Four ways a WDR5 result can be true and have nothing to do with KMT2A
also called WDR5 attribution rule · reading a WDR5 paper · WDR5 is not a KMT2A readout
WDR5 shows up in hundreds of cancer papers, and it is easy to read any of them as saying something about the leukemia fusion. Four documented facts say otherwise: the pocket is shared by several enzymes, the drugs work through ribosomes, the protein is split between two different complexes, and knocking it out actually hurts leukemias without the fusion slightly more.
The same thing, for a scientist
WDR5 has 611 PubMed records, 1,662 potent ChEMBL activity rows and the largest chemical-matter footprint of any protein in the KMT2A complex - and none of that is evidence about KMT2A fusions. This entry states the four specific reasons, each with a primary source, so that a future cycle reading a WDR5 paper has a checklist rather than an intuition. (1) **The WIN pocket is shared**: KMT2B, SETD1A and SETD1B each carry their own WIN motif competing for it [E1][E2]. (2) **The phenotype of WIN-site inhibition is ribosomal and p53-dependent**, and the authors of that work state explicitly that it displaces the earlier haematopoietic explanation [E3]. (3) **WDR5 is partitioned**: its binding to the NSL complex through a KANSL1 linear motif is mutually exclusive with its MLL/COMPASS binding, so a WIN-site agent redistributes WDR5 rather than removing it from one place [E4]. (4) **The dependency runs the wrong way**: WDR5 is depleted less in KMT2A-rearranged lines than in leukaemia lines generally, a lineage-controlled selectivity of +0.191, the worst of 41 candidates scored in this project [E5]. A fifth practical point: **occupancy and degradation of WDR5 give different answers**, with a VHL-recruiting degrader outperforming occupancy inhibitors in patient-derived MLL-rearranged AML models [E6], so 'WDR5 inhibitor' does not name one experiment.
The checklist
Before treating a WDR5 result as evidence about KMT2A fusions, answer four questions about the paper:
- Which surface? WIN site, MYC-binding site (WBM), NSL-facing surface, or unspecified. Only WIN-site engagement is even nominally a KMT2A-axis event, and it is shared with three other enzymes.
- Which readout? Ribosome protein gene expression, nucleolar stress, p53 targets and MDM4 splicing are the published signature of WIN-site inhibition and are not KMT2A readouts 3.
- Which modality? Occupancy inhibitor, degrader, genetic deletion or domain mutant - these disagree 6.
- Which comparator? KMT2A-rearranged versus KMT2A-wild-type leukaemia, or cancer versus normal. Without the first comparison, nothing in the result is fusion-specific.
A WDR5 paper that answers 'WIN site / ribosome and p53 readouts / occupancy inhibitor / no KMT2A-wild-type comparator' is a paper about WDR5, not about the fusion.
What this does NOT say
It does not say WDR5 is irrelevant to KMT2A-rearranged leukaemia. Genetic deletion of Wdr5 attenuates MLL-AF9 leukaemogenesis in vivo 7, a selective degrader inhibits patient-derived MLL-rearranged AML in mice 6, and resistance to a potent WDR5 inhibitor has been produced by a single target mutation, WDR5 P173L 8 - which is the signature of a target that is being genuinely engaged. The claim is narrower and harder: WDR5 is required in these cells, and nobody has shown that the requirement runs through the fusion protein. Those are different statements and the literature routinely collapses them.
| pubmed records wdr5 | 611 |
|---|---|
| chembl potent rows | 1662 |
| lineage controlled selectivity | 0.191 |
| selectivity rank | 41 of 41 candidate partners (worst) |
surfaces
[ "WIN site (shared with KMT2A, KMT2B, SETD1A, SETD1B)", "WBM/MYC-binding site", "NSL-facing surface engaged by KANSL1/KANSL2" ]
published win inhibitor signature
[ "ribosome protein gene suppression", "nucleolar stress", "p53 activation", "MDM4-inactivating splicing" ]
modalities that disagree
[ "occupancy inhibitor", "VHL-recruiting degrader", "genetic deletion" ]
- A WIN-site inhibitor engages every WIN motif in the cell at once, so any efficacy is unattributable without a separation-of-function control that nobody has built.
- The published mechanism requires functional p53, predicting failure in the TP53-altered relapsed disease where a new agent is most needed.
- Resistance by a single point mutation, WDR5 P173L, has already been demonstrated in vitro for the occupancy class.
- If WDR5 is limiting and partitioned between NSL and COMPASS, a WIN-site agent may cause an NSL gain of function, and NSL/KAT8 activity is itself implicated in this disease - so the intervention could drive the biology it is meant to block.
- Does a WIN-site inhibitor produce its ribosomal/p53 phenotype in KMT2A-wild-type leukaemia cells to the same degree? If yes, the phenotype has nothing to do with KMT2A rearrangement and the disease framing of the whole programme is wrong.
- Does WIN-site inhibition increase NSL-bound WDR5? That single measurement distinguishes COMPASS loss-of-function from NSL gain-of-function and has never been reported.
- Is the nucleolar convergence a coincidence? MLL-AF4 is reported to localise predominantly to the nucleolus and to interfere with ribosome biogenesis, and WIN-site inhibition produces nucleolar stress. Two independent routes to the same compartment is either the most interesting thing in this entry or nothing.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review
WDR5 is the shared WD40 hub of the WRAD module common to all six SET1-family H3K4 methyltransferases, so its WIN pocket is engaged by several catalytic subunits besides KMT2A.
WRAD: enabler of the SET1-family of H3K4 methyltransferases. · Ernst P, Vakoc CR · Brief Funct Genomics 2012What this evidence cannot show
A review, cited for complex composition only, not for any experimental claim.
Evidence that argues against this (1)
- Sharing a module does not mean the subunit is functionally interchangeable between complexes; BOD1L supports a non-catalytic SETD1A-specific function with H3K4me3 unchanged, which shows the shared-core picture hides enzyme-specific requirements. Hoshii T et al. 2024
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
WDR5 binds the KMT2A WIN motif through an arginine pocket occupied by Arg3765, and the same pocket recognises WIN motifs from other SET1-family enzymes preferentially over histone H3.
Structure of WDR5 bound to mixed lineage leukemia protein-1 peptide. · Patel A, Dharmarajan V, Cosgrove MS · J Biol Chem 2008What this evidence cannot show
A peptide co-crystal; it establishes the chemistry of the pocket, not the competition between partners at cellular concentrations.
Evidence that argues against this (1)
- Cryo-EM of MLL1-WRAD on nucleosomes shows variable docking of the whole module, so a single high-resolution peptide contact overstates how deterministic the assembly is in situ. Rahman S et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In MLL-rearranged cancer cells, WDR5 WIN-site inhibition acts by suppressing ribosome protein gene transcription, inducing nucleolar stress and activating p53, with ribosome-subunit attrition, a broad modest translational choke and MDM4-inactivating alternative splicing; the authors state their data displace the earlier haematopoietic-stem-cell-expansion explanation.
Ribosome subunit attrition and activation of the p53-MDM4 axis dominate the response of MLL-rearranged cancer cells to WDR5 WIN site inhibition. · Howard GC, Wang J, Rose KL, et al. · Elife 2024What this evidence cannot show
Cell-line pharmacology; the mechanism requires functional p53 and therefore cannot be assumed in TP53-altered relapsed disease.
Evidence that argues against this (1)
- An independent group reports that MLL-AF4 itself localises predominantly to the nucleolus and interferes with Pol I transcription and ribosome biogenesis, which would make the ribosomal axis a fusion-proximal one rather than an off-axis artefact of the drug. Siemund AL et al. 2022
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
WDR5 is required for NSL complex assembly and promoter targeting, and its NSL and MLL/COMPASS interactions are mutually exclusive, so WDR5 is partitioned between two complexes rather than dedicated to one.
Hijacking a key chromatin modulator creates epigenetic vulnerability for MYC-driven cancer. · Yang Z, Shah K, Busby T, et al. · J Clin Invest 2018What this evidence cannot show
This citation establishes DPY30/WDR5-family hijacking in MYC-driven cancer; the mutual exclusivity of the KANSL1 and MLL WIN-motif interactions is recorded in this graph's hyp-wdr5-partition-nsl-versus-compass entry from structural work not re-fetched here, so the mutual-exclusivity half of this claim should be read at the confidence of that entry rather than of this citation.
Evidence that argues against this (1)
- If WDR5 is limiting and partitioned, a WIN-site inhibitor releases WDR5 into the other pool rather than inactivating it, so the observed phenotype could be an NSL gain-of-function rather than a COMPASS loss-of-function - a possibility no published WIN-site study controls for. Howard GC et al. 2024
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
WDR5 is DepMap common-essential (mean Chronos -1.159, 98.3% of lines depleted) and is depleted LESS in KMT2A-rearranged lines than in leukaemia lines generally, giving a lineage-controlled selectivity of +0.191, the worst of 41 candidate partners scored in this project.
DepMap CRISPR (Chronos) gene effect and essentiality flags via the Open Targets Platform GraphQL API; lineage-controlled selectivity ranking computed by this project · Broad Institute DepMap; Open Targets Platform; analysis run by this project · Database 2026What this evidence cannot show
Restated from an earlier run in this project, not recomputed. Floor compression: a gene depleted in 98.3% of lines has almost no dynamic range in which selectivity could appear, so +0.191 is close to uninterpretable rather than strongly anti-selective. The KMT2A-rearranged arm is myeloid-enriched and the KMT2A::AFF1 subset is 3-4 lines.
Evidence that argues against this (1)
- Genetic Wdr5 deletion in an MLL-AF9 model impairs leukaemogenesis in vivo, so a flat or mildly anti-selective CRISPR profile across a cell-line panel is not evidence that WDR5 loss is tolerated by a KMT2A-rearranged leukaemia in an animal. Liu L et al. 2023
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A selective VHL-recruiting WDR5 degrader outperformed WDR5 occupancy inhibitors on WDR5-regulated transcription, chromatin-bound MLL complex, c-MYC and MLL-rearranged AML patient cells in vivo, in patient-derived mouse models.
A selective WDR5 degrader inhibits acute myeloid leukemia in patient-derived mouse models. · Yu X, Li D, Kottur J, et al. · Sci Transl Med 2021What this evidence cannot show
A single degrader chemotype; degradation removes all of WDR5's functions at once, so a degrader result is even less attributable to the KMT2A axis than an occupancy result.
Evidence that argues against this (1)
- Because degradation removes WDR5 from NSL and from all six SET1-family complexes simultaneously, superior efficacy of a degrader is an argument that the relevant function is NOT the WIN-site-mediated KMT2A one. Yang Z et al. 2018
- E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Genetic deletion of Wdr5 impairs growth and colony formation of MLL-AF9 leukaemia cells and attenuates leukemogenesis in vivo, acting through direct regulation of ribosomal genes.
Loss of Wdr5 attenuates MLL-rearranged leukemogenesis by suppressing Myc targets. · Liu L, Guo X, Wang Y, et al. · Biochim Biophys Acta Mol Basis Dis 2023What this evidence cannot show
One murine model, no lymphoid or t(4;11) system, and the ribosomal mechanism is inferred from expression data rather than from a separation-of-function allele.
Evidence that argues against this (1)
- The same ribosomal mechanism is reported for pharmacological WIN-site inhibition, where it is accompanied by p53 activation; if the genetic and pharmacological phenotypes share a p53-dependent route, neither is a KMT2A-programme result. Howard GC et al. 2024
- E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A drug-adapted cancer cell line selected against a highly potent WDR5 inhibitor acquired the WDR5 P173L mutation, which confers resistance by preventing target engagement.
Mechanism of Resistance to the WDR5 Inhibitor in MLL-Rearranged Leukemia. · Liu L, Shen L, Ding Z, et al. · ACS Chem Biol 2023What this evidence cannot show
An in vitro resistance selection in a single line; it demonstrates on-target engagement and an occupancy-inhibitor liability, not clinical relevance.
Evidence that argues against this (1)
- A target mutation that blocks compound binding would not confer resistance to a degrader acting through a different surface, which is the case for switching modality rather than an argument against the target. Yu X et al. 2021
MLL3 WIN peptide binds 51-fold tighter than MLL1's at the same site
in MLL-rearranged cells the dominant WIN-site response is ribosome attrition and p53-MDM4, not a fusion-complex effect
a WDR5 result in which WDR5 REPRESSES, in primary human erythroid cells, with an arginine mark in the mechanism
the same discipline applied to a different WRAD-adjacent protein: which allele, which complex, which region
The reading of every WDR5 result in this entry depends on the four attribution rules recorded there.
- related to →WDR5 — the partner that is not a partner: it binds the half of KMT2A the fusion throws away
A four-point checklist - which surface, which readout, which modality, which comparator - for reading any WDR5 result as evidence about KMT2A fusions.
The partition hypothesis is rule three of the checklist: a WIN-site agent redistributes WDR5 between NSL and COMPASS rather than removing it.
- related to →Lost from the KMT2A half: the WIN motif and the whole WRAD interface (aa 3762-3767, Arg3765)
Constrains what the WIN-site dossier can claim: WIN-site engagement is shared with KMT2B, SETD1A and SETD1B and its published cellular signature is ribosomal and p53-dependent.
SETD1A carries its own WIN motif competing for the same WDR5 pocket, which is why WIN-site engagement is not a KMT2A-axis event.