BAF inhibition against the menin axis: on the evidence it is a complement, not an alternative - and the case rests on one laboratory
also called SWI/SNF versus menin inhibition · FHD-286 plus revumenib
Menin inhibitors and SWI/SNF inhibitors attack MLL-rearranged leukaemia in different places, so the obvious question is whether one could replace the other or whether they should be used together. Everything published says together - but almost all of it comes from a single research group.
The same thing, for a scientist
Four independent lines say complement rather than alternative. 1. **Different loci.** Brg1 loss suppresses MYC strongly and Hoxa9 weakly, and most direct MLL-AF9 target genes do not respond to it at all; menin inhibition is defined by HOXA9/MEIS1 repression [E1][E2]. 2. **Different dependency profiles.** MEN1 ranks 1 of 17,728 for KMT2A-rearranged selectivity (difference -0.8656); SMARCA4 ranks 16,333 with a positive difference. They are at opposite ends of the same scan [E3]. 3. **Synergy, preclinically.** FHD-286 plus a menin inhibitor is synergistic in vitro and in patient-derived xenografts of MLL1-rearranged and mutant-NPM1 AML [E4]. 4. **The resistant state names SWI/SNF.** Menin-inhibitor-resistant cells derived without MEN1 mutation carry BRD4, SMARCA4 and CREBBP as CRISPR co-dependencies, and FHD-286 restores lethality against them and against menin M327I [E5]. **The caution that has to travel with this: points 3 and 4 are the same laboratory** (Fiskus, Bhalla and colleagues), across two papers two years apart [E4][E5]. Independent replication of the FHD-286-plus-menin-inhibitor combination was not found by this beat. And the only B-lineage dataset in the literature reports KMT2A-R B-ALL resistant to this compound class [E6]. So the combination case is strong in myeloid models from one group and untested everywhere else.
The two axes side by side
| menin-KMT2A axis | BAF / SWI/SNF | |
|---|---|---|
| target | protein-protein interface, MEN1 | ATPase module, SMARCA4 + SMARCA2 |
| DepMap KMT2A-r rank | 1 / 17,728 (-0.8656) | 16,333 (+0.0791) |
| pharmacodynamic readout | HOXA9, MEIS1 down | c-Myc, PU.1, CDK4/6 down |
| loci | fusion-bound target genes | distal MYC enhancers E1-E5, 1.7 Mb away |
| clinical status in KMT2A-r | revumenib FDA-approved, >=1 year of age | FHD-286 phase 1, minimum age 16 |
| resistance route | MEN1 mutation (M327I and others), plus non-genetic states | unknown; paralogue-slot restoration is theoretical |
| evidence in B-lineage | extensive | one preprint, negative |
The asymmetry in the DepMap column is the cleanest statement of the difference: on an unbiased genome-wide scan of the same lines, the menin axis is first and the SWI/SNF ATPase is in the bottom 8% 3. Whatever BAF is doing in this disease, it is not what MEN1 is doing.
What would make BAF an alternative rather than a complement
Only one thing: evidence that BAF inhibition works where menin inhibition has failed, in patients rather than in derived lines. The published version of that is E5 - derived menin-inhibitor-resistant cells, cells carrying menin M327I, and patient-derived cells with ex vivo menin-inhibitor resistance, all killed by the combination 5. What is missing is samples from patients who progressed on a menin inhibitor, which the paper itself does not claim to have.
So the correct positioning today is: a combination partner and a candidate for the post-menin-inhibitor setting, on preclinical evidence from one group, with no clinical data and no paediatric access.
| verdict | complement, not alternative, on present evidence |
|---|---|
| independence of evidence | the two strongest supporting papers (PMID 38437498, PMID 41734382) share a senior laboratory; no independent replication of the combination was found on 2026-09-20 |
| untested clinically | no registered protocol combines a BRG1/BRM inhibitor with a menin inhibitor |
axes
{
"menin": {
"target": "gene-men1",
"depmap_rank": 1,
"difference": -0.8656,
"readout": "HOXA9/MEIS1",
"clinical": "revumenib approved, >=1 year"
},
"baf": {
"target": "gene-smarca4",
"depmap_rank": 16333,
"difference": 0.0791,
"readout": "c-Myc/PU.1/CDK4-6",
"clinical": "FHD-286 phase 1, >=16 years"
}
}- Additive differentiation syndrome: both classes induce differentiation, and menin inhibitors already carry this as a labelled risk; a combination trial would need a management plan before it starts.
- Combining an approved agent available from one year of age with an agent whose only protocol starts at sixteen is not a combination a paediatric patient can currently receive, whatever the preclinical data say.
- If the BAF contribution is really MYC suppression, and the MYC node is disputed, the combination rationale rests on a contested mechanism.
- A single-laboratory evidence base is itself a risk: the most common reason a preclinical combination fails is that it was never independently reproduced.
- Does an independent laboratory reproduce FHD-286 plus menin-inhibitor synergy? Everything rests on two papers from one group.
- Do blasts from patients who progressed clinically on revumenib retain FHD-286 sensitivity? The derived-resistance model predicts yes and no patient-progression sample has been tested.
- Is the combination tolerable? Both agents induce differentiation, so differentiation syndrome risk is additive and no combination toxicology is published.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Brg1 loss suppresses Myc strongly and Hoxa9 significantly less, and the majority of direct MLL-AF9 target genes are not down-regulated by it, so SWI/SNF and the fusion act largely at different loci.
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
Mouse RN2 microarray at 96 h of knockdown, by which point cells are differentiating; the 'majority of direct targets' statement is reported as data not shown.
Evidence that argues against this (1)
- A second paper finds no MYC change at all on SWI/SNF subunit knockdown in the same disease model and could not rescue with MYC, which undermines the MYC half of this contrast. Cruickshank VA et al. 2015
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Menin inhibitors act by disrupting menin-MLL1 binding and repressing MLL1-fusion target genes including HOXA9 and MEIS1.
Overcoming menin inhibitor resistance in AML cells with combinations including BET proteins and a dual BRG1/BRM inhibitor. · Fiskus W, Mill CP, Issa GC, Piel J, et al. · Blood 2026What this evidence cannot show
Stated as background in a primary paper; this beat cites it for the mechanism of the comparator class, which this graph documents extensively elsewhere.
Evidence that argues against this (1)
- HOXA9/MEIS1 repression is a pharmacodynamic readout, not the full mechanism; menin inhibition also has effects this comparison does not enumerate. Perner F et al. 2026
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In the same genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia lines, MEN1 ranks 1 of 17,728 (difference -0.8656, KMT2A-r mean -1.2629) while SMARCA4 ranks 16,333 (difference +0.0791).
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
CRISPR knockout cannot see paralogue-buffered dependencies, which systematically disadvantages SMARCA4 relative to MEN1, a gene with no paralogue. The comparison is therefore not like-for-like and this caveat is the main reason the rank gap should not be read as a straight ranking of importance.
Evidence that argues against this (1)
- FHD-286 works preclinically in exactly these models, which a rank of 16,333 would not predict - so the rank is measuring gene dispensability, not target value. Fiskus W et al. 2024
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
FHD-286 cotreatment with a menin inhibitor synergistically induces lethality in AML cells with MLL1 rearrangement or mutant NPM1 and, compared with either drug alone, significantly reduces AML burden and improves survival in patient-derived xenografts without significant toxicity.
BRG1/BRM inhibitor targets AML stem cells and exerts superior preclinical efficacy combined with BET or menin inhibitor. · Fiskus W, Piel J, Collins M, Hentemann M, et al. · Blood 2024What this evidence cannot show
Single laboratory; immune-depleted mice; 'without significant toxicity' in that model does not address marrow reserve in treated children.
Evidence that argues against this (1)
- No independent group has published the FHD-286-plus-menin-inhibitor combination as far as this beat could find, and neither registered FHD-286 leukaemia protocol includes a menin inhibitor arm. ClinicalTrials.gov et al. 2026
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin-inhibitor-resistant AML cells derived without menin mutations carry BRD4, SMARCA4 and CREBBP as CRISPR co-enrichments; menin inhibitor plus FHD-286 kills those cells, cells expressing menin M327I, and patient-derived cells with ex vivo menin-inhibitor resistance, and revumenib plus FHD-286 reduces burden and improves survival in vivo.
Overcoming menin inhibitor resistance in AML cells with combinations including BET proteins and a dual BRG1/BRM inhibitor. · Fiskus W, Mill CP, Issa GC, Piel J, et al. · Blood 2026What this evidence cannot show
Resistance derived in MV4-11 and OCI-AML3 by repeated drug shocks; the patient-derived cells were ex vivo resistant, not from patients who progressed on a menin inhibitor. Same laboratory as E4.
Evidence that argues against this (1)
- Derived resistance selects whatever the culture permits; clinical menin-inhibitor resistance includes MEN1 mutations and lineage switch, and neither is modelled by repeated shocks in two lines. Perner F et al. 2026
- E6weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.preprint
The only B-lineage dataset on this compound class reports KMT2A-R B-ALL lines resistant to FHD-286 and BRM014 while Ph-like B-ALL is sensitive.
Selective Sensitivity of Ph-like B-ALL to BRG1 Inhibition Reveals a Novel Targeted Therapy Strategy. · Abhinav Ayyadevara VSS, Gaur S, Paik A, Perencsik R, et al. · bioRxiv 2025What this evidence cannot show
bioRxiv preprint, abstract only, line identities not stated. Graded weak and flagged for re-check on publication; it is nonetheless the only evidence that exists in the lineage this graph is about.
Evidence that argues against this (1)
- A negative in adolescent/adult B-ALL lines does not predict infant KMT2A::AFF1 biology, which differs in cell of origin and mutational burden and is not represented by any available line. Fiskus W et al. 2024
the comparator axis; rank 1 against SMARCA4's 16,333 in the same scan
the BAF-side agent the comparison is about
synergistic with FHD-286 preclinically; no combination protocol exists
the loci argument for non-redundancy