Put both fusion proteins in the same cell and you get something neither produces alone
also called MLL-AF4 and AF4-MLL cooperation · double fusion synergy · Nanog Oct4 Bmi1 t(4;11)
Cells carrying both fusion proteins resist death, keep dividing and switch on stem-cell genes in a way that cells with only one of them do not. The pair does something the singles do not.
The same thing, for a scientist
Three systems report cooperation. In a conditional cell-culture model, co-transfected cells showed the highest resistance to apoptotic triggers, the greatest cell-cycling capacity and the greatest loss of contact inhibition, with strong upregulation of the pluripotency gene NANOG; transcription of NANOG, POU5F1/Oct4 and BMI1 was verified in biopsy material from t(4;11) patients expressing both reciprocal fusion genes [E1]. In human embryonic stem cell differentiation, co-expression of AF4-MLL and MLL-AF4 strongly promoted emergence of hemato-endothelial precursors, both endothelial- and hemogenic-primed, without hijacking survival or proliferation, with H3K79 methylation specifically enriched at HOX-A cluster genes in double-fusion-expressing cells [E2]. And in the Frankfurt HEK293T model, 48 hours of reciprocal expression on a constitutive MLL-AF4 background produced roughly eight times the differential expression of MLL-AF4 alone, an effect the authors state cannot be maintained by either fusion protein alone [E3]. The pattern is consistent across three unrelated systems, which is more than most claims in this beat have. What none of them has is a leukaemia: the hESC model produces precursors, the culture model produces transformed phenotypes, and the HEK293T model produces a transcriptional signature.
Why the hESC result is the most interesting of the three
It is the only one in a human developmental system with a plausible relationship to the prenatal origin of t(4;11) infant ALL, and it reports a specific epigenetic finding - H3K79 methylation enriched at HOX-A genes specifically in double-fusion-expressing differentiating haematopoietic cells 2. That connects cooperation to the HOXA axis, which is also where the clinical collinearity between reciprocal status and HOXA expression sits. Two independent systems now put HOXA output and reciprocal presence together.
The result that does not fit
A 2026 report has the reciprocal product counteracting the direct product's target-gene activation rather than amplifying it, with IGF2BP3 lowering KMT2A::AFF1 protein and raising AFF1::KMT2A 4. That is the opposite sign to 1 and 3. It is a single recent report in engineered systems and should not overturn three older ones, but it is on the record and the sign of the interaction is therefore not settled.
| kind | genetic interaction between the two products of one translocation |
|---|---|
| epigenetic readout | H3K79 methylation enriched at HOX-A genes in double-fusion cells |
| leukaemia produced | false |
systems reporting cooperation
[ "conditional cell-culture model (2007)", "human embryonic stem cell haematopoietic differentiation (2019)", "HEK293T transient induction (2021)" ]
systems reporting antagonism
[ "engineered model systems with IGF2BP3 (2026)" ]
genes implicated
[ "NANOG", "POU5F1", "BMI1", "HOXA cluster" ]
- If the two products cooperate, then removing only the menin-dependent one leaves a partner that was contributing to the phenotype - which is the mechanism by which a menin inhibitor could produce partial rather than complete responses in reciprocal-positive disease.
- If instead the reciprocal product antagonises the direct one, a drug that lowered it would make the disease worse, and IGF2BP3 inhibition has been proposed on grounds that assume the opposite sign.
- Pluripotency gene activation in a leukaemia is a plasticity mechanism, and lineage switch under targeted therapy is already a documented failure mode in KMT2A-rearranged disease.
- Is the interaction cooperative or antagonistic? Three older systems say one thing and one 2026 report says the other, and nobody has run them side by side.
- Does the H3K79/HOXA enrichment in double-fusion cells depend on DOT1L, and does it persist under menin inhibition? The reciprocal product is menin-blind, so this is a direct test of whether HOXA output survives the drug.
- Why does NANOG appear in a leukaemia model at all, and is it detectable in reciprocal-negative t(4;11) blasts?
- None of these systems produces leukaemia. What is missing - time, cell type, a second hit, or expression level?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a conditional in vitro model, cells co-expressing MLL.AF4 and AF4.MLL displayed the highest resistance to apoptotic triggers, the greatest cell-cycling capacity and loss of contact inhibition compared with either fusion alone, strongly upregulated NANOG, and transcription of NANOG, Oct4 and BMI1 was verified in biopsy material from t(4;11) patients expressing both reciprocal fusion genes.
Combined effects of the two reciprocal t(4;11) fusion proteins MLL.AF4 and AF4.MLL confer resistance to apoptosis, cell cycling capacity and growth transformation. · Gaussmann A, Wenger T, Eberle I, et al. · Oncogene 2007What this evidence cannot show
2007 conditional transfection model in a non-haematopoietic background; the patient verification is transcript-level on an unstated number of biopsies; gene-expression signatures were established per cell line rather than across replicates of primary material.
Evidence that argues against this (1)
- A study that removed the reciprocal transcript from established human t(4;11) lines found no effect on growth, cell cycle or apoptosis, which is the same phenotypic axis this study reports cooperation on. Kumar AR et al. 2011
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Co-expression of AF4-MLL and MLL-AF4 in human embryonic stem cell hematopoietic differentiation strongly promoted emergence of hemato-endothelial precursors, both endothelial- and hemogenic-primed, specified higher numbers of hematopoietic and endothelial cells irrespective of differentiation protocol and without hijacking survival or proliferation, and produced significant enrichment of H3K79 methylated regions at HOX-A cluster genes in double fusion-expressing differentiating hematopoietic cells.
Enhanced hemato-endothelial specification during human embryonic differentiation through developmental cooperation between AF4-MLL and MLL-AF4 fusions. · Bueno C, Calero-Nieto FJ, Wang X, et al. · Haematologica 2019What this evidence cannot show
A human embryonic stem cell differentiation model, not a leukaemia model; lentiviral expression levels are not matched to an endogenous translocation; a developmental phenotype is not transformation and the authors do not claim it is.
Evidence that argues against this (1)
- The same laboratory's xenograft work with AF4-MLL in human cord-blood CD34+ cells found enhanced engraftment but no serial self-renewal and no leukaemia within 36 weeks, so an enhanced developmental output does not become leukaemia in the human system they tested. Prieto C et al. 2017
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Co-expression of both fusion proteins, even for only 48 hours, was necessary to induce chromatin changes that persisted over longer time periods; constitutive MLL-AF4 alone produced a relatively small signature and the addition of transient AF4-MLL produced roughly eight times more protein-coding genes plus hundreds of pseudogenes and non-annotated genes.
The role of reciprocal fusions in MLL-r acute leukemia: studying the chromosomal translocation t(4;11). · Wilhelm A, Marschalek R · Oncogene 2021What this evidence cannot show
HEK293T cells; a transcriptional and chromatin-accessibility readout with no cellular phenotype; the effect size depends on a stringent post-hoc filter (more than 10 reads, p < 0.05, log2 fold change beyond 2).
Evidence that argues against this (1)
- Oncogenic fusion proteins expressed in immature haematopoietic cells have been shown to fail to recapitulate the transcriptional changes observed in human AML, so a model-system signature is weak evidence about the patient state. Rapin N et al. 2014
- E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In engineered model systems KMT2A::AFF1 alone drives aberrantly high expression of only a few target genes while AFF1::KMT2A counteracts that effect, and IGF2BP3 binding reduces KMT2A::AFF1 protein while increasing AFF1::KMT2A levels.
Influence of IGF2BP3 on t(4;11) fusion mRNA and protein abundance. · Hanewald T, Benz T, Marschalek R · Leuk Res 2026What this evidence cannot show
A 2026 report in engineered systems without independent replication; the direction of the interaction is opposite to three earlier studies and the phenotypic consequence is not measured.
Evidence that argues against this (1)
- Three independent systems - a conditional culture model, a human embryonic stem cell differentiation model and a transient-induction transcriptomic model - all report cooperation rather than antagonism between the two products. Bueno C et al. 2019
The reciprocal product's phenotype is reported only in combination with the direct one.
- co-occurs with →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
Cooperation between the two products of one translocation, reported in three systems.
Independently connects reciprocal presence to HOXA output, via H3K79 methylation in double-fusion human cells.
The one report of antagonism rather than cooperation runs through IGF2BP3 control of the two transcripts.