Put the same fusion in two cell types and degrade it, rather than cutting the gene
also called fusion degron isogenic cell type experiment
Two experiments in one: grow the identical fusion in a myeloid and a B-cell background, and switch the fusion protein off directly instead of deleting the whole gene. Together they would answer whether cell type changes what the fusion needs.
The same thing, for a scientist
Two instrument failures identified in this beat have the same fix. First, the cell-type-versus-fusion contrast currently rests on ONE myeloid KMT2A::AFF1 line (MV4-11) against three B-lineage ones, and MV4-11 additionally carries FLT3-ITD. Second, whole-gene CRISPR cannot test a 3' fusion partner: AFF1, MLLT1, MLLT3 and MLLT10 all return near-null gene effects because the library cuts the wild-type and fusion alleles together [E1]. An allele-selective degron on the fusion, deployed in an isogenic pair of cell backgrounds, fixes both. **ARM 1 WEAKENED 2026-09-21 (Q68).** The cell-background half of this design is weaker than the non-isogenic caveat allows. Cellosaurus annotates MV4-11 as monocytic, but the paper that established the line's cytokine dependence describes it in its own text as 'the biphenotypic B-myelomonocytic leukemia (MV4-11) cells' (PMID 3500218), distinguishing it in the same sentence from AML-193, 'the monocytic leukemia' line. An experiment whose premise is 'the same fusion in two cell types' needs the two cell types to be two. **This strengthens rather than weakens the case for Arm 2's isogenic knock-in pair, and the argument should now be made from the lineage ambiguity as well as from the FLT3-ITD** - an isogenic pair fixes both at once, and no cell-line pair available today fixes either.
Design
Arm 1, isogenic cell type. Knock a dTAG or auxin degron tag into the endogenous KMT2A::AFF1 fusion allele in MV4-11 (myeloid, CVCL_0064) and in SEM (B-lineage, CVCL_0095). Degrade acutely and profile at 4, 12 and 24 hours.
Arm 2, allele selectivity. In the same lines, compare degron-mediated fusion loss against whole-gene CRISPR of AFF1. The prediction is that the degron is lethal and the CRISPR is not, and that the gap is the size of the instrument error this beat measured 1.
Readout. Viability; the core dependency set (MEN1, DOT1L, PSIP1) by rescue or by co-perturbation; and a state panel — the CD24 pre-leukaemic marker, quiescence by Ki-67/DNA content, and the differentiation-block markers already used in this graph.
Controls. An untagged parental line for the degrader; a degron on a pan-essential gene (RBBP5) as a positive control for the system; and the FLT3-ITD status of MV4-11 recorded as a covariate, since it is the known confound in Arm 1.
What each outcome would mean
If MEN1 and DOT1L dependency are equal in the two backgrounds after fusion degradation, hyp-cell-type-is-set-before-the-fusion-and-the-fusion-does-not-move-it survives and menin inhibition across lineages is licensed. If they differ, the hypothesis dies and the lineage-stratified trial analysis becomes urgent.
If the degron is lethal where CRISPR is not, this graph should re-read every near-null DepMap number for a 3' fusion partner as an artefact, which is at least four entries.
Correction, 2026-09-21 (Q68): MV4-11's lineage is itself contested
The design's Arm 1 rested on MV4-11 being myeloid and SEM being B-lineage. PMID 3500218 (Santoli et al., J Immunol 1987), the founding characterisation of MV4-11's GM-CSF dependence, calls it biphenotypic B-myelomonocytic. Cellosaurus's NCIt code C9163 (childhood acute monocytic leukemia) is curated from depositor and literature and is coarse.
Consequence for the experiment as designed: Arm 1 should either be dropped in favour of Arm 2, or re-run with the immunophenotype of the specific MV4-11 stock measured and reported rather than taken from an annotation - which is the same lesson the U-937 line entry produced in cycle 38 for the comparison arm.
| design | Endogenous dTAG or AID degron knocked into the KMT2A::AFF1 fusion allele in MV4-11 (myeloid) and SEM (B-lineage); acute degradation compared against whole-gene AFF1 CRISPR in the same lines. |
|---|---|
| readout | Viability; MEN1/DOT1L/PSIP1 co-dependency; state panel (CD24, Ki-67/DNA content for quiescence, differentiation-block markers) at 4, 12 and 24 h. |
| controls | Untagged parental line plus degrader; RBBP5 degron as a pan-essential positive control; MV4-11 FLT3-ITD recorded as a covariate; both lines' identity re-verified by STR. |
| cost tier | med |
| timeline | 6-9 months for the knock-ins, 3 months for the profiling |
| tests hypothesis | hyp-cell-type-is-set-before-the-fusion-and-the-fusion-does-not-move-it |
lines
[
{
"name": "MV4-11",
"cvcl": "CVCL_0064",
"depmap": "ACH-000045",
"lineage": "myeloid"
},
{
"name": "SEM",
"cvcl": "CVCL_0095",
"depmap": "ACH-000782",
"lineage": "b-lineage"
}
]- A degron result showing the fusion is essential where the gene is not would not by itself say the fusion is druggable - degradation is not a pharmacology.
- Acute degradation and chronic loss give different answers routinely, so a lethal degron does not predict that a slow-acting drug would work.
- If the two cell backgrounds give the same answer, the experiment licenses cross-lineage dosing of menin inhibitors on n = 2 cell lines, which is a thin basis for a clinical inference.
- Can a degron be knocked into the fusion allele without also tagging the wild-type KMT2A allele? The 5' portion is shared, so the tag has to go at the 3' end, in AFF1 sequence that the wild-type AFF1 allele also carries.
- Is MV4-11's FLT3-ITD dispensable enough that the comparison survives it, or does Arm 1 need an FLT3-ITD-corrected MV4-11?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
In DepMap fetched 2026-09-20 the 3' fusion partners AFF1, MLLT3, MLLT10 and AFF4 are depleted below -0.5 in 0.2%, 0.4%, 0.6% and 0.4% of 1,258 screened lines respectively, in a panel containing four KMT2A::AFF1 lines and four KMT2A::MLLT3 lines.
Open Targets Platform - DepMap CRISPR gene essentiality (depMapEssentiality), queried by Ensembl gene id 2026-09-20 · Open Targets Platform GraphQL API 2026What this evidence cannot show
Whole-gene CRISPR cuts wild-type and fusion alleles together, which is the proposed explanation for the nulls and is itself the thing the experiment would test. Cancer lines in culture.
Evidence that argues against this (1)
- The experiment assumes the whole-gene nulls are instrument artefacts. They may not be: MLLT3 is required in normal human HSCs (PMID 31776511) and in no screened leukaemia line, which is a coherent biological result rather than an artefact, and a degron might simply confirm the null at higher cost. Calvanese V et al. 2019
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
MV4-11 (CVCL_0064, DepMap ACH-000045) carries KMT2A::AFF1 in a monocytic leukaemia and additionally carries an FLT3 internal tandem duplication, while SEM (CVCL_0095, ACH-000782) carries KMT2A::AFF1 in B-cell acute lymphoblastic leukaemia.
Cellosaurus cell lines MV4-11 (CVCL_0064) and SEM (CVCL_0095) · Cellosaurus, SIB Swiss Institute of Bioinformatics 2026What this evidence cannot show
Read back live on 2026-09-20. The two lines are not isogenic - they differ in every other genetic respect as well as in lineage - which is the principal weakness of Arm 1 and the reason a knock-in isogenic pair would be better if slower.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT that weakens Arm 1 further than the non-isogenic caveat already does. Cellosaurus annotates MV4-11 as monocytic, but the paper that established the line's cytokine dependence describes it in its own text as 'the BIPHENOTYPIC B-MYELOMONOCYTIC leukemia (MV4-11) cells', distinguishing it in the same sentence from AML-193, 'the monocytic leukemia' line. An experiment whose design is 'put the same fusion in two cell types' needs the two cell types to be two; if the myeloid arm is a B-myelomonocytic line, the contrast is partly within one mixed phenotype. This strengthens rather than weakens the entry's own argument for an isogenic knock-in pair, and the argument should be made from this rather than only from the FLT3-ITD. Santoli D et al. 1987
an isogenic degron pair would replace the n = 1 against n = 3 look with a test
an endogenous degron on the fusion allele in MV4-11 and SEM
degron against whole-gene CRISPR, to measure the instrument error directly
LINEAGE CORRECTED: MV4-11's founding characterisation calls it 'the biphenotypic B-myelomonocytic leukemia' line (PMID 3500218), so contrast C is a mixed-phenotype line against three B-lineage lines - which is a better argument for the isogenic knock-in pair than the FLT3-ITD is
the same failure the comparison arm found in U-937: a line's lineage and variant are taken from an annotation rather than measured in the stock used. MV4-11 is annotated monocytic and described by its founding paper as biphenotypic B-myelomonocytic