Experiment: does marrow stroma protect KMT2A-rearranged cells from a menin inhibitor? Nobody has run the week-long experiment that would say
also called menin inhibitor stromal co-culture assay · niche protection against revumenib
Grow this leukaemia on a layer of bone-marrow support cells, add a menin inhibitor, and see whether the support cells keep the leukaemia alive. It is a standard laboratory test, it takes about a week, and despite menin inhibitors being the most important new drug class in this disease, nobody has published it.
The same thing, for a scientist
Direct test of the largest single gap identified in this beat. Niche-mediated chemoprotection is demonstrated in KMT2A-rearranged disease against cytarabine and against a FLT3 inhibitor, and a controlled PubMed intersection run for this beat finds **no primary study testing stromal protection against any menin inhibitor**: a 216-record menin-inhibitor half against a 543,067-record niche/stroma half intersects in **9 records**, all read, none of which does the experiment [E1]. Design: KMT2A-rearranged lines and infant patient-derived xenograft cells cultured with and without a marrow stromal layer, treated with revumenib and with ziftomenib across a dose range, read out as viability, apoptosis and differentiation, with cytarabine as the **positive control for protection** (because protection against cytarabine in this system is already established [E2]). The positive control is the point: without it a null result cannot be distinguished from a co-culture that does not confer protection against anything.
Design
Cells. KMT2A-rearranged lines spanning both fusion families and both lineages (a KMT2A::AFF1 B-ALL line, a KMT2A::MLLT3 AML line), plus at least two infant KMT2A::AFF1 patient-derived xenograft samples, because the entire human stromal-protection literature in this genotype is in infant ALL and the entire menin-inhibitor preclinical literature is in adult-derived myeloid lines.
Comparator arm — and this is a standing rule this project added after failing it. The non-KMT2A comparator group must be screened for pathway-equivalent lesions, not merely for absence of a KMT2A rearrangement: PICALM::MLLT10, NUP98 fusions, NPM1c and any other HOXA/MEIS1 driver. A comparator carrying NPM1c is a menin-dependent leukaemia and would silently destroy the specificity claim. Every line's identity to be confirmed in Cellosaurus by accession.
Stroma. A marrow stromal monolayer, and in a second arm the multicell system used in the KMT2A-r AML chemoprotection work (stromal, endothelial and mesenchymal stem cells plus extracellular matrix), since direct contact was what mattered there 2. Include a transwell arm to separate soluble from contact-dependent protection, because in infant KMT2A-rearranged ALL stroma-conditioned medium alone was sufficient.
Drugs. Revumenib and ziftomenib across a full dose range. Cytarabine as positive control. A menin-inhibitor-insensitive line as negative control.
Readouts. Viability and apoptosis; plus differentiation markers, because a menin inhibitor's mechanism is differentiation and a niche that blocks differentiation would protect in a way a viability assay at 72 hours could miss entirely. Add a redox readout (mitochondrial ROS, glutathione), because every documented niche-protection mechanism in this genotype is metabolic.
Duration. Menin inhibitors act over days, not hours; a 72-hour endpoint is likely to be too short and a 7-10 day endpoint with re-dosing is the honest design.
What each outcome would mean
- Protection comparable to cytarabine. The niche is a general resistance mechanism in this disease, the resistance beat's rejection of the candidate was a false negative caused by nobody having looked, and stromal co-culture should become a standard arm in menin-inhibitor preclinical work.
- No protection, with the cytarabine positive control working. This is the informative null. It would mean menin inhibition is niche-independent, which is a genuine advantage of the drug class and an argument for combining it with agents whose activity is niche-dependent rather than with more of the same.
- Protection only in the differentiation readout. The most interesting outcome: a niche that permits killing but blocks differentiation would explain why responses to menin inhibitors are often partial, and would point at the stroma as the reason.
- Protection only under contact, not transwell. Directs attention to adhesion, which connects to
mechanism-itga4-niche-adhesion-kmt2aand makes natalizumab a rational partner.
| design | KMT2A-rearranged lines (KMT2A::AFF1 B-ALL and KMT2A::MLLT3 AML) plus >=2 infant KMT2A::AFF1 PDX samples, cultured with and without a marrow stromal monolayer, with a multicell (stromal + endothelial + MSC + ECM) arm and a transwell arm, treated with revumenib and ziftomenib across a dose range for 7-10 days with re-dosing |
|---|---|
| readout | viability, apoptosis, differentiation markers (CD11b/CD14 for myeloid, immunophenotypic shift for lymphoid), HOXA9/MEIS1 transcript suppression, and a redox panel (mitochondrial ROS, reduced glutathione) |
| controls | cytarabine as POSITIVE control for stromal protection (established in this assay system); a menin-inhibitor-insensitive line as negative control; comparator non-KMT2A lines screened for PATHWAY-EQUIVALENT lesions (PICALM::MLLT10, NUP98 fusions, NPM1c, other HOXA/MEIS1 drivers), not only for absence of KMT2A rearrangement; all lines verified in Cellosaurus by accession |
| cost tier | low |
| timeline | 6-8 weeks including PDX expansion |
| why it matters | the resistance beat rejected niche-mediated protection as a menin-inhibitor resistance candidate because no primary study supports it; that rejection is correct on current evidence and would be overturned or confirmed by one week of bench work |
| decision value | a clean null with a working positive control would establish menin inhibition as niche-independent, which is a marketable advantage of the class and changes which drugs it should be partnered with |
tests hypotheses
[ "hyp-niche-remodelling-precedes-relapse-kmt2a" ]
- A negative result in vitro would not exclude niche protection in vivo, where the compartments, oxygen tension and cell types are all different; the experiment can confirm protection but cannot definitively exclude it.
- Running this experiment and finding nothing would still leave the 2026 review's category heading in the literature, and reviews are what clinicians read.
- Co-culture assays reward whichever readout the experimenter chooses; a viability-only design would probably return a null and be wrong.
- Is 7-10 days long enough? Menin-inhibitor differentiation responses in vivo take weeks, and no in vitro duration fully models that.
- Does murine stroma protect human KMT2A-rearranged cells the way human stroma would? Most co-culture work uses murine lines, and the adhesion ligands are not identical.
- Should the infant PDX arm use a neonatal rather than adult host, given that niche developmental stage changes MLL-rearranged leukaemia lineage?
- If protection appears only in the differentiation readout, what is the right clinical correlate - depth of response rather than response rate?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
A controlled PubMed intersection run on 2026-09-19 finds no primary study testing niche or stromal protection against a menin inhibitor: the menin-inhibitor half `(menin inhibitor OR menin inhibition OR revumenib OR ziftomenib OR bleximenib OR MI-503 OR VTP-50469)[Title/Abstract]` returns 216 records, the niche half `(bone marrow stroma OR stromal OR mesenchymal OR co-culture OR microenvironment OR niche)[Title/Abstract]` returns 543,067, and the intersection is 9 records whose titles were all read - a GIST paper in which 'stromal' denotes gastrointestinal stromal tumour, a renal fibrosis paper, a urologic cancer paper, two reviews, a differentiation-syndrome proteomics paper, a bleximenib immune-evasion paper, a CDK9-menin synergy paper in infant leukaemia, and a 2026 resistance review that lists microenvironmental resistance as a category without a primary study behind it.
PubMed eutils esearch, controlled intersection: menin-inhibitor half (216 records) x niche/stroma half (543,067 records) = 9 records, all titles read, none testing niche protection against a menin inhibitor · NCBI E-utilities · NCBI Entrez 2026What this evidence cannot show
Title/Abstract restriction misses studies that ran a stromal co-culture arm without saying so in the abstract, which is common for a control condition; this is an upper bound on absence. PubMed matches a multi-word Title/Abstract term as an unstemmed phrase, so singular and plural forms were varied where it mattered. Counts drift.
Evidence that argues against this (1)
- A negative literature search is weak evidence that the experiment has not been done, because a co-culture arm is exactly the kind of supporting result that appears in a supplementary figure and never in an abstract. Several menin-inhibitor efficacy papers use ex vivo primary patient cells that are routinely cultured on stroma to keep them alive, so protection data may exist unlabelled in the field's methods sections. The claim here is that no study reports it as a finding, not that no one has observed it. Krivtsov AV et al. 2019
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In KMT2A-rearranged paediatric AML, direct contact between AML cells and bone-marrow components mediates chemoprotection modulated by stromal, endothelial and mesenchymal stem cells and the extracellular matrix, and azacitidine plus panobinostat synergistically sensitises MV4;11 cells and KMT2A-rearranged paediatric patient-derived xenograft lines to cytarabine in multicell co-culture - establishing that this assay system detects protection in this genotype and can serve as the positive control.
Bone Marrow Microenvironment-Induced Chemoprotection in KMT2A Rearranged Pediatric AML Is Overcome by Azacitidine-Panobinostat Combination · Lehner KM, Gopalakrishnapillai A, Kolb EA, Barwe SP · Cancers (Basel) 2023What this evidence cannot show
In vitro reconstruction of marrow architecture; MV4;11 identity not verified against Cellosaurus in this beat. The protection demonstrated is against cytarabine and the epigenetic combination, and a positive control established for one drug does not guarantee sensitivity of the assay to another mechanism. Read: abstract only.
Evidence that argues against this (1)
- In infant KMT2A-rearranged ALL the protective effect was reproduced by stroma-produced SOLUBLE factors alone, without direct contact, which is the opposite conclusion about the requirement for contact. The two systems may be measuring different things, and a co-culture design that assumes contact dependence could miss the protection that matters in the lymphoid disease. Read: abstract only. Sison EA et al. 2013
prediction 2: stroma confers little or no protection against a menin inhibitor
does marrow stroma protect KMT2A-rearranged cells from revumenib? No primary study exists
extends the documented cytarabine chemoprotection to the menin-inhibitor class