WDR5 maintains a differentiation block — in CEBPA p30 AML, not in KMT2A-rearranged disease, and a title check nearly buried it
also called Wdr5 differentiation block · OICR-9429 cell state
WDR5 holds a particular kind of leukaemia cell in an immature state: take WDR5 away and the cells start maturing again. That result is in a different genetic subtype from the one this project studies, and it was invisible to a search that only read titles.
The same thing, for a scientist
In murine CEBPA p30 AML models, the p30-dependent increase in self-renewal and inhibition of myeloid differentiation REQUIRE Wdr5: inducible Wdr5 knockdown restores myeloid surface markers and terminal granulocytic maturation, and the small-molecule Wdr5-MLL antagonist OICR-9429 reproduces the effect [E1]. The system is CEBPA N-terminal mutant AML, not KMT2A-rearranged leukaemia, and saying so in the claim rather than the caveat is brief rule 73. In DepMap, WDR5 is pan-essential (98.3% of 1,258 lines depleted) with the KMT2A-rearranged arm marginally LESS dependent than average (+0.081), so the dependency screen contributes nothing to the cell-context question [E2].
Rule 80 changed this entry's answer, and the mechanism is worth recording
The bounded cell-state query for WDR5 returned four records and a title-only check scored all four off-topic: an embryonic-stem-cell condensate paper, a KMT2A review, a study of WDR5 in porcine preimplantation embryos, and this one, whose title is 'Pharmacological targeting of the Wdr5-MLL interaction in C/EBPα N-terminal leukemia' — a title that names an interaction and a genotype and no state at all.
Opening the full text (PMC4511833, 72,114 characters, 2026-09-20) found a results section headed 'The p30-mediated differentiation block is Wdr5-dependent', nine occurrences of self-renewal and four of differentiation block 1. The absence was withdrawn.
This is the second time in this single beat that a title check discarded a real cell-state result; the other was MLLT10. A title check is not a substitute for reading, and for cell-state questions specifically it fails often, because a state is usually shown in a figure panel and named in a results heading rather than in a title.
What survives and what does not
Survives: WDR5 has never been placed in a defined cell state in KMT2A-rearranged disease. Withdrawn: the broader claim that WDR5 has no cell-state function on record. The distinction is the whole value of the correction, and it is the same shape as the H3K27me3 narrowing this graph performed twice in earlier cycles.
A rule-66 datum from the same paper
The paper tests MLL1, MLL2, MLL3 and 'MLL4/Wbp7'. WBP7 is a UniProt gene synonym of KMT2B (Q9UMN6), not of KMT2D, so 'MLL4' in this paper means KMT2B. Only Mll1 knockdown, and to a lesser extent Mll3, phenocopied Wdr5 loss. Recorded here so the next reader does not resolve it the other way.
| gene | gene-wdr5 |
|---|---|
| uniprot | P61964 |
| evidence class | 3 |
| rule 80 outcome | absence withdrawn after opening PMC4511833 |
| chemical probe | OICR-9429 (Wdr5-MLL antagonist) |
cell type
{
"required_in": "all screened lines (floor-compressed)",
"state_result_system": "murine CEBPA p30 AML"
}cell state
{
"maintains": [
"cell-state-differentiation-block",
"cell-state-stem-like-self-renewal"
],
"context": "CEBPA p30 AML, not KMT2A-rearranged"
}depmap
{
"pct_all_lines_depleted": 98.3,
"mean_all_lines": -1.161,
"mean_kmt2ar": -1.08,
"floor_compressed": true
}- WDR5 is pan-essential in 98.3% of screened lines, so the therapeutic argument has to rest entirely on the WIN-site interface rather than on the protein, and an interface antagonist's selectivity is a chemistry claim that needs its own evidence.
- Releasing a differentiation block is not automatically good: differentiation therapy in this disease carries differentiation-syndrome risk, already recorded against the menin inhibitors.
- Transferring the CEBPA p30 result to KMT2A-rearranged disease is precisely the generalisation brief rule 73 exists to prevent.
- Does WDR5 maintain the differentiation block in KMT2A-rearranged cells as it does in CEBPA p30 cells? The experiment has not been reported and OICR-9429 makes it cheap.
- How many of this graph's other cell-state absences were established by a title check alone and would not survive opening the papers?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In murine CEBPA p30 AML models, WDR5 is required to maintain the p30-driven myeloid differentiation block and the associated increase in self-renewal; inducible Wdr5 knockdown restores granulocytic maturation and the Wdr5-MLL antagonist OICR-9429 reproduces this.
Pharmacological targeting of the Wdr5-MLL interaction in C/EBPα N-terminal leukemia. · Grebien F, Vedadi M, et al. · Nature chemical biology 2015What this evidence cannot show
Murine CEBPA N-terminal mutant AML (32D cells and primary murine models) - NOT KMT2A-rearranged, NOT human, NOT B-lineage (brief rule 73). Full text opened and read in this beat via PMC4511833. The OICR-9429 arm is a chemical-probe result and inherits probe-selectivity caveats.
Evidence that argues against this (1)
- The result is in CEBPA p30 AML and the paper's own knockdown series finds that Mll1, not Wdr5's association with the KMT2A fusion, is what phenocopies Wdr5 loss. Transferring it to KMT2A-rearranged disease is unsupported, and the DOT1L precedent shows how a mechanism that is real in MLL-rearranged AML can still fail to produce clinical benefit. Bernt KM et al. 2011
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
In DepMap fetched 2026-09-20, WDR5 is depleted below -0.5 in 98.3% of 1,258 screened lines (mean -1.161) and the KMT2A-rearranged arm is +0.081 relative to all 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
Floor compression: the positive difference is uninterpretable rather than evidence of protection. Cancer lines in culture.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT: perturbing the WDR5 INTERFACE is MLL-selective where deleting the WDR5 GENE is pan-essential, so +0.081 does not license 'the KMT2A-rearranged arm is marginally less dependent'. Cao et al. showed that MM-401, which blocks the MLL1-WDR5 interaction and thereby MLL1 complex assembly without touching other SET1-family methyltransferases, specifically blocks proliferation of MLL cells by inducing cell-cycle arrest, apoptosis and myeloid differentiation, WITHOUT general toxicity to normal bone marrow or non-MLL cells, and produces transcriptome changes similar to MLL1 deletion. The difference is not culture versus animal or short versus long - it is that knocking out a shared scaffold removes six complexes and blocking one interface removes one. This is already the position of `gene-wdr5` in `interactome/` - 'the window is in the interface and not in the protein' - and it had not reached this entry's DepMap reading. Cao F et al. 2014
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
A bounded PubMed query for WDR5 and an eighteen-term cell-state clause scoped to leukaemia returns four records, of which a title-only check scores zero on topic and opening one full text scores one on topic.
NCBI PubMed E-utilities esearch, db=pubmed, queried 2026-09-20 · NCBI Entrez Programming Utilities 2026What this evidence cannot show
The count is a retrieval measurement as of 2026-09-20. The point of the evidence item is the discrepancy between the title check and the full text, not the count.
Evidence that argues against this (1)
- NAMED METHODOLOGICAL DEFEATER for any WDR5 result reported without naming the complex. Alicea-Velazquez et al. showed that the Win motif is necessary for WDR5's interaction with EVERY member of the human SET1 family, and that mutating the Win-WDR5 interface severely disrupts assembly and activity of MLL1 and SETD1A complexes but only modestly disrupts MLL2/4 and SETD1B without significantly altering their in vitro activity - and that in the absence of WDR5, MLL3 interacts with RAD and shows ENHANCED activity. So a WDR5 perturbation can raise one methyltransferase's activity while abolishing another's. The discrepancy this item reports between a title check and a full text is real, but the deeper problem is that neither a title nor an abstract usually states which SET1-family complex a WDR5 result concerns, and this entry's own headline result (CEBPA p30 AML, OICR-9429) does not either. Alicea-Velázquez NL et al. 2016
- part of →WDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away
the cell-context dossier for this gene
floor-compressed at 98.3% of lines; the KMT2A-r arm is marginally LESS dependent than average
- refutes →WDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away
blocking the MLL1-WDR5 INTERFACE with MM-401 is MLL-selective - cell-cycle arrest, apoptosis and myeloid differentiation, sparing normal marrow and non-MLL cells (PMID 24389101) - where deleting the WDR5 GENE is pan-essential, because the scaffold is shared by all six SET1-family complexes