TP53 inactivation and the apoptotic threshold — the drug still switches the program off, and the cell still does not die
also called TP53-mutant de novo revumenib resistance · MCL-1 dependence in TP53-mutant menin-inhibitor resistance
A menin inhibitor kills leukaemia cells only if those cells are still able to carry out the self-destruct program. Leukaemias that have lost the p53 gene keep the self-destruct machinery switched off, so the drug can do its job on the genes and the cell simply survives anyway.
The same thing, for a scientist
In a panel of AML cell lines and an isogenic model, TP53 mutation was associated with de novo resistance to revumenib, with impaired induction of TP53 transcriptional targets and deregulated expression of the BH3 proteins BCL-2 and MCL-1; the MCL-1 inhibitor MIK665, but notably **not** venetoclax, preferentially sensitised TP53-mutant AML cells to revumenib [E1]. This is a different class of mechanism from the four already in this graph: the fusion program is still being shut off, the drug is still on target, and the cell survives because the execution step downstream of the transcriptional effect is broken. The clinical literature is consistent but weaker and, importantly, partially contradictory in the direction of causality: TP53 mutations emerged alongside acquired menin-inhibitor resistance in serial patient-derived xenografts of multiply pretreated KMT2A::AFF1 ALL but were **not sufficient to induce resistance in vitro** in that study [E2]. The honest position is that TP53 status is a strong candidate de novo biomarker with one direct isogenic demonstration that remains a preprint, and one xenograft study that could not reproduce sufficiency.
Why this is a different kind of resistance
The four resistance entries that pre-date this one all describe ways the transcriptional output survives the drug: the drug cannot bind (MEN1 interface mutation), the output is re-created from a different set of loci (PRC1.1/MYC), a parallel signalling input substitutes for it (RAS/FLT3), or the maintenance of the program is rewired (adaptive non-genetic). TP53 loss is not any of these. The drug binds, menin is evicted, the canonical program goes down — and the cell does not die, because killing a leukaemia blast by removing its transcriptional driver requires an intact apoptotic apparatus to finish the job.
That distinction matters operationally. Every other mechanism can in principle be countered by getting the target engagement deeper or more durable. This one cannot. A perfect menin inhibitor would not help.
The specific finding, and why the negative arm is the interesting part
D'Souza and colleagues analysed a panel of AML lines and built an isogenic TP53 model. TP53 mutation was associated with de novo revumenib resistance, impaired induction of TP53 transcriptional targets, and deregulated BCL-2 and MCL-1 expression. Then the pharmacology: the MCL-1 inhibitor MIK665, but not venetoclax, preferentially sensitised TP53-mutant AML cells to revumenib 1.
The venetoclax negative is the load-bearing part for this project. Venetoclax is the default combination partner for menin inhibitors — it is the countermeasure named in this graph's PRC1.1 entry, it produced the only complete remissions after menin-inhibitor failure in the largest clinical series 3, and menin-plus-venetoclax-plus-hypomethylating-agent triplets now give the best pooled complete-response rates in relapsed/refractory AML 4. If TP53-mutant disease is MCL-1-dependent rather than BCL-2-dependent, then the standard combination is the wrong combination in exactly the subgroup with the worst outcomes, and the field would not notice because TP53-mutant patients do badly on everything.
The contradiction with the xenograft series, stated plainly
This entry cannot be written as a settled mechanism, because a second primary study points the other way on causality. In serially sampled patient-derived xenografts from two patients with multiply relapsed KMT2A::AFF1 ALL, late samples were markedly less menin-inhibitor-sensitive than early samples from the same patients, and RAS pathway and TP53 co-mutations emerged alongside that resistance — but neither was sufficient to induce menin-inhibitor resistance in vitro 2.
The two results are reconcilable but only with assumptions. 1 is AML, mostly cell lines plus an isogenic pair, testing de novo (intrinsic) resistance; 2 is KMT2A::AFF1 ALL patient material testing whether introducing the lesion confers resistance. A lesion can be a reliable marker of an intrinsically resistant state without being able to create that state when introduced alone. But that is an interpretation, not a finding, and this graph should not record it as one.
The existing entry resistance-ras-flt3-wt1-comutation-menin already lists TP53 among its co-mutation genes and records the insufficiency result. This entry does not contradict it; it separates out the apoptotic-threshold mechanism, which that entry does not describe, and adds the MCL-1 versus BCL-2 pharmacology, which is the decision-relevant part.
Status of the source
The central citation 1 is a bioRxiv preprint (2025), indexed in PubMed as PMID 40196700. A search of PubMed for a published version (TP53[tiab] AND revumenib[tiab], and an author search on the first author combined with menin/revumenib) returned only the preprint as of 2026-09-19; the other two hits were reviews. So this mechanism rests on unrefereed work and one associative clinical signal, and the entry is marked moderate accordingly. If the peer-reviewed version appears with the isogenic result intact, this becomes one of the most actionable biomarkers in the beat, because TP53 status is already measured in every patient.
| kind | resistance |
|---|---|
| mechanism class | downstream / apoptotic-threshold (on-target transcriptional effect preserved) |
| seen in patients | associative only — TP53 mutations co-emerge with resistance in patient-derived xenografts; no prospective patient cohort has reported menin-inhibitor response stratified by TP53 status |
| seen in models | true |
| causality | demonstrated in an isogenic AML model in a preprint [E1]; NOT reproduced as sufficient in KMT2A::AFF1 ALL cells [E2]. Contested. |
| timing | de novo / intrinsic rather than acquired |
| detectability | trivially detectable — TP53 is on every myeloid and paediatric leukaemia panel already; the gap is that nobody reports menin-inhibitor outcome stratified by it |
| predicted biomarker | TP53 mutation or 17p loss at diagnosis predicts shallow response to single-agent menin inhibition and predicts MCL-1 rather than BCL-2 dependence |
| unresolved contradiction | E1 reports TP53 inactivation confers resistance in an isogenic AML model; E2 reports TP53 mutation is not sufficient to induce resistance in KMT2A::AFF1 ALL cells. Neither study tested the other's system. This entry records the disagreement rather than resolving it. |
affects
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib", "drug-enzomenib", "drug-icovamenib" ]
genes
[ "TP53", "MCL1", "BCL2" ]
countermeasures
[
{
"agent": "MCL-1 inhibition (MIK665 / S64315)",
"rationale": "preferentially sensitised TP53-mutant AML cells to revumenib where venetoclax did not",
"tier": "preclinical, preprint"
},
{
"agent": "venetoclax",
"rationale": "the field default, and specifically reported NOT to sensitise TP53-mutant cells to revumenib",
"tier": "preclinical negative — this is a counter-indication, not a countermeasure"
},
{
"agent": "allogeneic transplant or immunotherapy",
"rationale": "cell-extrinsic killing does not require intact p53-dependent apoptotic priming",
"tier": "clinical rationale, untested in this specific setting"
}
]pre emption strategies
[ "stratify every reported menin-inhibitor cohort by TP53 status retrospectively — the data already exist in trial databases and no new experiment is needed", "do not assume the menin+venetoclax+HMA triplet covers TP53-mutant disease; test menin plus an MCL-1 inhibitor separately", "in KMT2A-rearranged infant ALL, check the base rate first: TP53 mutation is uncommon at diagnosis in this genotype, which may make the whole mechanism a relapsed-adult-AML problem" ]
- This mechanism defeats the interface hypothesis without touching the interface. Perfect, complete, durable eviction of menin from the fusion would not rescue a cell that cannot execute apoptosis, so improving the drug is not the answer here.
- The standard combination may be actively wrong in this subgroup. Venetoclax is the field's default menin partner and is specifically reported not to sensitise TP53-mutant cells, so the triplet now being rolled out could be delivering venetoclax toxicity with no venetoclax benefit to precisely the patients with the fewest options.
- MCL-1 inhibition has a known essential-function problem. MCL-1 is required for cardiomyocyte and haematopoietic stem cell survival, and MCL-1 inhibitor development has been repeatedly interrupted by cardiac toxicity signals; proposing it for an infant needs that stated up front.
- The evidence base is a preprint. Writing this entry at all is a judgement call; if the isogenic result does not survive review, the graph will carry a mechanism that was never established, which is why the confidence is moderate and the contradiction with the xenograft study is in the body rather than only in the counter-evidence array.
- Selection ambiguity: if TP53-mutant clones are selected by prior chemotherapy rather than by the menin inhibitor, then the observed association between TP53 and menin failure is partly an artefact of who gets offered a menin inhibitor, and a frontline trial would show a different picture.
- What is the menin-inhibitor response rate in TP53-mutant versus TP53-wild-type KMT2A-rearranged disease? Every registration trial has the sequencing to answer this and none has reported it.
- Is the MCL-1 versus BCL-2 dependence switch real in KMT2A-rearranged cells, or is it an NPM1/AML-context finding? The whole clinical implication turns on this and it has been tested in one preprint in AML lines.
- What is the prevalence of TP53 alteration in KMT2A-rearranged infant ALL at diagnosis? If it is near zero, this mechanism is irrelevant to the population this project cares most about, and saying so would be a useful negative.
- Can a functional assay replace the genotype? Dynamic BH3 profiling would measure apoptotic priming directly and would capture TP53-wild-type cells that are unprimed for other reasons, which the genotype misses.
- Does the emergence of TP53 mutation in pretreated samples reflect selection by prior chemotherapy rather than by menin inhibition? Both PDX patients were multiply relapsed and menin-naive, which points at chemotherapy, and that would argue for moving menin inhibitors earlier.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.preprint
In a panel of AML cell lines and an isogenic model, TP53 mutations were associated with de novo resistance to revumenib, impaired induction of TP53 transcriptional targets, and deregulated expression of the BH3 proteins BCL-2 and MCL-1; the MCL-1 inhibitor MIK665, but not venetoclax, preferentially sensitised TP53-mutant AML cells to revumenib.
TP53 Inactivation Confers Resistance to the Menin Inhibitor Revumenib in Acute Myeloid Leukemia. · D'Souza J, Leung CJ, Ballapuram AC, et al. · bioRxiv 2025What this evidence cannot show
Preprint, not peer reviewed; no published version located in PubMed as of 2026-09-19. AML cell lines plus an isogenic pair; no patient cohort and no KMT2A-rearranged ALL. Abstract read in full via NCBI efetch; full text not obtained, so the number of lines, the specific TP53 alleles and the effect sizes are not recorded here.
Evidence that argues against this (1)
- In serial patient-derived xenografts of multiply relapsed KMT2A::AFF1 ALL, TP53 mutations emerged alongside menin-inhibitor resistance but were NOT sufficient to induce resistance in vitro. This is a direct challenge to a causal reading of [E1] in the ALL setting. Mahdavi L et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In serially sampled patient-derived xenografts from two patients with multiply relapsed KMT2A::AFF1 ALL, later samples were much less sensitive to menin inhibition than earlier samples from the same patients despite no menin-inhibitor exposure; transcriptomic analysis documented sustained on-target efficacy against the canonical menin-inhibitor targets in the resistant cells, and multiple co-mutations emerged including RAS pathway and TP53 mutations, although neither was sufficient to induce menin-inhibitor resistance in vitro.
Upfront menin-inhibitor resistance in multiply pretreated leukemias. · Mahdavi L, Alikarami F, Goodrow H, et al. · Exp Hematol 2026What this evidence cannot show
Serial samples from only two patients for the resistance arm. Abstract read in full via NCBI efetch. 'Not sufficient in vitro' is a statement about single-lesion introduction into a permissive background, which is a weaker test than it sounds: a lesion can mark a resistant state without creating it.
Evidence that argues against this (1)
- The PDX result shows resistance arising WITHOUT menin-inhibitor exposure in multiply pretreated patients, which makes it a poor model of acquired resistance to menin inhibition specifically. Prior therapy is the exposure that differs, so what is measured is upfront resistance in heavily treated disease, and attributing it to apoptotic priming requires ruling out everything else those patients received. Experimental Hematology 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 84 adults with relapsed/refractory AML after menin-inhibitor failure, all complete remissions on subsequent therapy occurred with venetoclax-containing regimens or with menin-inhibitor switching, the CR/CRi rate for non-trial therapies was 19% and overall response 32%, and 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. · Chin KK, Ball BJ, Abaza Y, et al. · Blood Adv 2026What this evidence cannot show
Retrospective, four US centres, adults only; 84 patients received a menin inhibitor and only 50 were treated after failure, so all response figures rest on very small numbers. Abstract read in full via NCBI efetch on 2026-09-19; title, authors and DOI taken from that record. TP53 status is not reported separately.
Evidence that argues against this (1)
- Same selection problem as the cross-resistance entry, and it bounds this one too: a post-failure cohort in which 77% were venetoclax-exposed cannot establish that venetoclax-containing regimens are what rescues menin-inhibitor failure. The CR/CRi rate for non-trial therapy in that setting was low, which is the more transferable number. Blood Advances 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review
In a systematic review and meta-analysis of 14 studies and 784 treated adults with relapsed/refractory AML, pooled overall response to menin-inhibitor-based therapy was 54.6%, pooled CR 29.3% and CR+CRh 28.5%; combination with a hypomethylating agent plus venetoclax produced significantly higher CR (43.3% vs 19.5%, p=0.002) and CR+CRh (48.6% vs 25.8%, p=0.007) than monotherapy.
Menin inhibitors for patients with relapsed/refractory acute myeloid leukemia (AML): a systematic review and meta-analysis. · Alhajahjeh A, Beke KE, Grimshaw AA, et al. · Leuk Lymphoma 2026What this evidence cannot show
Meta-analysis across heterogeneous studies (I2 = 63.5% for ORR), adults only, no paediatric data, and the monotherapy versus triplet comparison is across studies rather than randomised. Cited here only to establish that venetoclax-containing combination is the field's default, which is what makes the TP53/MCL-1 negative result consequential.
Evidence that argues against this (1)
- The pooled meta-analytic response rate hides the distinction the clinic turns on. Pooled overall response was 54.6% but pooled CR only 29.3%, so most responses are not remissions, and the pooling runs across 14 studies with different agents, genotypes, lines of therapy and response criteria. A single headline ORR across that heterogeneity is not an estimate of what any one patient faces. Leukemia & Lymphoma 2026
the priming explanation of menin-inhibitor failure: transcriptional effect intact, execution step broken; MCL-1 inhibition but not venetoclax sensitises TP53-mutant cells
- ← supportsSix countermeasures, no head-to-head — so rank their targets: CDK6 is 13th and CDK4 is 16,052nd
Both its countermeasures clear the top 1% - MCL1 rank 18, BCL2 rank 36 - which is the strongest single-gene support any countermeasure in the beat has.
the IntAct MEN1-TP53 row is a TP53 R282W neo-interaction, unrelated to apoptotic priming
- confers resistance to →Revumenib (SNDX-5613, Revuforj) — first approved menin–KMT2A interface inhibitor
TP53 inactivation confers de novo resistance despite preserved on-target transcriptional effect
- related to →BCL2 — the one co-dependency in this beat that already has an approved drug pointed at it
BCL-2 and MCL-1 expression deregulated; venetoclax specifically did NOT resensitise
counter-indication: venetoclax did not sensitise TP53-mutant AML cells to revumenib, where an MCL-1 inhibitor did
same TP53 observation, different mechanism: that entry records co-mutation association, this one the apoptotic-threshold mechanism
the venetoclax countermeasure named there is reported not to work in TP53-mutant cells
adds a downstream/apoptotic route to the route-by-route co-occurrence matrix
entry produced by this enumeration