Oncofusion

start / graph / fusion-kmt2a-afdn

Fusionstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q11-comparative

KMT2A::AFDN (MLL-AF6), t(6;11)(q27;q23)

also called MLL-AF6 · MLL::AF6 · KMT2A-AFDN · MLL-MLLT4 · MLL/AF6 · t(6;11)(q27;q23)

This is the odd one out: the broken KMT2A gene joins afadin, a protein that normally lives outside the nucleus holding cells together at their junctions. Afadin contributes nothing but the ability to stick to itself, which forces two copies of the fusion together, and that alone is enough to switch the leukemia program on - the same trick the curable kinase fusions use.

The same thing, for a scientist

KMT2A::AFDN accounts for 138/3401 (4.1%) of genomically characterised KMT2A rearrangements, ~8.1% of KMT2A-rearranged AML and 9.0% of paediatric KMT2A-r AML, and it breaks predominantly in KMT2A intron 9, being depleted in intron 11 [E1]. It is the family's mechanistic outlier. AFDN/afadin is a **cytoplasmic adherens-junction scaffold**, not a nuclear transcription protein, and the only function it contributes is **self-association through its Ras-association 1 (RA1) domain**, which is necessary to activate the oncogenic potential of MLL-AF6 [E3]. Forced dimerisation of the retained KMT2A N-terminus is itself sufficient to transform haematopoietic progenitors, so AFDN belongs to the dimerisation class of KMT2A partners rather than to the SEC/DOT1L recruitment class [E2]. A second, partly separate mechanism is that the chimera **sequesters wild-type AF6 into the nucleus**, depleting the cytoplasmic pool and triggering RAS activation [E4]. KMT2A::AFDN is also the one partner in this set with a substantial **minor-BCR** presence: 9 of the 47 BCR2 cases in the recombinome were KMT2A::AFDN and 8 of those were T-ALL, and BCR2 breaks place the complete PHD1-3, bromodomain and PHD4 modules into the direct fusion protein - the one documented exception to the rule that all KMT2A fusions lose the PHD fingers [E1]. Clinically t(6;11) was the **worst** independently prognostic subgroup in paediatric KMT2A-r AML (HR 2.2, p<0.001) against an overall 5-year EFS of 44% +/- 5% [E5]. Menin inhibition applies partner-agnostically [E6].

What it looks like
loading 1XZ9…
PDB 1XZ9 ↗
shape

This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.

I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
What is kept and what is lost on each side1 deeper

Kept from KMT2A (major-BCR cases): AT-hooks 1-3, SNL1/2, MBM1/MBM2, the LEDGF/PSIP1-binding region, CXXC. Lost: PHD1-3, bromodomain, PHD4, FYRN/FYRC, TAD, WIN, SET 1.

Kept from AFDN/afadin: nearly the whole 1824-residue protein, because the AFDN break lies very close to its N-terminus. That includes RA1 and RA2 (the Ras-association domains), the FHA, DIL and PDZ domains and the C-terminal proline-rich region. The functionally decisive element is RA1, which mediates self-association 3.

Lost from AFDN: only the extreme N-terminal ~35 residues upstream of the junction.

Kept from KMT2A in the minor-BCR (T-ALL) cases - the exception that matters. Nine of the 47 minor-BCR (intron 20 to exon 24) cases in the recombinome were KMT2A::AFDN, and 8 of those 9 were T-ALL 1. A BCR2 break puts the complete PHD1-PHD2-ePHD3 cluster, the bromodomain and the complete ePHD4 into the direct fusion protein. That restores the CYP33/PPIE-responsive activator-to-repressor switch, the H3K4me2/3 reader and the capacity to recruit a BMI1-containing Polycomb repressor to CXXC - modules that are otherwise thought to be incompatible with transformation. So 'KMT2A::AFDN' names two architecturally different oncoproteins with different lineage outputs, and the minor-BCR form is the single best natural experiment in the family for asking what the PHD module actually does to a fusion.

A note on a discrepancy in this graph

The recombinome reports that the T-ALL group is 'mainly characterized by KMT2A fusions with MLLT1 (n = 43) and AFDN (n = 40)'. An existing node in this graph (mechanism-different-breaks-different-biology, in the atlas folder) instead records the second T-ALL partner as MLLT10 (n = 40). We have not edited that file. Our reading of the primary source, cross-checked against the independent minor-BCR finding that 8 of 9 BCR2 AFDN cases were T-ALL (which the atlas breakpoint-kmt2a-minor-bcr node also records), supports AFDN rather than MLLT10 as the second T-ALL partner. We flag this as a discrepancy to be resolved by the King rather than silently overwriting either version.

Mechanism: dimerisation, not recruitment - the bridge to the kinase fusions

Class: chromatin-elongation-hijack driven by forced oligomerisation.

Most KMT2A partners work by recruitment: AFF1 brings a SEC scaffold, AF9/ENL bring DOT1L, AF10 brings DOT1L. AFDN brings none of these. It is cytoplasmic and has no transcriptional activity at all. What it brings is self-association.

Liedtke and colleagues showed that the AF6 RA1 domain mediates self-association and that this is what activates MLL-AF6's oncogenic potential: RA1-dependent dimerisation is required for transformation 3. This sits inside a broader principle - dimerisation contributes to oncogenic activation of MLL chimeras generally, and artificially dimerising the retained KMT2A N-terminus is sufficient to transform progenitors without any partner-derived transcriptional module 2.

Why this is the most important node in the comparative set. Forced oligomerisation is exactly the mechanism that NPM1::ALK and ETV6::NTRK3 use - NPM1 contributes an oligomerisation domain, ETV6 contributes a SAM/PNT polymerisation domain, and in both cases the N-terminal partner's only job is to force the C-terminal module into a constitutively active multimer. KMT2A::AFDN is the same logic applied to a chromatin module instead of a kinase. That makes it the direct mechanistic bridge between the 'undruggable' chromatin fusions and the 'curable' kinase fusions, and it sharpens the question the lab is asking: if the mechanism is the same, why is only one of them drugged? The answer is not the mechanism, it is what the multimerised module is. A multimerised kinase has an ATP pocket; a multimerised AT-hook/CXXC module has nothing.

A second, cytoplasmic mechanism. MLL-AF6 also sequesters wild-type AF6 into the nucleus, depleting the cytoplasmic afadin pool and thereby triggering RAS activation, which contributes to the leukemic phenotype in paediatric AML 4. This is a genuine dominant-negative/gain-of-signalling arm on top of the chromatin arm, and it is unique to this partner. It also means KMT2A::AFDN leukemia has a signalling dependency that the other KMT2A fusions lack - and therefore a potential MEK/RAS-pathway vulnerability that the others do not.

Clinical phenotype, age and prognosis

Lineage: bimodal by breakpoint. Major-BCR KMT2A::AFDN is predominantly AML (~8.1% of KMT2A-r AML, 9.0% of paediatric KMT2A-r AML). Minor-BCR KMT2A::AFDN is predominantly T-ALL (8 of 9 BCR2 AFDN cases) 1. The same partner therefore produces myeloid or T-lymphoid disease depending on where KMT2A broke - the strongest example in the family of breakpoint position, not partner, determining lineage.

Age: childhood and adult AML, with the T-ALL cases skewing older than infancy.

Prognosis: the worst in KMT2A-rearranged AML. In the 756-patient international paediatric series, t(6;11)(q27;q23) carried the highest hazard of any KMT2A-r AML subgroup (HR 2.2, p<0.001), against an overall 5-year EFS of 44% +/- 5% and a subgroup range of 11% to 92% 5. Where MLLT3 is intermediate and MLLT10 adverse, AFDN is the floor.

The comparative ladder, assembled. Within KMT2A-r AML: t(1;11) favourable, t(9;11)/MLLT3 intermediate, t(11;19)(p13.1)/ELL not independently prognostic, t(10;11)/MLLT10 adverse (HR 1.5), t(6;11)/AFDN worst (HR 2.2) 5. Within KMT2A-r ALL: AFF1 uniformly poor, MLLT1 poor in infants and excellent in 1-9-year-olds. The dispersion across a single retained 5' half is the central comparative fact of this beat.

Druggability and therapeutic status

Menin inhibition: clinically validated for the class, partner-agnostic (CR+CRh 22.8%, ORR 63.2%) 6. The retained MBM module is identical in major-BCR cases.

Partner-specific opportunities, and they are unusually concrete here:

  • The RA1 self-association interface. This is the fusion's transforming requirement 3. A dimerisation interface is in principle a tractable small-molecule target - and unlike DOT1L recruitment, disrupting it would not touch a general transcription enzyme. No molecule exists.
  • The AF6 PDZ domain is the only KMT2A-partner domain with published fragment-based ligand discovery. The PDB contains AF6 PDZ structures alone (1T2M, 1XZ9), bound to a BCR C-terminal peptide (2AIN) and - notably - bound to low-molecular-weight ligands from a campaign explicitly framed as 'making protein-protein interactions drugable' (2EXG). The PDZ domain is retained in the fusion. Whether it is required is unknown, which makes this a real and cheap experiment.
  • RAS-pathway inhibition. Because the chimera sequesters wild-type AF6 and activates RAS 4, MEK inhibitors are a rational, testable, already-available combination partner specific to this fusion. This is the only KMT2A partner for which a signalling-directed drug has a mechanistic rationale.

Status: emerging. Clinically validated only through menin, but with more structurally defined partner-side handles than any other KMT2A partner in this set.

Facts and figures
gene5KMT2A
gene3AFDN
cytobandt(6;11)(q27;q23)
prognosisWorst subgroup in paediatric KMT2A-rearranged AML: t(6;11)(q27;q23) hazard ratio 2.2 (p<0.001) in 756 children whose overall 5-year EFS was 44% +/- 5%. No separate survival estimate exists for the minor-BCR T-ALL form.
uniprot5Q03164
uniprot3P55196
uniprot5 verifiedQ03164 = KMT2A_HUMAN, 3969 aa (UniProt REST, checked 2026-09-18)
uniprot3 verifiedP55196 = AFAD_HUMAN, gene AFDN, afadin, 1824 aa, Homo sapiens (UniProt REST, checked 2026-09-18)
exon numbering noteKMT2A exon numbers follow the Meyer/Marschalek DCAL consensus (38 exons); the minor BCR spans intron 20 to exon 24 in that convention. The AFDN exon number is indicative; the AFDN break lies very close to the gene's 5' end so nearly the whole protein is retained.
PDB noteNo structure of the KMT2A::AFDN fusion protein exists, and - importantly - there is NO structure of the RA1 self-association interface, which is the fusion's actual transforming element. What exists covers the retained but functionally unproven PDZ domain: 1T2M and 1XZ9 = AF-6 PDZ domain alone (NMR and crystal); 2AIN = AF-6 PDZ bound to the BCR C-terminal peptide; 2EXG = AF6 PDZ bound to low-molecular-weight ligands from a fragment campaign framed as making protein-protein interactions drugable - the only published small-molecule ligand discovery against any KMT2A-partner domain. 7UJ4 and 4GQ6 = menin with revumenib and with an MLL peptide, the partner-agnostic drugged node. All verified against the RCSB entry API on 2026-09-18.
mechanism class noteThe primary class is a chromatin hijack driven by FORCED OLIGOMERISATION rather than by recruitment of an elongation machine - mechanistically the same logic NPM1::ALK and ETV6::NTRK3 use, applied to a chromatin module instead of a kinase. 'condensate-rewiring' is used here loosely to denote RA1-driven self-association/multimerisation; there is no phase-separation evidence for this fusion and the label should not be read as such. A separate, genuine second mechanism is nuclear sequestration of wild-type afadin with consequent RAS activation, which does not fit any of the listed mechanism classes.
druggability classemerging
druggability noteClinically validated only through the partner-agnostic menin node, but this partner has the most concrete untried handles in the KMT2A set: the RA1 dimerisation interface (the transforming requirement, no structure, no ligand), the retained PDZ domain (the only KMT2A-partner domain with published fragment-based ligand discovery, 2EXG, but of unproven functional requirement), and RAS/MEK-pathway inhibition rationalised by nuclear sequestration of wild-type afadin.
stage of hematopoiesishaematopoietic stem cell or myeloid progenitor for the major-BCR AML form; a T-lineage-committed thymic precursor for the minor-BCR T-ALL form
age peakchildhood to adult AML; the minor-BCR T-ALL cases skew beyond infancy
developmental order rank7
developmental order noteRanked 7 of 10, the latest of the six KMT2A fusions: this partner has no infant-lymphoid representation, its AML cases span childhood to adulthood, and its distinctive minor-BCR form arises in a T-lineage-committed thymic precursor, which is the most mature cell of origin in the KMT2A set. The rank averages two breakpoint-defined diseases with different origins.
frequency
all_kmt2a_r_acute_leukemi…4.100all_kmt2a_r_acute_leukemia_pct: 4.1minor_bcr_t_all_n8.000minor_bcr_t_all_n: 8kmt2a_r_aml_pct8.100kmt2a_r_aml_pct: 8.1kmt2a_r_paediatric_aml_pct9.000kmt2a_r_paediatric_aml_pct: 9minor_bcr_cases_n9.000minor_bcr_cases_n: 9kmt2a_r_t_all_n40.0kmt2a_r_t_all_n: 40n_cases_dcal_cohort138.0n_cases_dcal_cohort: 138
diseases
[
  "paediatric AML",
  "adult AML",
  "T-ALL (minor-BCR cases)",
  "B-ALL (rare)"
]
frequency — the numbers behind the chart
{
  "all_kmt2a_r_acute_leukemia_pct": 4.1,
  "kmt2a_r_aml_pct": 8.1,
  "kmt2a_r_paediatric_aml_pct": 9,
  "kmt2a_r_t_all_n": 40,
  "minor_bcr_cases_n": 9,
  "minor_bcr_t_all_n": 8,
  "n_cases_dcal_cohort": 138
}
breakpoints
[
  {
    "id": "bp-kmt2a-intron-9-afdn-major-bcr",
    "gene5_exon_last": 9,
    "gene3_exon_first": 2,
    "frequency_pct": 37.2,
    "population": "the majority of KMT2A::AFDN cases, predominantly paediatric and adult AML; frequency_pct is the share of all KMT2A breaks in BCR subregion A (exon 9 + intron 9), the subregion AFDN strongly prefers, not the share of AFDN fusions",
    "domains_retained_5": [
      "MBM1",
      "MBM2",
      "AT-hook 1",
      "AT-hook 2",
      "AT-hook 3",
      "SNL1/2",
      "CXXC"
    ],
    "domains_lost_5": [
      "PHD1",
      "PHD2",
      "PHD3",
      "Bromodomain",
      "PHD4",
      "FYRN",
      "TAD",
      "FYRC",
      "WIN",
      "SET"
    ],
    "domains_retained_3": [
      "RA1 (Ras-association 1; mediates self-association)",
      "RA2",
      "FHA",
      "DIL",
      "PDZ",
      "proline-rich C-terminal region"
    ],
    "domains_lost_3": [
      "AFDN extreme N-terminal segment upstream of the junction (no named domain)"
    ],
    "phenotype_notes": "Myeloid, HOXA-high, worst-prognosis KMT2A-r AML subgroup. The transforming element is RA1-mediated dimerisation, not a transcriptional module. The retained KMT2A half also contains the LEDGF/PSIP1-binding region, which has no lane in the parent gene-kmt2a domain map and is therefore recorded here rather than in the domain arrays."
  },
  {
    "id": "bp-kmt2a-minor-bcr-afdn-t-all",
    "gene5_exon_last": 21,
    "gene3_exon_first": 2,
    "frequency_pct": 1.4,
    "population": "minor-BCR (BCR2) KMT2A::AFDN, 9 of 47 BCR2 cases in the recombinome, of which 8 were T-ALL; frequency_pct is the BCR2 share of all KMT2A breaks (rising to about 6% in the targeted-NGS-only subset)",
    "domains_retained_5": [
      "MBM1",
      "MBM2",
      "AT-hook 1",
      "AT-hook 2",
      "AT-hook 3",
      "SNL1/2",
      "CXXC",
      "PHD1",
      "PHD2",
      "PHD3",
      "Bromodomain",
      "PHD4"
    ],
    "domains_lost_5": [
      "FYRN",
      "TAD",
      "FYRC",
      "WIN",
      "SET"
    ],
    "domains_retained_3": [
      "RA1",
      "RA2",
      "FHA",
      "DIL",
      "PDZ",
      "proline-rich C-terminal region"
    ],
    "domains_lost_3": [
      "AFDN extreme N-terminal segment upstream of the junction"
    ],
    "phenotype_notes": "T-ALL. The one documented KMT2A fusion architecture that RETAINS the complete PHD1-PHD2-ePHD3 cluster, the bromodomain and ePHD4, restoring the CYP33/PPIE activator-to-repressor switch, H3K4me2/3 reading and BMI1/Polycomb recruitment capacity to CXXC. gene5_exon_last = 21 is indicative of the intron 20 - exon 24 BCR2 interval rather than an exact universal junction. The retained KMT2A half also contains the LEDGF/PSIP1-binding region, recorded here rather than in the domain arrays because the parent gene-kmt2a domain map has no corresponding lane."
  }
]
PDB IDS
[
  "1XZ9",
  "1T2M",
  "2AIN",
  "2EXG",
  "7UJ4",
  "4GQ6"
]
mechanism class
[
  "chromatin-elongation-hijack",
  "condensate-rewiring"
]
drugs
[
  "drug-revumenib",
  "drug-ziftomenib"
]
key interactors
[
  "gene-men1",
  "gene-psip1",
  "gene-hoxa9",
  "gene-meis1"
]
trials
[
  "NCT04065399"
]
How this could bite us back
  • Resistance: menin-inhibitor resistance via MEN1 binding-site mutation applies to the major-BCR form identically. For the minor-BCR PHD-retaining form, menin dependence itself is unproven, so those patients could be primary non-responders for reasons nobody has looked for.
  • Compensation: if RA1-mediated dimerisation is blocked, other retained AFDN surfaces (RA2, FHA, DIL, the proline-rich region) could plausibly support alternative self-association, and the chimera would also retain the option of heterodimerising with endogenous afadin.
  • Essential-function toxicity: afadin is required for adherens-junction integrity in epithelia, endothelium and the developing nervous system, and afadin-null mice are embryonic lethal. Any ligand against RA1 or PDZ would hit wild-type afadin throughout the body, in a child.
  • Second-driver escape: this fusion carries a RAS-activating arm on top of its chromatin arm. A therapy aimed only at the chromatin arm (menin) leaves the signalling arm intact - a plausible contributor to the subgroup's poor prognosis and a reason to expect single-agent menin inhibition to underperform here specifically.
  • Delivery and diagnosis: the minor-BCR form is detected only by targeted NGS or RNA sequencing (BCR2 rises from 1.4% to ~6% of cases in the NGS-only subset), so the T-ALL patients whose fusion architecture is most unusual are also the ones most likely to be missed entirely.
  • If AFDN's contribution is partly loss of cytoplasmic AF6 and consequent RAS activation, then menin inhibition leaves a RAS-pathway driver untouched in this partner group -- and t(6;11) was already the highest-hazard KMT2A-r AML subgroup before any targeted therapy existed.
Still open
  1. KMT2A::AFDN and NPM1::ALK use the same mechanistic trick - an N-terminal partner whose only contribution is forced oligomerisation. One is the worst-prognosis KMT2A-r AML and the other is largely curable. Is the entire difference that a multimerised kinase has an ATP pocket while a multimerised AT-hook/CXXC module has none? If so, the lab's central question becomes: can a dimerisation interface itself be drugged, and RA1 is the test case.
  2. Minor-BCR KMT2A::AFDN retains the complete PHD1-3 cluster, which is supposed to be incompatible with transformation, and it makes T-ALL rather than AML. Is the PHD-retaining fusion a weaker oncoprotein that needs a more permissive cell, or a different oncoprotein altogether? Isogenic comparison of PHD-retaining and PHD-lacking KMT2A::AFDN is a clean, cheap experiment nobody appears to have done.
  3. Is the retained AF6 PDZ domain required for transformation? It is the only KMT2A-partner domain with published small-molecule ligands (2EXG), so if it is required, this is the nearest thing to a ready-made partner-specific chemical starting point in the whole family.
  4. Does RAS-pathway activation via nuclear sequestration of wild-type afadin make t(6;11) AML MEK-inhibitor-sensitive, and would that explain its uniquely poor prognosis (a second, signalling oncogenic driver the other partners lack)?
  5. The T-ALL partner counts in the recombinome need adjudicating: is the second T-ALL partner AFDN (our reading, n=40) or MLLT10 (as recorded in the atlas mechanism node)? The answer changes how the family's lineage map is drawn.
  6. Does an RA1 self-association point mutant of MLL-AF6 also fail to sequester wild-type AF6? One construct tested against both readouts would settle a dispute that has stood since 2014.
Every source behind this entry (6)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    KMT2A::AFDN accounts for 138/3401 (4.1%) of genomically characterised KMT2A rearrangements and about 8.1% of KMT2A-rearranged AML and 9.0% of paediatric KMT2A-r AML, breaks predominantly in KMT2A intron 9, and contributed 9 of the 47 minor-BCR cases of which 8 were T-ALL; minor-BCR breaks place the complete PHD1-3, bromodomain and ePHD4 modules into the direct fusion protein.

    The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023
    PMID 37019990DOI
    What this evidence cannot show

    Referral-based single-centre cohort; the minor-BCR AFDN/T-ALL finding rests on 9 cases and the BCR2 share rises from 1.4% to about 6% in the targeted-NGS-only subset, so historical under-detection is substantial and the true T-ALL share may be higher. The claim that minor-BCR breaks make PHD-retaining fusions is the authors' structural inference from exon boundaries, not a protein-level demonstration. Our reading of the T-ALL partner counts differs from another node in this graph (see the discrepancy note in the sections).

    Evidence that argues against this (1)
    • The bias here runs the opposite way from the cryptic partners and is worth stating because it is easy to read this table as if all partners were ascertained alike. t(6;11) is a visible reciprocal translocation, so AFDN is detected by conventional karyotype and is relatively OVER-represented against partners that require molecular work to see. Its 4.1% is closer to a true share than a cryptic partner's, which means the two cannot be compared directly within the same table. Leukemia 2023
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Dimerisation contributes to oncogenic activation of MLL chimeras, and forced dimerisation of the retained KMT2A N-terminus can transform haematopoietic progenitors without a partner-derived transcriptional module.

    Dimerization contributes to oncogenic activation of MLL chimeras in acute leukemias · So CW, Lin M, Ayton PM, Chen EH, Cleary ML · Cancer Cell 2003
    PMID 12957285DOI
    What this evidence cannot show

    Retroviral overexpression in murine progenitors with engineered dimerisation modules; 'contributes to' is the authors' careful wording - dimerisation is sufficient in these assays but this does not establish that it is the whole mechanism in a human patient, nor that dimerisation-class and recruitment-class fusions produce identical diseases.

    Evidence that argues against this (1)
    • Dimerisation is the best explanation for the cytoplasmic partners and it is not a complete class account, which matters because this entry uses AFDN as the exemplar that partner-supplied transcriptional modules are dispensable. Purification of several different KMT2A chimeras recovers the same super elongation complex - ELL, P-TEFb and the AFF1/AFF4 scaffold - and AFF4 knockdown reduces KMT2A-chimera target gene expression. The two accounts describe different partner sets rather than competing for the same one, and neither has been shown to cover both. The open question this leaves is whether forced dimerisation of the KMT2A N-terminus recruits SEC indirectly, in which case the two are one mechanism, or whether it produces leukaemia by a route that does not involve SEC at all. Lin C et al. 2010
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Self-association mediated by the Ras-association 1 (RA1) domain of AF6 activates the oncogenic potential of MLL-AF6, making AFDN a dimerisation partner rather than a transcriptional-module partner.

    Self-association mediated by the Ras association 1 domain of AF6 activates the oncogenic potential of MLL-AF6 · Liedtke M, Ayton PM, Somervaille TC, Smith KS, Cleary ML · Blood 2010
    PMID 20395419DOI
    What this evidence cannot show

    Murine transformation assays with domain-deletion and substitution constructs; there is no structure of the RA1 self-association interface, so its druggability is unassessed. The study does not exclude additional contributions from the retained PDZ, FHA or DIL domains.

    Evidence that argues against this (1)
    • The self-association model competes with an alternative mechanism that this entry carries as a separate evidence item without noting the conflict. In the sequestration model, MLL-AF6 draws wild-type AF6 into the nucleus, depletes the cytoplasmic pool and triggers RAS activation -- a loss-of-function on the partner's normal job, not a gain-of-function self-association of the fusion. The two accounts make different predictions for an RA1 interface mutant and have not been reconciled. Manara E et al. 2014
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    MLL-AF6 sequesters wild-type AF6 into the nucleus, depleting the cytoplasmic pool and triggering RAS activation in myeloid leukemia.

    PMID 24695851DOI
    What this evidence cannot show

    Based on a small number of paediatric t(6;11) primary samples plus cell-line work; the causal chain from nuclear sequestration to RAS activation to leukemic phenotype is inferred from correlation plus knockdown rather than from an in vivo rescue. Whether RAS-pathway inhibition is therapeutically useful in t(6;11) AML has not been tested clinically.

    Evidence that argues against this (1)
    • And conversely: the sequestration model has to explain why an RA1 self-association mutant loses oncogenic potential, which is the result the dimerisation model was built on. Neither paper tests the other's construct, so the mechanism of this fusion is genuinely unresolved and the entry should present it as a dispute. Liedtke M et al. 2010
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    t(6;11)(q27;q23) was the highest-hazard independently prognostic subgroup in 756 children with 11q23/KMT2A-rearranged AML (HR 2.2, p<0.001), against an overall 5-year EFS of 44% +/- 5%.

    PMID 19528532DOI
    What this evidence cannot show

    Retrospective pooling across 11 study groups and multiple protocol eras with karyotype-based subgroup assignment; t(6;11) is a small subgroup so the hazard ratio has wide confidence intervals, and minor-BCR AFDN cases (predominantly T-ALL) are by definition absent from an AML cohort. Pre-dates menin inhibitors and modern MRD stratification.

    Evidence that argues against this (1)
    • The hazard ratio is attached to a karyotypic category in a retrospective cohort, and the molecular denominators have since been resolved at breakpoint level. Genomic characterisation of 3401 acute leukaemia patients catalogued 107 in-frame KMT2A fusions, 37 recurrent and 63 seen only once, with seven partners plus partial tandem duplications covering over 90% of recombinations - so subgroup sizes in any partner-stratified outcome analysis are extremely unequal, and t(6;11) is not among the largest. An HR of 2.2 computed on a small subgroup of a 756-patient cohort treated before 2006 is a reasonable signal and a poor basis for a present-day risk assignment, which is how partner-level prognostic statements tend to get used. Meyer C et al. 2023
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Revumenib was studied and approved in KMT2A-translocated relapsed/refractory acute leukemia without restriction by partner gene, achieving CR+CRh 22.8% and ORR 63.2%.

    Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101) · Issa GC, Aldoss I, Thirman MJ, et al. · Journal of Clinical Oncology 2025
    PMID 39121437DOINCT04065399
    What this evidence cannot show

    No partner-stratified efficacy analysis. Minor-BCR KMT2A::AFDN fusions retain the PHD1-3 repressor switch and might behave differently from major-BCR fusions under menin blockade; nobody has tested this, and the trial did not record breakpoint position.

    Evidence that argues against this (1)
    • The practical consequence for the beat: revumenib was approved across KMT2A-translocated acute leukaemia WITHOUT restriction by partner gene, so the actionable obligate interface for KMT2A::AFDN is menin -- the same host adaptor as for every other KMT2A fusion -- and not the AFDN RA1 self-association surface. A partner-specific interface can be real, mechanistically important and therapeutically irrelevant at the same time. Meyer C et al. 2023
Connected to