Oncofusion

start / graph / gene-mettl3

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6-interactome-rna-layer

METTL3 — the m6A writer with a clinical-grade inhibitor, an MLL-rearranged AML mechanism, and a B-lymphoid signal that goes the wrong way

also called MTA70 · M6A · methyltransferase-like 3 · hMETTL3 · N6-adenosine-methyltransferase catalytic subunit

METTL3 is the enzyme that puts a chemical tag on messenger RNA, and there is already a drug-like molecule that blocks it. In myeloid leukaemia driven by MLL it is needed; in the B-cell leukaemia this project cares about, the dependency data point the other way.

The same thing, for a scientist

METTL3 is the catalytic subunit of the m6A methyltransferase complex. In AML it is over-expressed relative to normal HSPCs, and depletion drives differentiation and apoptosis and delays leukaemia in vivo [E1]; a selective catalytic inhibitor, STM2457, co-crystallised with METTL3-METTL14, reduces AML growth and prolongs survival in mouse models [E2]. The closest evidence to this graph's index disease is a 2025 report that in MLL-rearranged AML, METTL3 is **transcriptionally activated by MLL** and forms a complex with YTHDC1 and CTCF, stabilising architectural RNAs and R-loops that maintain TAD integrity at leukaemic loci [E3]. **All of it is myeloid.** `MLL-AF4 AND (m6A OR N6-methyladenosine)` returns 0 records and `KMT2A-AFF1 AND (m6A OR N6-methyladenosine)` returns 0 [E4]. This project's dependency scan is consistent with that boundary: METTL3 is -0.229 in the seven myeloid KMT2A-rearranged lines and **+0.097** in the four B-lymphoid ones [E5].

Go deeper
The lineage line, and why this entry refuses to cross it

This is the best-developed target in the whole RNA layer - real biology, a co-crystal structure, a selective inhibitor - and it is the entry most at risk of being imported wholesale into infant t(4;11) disease on the strength of the phrase 'MLL-rearranged'.

Two independent reasons not to.

The literature has no B-lymphoid arm. The MLL-rearranged METTL3 mechanism is stated in MLLr**+ AML** cells throughout 3, the founding paper is about myeloid differentiation 1, and the inhibitor paper's title says myeloid leukaemia 2. The counted absence is stark: zero records pairing this disease's fusion with m6A at all 4.

The dependency data flip sign at the lineage boundary. METTL3's index-arm difference is -0.229 in myeloid lines and +0.097 in B-lymphoid ones 5. Pooling the eleven index lines would have produced a modest negative number that read as a fusion effect. It is a lineage effect, and this graph made exactly that mistake in cycle 41 with CEBPA and SPI1.

What the MLL-rearranged mechanism actually says

Worth stating precisely, because it is the only mechanism in the m6A literature that puts the fusion upstream. METTL3 is transcriptionally activated by MLL; METTL3 complexes with YTHDC1 and CTCF and co-localises at promoters and enhancers; METTL3 depletion disrupts CTCF binding sites and reduces chromatin accessibility at MYB and RUNX1; YTHDC1 loss compromises CTCF-dependent 3D genome organisation; and YTHDC1 recognition of m6A-modified architectural RNAs such as MALAT1 enhances R-loop formation and sustains TAD activity 3.

If that is right, m6A is not a parallel layer in MLL-rearranged leukaemia - it is a downstream effector of the fusion that loops back onto chromatin architecture. That is a more interesting claim than 'RNA modification matters in cancer', and it is also the claim most exposed to the objection that a broadly required enzyme will look important in any assay applied to a fast-growing blast.

Facts and figures
uniprotQ86U44
length AA580
gene symbolMETTL3
uniprot resolutiongene_exact:METTL3 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as METTL3; length 580 aa checked; entry name MTA70_HUMAN
rna functioncatalytic subunit of the m6A methyltransferase complex (with METTL14, WTAP and others)
chemical matterSTM2457, selective catalytic inhibitor, co-crystallised with METTL3-METTL14
subnuclear localizationnuclear speckles, WTAP-dependent - METTL3/14 localises there and WTAP is required for the pattern (PMID 29348140). The subnuclear answer is therefore a property of the three-subunit complex and not of METTL3 alone. A second, non-nuclear pool on ribosomes carries a catalysis-independent translational activity (PMID 27117702), which is the pool a catalytic inhibitor would not reach.
shown for fusionNO. Measured 2026-09-21 with NCBI esearch: (METTL3[tiab] OR MT-A70[tiab]) AND (MLL[tiab] OR KMT2A[tiab] OR MLL-AF4[tiab] OR MLL-AF9[tiab] OR mixed lineage leukemia[tiab]) AND (localiz*[tiab] OR localis*[tiab] OR subcellular[tiab] OR nuclear body[tiab] OR nuclear bodies[tiab] OR speckle*[tiab] OR condensat*[tiab] OR nucleolar[tiab] OR immunofluorescence[tiab] OR compartment*[tiab]) returns 0. Operands: gene clause 3,581, gene AND fusion 4 - so a METTL3-and-KMT2A literature exists but is tiny, and none of its four records makes a spatial claim. Positive control through the identical shape with (MLLT1[tiab] OR ENL[tiab]) returns 23. Zero in titles and abstracts; supporting figures not searched (brief rule 80). This is a conspicuous gap given that this entry's own E3 places METTL3 in a complex with YTHDC1 and CTCF on architectural RNA in MLL-rearranged AML - an explicitly spatial mechanism with no spatial measurement behind it.
epigenetic readoutNOT A HISTONE WRITER - IT WRITES m6A ON RNA, AND THE ENTRY SHOULD SAY WHAT IT DOES WRITE RATHER THAN LEAVE THE FIELD EMPTY. METTL3 is the catalytic subunit of the m6A methyltransferase complex (with METTL14, WTAP and others) and its product is N6-METHYLADENOSINE ON mRNA, not a histone or DNA modification. THE DOMAIN ARGUMENT, FETCHED LIVE 2026-09-21: UniProt Q86U44, entry name MTA70_HUMAN, 580 aa, with a Pfam complement of EXACTLY ONE FAMILY - PF05063 (MT-A70). There is no SET, JmjC, HAT, HDAC, bromodomain, chromodomain, Tudor, PHD, CXXC or DNMT module, and the reader for its mark is a YTH-domain protein (YTHDC1, YTHDF1-3) rather than a chromatin reader. So the honest answer to 'which histone mark does METTL3 write' is NONE, and the reason is domain architecture rather than a failure to find a paper. BUT IT IS NOT AN EPIGENETIC NULL, AND THE CONNECTION RUNS THROUGH CHROMATIN ARCHITECTURE RATHER THAN THROUGH A MARK. This entry's E3 already records that in MLL-rearranged AML, METTL3 is transcriptionally activated by MLL, forms a complex with YTHDC1 and CTCF that co-localises at promoters and enhancers, and that METTL3 depletion DISRUPTS CTCF BINDING SITES AND REDUCES CHROMATIN ACCESSIBILITY at MYB and RUNX1, modulating the integrity of chromatin TADs. That is a chromatin readout with no histone mark in it: an RNA modification holding a loop anchor. It is also the third place in this tranche where the CTCF/loop-anchor layer surfaces - gene-top2b records CTCF/cohesin-bound loop anchors as the sites vulnerable to TOP2B-mediated breaks - so architecture, not marks, may be where the RNA and DNA layers of this graph actually meet. THE MARK-LEVEL QUESTION IS WHOLLY UNMEASURED IN THE INDEX DISEASE, and this entry already bounds it: MLL-AF4 or KMT2A-AFF1 paired with m6A or N6-methyladenosine returns ZERO records [E4]. Nothing read for this entry reports H3K4me3, H3K79me2, H3K27ac or any histone mark under METTL3 depletion or under STM2457 in any system, myeloid included - the chromatin readout is accessibility and CTCF occupancy, not methylation or acetylation. So there are two nested absences: no m6A work in the index fusion at all, and no histone-mark work under an m6A perturbation anywhere. REOPENING CONDITION: any ChIP/CUT&RUN for a histone mark under METTL3 depletion or STM2457, in any leukaemia, reopens the mark question; any m6A measurement in a KMT2A::AFF1 B-lymphoid cell reopens the disease question. Either would be new.
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": -0.708,
  "pct_lines_depleted": 79.5,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": 0.0968,
  "kmt2a_r_myeloid_diff": -0.229,
  "lineage_discordant": true,
  "null_percentile": 6.22
}
pubmed counts
{
  "METTL3 AND leukemia": 109,
  "(MLL OR KMT2A) AND METTL3": "4 raw, 1 direct after title check",
  "(MLL OR KMT2A) AND MTA70": 0,
  "MLL-AF4 AND (m6A OR N6-methyladenosine)": 0,
  "KMT2A-AFF1 AND (m6A OR N6-methyladenosine)": 0
}
localization
{
  "subcellular": "TWO compartments with two different mechanisms. Nucleus: the METTL3/METTL14 heterodimer is nuclear and speckle-localised, and WTAP is required for that pattern. Cytoplasm: METTL3 associates with ribosomes and promotes translation there, and this activity survives catalytic inactivation. UniProt Q86U44 annotates Nucleus, Nucleus speckle and Cytoplasm, all at ECO:0000269.",
  "subnuclear": "NUCLEAR SPECKLES, and the address belongs to the complex rather than to METTL3. METTL3/14 localises to nuclear speckles and the splicing regulator WTAP is REQUIRED for this distinct pattern; monomeric METTL3 is soluble and inactive. So 'METTL3 is in speckles' is shorthand for 'the METTL3-METTL14-WTAP complex is in speckles', and a cell with limiting WTAP is a cell in which the assignment has not been tested.",
  "how_shown": "recombinant reconstitution, binding-surface mapping, NLS identification and endogenous localisation of the METTL3-METTL14-WTAP complex (PMID 29348140); ribosome association, tethering of wild-type and catalytically dead METTL3 to a reporter, and translation assays for the cytoplasmic pool (PMID 27117702); UniProt Q86U44 SUBCELLULAR LOCATION comment.",
  "systems": "cultured human non-haematopoietic cells for both compartments; no leukaemia and no KMT2A-rearranged system has ever been used to localise METTL3",
  "node": "localization-nuclear-speckles-kmt2a-r"
}
cell type
{
  "expressed_in": "AML cells, more abundantly than healthy human HSPCs or other tumour types, at both mRNA and protein level [E_q66_cell1]. This is a bulk comparison against the HSPC compartment and is not resolved to a cell type within it.",
  "required_in": "Human myeloid leukaemia cells, including KMT2A::MLLT3 MOLM-13, and mouse models of AML. METTL3 depletion in human myeloid leukaemia lines induces differentiation and apoptosis and delays leukaemia progression in recipient mice [E_q66_cell1]; the catalytic inhibitor STM2457 reduces AML growth, impairs engraftment and prolongs survival across mouse AML models [E2]. DepMap makes METTL3 broadly required rather than selectively so - mean Chronos -0.708 across 1,208 lines with 79.5 percent of lines depleted [E5] - so the requirement is real and the selectivity is not established.",
  "not_required_in": "No KMT2A::AFF1 B-lineage arm exists in any of it. Every load-bearing METTL3 leukaemia result in this entry is MYELOID, and this entry's own title records a B-lymphoid signal that goes the wrong way.",
  "how_shown": "shRNA depletion and overexpression of wild-type versus catalytically dead METTL3 in human HSPCs and myeloid leukaemia lines with in vivo transplantation (PMID 28920958); selective catalytic inhibition with STM2457 in vitro and in vivo (PMID 33902106)",
  "nodes": [
    "cell-type-hsc-bone-marrow",
    "cell-type-monoblast"
  ]
}
cell state
{
  "maintains_state": "THE DIFFERENTIATION BLOCK, and it does so in normal cells too. Depleting METTL3 in human HSPCs promotes differentiation; overexpressing catalytically active METTL3 inhibits differentiation and increases growth; depleting it in myeloid leukaemia lines induces differentiation and apoptosis [E_q66_cell1]. Catalytic inhibition reproduces the differentiation and apoptosis in AML [E2]. This is a state claim in the strict sense - the readout is cell identity, not fitness.",
  "regulates_state": "Apoptotic priming, indirectly and through translation rather than through transcription: m6A promotes translation of c-MYC, BCL2 and PTEN in MOLM-13, so METTL3 loss lowers BCL2 protein by a route no BH3 mimetic shares [E_q66_cell1].",
  "expressed_in_state": "not established. Nobody has measured METTL3 across the quiescent, cycling or relapse-persister states of a KMT2A-rearranged blast.",
  "how_shown": "differentiation marker readouts, apoptosis assays, in vivo transplantation, and single-nucleotide m6A mapping coupled to ribosome profiling (PMID 28920958); differentiation and apoptosis under STM2457 (PMID 33902106)",
  "nodes": [
    "cell-state-differentiation-block",
    "cell-state-apoptotic-priming-bcl2"
  ]
}
How this could bite us back
  • m6A is a global mRNA modification. Reducing it pharmacologically affects every dividing tissue, and the mouse tolerability of STM2457 is not a statement about an infant.
  • METTL3 is depleted in 79.5% of 1,208 DepMap lines. A near-broadly-required enzyme is the classic essential-function toxicity risk.
  • If METTL3 is transcriptionally activated by MLL, then inhibiting it is downstream of the fusion and inherits every escape route that acts on the fusion's transcriptional output, including the differentiation-state changes menin inhibitors already produce.
  • The entire supporting literature is myeloid. Deploying this in infant t(4;11) disease would be an extrapolation across the lineage boundary that this graph's own dependency data cross in the wrong direction.
  • Essential-function toxicity, stated by the supporting paper rather than against it: METTL3 depletion drives differentiation of HEALTHY human HSPCs as well as leukaemic ones (PMID 28920958). The differentiation phenotype that makes METTL3 attractive is not leukaemia-specific, so the therapeutic window has to come from dose and duration, not from target biology, and no published experiment has measured a normal-HSPC margin for STM2457.
  • Compartment-versus-mechanism mismatch, which is a resistance route and not merely an annotation problem: METTL3 promotes translation from the CYTOPLASM in a manner that survives catalytic inactivation (PMID 27117702), while STM2457 is a catalytic-site inhibitor (PMID 33902106). A cell that shifts its dependence onto the catalysis-independent cytoplasmic pool would be resistant to the inhibitor with the target still present and still expressed, which no on-target biomarker would detect.
  • The speckle address is not METTL3's own: WTAP is required for the nuclear-speckle pattern of METTL3/14 (PMID 29348140). Perturbing WTAP relocates the enzyme without touching the drug's binding site, so WTAP-level variation is a plausible and unmeasured modifier of METTL3 inhibitor response.
  • Lineage: every METTL3 leukaemia result cited here is myeloid, and the index disease is infant B-lineage KMT2A::AFF1. Nothing in this entry licenses a B-lineage prediction, and the entry's own title records a B-lymphoid signal pointing the wrong way.
  • STM2457 removes a mark whose reader set is YTH-domain proteins and whose chromatin consequence in MLL-r AML is loss of CTCF binding and reduced accessibility at MYB and RUNX1 (E3). Destabilising loop anchors pharmacologically in a cell type already prone to KMT2A-locus breakage is a hazard this graph should state explicitly: gene-top2b records CTCF/cohesin-bound anchors as the sites where TOP2B-mediated double-strand breaks occur, and a therapy-related KMT2A rearrangement is the canonical toxicity of that mechanism.
Still open
  1. Does STM2457 do anything in a KMT2A::AFF1 B-lymphoid line? The compound exists and the experiment has apparently never been reported.
  2. Is the METTL3-YTHDC1-CTCF TAD mechanism present in B-lymphoid KMT2A-rearranged cells, or is it specific to the myeloid context in which it was found?
  3. Why does the B-lymphoid index arm give METTL3 a positive difference - is that four-line noise, or is m6A writing genuinely less required where the fusion is KMT2A::AFF1?
  4. Where is METTL3 in a KMT2A-rearranged blast? The entry already claims a METTL3-YTHDC1-CTCF module acting on architectural RNA in MLL-rearranged AML, which is a spatial mechanism, and no one has ever imaged METTL3 in such a cell. One co-stain against SC35 and against CTCF in an index line would test it.
  5. What fraction of METTL3 in an AML blast is on ribosomes rather than in speckles, and does STM2457 change that ratio? If the cytoplasmic pool is large, the catalytic inhibitor is hitting a minority of the protein's activity.
  6. Does any histone mark move under METTL3 depletion or STM2457 in a leukaemia cell? The reported chromatin consequences are CTCF occupancy and accessibility, and a histone-mark readout under an m6A perturbation appears in nothing read for this entry, in any system - which means the RNA and chromatin layers of this graph have never been connected at the mark level.
  7. Is the METTL3/YTHDC1/CTCF loop-anchor mechanism the same layer that makes CTCF/cohesin anchors vulnerable to TOP2B breaks? Two entries in this tranche reach the same architectural feature from opposite directions - an RNA modification holding anchors together and a topoisomerase breaking DNA at them - and nobody has asked whether they interact.
Every source behind this entry (9)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    METTL3 is more abundantly expressed in AML cells than in healthy haematopoietic stem/progenitor cells; its depletion induces differentiation and apoptosis and delays leukaemia progression in vivo, and m6A promotes translation of MYC, BCL2 and PTEN mRNAs in the KMT2A::MLLT3 line MOLM-13.

    PMID 28920958DOI
    What this evidence cannot show

    Abstract read; full text not read. Myeloid throughout. The MOLM-13 observation is in a KMT2A::MLLT3 myeloid line, not a KMT2A::AFF1 B-lymphoid one.

    Evidence that argues against this (1)
    • This project's lineage-split scan gives METTL3 a positive difference (+0.097) in the four B-lymphoid KMT2A-rearranged lines against -0.229 in the seven myeloid ones. The published direction is reproduced only where the papers are. Oncofusion knowledge graph (internal analysis) 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    STM2457 is a selective catalytic METTL3 inhibitor, co-crystallised with METTL3-METTL14, that reduces AML growth, increases differentiation and apoptosis, and prolongs survival in mouse models of myeloid leukaemia.

    PMID 33902106DOI
    What this evidence cannot show

    Abstract read; full text not read. Myeloid models. A co-crystal structure establishes that the catalytic site is ligandable; it does not establish a therapeutic window, and m6A is a global mRNA modification whose bulk reduction has consequences in every dividing tissue.

    Evidence that argues against this (1)
    • A tractable catalytic pocket is a statement about chemistry, not about selectivity for a disease. This graph's own database-scope rule applies: a PDB entry existing is not a structure suitable for design, and an inhibitor working in a mouse AML model says nothing about B-lymphoid infant disease where the enzyme's dependency data run the other way. Oncofusion knowledge graph (internal analysis) 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In MLL-rearranged AML, METTL3 is transcriptionally activated by MLL and forms a complex with YTHDC1 and CTCF that co-localises at promoters and enhancers; METTL3 depletion disrupts CTCF binding sites and reduces chromatin accessibility at MYB and RUNX1, YTHDC1 loss compromises CTCF-dependent 3D genome organisation, and METTL3-mediated m6A stabilises architectural RNAs such as MALAT1 and R-loops that maintain TAD integrity at leukaemic loci.

    PMID 41430607DOI
    What this evidence cannot show

    Abstract read; full text not read. MLLr+ AML throughout, with no B-lymphoid arm. A multi-omics study in which the causal chain runs through several inference steps, and R-loop and TAD measurements are sensitive to method.

    Evidence that argues against this (1)
    • Nothing in this mechanism is shown in KMT2A::AFF1 B-lymphoid cells, and the counted absence is total: zero PubMed records pair MLL-AF4 or KMT2A-AFF1 with m6A. Importing the mechanism would be the exact move this graph has had to strike repeatedly. NCBI PubMed (queries run 2026-09-20) 2026
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence, queries run this session: `MLL-AF4 AND (m6A OR N6-methyladenosine)` returns 0 records and `KMT2A-AFF1 AND (m6A OR N6-methyladenosine)` returns 0. `(MLL OR KMT2A) AND METTL3` returns 4 raw records, of which title-checking leaves one direct MLL-rearranged AML mechanism paper, one paediatric-ALL polymorphism association, one study of Mettl3 in bone marrow stroma in an MLL-AF9 model, and one about intervertebral disc degeneration. The alias `MTA70` returns 0.

    PubMed E-utilities esearch counts, alias-expanded and title-checked · NCBI PubMed (queries run 2026-09-20) 2026
    What this evidence cannot show

    Absence claims are only as good as the symbol set and the phrasing. METTL3, MTA70 and the m6A spellings were covered; a paper framing its work at complex level ('m6A methyltransferase complex') is reachable from that phrase, which returns 1 leukaemia record and not a KMT2A::AFF1 one.

    Evidence that argues against this (1)
    • An absence of testing is not a negative result. Nobody has asked whether m6A matters in KMT2A::AFF1 B-lymphoid leukaemia, and the zero records mean exactly that and nothing more. Fu R et al. 2025
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives METTL3 a mean gene effect of -0.708 across 1,208 lines with 79.5% of lines depleted below -0.5, and in the KMT2A-rearranged index arm a difference of +0.097 in four B-lymphoid lines and -0.229 in seven myeloid lines against 43 other leukaemia lines.

    What this evidence cannot show

    Four lymphoid and seven myeloid index lines; neither number is a result alone. METTL3 is depleted in nearly 80% of all lines, so it is close to broadly required and its selectivity window is correspondingly narrow.

    Evidence that argues against this (1)
    • A gene depleted in 79.5% of 1,208 lines is approaching the floor-compression regime this project flags at 90%. Broad requirement cuts both ways: it supports METTL3 mattering everywhere, and it means the small index-versus-comparator differences are measured on a compressed scale. Oncofusion knowledge graph (internal analysis) 2026
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q86U44 (MTA70_HUMAN) is human METTL3, 580 amino acids; resolved by gene_exact query with the primary gene name read back as METTL3 and the length checked. The UniProt entry name preserves the retired symbol MTA70.

    What this evidence cannot show

    Single reviewed human hit with matching primary gene name. Worth noting that UniProt's own entry name is MTA70_HUMAN while PubMed returns 0 records for MTA70 with MLL or KMT2A - so the alias exists in a database but is effectively dead in the literature.

    Evidence that argues against this (1)
    • The METTL3 accession is for the catalytic subunit alone. The functional enzyme is a METTL3-METTL14 heterodimer with WTAP and further subunits, so any statement about METTL3 that is really about complex activity should name the complex, and complex-name phrases are unreachable from subunit symbols in PubMed. UniProt Consortium 2026
  7. E_q66_loc1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    METTL3 is nuclear as a heterodimer and its speckle address is supplied by a third subunit: METTL3/14 is found in the nucleus where it is localised to NUCLEAR SPECKLES, and the splicing regulator WTAP is REQUIRED for this distinct nuclear localisation pattern; nuclear localisation signals were mapped on the endogenous proteins in the same work.

    Interactions, localization, and phosphorylation of the m6A generating METTL3-METTL14-WTAP complex. · Schöller E; Weichmann F; Treiber T; Ringle S; Treiber N; Flatley A; Feederle R; Bruckmann A; Meister G · RNA 2018
    PMID 29348140DOI
    What this evidence cannot show

    Recombinant protein plus endogenous mapping in cultured non-haematopoietic human cells; not leukaemia and not KMT2A-rearranged. The dependency claim is that WTAP is required for the PATTERN, so a speckle assignment for METTL3 in any cell where WTAP is limiting cannot be assumed. Monomeric METTL3 is soluble and inactive and METTL14's catalytic centre is degenerate, so the compartment being described belongs to the complex, not to METTL3 alone - the brief's rule-69b caution about complex entries applies to the localisation as well as to the accession. Read: abstract only. Identifiers copied field by field from the efetch record fetched 2026-09-21, ArticleId anchored per rule 81.

    Evidence that argues against this (1)
    • The functionally important METTL3 pool in cancer cells may not be the nuclear one at all. METTL3 associates with RIBOSOMES and promotes translation IN THE CYTOPLASM, and both wild-type and CATALYTICALLY INACTIVE METTL3 promote translation when tethered to a reporter mRNA - so a compartment assignment that records only the speckle pool understates the protein and points a reader at the wrong activity. Lin S; Choe J; Du P; Triboulet R; Gregory RI 2016
  8. E_q66_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    METTL3 has a catalysis-independent cytoplasmic pool: in human cancer cells METTL3 associates with ribosomes and promotes translation in the cytoplasm, METTL3 depletion inhibits translation, and both wild-type and catalytically inactive METTL3 promote translation when tethered to a reporter mRNA through an interaction with the translation initiation machinery.

    The m(6)A Methyltransferase METTL3 Promotes Translation in Human Cancer Cells. · Lin S; Choe J; Du P; Triboulet R; Gregory RI · Mol Cell 2016
    PMID 27117702DOI
    What this evidence cannot show

    Human lung adenocarcinoma cells, not leukaemia and not a KMT2A-rearranged system, and the compartment claim rests on ribosome association and tethering rather than on quantitative imaging. This matters directly for the drug: the clinical-stage inhibitor STM2457 is a CATALYTIC inhibitor, and the cytoplasmic activity described here is explicitly catalysis-independent, so the compartment and the mechanism of action are not aligned. Read: abstract only.

    Evidence that argues against this (1)
  9. E_q66_cell1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    METTL3 sits on the myeloid differentiation switch in NORMAL human haematopoiesis as well as in leukaemia: shRNA depletion of METTL3 in human haematopoietic stem/progenitor cells PROMOTES differentiation with reduced proliferation, overexpression of wild-type but not catalytically inactive METTL3 INHIBITS differentiation and increases growth, and METTL3 mRNA and protein are more abundant in AML cells than in healthy HSPCs or other tumour types.

    The N6-methyladenosine (m6A)-forming enzyme METTL3 controls myeloid differentiation of normal hematopoietic and leukemia cells. · Vu LP; Pickering BF; Cheng Y; Zaccara S; Nguyen D; Minuesa G; Chou T; Chow A; Saletore Y; MacKay M; Schulman J; Famulare C; Patel M; Klimek VM; Garrett-Bakelman FE; Melnick A; Carroll M; Mason CE; Jaffrey SR; Kharas MG · Nat Med 2017
    PMID 28920958DOI
    What this evidence cannot show

    Human HSPCs and human MYELOID leukaemia cell lines including MOLM-13, which is KMT2A::MLLT3; the index disease here is infant B-lineage KMT2A::AFF1 and no B-lineage arm is reported, so the lineage gap is real (brief rule 73). The differentiation readout is myeloid-specific by construction. The expression comparison is against healthy HSPCs in bulk, not resolved to a cell type within the HSPC compartment. Read: abstract only.

    Evidence that argues against this (1)
Connected to