Covalent or reversible is the axis on which ‘the menin inhibitor class’ most plausibly breaks
also called covalent menin inhibition · menin class effect · irreversible menin binding
This project talks about menin inhibitors as one kind of drug. Some of them stick to menin permanently and others let go, and that is the difference most likely to matter.
The same thing, for a scientist
The graph holds five menin-inhibitor drug entries and reasons about them as **one intervention** in its trial-design, go/no-go and resistance entries. It also holds, unread, a 2026 report that **menin inhibitors display differential in vitro sensitivity in KMT2A-rearranged infant ALL**. Covalent-versus-reversible is the most mechanistically obvious axis for such a difference and the graph has no entry on it: four of the five clinical compounds bind reversibly, icovamenib is covalent, and there is a published covalent medicinal-chemistry lineage behind it that this project had never cited.
Why the two modalities should behave differently, and what is actually known
| axis | reversible | covalent | status here |
|---|---|---|---|
| duration of target engagement | set by plasma exposure | set by protein resynthesis | mechanistically expected, not measured in this disease |
| interface-mutation resistance | attenuated binding restores function | depends on whether the mutation touches the modified residue | the checkable prediction; unknown |
| JUND co-occupancy | equilibrium shifts | occupancy removed for the protein's lifetime | structurally implied by the shared pocket |
| on-target endocrine risk | transient partial inhibition | permanent inactivation of a tumour suppressor until resynthesis | unaddressed |
The resistance prediction is the one worth testing. Clinical resistance to reversible menin inhibition arises through MEN1 mutations at the drug interface that attenuate small-molecule binding while preserving the interaction with MLL1. A covalent agent's fate under those same mutations is not the same question: if the modified residue is untouched and still accessible, covalency may survive a mutation that defeats reversible binding — and if the mutation lies in the covalent contact, it should defeat it completely. Nobody appears to have asked.
What this does not claim
It does not claim the class abstraction is wrong. The clinical data behave class-like where they have been compared: composite remission rates in the same range for the two approved agents, and differentiation syndrome across both. This graph separately established that QT prolongation is NOT a class effect, so the class already has known internal structure, and one more axis of difference would refine it rather than overturn it.
It also does not claim covalency is better. The one covalent compound that reached patients had its oncology programme terminated — for strategic reasons, with no safety or efficacy observation cited — and the published tool compound achieved partial rather than complete regression.
- Permanent inactivation of a tumour suppressor is a different risk profile from transient occupancy, and the endocrine beat has already had to separate germline penetrance over decades from weeks of partial inhibition.
- Covalent exposure is governed by protein resynthesis, so off-target modification accumulates across repeated dosing - a particular concern in an infant dosed for months.
- A covalent agent removes JUND co-occupancy of the shared pocket for the protein's lifetime, and the sign of that contribution is unknown.
- Which residue do the covalent menin inhibitors modify, and does it fall inside the MEN1 interface-mutation hotspots that confer clinical resistance to reversible agents?
- Is there any head-to-head of a covalent and a reversible menin inhibitor in the same KMT2A-rearranged material? The 2026 differential-sensitivity report may be exactly this and cannot currently be read.
- Why did the covalent medicinal-chemistry series stop? No published reason was found.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A covalent medicinal-chemistry lineage against the menin-MLL interaction exists and predates the clinical covalent compound: structure-based optimisation produced M-808, with low-nanomolar cellular activity, partial tumour regression in an MV4;11 xenograft and a co-crystal structure with menin defining the covalent interaction.
Discovery of M-808 as a Highly Potent, Covalent, Small-Molecule Inhibitor of the Menin-MLL Interaction with Strong In Vivo Antitumor Activity · Journal of Medicinal Chemistry 2020What this evidence cannot show
Abstract read; full text not read. The modified residue is not established here.
Evidence that argues against this (1)
- One published series is a lineage only in retrospect. Without knowing why it stopped, its existence is as consistent with covalent menin inhibition having been tried and found wanting as with it being underexplored. Oncofusion knowledge graph (internal analysis) 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Clinical resistance to reversible menin inhibition arises through somatic MEN1 mutations at the drug-menin interface that attenuate small-molecule binding while preserving the interaction with the natural ligand MLL1 and preventing inhibitor-induced eviction of menin and MLL1 from chromatin.
What this evidence cannot show
Established for the reversible agents in which it was described. Whether a covalent agent is defeated by the same alleles depends on where the modified residue sits relative to them, which is not published.
Evidence that argues against this (2)
- Interface mutations are not the dominant clinical resistance mechanism. This graph records resistance arising in patients with NO prior menin-inhibitor exposure and with on-target repression preserved, so a modality that defeated interface mutations would address a minority route. Journal of Clinical Oncology 2025
- The covalent-versus-reversible framing also assumes the modified residue is the relevant variable. If a covalent agent's advantage were mainly duration of engagement rather than mutation resistance, the comparison would need pharmacodynamic measurement rather than structural reasoning. Nature 2012
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
This graph holds five separate menin-inhibitor entries and reasons about them as one intervention in its trial-design, go/no-go and resistance entries, while holding in its unreadable register a 2026 report that menin inhibitors display differential in vitro sensitivity in KMT2A-rearranged infant acute lymphoblastic leukaemia.
Register of on-topic literature known to exist and not retrievable by this project's APIs · Oncofusion knowledge graph (internal register) 2026What this evidence cannot show
The 2026 report is title-only: no abstract was retrievable from NCBI efetch or Europe PMC on 2026-09-22, so which inhibitors differ and by how much is unknown here. Covalent-versus-reversible is this entry's hypothesis about that difference, not the paper's stated finding.
Evidence that argues against this (2)
- Differential in vitro sensitivity between agents is routine and usually reflects potency, protein binding and assay duration rather than modality. Attributing an unread result to the axis this entry finds interesting is exactly the reasoning the unreadable register exists to prevent, and the entry should be read as naming a hypothesis to test when the paper becomes available. Oncofusion knowledge graph (internal register) 2026
- Four of the five clinical compounds are reversible, so any in vitro spread among them cannot be a covalency effect at all. The modality hypothesis only engages if icovamenib is the outlier, and it is the one compound whose oncology programme was terminated and which is therefore least likely to be in a 2026 comparison. Journal of Clinical Oncology 2025