Menin dependence of KMT2A::MLLT3: the fusion the co-IP was done on, the largest named partner in the pivotal trial, and still no published response rate
also called MLL-AF9 menin · t(9;11) menin inhibitor
The experiment that showed a menin-blocking drug actually breaks the menin grip was done on this exact fusion, and this fusion is the commonest one in the trial that got the drug approved. Even so, nobody has published how well those patients did.
The same thing, for a scientist
KMT2A::MLLT3 is not one lane of a partner panel in the menin literature - it is repeatedly **the** lane. The co-immunoprecipitation establishing that MI-463 and MI-503 reach the target in mammalian cells and disrupt the menin-MLL interaction at sub-micromolar concentrations was performed on **MLL-AF9** transfected into HEK293 [E1]. The cellular potency anchor is **MLL-AF9-transformed murine bone marrow cells, GI50 0.23 uM (MI-463) and 0.22 uM (MI-503)**, against no effect in Hoxa9/Meis1-transformed controls [E1]. The human line in that panel is **MOLM-13 (MLL-AF9)**, in which MI-389 and MI-503 downregulate HOXA9, HOXA10, HOXA7, PBX3, MEIS1, DLX2, MEF2C and FLT3 and upregulate the myeloid differentiation marker MNDA [E1]. The partner-independence study built six murine lines and used MLL-AF9 as its ChIP arm [E2]. And in AUGMENT-101, **t(9;11) was the largest single named translocation in the safety population (21 of 94, 22.3%) and the second largest in the efficacy population (11 of 57, 19.3%)** [E3]. Against all of that, **no response rate for t(9;11) has ever been published**: the trial's own partner analysis sits in Data Supplement Table S3, which is not in the PMC record - re-tested this session, the Europe PMC `supplementaryFiles` endpoint for PMC11687943 returns a valid zip containing six figure files and no supplement [E3].
What was actually measured, and on which fusion
| result | fusion used | system |
|---|---|---|
| Menin-MLL co-IP disrupted by MI-463/MI-503 at sub-micromolar | MLL-AF9 | HEK293 transfection 1 |
| GI50 0.23 / 0.22 uM | MLL-AF9 | murine BMC, 7-day MTT 1 |
| No growth inhibition in the non-MLL control | Hoxa9/Meis1 (HM-2) | murine BMC 1 |
| Human line panel, MI-503 GI50 250-570 nM | MV4;11, MOLM-13, KOPN-8, SEM | human leukaemia lines 1 |
| Target-gene downregulation by qRT-PCR | MOLM-13 | MI-389 2 uM / MI-503 1 uM, 6 days 1 |
| ChIP for menin, MLL, H3K79me2, H3K4me3 under inhibitor | MLL-AF9 | murine BMC 2 |
| In vivo xenograft, tumour regression and survival | MV4;11 (MLL-AF4) | BALB/c nude and NSG 1 |
The pattern is worth naming. The biochemistry and the cell biology of menin inhibition were done on MLL-AF9; the in vivo efficacy was done on MLL-AF4. Anyone who believes the menin story rests on a single fusion should notice it rests on two, split by assay type.
The partner-independence claim, and what it is a claim about
The paper that addressed the partner question head-on transformed murine bone marrow cells with six MLL fusions spanning the mechanistic classes - MLL-AF9 and MLL-ENL (nuclear, commonest), MLL-CBP (histone acetyltransferase), MLL-AF6 (cytoplasmic, dimerising), MLL-GAS7 and MLL-AF1P (cytoplasmic, dimerising) - and found MI-2-2 blocked all of them, concluding that menin-MLL inhibitors act in a fusion partner-independent manner 2.
That is a strong result and it should be read for what it is: a statement about transformation of murine bone marrow by retrovirally expressed fusions, at one compound, with a GI50 comparison between arms. It is not a statement about clinical response rates, it does not include reciprocal alleles, and it does not address partner-specific depth or duration of response, which is where a partner effect in patients would most plausibly appear.
The trial composition, which is new here
AUGMENT-101's baseline table gives the KMT2A translocation breakdown, and this graph had not recorded it 3:
| translocation | efficacy population (n=57) | safety population (n=94) |
|---|---|---|
| t(11;19)(q23;p13) | 13 (22.8%) | 17 (18.1%) |
| t(9;11)(p22;q23) | 11 (19.3%) | 21 (22.3%) |
| t(10;11)(p12;q23) | 7 (12.3%) | 12 (12.8%) |
| t(6;11)(q27;q23) | 7 (12.3%) | 8 (8.5%) |
| t(4;11)(q21;q23) | 2 (3.5%) | 7 (7.4%) |
| t(1;11) | 2 (3.5%) | 2 (2.1%) |
| fusion partner unknown | 13 (22.8%) | 24 (25.5%) |
| other | 2 (3.5%) | 3 (3.2%) |
Two things stand out. t(9;11) is the largest named partner group in the safety population. And the partner is unknown in a quarter of patients - a group as large as the largest known one, which by itself limits how well any retrospective partner analysis of this trial could ever perform.
The report says only that 'exploratory analysis with limited sample sizes demonstrated efficacy for KMT2Ar across various translocation partners (Data Supplement, Table S3)' 3. This graph's absence-partner-stratified-response-in-menin-inhibitor-trials established that Table S3 is not retrievable through this project's routes; that was re-tested this session against PMC11687943 and the result is unchanged - the supplementaryFiles endpoint returns HTTP 200 and a valid zip containing jco-43-75-g001/g002/g003 in .jpg and .gif and nothing else 3.
Why partner-agnostic is the right default anyway
The menin-binding motif sits at KMT2A residues 6-25 and is retained identically in KMT2A::MLLT3, KMT2A::MLLT1 and KMT2A::AFF1. There is no structural reason to expect the partner to change the menin contact, and the founding genetic work showed that MLL fusion proteins retain stable menin association through a high-affinity amino-terminal motif and that this interaction is required for initiation of MLL-mediated leukemogenesis 4.
The live question is therefore not whether t(9;11) is menin-dependent - it is - but whether the magnitude of clinical benefit differs by partner, and that question is unanswered for the simple reason that no trial has published the stratified numbers.
| partner | MLLT3 |
|---|---|
| fusion | fusion-kmt2a-mllt3 |
| menin dependence | established; and unusually for this graph, the partner on which the key biochemistry was done |
| human line used | MOLM-13 (CVCL_2119) |
| in vivo model in that paper | MV4;11 (KMT2A::AFF1) xenograft, NOT MLL-AF9 |
| augment101 t911 safety n | 21 |
| augment101 safety total | 94 |
| augment101 t911 safety PCT | 22.3 |
| augment101 t911 efficacy n | 11 |
| augment101 efficacy total | 57 |
| augment101 t911 efficacy PCT | 19.3 |
| augment101 partner unknown safety n | 24 |
| augment101 partner unknown safety PCT | 25.5 |
| partner stratified response published | false |
| supplementary retrieval retested | 2026-09-20; https://www.ebi.ac.uk/europepmc/webservices/rest/PMC11687943/supplementaryFiles returns HTTP 200 and a valid PK zip of 1,261,425 bytes containing only jco-43-75-g001.gif/.jpg, g002.gif/.jpg and g003.gif/.jpg. No Table S3. |
| reciprocal stratified menin experiment exists | false |
evidence classes
[ "menin-MLL-AF9 co-IP in HEK293 under inhibitor", "GI50 in MLL-AF9-transformed murine BMC", "MOLM-13 in the human line panel and its target-gene qRT-PCR", "ChIP arm of the partner-independence study", "largest named partner group in the pivotal registration trial" ]
mll af9 bmc gi50 uM
{
"MI-463": 0.23,
"MI-503": 0.22
}human panel gi50 NM range
[ 250, 570 ]
- The menin contact is on the shared KMT2A half, so menin inhibition gives up any possibility of fusion-partner selectivity in exchange for class breadth - and it also hits wild-type KMT2A and the whole menin-dependent endocrine programme, which this graph documents at length.
- MOLM-13, the human line carrying most of the MLL-AF9 menin pharmacology, is FLT3-ITD positive and its response to menin inhibition includes FLT3 downregulation; a t(9;11) response measured in that line is partly a FLT3 result.
- If the partner does modulate depth of response, the trial population most exposed to that effect is the t(9;11) group, because it is the largest - so a partner effect would be diluted into, rather than distinguishable from, the headline number.
- What is the CR+CRh rate for t(9;11) in AUGMENT-101? It exists in Table S3 and 21 patients is a large enough group to be worth publishing.
- Who are the 24 partner-unknown patients, and is partner-unknown status associated with response? A missing stratum this large may carry its own signal.
- Does the menin dependency of MLL-AF9 measured in murine BMC hold in a t(9;11) line at endogenous expression? Every GI50 anchor in this literature is either an overexpression model or a human line with a second driver.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MI-463 and MI-503 inhibit the menin-MLL-AF9 interaction at sub-micromolar concentrations by co-immunoprecipitation in HEK293 cells; they give GI50 values of 0.23 uM and 0.22 uM in MLL-AF9-transformed murine bone marrow cells with no growth inhibition in Hoxa9/Meis1-transformed controls; MI-503 gives GI50 250-570 nM across a human panel including MOLM-13 (MLL-AF9); and in MOLM-13 treated with MI-389 or MI-503, HOXA9, HOXA10, HOXA7, PBX3, MEIS1, DLX2, MEF2C and FLT3 are downregulated and MNDA upregulated.
Pharmacologic inhibition of the Menin-MLL interaction blocks progression of MLL leukemia in vivo. · Borkin D, He S, Miao H, et al. · Cancer Cell 2015What this evidence cannot show
Read in full text from the PMC HTML page (PMC4415852) on 2026-09-20. The co-IP is in transfected HEK293, not in a leukaemic cell at endogenous stoichiometry. The GI50 comparison against Hoxa9/Meis1 cells controls for a downstream oncogene but not for lineage or proliferation rate. In vivo efficacy in this paper is MV4;11 (MLL-AF4) xenograft, not MLL-AF9, so the survival and tumour-regression data are from a different fusion.
Evidence that argues against this (1)
- MOLM-13 carries FLT3-ITD, and FLT3 is one of the transcripts the inhibitor downregulates, so part of the MOLM-13 response could be loss of a FLT3-ITD dependency rather than loss of a fusion-programme dependency. The paper does not separate the two. Borkin D et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Murine bone marrow cells were transformed with six MLL fusions representing the major translocation types - MLL-AF9 and MLL-ENL as the commonest nuclear partners, MLL-CBP, MLL-AF6, MLL-GAS7 and MLL-AF1P - and the menin-MLL inhibitor MI-2-2 blocked oncogenic activity across all of them, with chromatin immunoprecipitation performed in the MLL-AF9, MLL-AF6 and MLL-AF1p arms.
Menin-MLL inhibitors block oncogenic transformation by MLL-fusion proteins in a fusion partner-independent manner. · He S, Malik B, Borkin D, et al. · Leukemia 2016What this evidence cannot show
Read in full text from the PMC HTML page (PMC4830728) on 2026-09-20. Retroviral overexpression of fusions in murine progenitors, one tool compound, n=2 replicates for most panels, readout at 10 days. 'Partner-independent' is a statement about whether transformation is blocked, not about the magnitude, depth or durability of a clinical response.
Evidence that argues against this (1)
- A partner-independence result obtained by overexpressing fusions in a uniform murine background removes, by construction, everything that differs between partners in patients: lineage of origin, age, co-mutations, reciprocal allele and fusion expression level. A null for partner in that design is weak evidence of a null for partner in people. Meyer C et al. 2023
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In AUGMENT-101, t(9;11)(p22;q23) accounted for 21 of 94 (22.3%) of the safety population and 11 of 57 (19.3%) of the efficacy population - the largest single named translocation in the safety population - with the fusion partner unknown in 24 of 94 (25.5%); the report states only that an exploratory analysis with limited sample sizes demonstrated efficacy across various translocation partners, with the numbers in Data Supplement Table S3.
Menin Inhibition With Revumenib for KMT2A-Rearranged Relapsed or Refractory Acute Leukemia (AUGMENT-101). · Issa GC, Aldoss I, Thirman MJ, et al. · J Clin Oncol 2025What this evidence cannot show
Read in full text from the PMC HTML page (PMC11687943) on 2026-09-20. Baseline composition, not outcome: no response rate by partner is given anywhere in the report. The Europe PMC `supplementaryFiles` endpoint for PMC11687943 was re-tested this session and returns a 1.26 MB zip containing only six figure files (jco-43-75-g001/g002/g003 as .jpg and .gif), so Table S3 remains unreachable by this project's routes. Karyotype-based partner assignment means t(9;11) here is a cytogenetic call, not necessarily a molecularly confirmed KMT2A::MLLT3.
Evidence that argues against this (1)
- With the partner unknown in a quarter of patients, any partner-stratified analysis of this trial has a missing-data stratum as large as its largest observed one, and the missingness is unlikely to be at random - partners are identified when a standard probe or panel covers them. Issa GC et al. 2025
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Oncogenic MLL fusion proteins retain the ability to associate stably with menin through a high-affinity amino-terminal conserved binding motif, this interaction is required for the initiation of MLL-mediated leukemogenesis, and menin is essential for maintenance of MLL-associated but not other oncogene-induced myeloid transformation.
The menin tumor suppressor protein is an essential oncogenic cofactor for MLL-associated leukemogenesis. · Yokoyama A, Somervaille TC, Smith KS, et al. · Cell 2005What this evidence cannot show
The menin-binding motif is on the KMT2A side and is retained identically by every common fusion, so this result is partner-agnostic by construction and says nothing specific about MLLT3. The paper is not open access by any of the four routes tried (no PMC record), so this claim rests on the abstract rather than on the full text.
Evidence that argues against this (1)
- That the founding menin result is partner-agnostic by construction is itself the reason this graph's `hyp-menin-interface-evidence-is-partner-imported` exists: an interface on the shared half cannot distinguish partners, so partner-specific menin biology, if it exists, would have to arise from something other than the motif. He S et al. 2016
menin dependence of this partner
MLL-AF9 co-IP is the biochemical anchor of the class
largest named partner group in AUGMENT-101, 21/94
re-tested the supplementary route; Table S3 still unreachable
- analogous to →Menin dependence of KMT2A::MLLT1: real, repeatedly demonstrated, and always as one lane of a partner panel
the same question for the paralogous partner, with the opposite evidence profile
for MLLT3 the biochemistry is native, not imported: the co-IP is MLL-AF9