JUND liberation — the drug's mechanism of action and one of its mechanisms of resistance are the same event

Ozyerli-Goknar and colleagues asked what menin inhibitors do to JUND, a question the field had left open despite the shared pocket being known since 2012 4. In MLL-rearranged leukaemic cells they showed by quantitative mass spectrometry that menin-MLL inhibitors disrupt menin-JUND interactions as well as menin-MLL; RNA sequencing and chromatin binding assays then showed that inhibitor treatment increased JUND chromatin occupancy and upregulated JUND target genes, and that this contributes to resistance against menin-MLL inhibitors. CRISPR-mediated JUND inactivation or JNK inhibition synergistically enhanced the anti-leukaemic effect of the inhibitors — reduced proliferation, cell-cycle arrest, apoptosis — and JUND-deficient leukaemia cells produced a lower tumour burden in immunocompromised mice 1.

The logical shape of this result is unusual and worth stating plainly: the drug's therapeutic act (vacating the menin pocket) and this escape route (vacating the menin pocket) are the same physical event. There is no selection step, no mutation, no clonal outgrowth. Every treated cell receives the resistance mechanism at the same instant it receives the therapy, in proportion to how well the drug works.

Menin binds JunD directly and represses JunD-activated transcription, and MEN1 missense mutations found in patients disrupt that interaction 5. The structural work then showed one pocket, two peptides, opposite transcriptional consequences: menin-MLL1 promotes transcription, menin-JUND blocks JNK-mediated JUND phosphorylation and suppresses JUND-induced transcription 4. And in fibroblasts, JunD stripped of menin does not simply stop suppressing growth — it promotes it, whether the stripping is done by a JunD mutation that cannot bind menin or by a Men1-null background 3.

So the prediction was available for over a decade. What was missing, and what 1 supplies, is the demonstration in leukaemic cells under a menin-MLL inhibitor, with a chromatin and transcriptional readout and a genetic epistasis test.

PMID 42371796 currently appears in this graph inside the endocrine/ folder, where the framing is menin-inhibitor toxicity in hormone-producing tissue, and inside two analogy/ nodes as an illustration that the drug redistributes chromatin binding rather than only subtracting it. Both uses are correct. Neither records the paper's own headline finding, which is a therapy-resistance finding in leukaemia, in the disease this project is about, with an in vivo arm.

The consequence of the misfiling is practical. A reader of the four pre-existing resistance entries would conclude that escape from menin inhibition requires either a MEN1 mutation, a co-mutation, loss of PRC1.1, or a slow adaptive rewiring. JUND liberation is none of those: it is immediate, universal and dose-proportional.

The countermeasure is already named and already exists as a drug class. JNK inhibition phenocopied JUND knockout in synergising with menin-MLL inhibition 1, and the mechanism is specific — menin normally represses JUND by blocking JNK-mediated phosphorylation 4, so removing menin without removing JNK activity is exactly the condition that activates JUND.

This also reframes a combination question the field is currently arguing about. Most proposed menin-inhibitor partners (venetoclax, FLT3 inhibitors, hypomethylating agents) are orthogonal agents chosen to cover routes the menin inhibitor does not reach. A JNK inhibitor is different in kind: it is a partner chosen to neutralise a liability the menin inhibitor creates. If the effect size in patients resembles the preclinical effect size, a menin inhibitor given without such a partner is giving away part of its own efficacy.

One primary paper, published 2026. No patient data: the resistance contribution is established in cell lines and in a xenograft, not in a treated child or adult. The compounds are described as menin-MLL inhibitors rather than specifically revumenib, ziftomenib or bleximenib at clinical exposure, so the dose-response relationship between KMT2A displacement and JUND displacement — the number that would say how much of the drug's window is spent liberating JUND — remains unmeasured. That measurement is already specified as an experiment in this graph (exp-jund-displacement-coip-under-clinical-menin-inhibitor), which was written for endocrine safety reasons and would answer the efficacy question at the same time.