoring rather than by cell-state mechanism.

    NCBI PubMed E-utilities esearch, db=pubmed, queried 2026-09-20 · NCBI Entrez Programming Utilities 2026
    What this evidence cannot show

    The dominance was assessed by reading the forty most recent titles, not all 617, so the characterisation is a sample rather than a census.

    Evidence that argues against this (1)
    • A BOUNDED QUERY RETURNING LITTLE IS BOUNDED BY ABSTRACTS (rule 80), AND CELL-STATE LANGUAGE IS EXACTLY WHAT ABSTRACTS OMIT. Eighteen cell-state terms against FLT3 and a leukaemia clause will miss a flow-cytometry panel in a supplementary figure, which is where a state association would live. The absence should be read as 'nobody has written this in an abstract', not as 'no state association exists'. Oncofusion knowledge graph (internal entry) 2026
Connected to