Oncofusion

start / graph / resistance-men1-interface-mutations

Resistancestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q6b-therapeutics2

MEN1 drug-interface mutations — the first proof that a chromatin drug can drive Darwinian escape in patients

also called MEN1 resistance mutations · menin inhibitor on-target resistance · inhibitor-menin interface mutants

Under treatment with a menin inhibitor, some leukemias evolve a tiny change in the menin protein exactly where the drug sits. The drug can no longer bind, but menin can still hold on to the fusion protein, so the cancer program restarts while the drug does nothing.

The same thing, for a scientist

Somatic MEN1 mutations at the revumenib-menin interface were identified in patients with acquired resistance to menin inhibition on the phase 1 first-in-human trial. The same class of mutations is a conserved resistance mechanism in xenografts and in an unbiased base-editor screen. Mechanistically, these mutants generate structural perturbations that attenuate small-molecule binding while leaving the interaction with the natural ligand MLL1 intact, so the drug fails to evict menin and MLL1 from chromatin. This was the first demonstration that a chromatin-targeting drug exerts enough selection pressure in patients to drive the evolution of escape mutants, and it is the most direct threat to the durability of the menin-interface hypothesis.

Go deeper
What was shown

Perner and colleagues sequenced patients with acquired resistance on the revumenib first-in-human trial and found somatic MEN1 mutations located at the revumenib-menin interface 1. Three independent lines converge:

  1. Patients. Mutations arose in treated patients with acquired resistance, not in models.
  2. Xenografts. The same interface mutations confer resistance in vivo.
  3. Unbiased base-editor screen. Scanning mutagenesis across MEN1 recovered the same interface residues without being told where to look — which means the resistance solution space is narrow and predictable.

The mechanism is the important part for this project. The mutants attenuate drug-target binding while preserving the interaction with the natural ligand MLL1, and they prevent inhibitor-induced eviction of menin and MLL1 from chromatin. The leukemia does not lose its dependency on menin. It keeps the dependency and loses the drug.

A note on residue lists. The specific residues usually quoted in reviews (M327I, G331R/D, T349M) are not enumerated in the primary abstract. They should be read from the paper's figures before being asserted. This node deliberately does not list them, and neither should anything downstream.

What this means for the lab's central hypothesis1 deeper

The project's working hypothesis is that fusion oncoproteins are undruggable but their obligate partners expose druggable protein-protein interfaces. This result is the sharpest available test of that hypothesis, and it cuts both ways.

Supporting it: the drug pocket and the MLL1-binding surface are separable. Mutations can abolish drug binding without abolishing MLL1 binding. That is a positive structural finding — it means the interface strategy is not intrinsically doomed, and it tells you precisely which sub-surface a next-generation inhibitor must avoid depending on. It is also, incidentally, the best argument that menin inhibition is on-target: you do not evolve interface mutants against an off-target drug.

Threatening it: the same separability means the leukemia has a cheap escape. A single point mutation restores the whole oncogenic program. Any interface-disruption strategy — against menin, against LEDGF, against any obligate partner — inherits this liability structurally, because a protein-protein interface is large and a small molecule occupies only part of it. This is a different and worse situation than kinase-inhibitor resistance, where gatekeeper mutations usually also cost the enzyme some activity.

The practical corollary. Single-agent, sequential use of menin inhibitors in relapsed disease is the regimen most likely to select these mutants, and that is exactly how the drugs were first approved. Combination and frontline use, where the clone burden is smaller and multiple pressures are applied at once, is the rational response — and it is what the phase 3 programmes are now doing.

Clinical frequency, and what we do not know

In a multicentre retrospective study of 84 adults with relapsed/refractory AML after menin-inhibitor failure, MEN1 mutations were among the alterations associated with resistance, alongside FLT3-ITD and WT1 2. The proportion of failures that are MEN1-mutant, rather than resistant by some other route, is not something we can state from the sources verified here — and it is the single most decision-relevant number, because it determines whether next-generation inhibitors that bind differently would rescue most patients or few.

One encouraging observation from the same cohort: among patients treated after menin-inhibitor failure, the only complete remissions occurred with venetoclax-containing regimens or with switching between menin inhibitors (bleximenib to revumenib, 1 of 2) 2. If inhibitor switching works at all, some resistance is compound-specific rather than target-specific.

The residue list this entry asked for, and the part that changes clinical options

This entry noted that the residues usually quoted in reviews are not enumerated in the primary abstract and should not be repeated without a primary source. There is one, and it does more than supply a list.

PMID 42103719 profiles MEN1 mutations against five different menin inhibitors in clinical trials using CRISPR base editor screens, with co-crystal structures of menin bound to each inhibitor, and separates the results into two classes:

classresidues
shared across inhibitorsM327I/V/T, G331D
inhibitor-specificC334R, E368K/V, V372A

The second row is the clinically consequential one. If some resistance mutations affect only certain inhibitors, then switching a patient from one menin inhibitor to another is a rational move for some mutations and futile for others — and which case a patient is in is determined by sequencing a residue, not by trying the next drug. That is a decision this graph could not previously support.

It also sharpens what a pocket mutation does and does not defeat. A shared mutation at M327 or G331 defeats the class; an inhibitor-specific one does not. This bears directly on the metabolic route recorded elsewhere in this graph, where IMPDH2 blockade lowers LEDGF and menin protein rather than competing for the pocket — a mechanism that should be unaffected by either class of mutation, and which is therefore worth testing in exactly these resistant backgrounds.

Read the class assignment, not the residue. Reviews quote M327I, G331R/D and T349M without distinguishing shared from inhibitor-specific, which is how a list becomes misleading: quoting a residue without its class implies a generality the data do not support.

Facts and figures
kindresistance
mechanism classon-target, genetic
seen in patientstrue
seen in modelstrue
structural basisinterface mutations perturb small-molecule binding without disrupting the menin-MLL1 interaction; menin and MLL1 remain on chromatin
severityhigh — a single point mutation restores the full oncogenic program
affects
[
  "drug-revumenib",
  "drug-ziftomenib",
  "drug-bleximenib",
  "drug-enzomenib",
  "drug-icovamenib"
]
evidence tiers
[
  "patient sequencing on trial",
  "xenograft",
  "unbiased base-editor scanning mutagenesis"
]
pre emption strategies
[
  "frontline or combination use rather than single-agent salvage, to reduce the clone burden under selection",
  "combination with an orthogonal mechanism (venetoclax, T-cell engager, CDK9) so no single mutation restores fitness",
  "next-generation inhibitors with a different binding footprint; menin-inhibitor switching produced a response in 1 of 2 patients",
  "covalent inhibitors (icovamenib class) whose binding chemistry differs from the reversible pocket binders",
  "serial MEN1 sequencing on therapy to detect emerging clones before overt relapse"
]
unknowns
[
  "frequency among all menin-inhibitor failures",
  "frequency in frontline and combination settings",
  "any data in children or infants"
]
How this could bite us back
  • This IS the bite-back. A chromatin drug generated Darwinian escape in patients within a single trial, which had not previously been shown for this drug class.
  • The escape is cheap and structurally inevitable: a small molecule occupies part of a large protein-protein interface, so mutations exist that block the drug and not the partner.
  • Generalisation risk: any obligate-partner interface strategy this project pursues — LEDGF, AFF1, DOT1L scaffolding — inherits the same liability by construction, not by accident.
  • False reassurance risk: menin-inhibitor switching produced a response in 1 of 2 patients in one retrospective cohort. Treating that as evidence that resistance is compound-specific would be a serious error.
  • Detection: without serial MEN1 sequencing on therapy, the first sign of an interface mutant is overt relapse, at which point the therapeutic window has closed.
Still open
  1. What fraction of menin-inhibitor failures carry a MEN1 interface mutation? This determines whether a differently-binding next-generation inhibitor rescues most patients or a minority, and it is the most decision-relevant unknown in the whole programme.
  2. Do MEN1 interface mutations arise in children and infants at the same rate? Every dataset is adult, and the infant leukemia genome has a strikingly low mutation burden, which might make genetic escape rarer and non-genetic escape correspondingly dominant.
  3. Is there a covalent or allosteric chemistry that engages menin outside the mutable interface footprint? The base-editor screen maps the escape space, which means it also maps the region a next-generation inhibitor must not depend on.
  4. Does combination therapy actually lower the emergence rate of MEN1 mutants, or merely delay detection? The phase 3 programmes will answer this only if they sequence MEN1 serially, which should be insisted on.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Somatic MEN1 mutations at the revumenib-menin interface were identified in patients with acquired resistance to menin inhibition; these mutations are a conserved resistance mechanism in xenograft models and in an unbiased base-editor screen, and they attenuate drug-target binding through structural perturbations while preserving the interaction with the natural ligand MLL1 and preventing inhibitor-induced eviction of menin and MLL1 from chromatin.

    MEN1 mutations mediate clinical resistance to menin inhibition · Perner F, Stein EM, Wenge DV, et al. · Nature 2023
    PMID 36922589DOI
    What this evidence cannot show

    Derived from patients with acquired resistance on a phase 1 trial of single-agent revumenib in heavily pretreated relapsed/refractory disease. The frequency of these mutations among all menin-inhibitor failures is not stated in the abstract, and their frequency in frontline or combination settings is entirely unknown. The specific residue identities commonly quoted in reviews are not enumerated in the abstract and are not asserted here.

    Evidence that argues against this (2)
    • Interface mutation is one route among several, and in lymphoid disease it may not be the main one. In primary KMT2A-rearranged ALL driven to revumenib resistance in vitro, escape occurred EITHER through acquisition of MEN1 mutations OR independently of any MEN1 mutation, and the same study found KMT2A-rearranged ALL and AML respond to revumenib with different kinetics and different downstream phenotypes. A resistance-monitoring strategy built on MEN1 sequencing alone will therefore miss an unknown fraction of escapes, particularly in the ALL setting that matters most for infant KMT2A::AFF1 disease. Adriaanse FRS et al. 2024
    • Menin-inhibitor resistance is not exclusively acquired at the drug-protein interface: PRIMARY (pre-existing) resistance to menin inhibitors exists in MLL-fusion AML models and is overcome by combined KAT6/KAT7 catalytic inhibition, and KAT7 remains a targetable dependency in both genetic and non-genetic acquired resistance. If a substantial share of clinical failure is primary rather than mutational, MEN1 interface mutations describe the escape route of the subset that responded first, which biases any estimate of how often the interface itself is the problem. Gordon SJV et al. 2025
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a multicentre retrospective study of 84 adults with relapsed/refractory AML after menin-inhibitor failure (63% KMT2A-rearranged, 86% previously intensively treated, 77% venetoclax-exposed), FLT3-ITD, WT1 and MEN1 mutations were associated with resistance; 40% received supportive care only, the CR/CRi rate for non-trial therapies was 19%, overall response 32%, and all complete remissions occurred with venetoclax-containing regimens or with menin-inhibitor switching; median overall survival from the start of next therapy was 4.4 months.

    Outcomes of relapsed or refractory acute myeloid leukemia after menin inhibition failure. · See PubMed record PMID 41351513 for the full author list (not transcribed here) · Blood Advances 2026
    PMID 41351513DOI
    What this evidence cannot show

    Retrospective, four US centres, adults only, n=84 and only 50 treated after failure, so all response figures rest on very small numbers (for example 'menin-inhibitor switching worked' is 1 of 2 patients). Mutation-resistance associations in a retrospective cohort are not causal. No paediatric or infant data.

    Evidence that argues against this (1)
    • The cohort is too selected to support a general statement about post-failure outcomes. 86% had been intensively pretreated and 77% were already venetoclax-exposed, so the patients remaining eligible for any subsequent therapy are an unrepresentative remainder, and the reported response pattern describes what that remainder could still receive rather than what works after menin-inhibitor failure. Blood Advances 2026
  3. E-residue-classes-across-five-inhibitorsstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CRISPR base editor screens profiling MEN1 mutations against five menin inhibitors in clinical trials identify resistance mutations shared across inhibitors (M327I/V/T, G331D) and mutations specific to individual inhibitors (C334R, E368K/V, V372A), with co-crystal structures of menin bound to each inhibitor suggesting the resistance mechanisms relate to how each engages the KMT2A binding pocket.

    PMID 42103719DOI
    What this evidence cannot show

    Abstract fetched and read back on 2026-09-20; the paper is open access (PMC13377036) and the full text was **not** read for this addition, so the residue classes are taken from the abstract's own summary. **Strong as a statement of what the screen reports.** A base editor screen measures which engineered variants survive drug exposure in a model system - it is a prediction about patients, and the entry's other evidence covers mutations actually observed in treated people. The five inhibitors are not named in the abstract, so this graph cannot yet say WHICH drug each specific mutation spares.

    Evidence that argues against this (1)
    • An inhibitor-specific resistance mutation in a screen does not establish that switching drugs helps a patient: sequential exposure selects for the shared mutations preferentially, and a clone carrying an inhibitor-specific mutation may already carry or rapidly acquire a shared one. The screen ranks vulnerabilities; it does not model sequential therapy. Perner F et al. 2023
Connected to