MLLT3 runs normal stem-cell self-renewal and is required in essentially no leukaemia line — the inverted case
also called AF9 normal versus leukaemic · MLLT3 inverted dependency
AF9 is one of the commonest partners in this disease, and the gene it comes from turns out to be important for healthy blood stem cells rather than for the leukaemia. Deleting it from cancer cells in a large screen does almost nothing.
The same thing, for a scientist
MLLT3 governs human haematopoietic stem-cell self-renewal and engraftment: it is a wild-type, normal-cell function [E1]. In DepMap fetched for this beat, MLLT3 is depleted below -0.5 in 0.4% of 1,258 lines, with a mean gene effect of -0.002 overall and +0.052 across eleven KMT2A-rearranged lines, i.e. marginally protective [E2]. The inversion matters because MLLT3 sits at degree 29 in this graph on the strength of being a fusion partner, and a reader could easily carry the normal-cell self-renewal result across into a leukaemia dependency claim, which would invert the therapeutic argument.
The whole-gene caveat that governs the reading
DepMap cuts the whole MLLT3 locus. In a KMT2A::MLLT3 line — THP-1, MOLM-13, MOLM-14 and NOMO-1 are all in the matched panel — the guides hit the wild-type allele and the fusion allele together. So the null cannot be read as 'the AF9 moiety of the fusion is dispensable' 2. What it does establish is that no screened cell line, leukaemic or otherwise, needs wild-type MLLT3 to survive three weeks in culture.
The instrument for the fusion question is a degron on the fusion allele, not a genome-wide CRISPR library; this graph already carries concept-degron-on-the-fusion-allele-versus-on-an-overexpressed-construct on exactly that point.
And a second reason an MLLT3 expression measurement may mislead
A 2026 report describes tissue- and tumour-type-specific expression of internal-promoter-driven, YEATS-domain-devoid isoforms of MLLT1 and MLLT3 3. An antibody or a transcript count that does not distinguish them is not measuring the reader-competent protein, so class-1 expression evidence for these two genes is weaker than it looks.
| gene | gene-mllt3 |
|---|---|
| uniprot | P42568 |
| evidence class | 2 |
cell type
{
"required_in_normal": "human haematopoietic stem cells",
"required_in_leukaemia": "no screened line"
}cell state
{
"maintains": "cell-state-stem-like-self-renewal",
"context": "normal cells, not leukaemic"
}depmap — the numbers behind the chart
{
"pct_all_lines_depleted": 0.4,
"mean_all_lines": -0.002,
"mean_kmt2ar": 0.052
}- Targeting MLLT3 itself would hit the normal HSC self-renewal machinery while, on this evidence, sparing the leukaemia — the worst available ratio.
- Any therapeutic argument built on 'AF9 is central to this disease' should say whether it means the fusion protein or the gene; the two have opposite dependency profiles.
- Would an allele-selective degron on KMT2A::MLLT3 produce the dependency that whole-gene CRISPR does not?
- Do the YEATS-devoid MLLT3 isoforms change the interpretation of any occupancy measurement already in this graph?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MLLT3 governs human haematopoietic stem-cell self-renewal and engraftment — a function of the wild-type protein in normal cells.
MLLT3 governs human haematopoietic stem-cell self-renewal and engraftment. · Calvanese V, Nguyen AT, et al. · Nature 2019What this evidence cannot show
Human cord-blood and fetal HSCs with overexpression and knockdown; normal haematopoiesis, not leukaemia. Abstract read in this beat; full text not opened.
Evidence that argues against this (1)
- METHODOLOGICAL DEFEATER for attributing the self-renewal phenotype to MLLT3 as a protein rather than to the complexes it sits in. Luo et al. biochemically isolated three SEC-family complexes - SEC, SEC-L2 and SEC-L3 - all containing AF9/MLLT3 together with P-TEFb and ENL, and showed by genome-wide mRNA-Seq that they control DIFFERENT subsets of genes, with only SEC required for proper HSP70 induction on stress. The human HSC result rests on overexpression and knockdown, both of which change the stoichiometry of all three at once. That does not make the phenotype wrong, but it does mean 'MLLT3 governs self-renewal' is a subunit label on a complex-level perturbation, and the inversion this entry warns against - carrying the normal-cell result into a leukaemia dependency claim - is not the only misreading available. Luo Z et al. 2012
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
In DepMap fetched 2026-09-20, MLLT3 is depleted below -0.5 in 0.4% of 1,258 lines (mean -0.002) and its mean across eleven KMT2A-rearranged lines is +0.052.
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: in KMT2A::MLLT3 lines the wild-type and fusion alleles are cut together, so this is not a test of the fusion's 3' moiety. Cancer lines in culture.
Evidence that argues against this (1)
- The DepMap null is measured on whole-gene loss, and in the four KMT2A::MLLT3 lines in the panel the wild-type and fusion alleles are cut together. A degron on the fusion allele could show a requirement that this measurement is structurally unable to detect, so 'required in no leukaemia line' is a statement about the assay as much as about the biology. Open Targets Platform GraphQL API 2026
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Internal-promoter-driven, YEATS-domain-devoid isoforms of MLLT1 and MLLT3 are expressed in a tissue- and tumour-type-specific manner.
Tissue- and tumor-type-specific expression of internal-promoter-driven YEATS-domain-devoid isoforms of MLLT1 and MLLT3. · Dailhau S, Zamani A, et al. · Biomarker research 2026What this evidence cannot show
2026 report; abstract read in this beat. The functional consequence for chromatin reading is asserted by the authors and was not independently checked here.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT about what a YEATS-devoid isoform would mean, which the abstract's assertion of functional consequence does not settle. Komata et al. established that the ENL YEATS domain's interaction with MOZ/MORF acetyltransferases is CRITICAL for CALM-AF10-mediated leukaemic transformation and that this complex maintains chromatin retention of the fusion complex. On that account a YEATS-devoid MLLT1 or MLLT3 isoform is not a variant reader with altered specificity but a species that cannot perform the chromatin-retention step at all, which makes its tissue- and tumour-type-specific expression a potential CONFOUND for every MLLT1 or MLLT3 expression or dependency measurement in this entry rather than a finding alongside them - including the DepMap null in E2, which averages over isoforms. Komata Y et al. 2023
the cell-context dossier for this gene
- related to →KMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)
whole-gene CRISPR cuts the wild-type and fusion alleles together, so the null is not a statement about the AF9 moiety
AF9/MLLT3 and ENL/MLLT1 both sit in three distinct SEC-family complexes controlling different gene subsets (PMID 22547686), so overexpression or knockdown of either changes the stoichiometry of three complexes at once - and YEATS-devoid internal-promoter isoforms are a confound for every MLLT1/MLLT3 expression or dependency measurement, since the ENL YEATS-MOZ contact is what maintains fusion chromatin retention (PMID 37031220)