How BAF maintains KMT2A-rearranged leukaemia: it holds the blast undifferentiated, and it does so in other leukaemias too
also called SWI/SNF leukaemia maintenance · BRG1 maintenance function
Take the SWI/SNF motor away from an MLL-rearranged leukaemia cell and it stops dividing, turns into something resembling a normal white blood cell, and dies. The same happens in several other kinds of leukaemia, which is the part usually left out.
The same thing, for a scientist
The maintenance phenotype is reproducible and specific in its shape. Depleting Brg1/SMARCA4 in MLL-AF9/NrasG12D mouse AML (RN2) causes G1/G0 arrest and death, loss of the LSC marker c-Kit, gain of the macrophage marker Mac1, blast-to-macrophage morphology, GSEA loss of an LSC signature and gain of macrophage-development genes, and in vivo inhibition of disease progression with extended survival [E1]. The identical readout - c-Kit down, Mac1 up, monocytic/granulocytic morphology, apoptosis after S-phase loss - is obtained independently for SMARCA4, SMARCD2 and DPF2 in mouse MLL-AF9 cells and human THP-1, with human-cDNA rescue [E2]. Buscarlet et al. add that BRG is essential for leukaemia maintenance while dispensable for immunophenotypic long-term repopulating HSCs [E3]. **The mechanism is differentiation blockade, and the differentiation blockade is what KMT2A fusions are for - which is why it is so easy to read this as a fusion mechanism.** It is not: the same dependency appears in BCR-ABL B-ALL, RUNX1::RUNX1T1 AML, T-ALL and CML lines, and is absent in others (HEL, CMK) [E1][E2].
The phenotype, assay by assay
| readout | Shi 2013 (RN2, mouse MLL-AF9) | Cruickshank 2015 (mouse MLL-AF9, THP-1) |
|---|---|---|
| proliferation | outcompeted within 6 days, near-Brd4 magnitude | reduced from day 4-5 post-selection |
| cell cycle | G1/G0 arrest | S-phase fraction falls before apoptosis |
| death | non-viable cells rise by 72 h | Annexin V rises days 2-4 |
| stemness marker | c-Kit down | c-Kit down |
| myeloid marker | Mac1 up | Mac1/Cd11b up |
| morphology | blast to macrophage | monocytic and granulocytic forms |
| in vivo | reduced bioluminescent burden, survival extension | all 21 mice survived to termination |
Two laboratories, two vector systems, mouse and human, with an orthogonal on-target control in each (Tet-On conditional shRNA in one, human cDNA rescue of mouse-specific shRNA in the other) 12. The maintenance phenotype itself is as solid as anything in this beat.
The part that is routinely dropped
Shi et al. state it in their own words: "This indicates that the Brg1 requirement for leukemia proliferation is not restricted to myeloid leukemia or MLL fusion subtypes of disease." Brg1 shRNAs inhibited a BCR-ABL/p19Arf-null mouse B-ALL, and KASUMI-1 (RUNX1::RUNX1T1), JURKAT (T-ALL), REH (B-ALL) and K-562 (CML blast crisis) were all sensitive - while HEL and CMK were much less affected, and melanoma, breast cancer lines and immortalised MEFs were unaffected 1. Cruickshank et al. independently found Jurkat and T-ALL1 require full SMARCA4 2.
So the correct statement is "a subset of leukaemias, including but not limited to KMT2A-rearranged, require BAF for maintenance; the determinant of which ones is unknown." Every downstream claim in this graph that treats SWI/SNF as a KMT2A-fusion mechanism should be read against that sentence.
Normal-tissue window: narrower than the headline
Three statements, in increasing order of caution:
- BRG is dispensable for immunophenotypic long-term repopulating HSCs 3.
- Brg1 knockdown had no effect on immortalised MEFs, melanoma or breast lines 1.
- But in the anchor paper, the same shRNAs that spared c-Kit-enriched bone marrow proliferation did reduce LSK colony number and LSK proliferation 2.
The third is the one to carry forward: the window exists but is not clean, and the assay that found the problem is the colony assay, not the immunophenotype. Any preclinical package for this target that reports only marrow immunophenotype is under-testing.
| mechanism class | transcriptional maintenance of an undifferentiated blast state via ATP-dependent chromatin remodelling |
|---|---|
| fusion specificity verdict | NOT fusion-specific; a subset of leukaemias including KMT2A-rearranged, determinant unknown |
readouts on loss
[ "G1/G0 arrest", "apoptosis after S-phase loss", "c-Kit (CD117) down", "Mac1/CD11b up", "monocytic and granulocytic morphology", "LSC signature lost", "macrophage-development genes gained" ]
models
[ "mouse MLL-AF9/NrasG12D (RN2)", "mouse MLL-AF9-transformed spleen cells", "human THP-1 (KMT2A::MLLT3)", "human NOMO-1 (KMT2A::MLLT3)", "human MV4-11 (KMT2A::AFF1, myeloid)" ]
same dependency outside kmt2a
[ "BCR-ABL/p19Arf-null mouse B-ALL", "KASUMI-1 (RUNX1::RUNX1T1)", "JURKAT and T-ALL1 (T-ALL)", "REH (B-ALL)", "K-562 (CML blast crisis)" ]
insensitive leukaemia lines
[ "HEL", "CMK" ]
insensitive non leukaemia
[ "immortalised MEFs", "mouse melanoma", "breast cancer lines" ]
- Differentiation induction is the mechanism, which in AML means differentiation syndrome is a foreseeable clinical toxicity - the same liability that menin inhibitors carry, and the two would be combined.
- Because the dependency is shared with BCR-ABL, RUNX1::RUNX1T1 and T-ALL lines, a KMT2A-rearrangement biomarker would not enrich for responders, and a trial selected on the rearrangement could fail while the drug works.
- LSK colony formation and proliferation fall with the same knockdowns, so marrow reserve is a live concern in a population that also receives intensive chemotherapy.
- What distinguishes the Brg1-dependent leukaemia lines from HEL and CMK? Nobody has looked, and the answer would be the biomarker this target lacks.
- Does the differentiation phenotype require the ATPase activity, or does an ATPase-dead SMARCA4 that still assembles suffice? Somatic ATPase-inactivating mutations exist in cancer, so the separation is biologically real and has never been made in a leukaemia line.
- Is the LSK colony defect on-target? A CRISPR or degron version of the same experiment in normal human CD34+ cells has not been reported.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Brg1 knockdown in MLL-AF9/NrasG12D RN2 cells causes G1/G0 arrest, death, c-Kit loss, Mac1 gain, macrophage morphology and loss of an LSC signature, and inhibits disease progression in vivo; the requirement extends to BCR-ABL B-ALL, KASUMI-1, JURKAT, REH and K-562 and is explicitly not restricted to MLL fusion subtypes, while HEL and CMK are much less affected.
Role of SWI/SNF in acute leukemia maintenance and enhancer-mediated Myc regulation. · Shi J, Whyte WA, Zepeda-Mendoza CJ, Milazzo JP, et al. · Genes Dev 2013What this evidence cannot show
shRNA competition assays; the cross-subtype sensitivity panel is supplementary and qualitative rather than a matched quantitative comparison. Read in full via PMC3877755.
Evidence that argues against this (1)
- If HEL and CMK are insensitive, the determinant of Brg1 dependence is neither lineage nor fusion status, and no published work identifies it - so 'a subset of leukaemias' is as far as the data go. Cruickshank VA et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
SMARCA4, SMARCD2 and DPF2 knockdown each produce apoptosis, S-phase loss, c-Kit loss, Mac1 gain and myeloid morphology in mouse MLL-AF9 cells, with a growth defect in THP-1 and MV4-11; knockdown of one subunit does not change the levels of the other two; human cDNA rescues; LSK colony formation and proliferation are also reduced.
SWI/SNF Subunits SMARCA4, SMARCD2 and DPF2 Collaborate in MLL-Rearranged Leukaemia Maintenance. · Cruickshank VA, Sroczynska P, Sankar A, Miyagi S, et al. · PLoS One 2015What this evidence cannot show
shRNA only, all models myeloid, no B-lineage or infant model. Read in full via PMC4646637.
Evidence that argues against this (1)
- Jurkat and T-ALL1 in the same paper also require SMARCA4, so the paper's own data remove fusion specificity from its title's claim of MLL-rearranged leukaemia maintenance. Xiu Y et al. 2018
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
BAF complexes in leukaemia are specifically assembled around the BRG ATPase; BRG is essential for leukaemia maintenance, with leukaemic cells lacking BRG rapidly undergoing cell-cycle arrest and apoptosis, and is dispensable for immunophenotypic long-term repopulating HSCs.
Essential role of BRG, the ATPase subunit of BAF chromatin remodeling complexes, in leukemia maintenance. · Buscarlet M, Krasteva V, Ho L, Simon C, et al. · Blood 2014What this evidence cannot show
Mouse AML model; abstract read, full text not retrieved by this beat, so the driver oncogene of that model is not asserted here. 'Dispensable for LT-HSC' is immunophenotypic, not a serial-transplant functional readout.
Evidence that argues against this (1)
- The anchor paper's LSK colony and proliferation defects contradict a clean HSPC-sparing claim, and a colony assay is the more functional of the two readouts. Cruickshank VA et al. 2015
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
The dependency profile in human cell lines does not show KMT2A selectivity for any cBAF or PBAF subunit: SMARCA4 ranks 16,333/17,728 (+0.0791), ARID1A 12,593 (+0.0276), SMARCB1 14,024 (+0.0415), ACTL6A 17,410 (+0.1454).
Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Oncofusion knowledge graph, analyses/depmap-kmt2a-selective-scan-full/run.py · DepMap Chronos gene effect via Open Targets GraphQL 2026What this evidence cannot show
11 index against 43 comparator leukaemia lines, none infant; CRISPR knockout cannot see paralogue-buffered dependencies (SMARCA4/SMARCA2, ARID1A/ARID1B, SMARCD1/2/3, DPF1/2/3), which is a systematic bias against exactly this complex.
Evidence that argues against this (1)
- The maintenance phenotype is real in shRNA and in vivo work; a flat DepMap rank says the dependency is shared with other leukaemias, not that it is absent. Shi J et al. 2013
the ATPase whose loss produces the phenotype
BAF holds the differentiation block open; loss releases it
demonstrated in MLL-AF9 mouse and human KMT2A-rearranged myeloid lines
same phenotype on SMARCD2 loss
same phenotype on DPF2 loss
the loci through which the maintenance phenotype is produced