Step 6 — The consequence: self-renewal held on, differentiation never fires
also called differentiation arrest · aberrant self-renewal · leukemia stem cell program · maturation block
The end result is a cell stuck in a loop: it keeps making copies of itself and never grows up into a working blood cell. Drugs that switch the fusion off let those cells finally mature, which is why they work — and why they cause differentiation syndrome.
The same thing, for a scientist
The terminal phenotype is a committed progenitor that has re-acquired a limited stem-cell self-renewal program without global reprogramming, together with an arrest in maturation. Leukemia stem cells isolated from MLL-AF9 leukemias initiated in granulocyte-macrophage progenitors retained the global identity and immunophenotype of normal GMPs while re-activating a subset of HSC-expressed genes, and transferred leukemia with as few as four cells. Acute removal of the fusion's obligate cofactor menin reverses aberrant Hox expression and specifically abrogates the differentiation arrest. The therapeutically decisive detail is that only a subset of interventions restores differentiation: menin, DOT1L and DHODH inhibitors induce differentiation, whereas BET and CDK9 inhibitors primarily induce apoptosis.
The leukemia stem cell is a progenitor wearing a stem-cell subroutine
This was the result that changed how the field thinks about the phenotype. Leukemia stem cells were isolated from leukemias initiated in committed granulocyte-macrophage progenitors by MLL-AF9. They transferred leukemia to secondary recipients from as few as four cells, and possessed an immunophenotype and global gene-expression profile very similar to normal GMPs — but with a subset of genes normally restricted to haematopoietic stem cells re-activated 1.
The conclusion the authors drew is the one that matters here: LSCs can be generated from committed progenitors without widespread reprogramming, by activating a self-renewal-associated programme in an abnormal context. The cell is not de-differentiated back to a stem cell; it is a progenitor running a stem-cell subroutine it should have shut down.
That is a strategically hopeful framing. A small, specific, aberrantly-running program is a better therapeutic proposition than wholesale identity loss.
The arrest is actively maintained and is reversible1 deeper
The differentiation block is not a structural consequence of the translocation — it is actively held in place and released when the complex is disrupted. Acute genetic ablation of menin reverses aberrant Hox expression and specifically abrogates the differentiation arrest and oncogenic properties of KMT2A-transformed blasts 2. Consistent with this, the fusion keeps Hox loci open and productive and prevents accumulation of inhibitory histone methylation even under conditions of forced differentiation 3 — i.e. the cell is receiving differentiation signals and the fusion is overriding them.
The dose-sensitivity of the block is informative too. Meis1 knockdown in murine Mll-AF9 cells decreased proliferation and clonogenicity, increased monocytic differentiation, and delayed rather than prevented leukemia in transplant recipients 4. So the block is a graded, titratable state, not a binary switch — which is why partial target engagement can produce partial differentiation rather than nothing.
Not all effective drugs release the block — and that distinction is clinically real
A head-to-head comparison across a large cell-line panel found that menin-MLL and DOT1L inhibitors act very specifically on KMT2A-fused lines, whereas BET, DHODH and P-TEFb/CDK9 inhibitors have strong effects well beyond KMT2A fusions. Crucially, menin, DOT1L and DHODH inhibitors produced significant differentiation induction, while BET and CDK9 inhibitors primarily induced apoptosis 5.
Three consequences follow.
First, this is the mechanistic basis of differentiation syndrome, the mechanism-linked and potentially fatal toxicity of menin inhibitors: the toxicity scales with how well the drug works, because a large mass of blasts maturing at once is itself the adverse event. An anti-leukemic drug whose mechanism is differentiation cannot have its efficacy and its principal toxicity decoupled by better dosing alone; it needs a differentiation-syndrome management strategy built in.
Second, differentiation and apoptosis are different therapeutic logics. A differentiating agent should be judged on maturation markers and time, an apoptotic agent on depth of clearance. Combining one of each (menin inhibitor plus venetoclax) is mechanistically coherent in a way that combining two apoptotic agents is not.
Third, the fact that CDK9 and BET inhibitors kill KMT2A-fused cells without releasing the block confirms that cytotoxicity in this disease is not evidence of on-mechanism activity.
What is not known, and it is the part the lab should care about
Everything above is myeloid. MLL-AF9-in-GMP is the model in which the LSC phenotype was defined, and menin-ablation differentiation was shown in myeloid blasts. The equivalent experiment in KMT2A::AFF1 infant B-ALL — what lineage do the cells go to when the block is released, and do they go anywhere at all — is not established in the primary literature this Queen verified.
That gap is not academic. In B-lineage disease there is no obvious 'differentiate and die' endpoint equivalent to a neutrophil, and KMT2A-r leukemias are notorious for lineage switching under pressure (including under blinatumomab and CAR-T). If releasing the block in an infant B-ALL produces a lineage switch rather than terminal maturation, the differentiation strategy could convert one disease into another rather than curing it.
| step | 6 |
|---|---|
| chain | mechanism-kmt2a-fusion-causal-chain |
| terminal phenotype | committed progenitor running an aberrant self-renewal subroutine plus maturation arrest |
| lsc transfer dose mouse | 4 |
| reversibility | differentiation arrest is actively maintained and released by menin ablation |
| linked toxicity | differentiation syndrome — scales with efficacy, mechanism-linked |
| cell state node | cell-state-differentiation-block |
| gap | lymphoid KMT2A::AFF1 differentiation fate on block release is not established |
differentiation inducing classes
[ "menin-MLL inhibitors", "DOT1L inhibitors", "DHODH inhibitors" ]
apoptosis inducing classes
[ "BET inhibitors", "CDK9/P-TEFb inhibitors" ]
- Differentiation syndrome is mechanism-linked and potentially fatal, and it scales with how well the drug works — efficacy and toxicity cannot be separated by dose alone.
- KMT2A-r leukemias lineage-switch under immunotherapy pressure; releasing a lymphoid differentiation block may produce a myeloid switch rather than maturation, converting the disease rather than curing it.
- Because the self-renewal program is a normal HSC program run in the wrong place, any therapy that suppresses it risks suppressing genuine stem-cell self-renewal — the stress-haematopoiesis toxicity seen with menin loss.
- Cytotoxicity is not evidence of on-mechanism activity: CDK9 and BET inhibitors kill these cells without releasing the block, so preclinical viability screens systematically mis-rank candidates.
- When the block is released in KMT2A::AFF1 infant B-ALL, where do the cells go? Terminal B maturation, apoptosis, or a myeloid lineage switch? This determines whether differentiation therapy cures or converts the disease, and it is unanswered.
- Is differentiation syndrome separable from efficacy at all, or is it an obligate cost of the mechanism? If obligate, the research question is management and timing, not avoidance.
- Does the 'progenitor with a stem-cell subroutine' framing hold in human infant disease, where the cell of origin is a fetal-derived lympho-myeloid progenitor rather than a murine GMP?
- Can the residual self-renewal program be measured directly in patients as a response biomarker, so that differentiation and apoptosis contributions to a combination can be attributed?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Leukemia stem cells from MLL-AF9 leukemias initiated in committed granulocyte-macrophage progenitors transferred leukemia from as few as four cells and retained a GMP-like immunophenotype and global expression profile, with a subset of HSC-restricted genes re-activated — showing LSC generation from a committed progenitor without widespread reprogramming.
Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9 · Krivtsov AV, Twomey D, Feng Z, et al. (Armstrong SA, senior author) · Nature 442(7104):818-22 2006What this evidence cannot show
Retroviral MLL-AF9 in mouse GMPs, a myeloid model with supraphysiological fusion expression; the four-cell transfer is a limiting-dilution transplant result in a permissive recipient and is not a statement about human LSC frequency. Says nothing about lymphoid KMT2A::AFF1 disease.
Evidence that argues against this (1)
- This is murine MLL-AF9 in the myeloid compartment and the index disease of this graph is human infant KMT2A::AFF1 B-lymphoid leukaemia. The two differ in the partner, in the lineage, in the cell of origin and — on this beat's own evidence — in HOXA output, since infant t(4;11) is HOXA-LOW while the mouse models are built on HOXA activation. Retroviral MLL-AF9 also expresses the fusion far above the single rearranged allele a patient carries. Transfer from as few as four cells is measured in syngeneic recipients with a retroviral fusion, conditions that maximise engraftment; the re-activation of a subset of HSC-restricted genes in a GMP-like cell is also what forced expression of a self-renewal-driving oncogene produces, which does not identify the physiological cell of origin. Oncofusion knowledge graph (internal analysis) 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Acute genetic ablation of menin reverses aberrant Hox gene expression driven by KMT2A/menin promoter-associated complexes and specifically abrogates the differentiation arrest and oncogenic properties of KMT2A-transformed leukemic blasts.
The menin tumor suppressor protein is an essential oncogenic cofactor for MLL-associated leukemogenesis · Yokoyama A, Somervaille TCP, Smith KS, Rozenblatt-Rosen O, Meyerson M, Cleary ML · Cell 123(2):207-18 2005What this evidence cannot show
Conditional Men1 deletion in mouse myeloid transformation; complete genetic loss rather than partial pharmacological occupancy, and a myeloid rather than lymphoid context.
Evidence that argues against this (1)
- Menin ablation reverses the differentiation arrest in models, but the phenotype produced by pharmacological menin-MLL disruption is lineage-dependent and in lymphoid disease is not differentiation at all. In primary KMT2A-rearranged samples, ALL showed rapid dose-dependent induction of cell DEATH with revumenib whereas AML responded much more slowly and by myeloid differentiation. A mechanism entry that treats 'release of the differentiation block' as the universal mode of action therefore mis-describes what the drug does in the B-lineage disease that dominates infant KMT2A::AFF1 leukaemia. Adriaanse FRS et al. 2024
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KMT2A fusions keep RNA polymerase II in an actively elongating state and prevent accumulation of inhibitory histone methylation, with Hox loci remaining open and productive even under conditions of forced differentiation.
Misguided transcriptional elongation causes mixed lineage leukemia · Mueller D, Garcia-Cuellar MP, Bach C, Buhl S, Maethner E, Slany RK · PLoS Biology 7(11):e1000249 2009What this evidence cannot show
Conditional engineered fusion in a myeloid cell system; 'forced differentiation' is a chemical induction protocol whose relationship to physiological differentiation signals in vivo is loose.
Evidence that argues against this (1)
- Polymerase remaining in an elongating state and inhibitory methylation failing to accumulate are two descriptions of sustained transcription, so the claim is close to restating the phenotype as its mechanism. "Open and productive under forced differentiation" also depends entirely on which differentiation stimulus was forced. eLife 2021
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Meis1 knockdown in murine Mll-AF9 cells decreased proliferation, survival and clonogenicity, increased monocytic differentiation, and significantly delayed rather than prevented leukemia establishment in transplant recipients, with reduced expression of cell-cycle-entry genes.
A role for MEIS1 in MLL-fusion gene leukemia · Kumar AR, Li Q, Hudson WA, et al. (Kersey JH, senior author) · Blood 113(8):1756-8 2009What this evidence cannot show
shRNA knockdown with residual protein; the outcome is delay, not prevention, so this cannot distinguish a dose threshold from partial-penetrance effects. Short report with limited data.
Evidence that argues against this (1)
- This is murine MLL-AF9 in the myeloid compartment and the index disease of this graph is human infant KMT2A::AFF1 B-lymphoid leukaemia. The two differ in the partner, in the lineage, in the cell of origin and — on this beat's own evidence — in HOXA output, since infant t(4;11) is HOXA-LOW while the mouse models are built on HOXA activation. Retroviral MLL-AF9 also expresses the fusion far above the single rearranged allele a patient carries. Delayed rather than prevented, from shRNA knockdown, in a transplant that selects for cells retaining expression, is the expected outcome for a gene that is contributory rather than required — and the increase in monocytic differentiation is a myeloid readout with no B-lymphoid counterpart. Oncofusion knowledge graph (internal analysis) 2026
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a head-to-head cell-line panel, menin-MLL and DOT1L inhibitors acted very specifically on KMT2A-fused lines while BET, DHODH and P-TEFb/CDK9 inhibitors had strong effects well beyond KMT2A fusions; menin, DOT1L and DHODH inhibitors induced significant differentiation, whereas BET and CDK9 inhibitors primarily induced apoptosis.
Functional diversity of inhibitors tackling the differentiation blockage of MLL-rearranged leukemia · Brzezinka K, Nevedomskaya E, Lesche R, et al. (Stresemann C, senior author) · Journal of Hematology & Oncology 12(1):66 2019What this evidence cannot show
Industry cell-line panel with no in-vivo confirmation; compound-specific potency and off-target profiles confound class-level conclusions, and differentiation was scored by surface markers and morphology rather than functional maturation.
Evidence that argues against this (1)
- Cell-line specificity has not translated into clinical specificity. Pooled across 14 studies and 784 treated adults with relapsed/refractory AML, menin-inhibitor-based therapy produced differentiation syndrome in 14.6%, liver-function-test elevation in 38.7%, febrile neutropenia in 33.6% and treatment-related mortality in 5.0%, with pooled CR+CRh of 28.5% and monotherapy CR of only 19.5%. The DOT1L arm of the same specificity claim performed worse still: pinometostat reduced H3K79 methylation on target in patients and produced 2 complete remissions in 51. A panel showing clean on-target selectivity in KMT2A-fused lines is not evidence of a wide therapeutic window in people. Alhajahjeh A et al. 2026
- ← related toThe differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia
mechanism-level account of the same phenomenon described here as a cell state
- ← related toCDK4/6 inhibition in KMT2A-rearranged leukemia — the class was already tried, and the trial never reported
CDK6 depletion in MLL-AF9 AML acts by enhancing myeloid differentiation
step 6 - self-renewal on, differentiation blocked