Competitive release and adaptive therapy — what ecology says happens when you remove the dominant species
also called mesopredator release · competitive release · adaptive therapy · refugia
In an ecosystem, removing the dominant species does not leave an empty space; the species it was holding down expand into it. Wiping out the B-cell leukemia with a CD19 drug and watching a myeloid leukemia appear has exactly that shape.
The same thing, for a scientist
Community ecology has two results this disease keeps reproducing. Competitive release: suppressing a dominant competitor allows subordinate species already present at low abundance to expand, rapidly and predictably, into the vacated niche. And the therapeutic corollary developed in oncology from the same mathematics — adaptive therapy — which holds that maintaining a drug-sensitive population to competitively suppress a resistant one can extend time to progression, and which produced a pilot result in metastatic castrate-resistant prostate cancer with median time to progression of at least 27 months at 47% of standard cumulative dose [E1]. Lineage switch after CD19-directed therapy in KMT2A-rearranged leukemia has the exact signature of competitive release: median 1.5 months to switch, more than half from documented MRD-negative remission of the B-ALL, and concentrated in one genotype [E2]. What ecology adds is the claim that the expanding population was already present, and the prerequisite that adaptive therapy needs — a fitness cost to resistance — which in this disease has never been measured.
Competitive release, and what it predicts that transdifferentiation does not
Two accounts of lineage switch make the same clinical observation and different predictions.
- Transdifferentiation. A B-lineage blast reprograms into a myeloid one under pressure. Predicts no detectable myeloid-primed population before therapy, and that the switch clone carries somatic changes acquired after treatment began.
- Competitive release. A myeloid-primed subpopulation of the same clone was present at low abundance and expanded once its competitor was removed. Predicts the population is detectable at diagnosis at low frequency and that its phylogeny branches before therapy.
The phylogenetic evidence leans toward the second. Whole-genome reconstruction in a post-immunotherapy lineage switch showed the AML derived from the initial ALL, carrying distinct driver mutations that were detectable before the AML emerged, interpreted as extensive prior clonal diversification followed by selection 3 — though that case was KMT2A wild-type, which matters. In MLL/AF4 specifically, myeloid relapses shared fusion breakpoints with their matched lymphoid presentations and originated from a range of differentiation stages 4, and in one infant case ClonoSEQ found dominant IGH/IGL sequences present in the original B-ALL, the switched AML and the post-transplant relapse, with three later sequences already present at low abundance in the diagnostic sample 5.
None of that is a single-cell demonstration of a myeloid-primed population at diagnosis, which is the missing piece and the target of the generated experiment.
Adaptive therapy, and the prerequisite nobody has checked1 deeper
Adaptive therapy works only if the resistant population is competitively inferior in the absence of drug. In the prostate pilot, on-off cycling guided by tumour dynamics suppressed the androgen-independent population and lowered cumulative dose 1. In melanoma, resistant tumours were actively drug-dependent and regressed when dosing stopped.
For menin inhibitors that prerequisite is simply unknown. The nearest data point comes from CML, where BCR-ABL T315I and Y253H showed no growth advantage over wild type in the absence of imatinib in competitive repopulation assays 6 — neutrality rather than cost. If MEN1 interface mutants are likewise neutral, adaptive scheduling suppresses nothing and the entire ecological dosing argument collapses into the much weaker toxicity argument. That measurement is exp-men1-mutant-fitness-cost-competition, and it gates two hypotheses in this graph.
Refugia, briefly
Ecology also supplies refugia: spatial locations where a suppressed population escapes the pressure acting elsewhere. The central nervous system and extramedullary sites are the obvious counterpart, and lineage switch has been reported presenting as an extramedullary myeloid sarcoma during blinatumomab infusion. This is noted rather than developed, because it generates no prediction that CNS-directed therapy practice does not already embody.
Where the ecology breaks
Species in a community are genetically distinct lineages that cannot become one another. The populations here are clonally related and, on the transdifferentiation account, one can become the other — a transformation with no ecological counterpart. This is not a minor mismatch: it is the difference between the two competing accounts of lineage switch, and it means the ecological frame presupposes the answer to the question it is being used to investigate. The frame is therefore useful for generating the test and must not be used to argue for its outcome.
Competition in ecology is also mediated by measurable shared resources. What a B-lineage blast and a myeloid-primed sibling compete for in a marrow — niche space, stromal contact, a cytokine — is unspecified, and without it 'competitive release' is a description rather than a mechanism.
| domain | ecology |
|---|---|
| source concept | competitive release, mesopredator release, refugia, and adaptive therapy derived from Lotka-Volterra competition |
| principal disanalogy | Species cannot become one another; clonally related leukemic populations may be able to. The ecological frame presupposes the answer to the very question about lineage switch it is being used to investigate, so it may generate the test but must not be used to argue for its outcome. |
mapping
[
{
"bio": "the dominant CD19-positive B-lineage blast population",
"analog": "the dominant competitor in a community"
},
{
"bio": "a myeloid-primed subpopulation of the same clone at low abundance",
"analog": "a subordinate species held down by competition"
},
{
"bio": "CD19-directed therapy removing the dominant population",
"analog": "selective removal of the dominant competitor"
},
{
"bio": "lineage switch to AML within a median of 1.5 months",
"analog": "competitive release — rapid expansion into the vacated niche"
},
{
"bio": "the marrow niche, stromal contact and cytokine supply",
"analog": "the shared limiting resource, which is unspecified here and is a real weakness of the mapping"
},
{
"bio": "central nervous system and extramedullary sites",
"analog": "refugia where a suppressed population escapes the local pressure"
},
{
"bio": "intermittent dosing that deliberately preserves a drug-sensitive population",
"analog": "adaptive therapy, which requires resistance to carry a fitness cost"
},
{
"bio": "MEN1 interface mutants, whose fitness off drug has never been measured",
"analog": "the unmeasured competitive coefficient on which the whole adaptive-therapy branch depends"
}
]generated hypotheses
[ "hyp-lineage-switch-is-competitive-release-of-preexisting-clone", "hyp-adaptive-therapy-requires-a-fitness-cost" ]
- If lineage switch is competitive release of a pre-existing population, then deepening the response to CD19-directed therapy actively causes it, and MRD-negativity is a risk factor rather than a success.
- Adaptive therapy deliberately leaves leukemia in a child in order to suppress a resistant clone. In a disease where cure requires eradication and the hazard is front-loaded, that is a far harder trade than in metastatic prostate cancer, and the ecological mathematics does not capture it.
- Is a myeloid-primed, fusion-positive subpopulation detectable at diagnosis in KMT2A-rearranged infant ALL by single-cell multi-omics, and at what frequency?
- What is the shared limiting resource that a B-lineage blast and a myeloid-primed sibling compete for in a marrow, without which competitive release is a description rather than a mechanism?
- Does acute leukemia have any real-time burden signal precise enough to guide adaptive scheduling, or is the prostate precedent untransferable for that reason alone?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
An evolutionary game-theory model with three competing prostate cancer populations predicted that standard continuous abiraterone dosing strongly selects androgen-independent cells and fails at a median of about 16.5 months, whereas adaptive therapy using patient-specific tumour dynamics to time on-off cycles suppressed the resistant population; in a pilot trial 10 of 11 patients maintained stable oscillations of tumour burden with median time to progression of at least 27 months at 47% of standard cumulative dose.
Integrating evolutionary dynamics into treatment of metastatic castrate-resistant prostate cancer · Zhang J, Cunningham JJ, Brown JS, Gatenby RA · Nature Communications 2017What this evidence cannot show
A pilot of 11 patients with historical and contemporaneous comparison, in a solid tumour with a continuously measurable biomarker (PSA). Acute leukemia has no equivalent real-time burden signal at the resolution adaptive scheduling requires, which may be the binding obstacle to transferring the approach at all.
Evidence that argues against this (1)
- REFUTES A STATED PREREQUISITE OF THIS ANALOGY. The entry asserts that adaptive therapy requires a fitness cost to resistance, and treats the absence of a measured cost in this disease as the missing precondition. The mathematical-oncology literature says otherwise: reviewing new modelling work, Wodarz concludes that a clinical benefit from adaptive therapy 'can be achieved even in the absence of a fitness cost for resistant cells, which broadens the applicability of adaptive therapy.' The competitive-suppression mechanism can work through spatial and resource competition among equally fit populations. This cuts both ways for the graph: it removes the objection that adaptive therapy in KMT2A-rearranged leukaemia is blocked by an unmeasured cost, and it simultaneously removes the entry's reason for treating `exp-men1-mutant-fitness-cost-competition` as a gating experiment. The cost measurement is still worth having — it sets the size of the benefit — but it is not a precondition, and the sibling hypothesis `hyp-adaptive-therapy-requires-a-fitness-cost` is, as stated, wrong. Wodarz D 2021
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Post-immunotherapy lineage switch occurred at a median of 1.5 months (1 month after blinatumomab, 2.8 months after CAR T cells), 64.3% of a 70-case cohort carried a KMT2A rearrangement, 81.4% of switches occurred within six months, and 54.3% of patients had achieved MRD-negative complete remission of their B-ALL beforehand.
Project EVOLVE: an international analysis of postimmunotherapy lineage switch, an emergent form of resistance · Project EVOLVE consortium · Blood 2025What this evidence cannot show
A case series without a denominator; it characterises switches rather than estimating their rate.
Evidence that argues against this (1)
- Competitive release requires that the expanding population already existed. An equally good account of the same observations is direct reprogramming of the resident clone: in MLL/AF4 lineage switch, myeloid relapses shared fusion breakpoints with their lymphoid presentations, switching involved substantial chromatin-accessibility change and transcriptional rewiring, and the myeloid phenotype was recurrently associated with altered chromatin modifiers — with CHD4 perturbation alone sufficient to induce myeloid gene expression. Under that account there is no subordinate species waiting in the niche; the dominant species changes its own phenotype. The two accounts make the same clinical prediction but opposite experimental ones, which is precisely what `exp-diagnostic-subclone-archaeology-lineage-switch` has to resolve, and the entry should not present competitive release as the established reading. Tirtakusuma R et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Whole-genome phylogenetic reconstruction in a child with multiply relapsed ALL who developed AML after anti-CD19 cellular and antibody therapy showed the AML derived from the initial ALL and carried distinct driver mutations that were detectable before the AML emerged, interpreted as extensive prior clonal diversification followed by selection.
Clonal origin of KMT2A wild-type lineage-switch leukemia following CAR-T cell and blinatumomab therapy · Coorens THH, Behjati S, et al. · Nature Cancer 2023What this evidence cannot show
This case is explicitly KMT2A wild-type, as the title states. It is the cleanest phylogenetic support for pre-existing diversification and it is not in the genotype of interest.
Evidence that argues against this (1)
- Searched for published cases of post-immunotherapy lineage switch in which no pre-existing subclone could be found by deep sequencing — the observation that would separate competitive release from reprogramming. None was found; the phylogenetic reconstructions published to date (n=1 whole-genome case, plus one ClonoSEQ case report) both report pre-existing clonal structure, and no negative case has been published. That is a thin and possibly publication-biased evidence base for an analogy this entry treats as a close fit. Recorded so the gap is known to have been checked. Q12-analogy-counter (search record) 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Myeloid relapses after lineage switch in MLL/AF4 leukemia shared oncogene fusion breakpoints with matched lymphoid presentations and originated from a range of differentiation stages from immature progenitors through committed B-cell precursors, with switching involving chromatin-accessibility and transcriptional rewiring recurrently associated with altered chromatin modifiers including CHD4.
Epigenetic regulator genes direct lineage switching in MLL/AF4 leukemia · Tirtakusuma R, Szoltysek K, Milne P, et al. · Blood 2022What this evidence cannot show
Shared clonal origin is established; whether the myeloid population pre-existed as a distinct subpopulation or was reprogrammed is not resolved, and the CHD4 finding favours reprogramming.
Evidence that argues against this (1)
- Shared breakpoints establish common ancestry and not the direction of the transition, which is the whole question for a competitive-release account: outgrowth of a pre-existing myeloid sibling and switching of the lymphoid clone both produce matched breakpoints. Journal of Clinical Oncology 2025
- E5weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
In a 6-month-old infant with t(1;11) KMT2A::EPS15 B-ALL who switched to AML after blinatumomab, ClonoSEQ IGH/IGL sequencing found three dominant sequences present in the original B-ALL, the switched AML and the post-transplant relapsed B-ALL, with the relapse showing four additional sequences, three of which were present at low abundance in the original diagnostic sample.
Lineage Switch in an Infant B-Lymphoblastic Leukemia With t(1;11)(p32;q23); KMT2A/EPS15, Following Blinatumomab Therapy. · Du J, Chisholm KM, Tsuchiya K, et al. · Pediatric and Developmental Pathology 2021What this evidence cannot show
A single case report, and immunoglobulin-rearrangement clonality rather than single-cell lineage profiling. The authors themselves interpret it as reprogramming rather than as expansion of a pre-existing myeloid population.
Evidence that argues against this (1)
- Immunoglobulin rearrangement happens once, early, and is inherited by every descendant including genetically divergent ones, so three shared dominant sequences across three phases are consistent with either model. This is one patient. Blood Cells, Molecules and Diseases 2010
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In competitive repopulation assays in primary murine bone marrow and a murine CML transplantation model, BCR-ABL T315I and Y253H conferred no growth advantage over unmutated BCR-ABL in the absence of imatinib, while imatinib clearly selected for them.
The Bcr-Abl mutations T315I and Y253H do not confer a growth advantage in the absence of imatinib · Miething C, Feihl S, Mugler C, et al. · Leukemia 2006What this evidence cannot show
Neutrality, not cost. It is cited to show that the prerequisite for adaptive therapy fails, or at least is not established, in the closest leukemia precedent.
Evidence that argues against this (1)
- This evidence item is offered as the template for measuring a fitness cost, but read carefully it reports the *absence* of a differential in either direction: T315I and Y253H conferred no growth advantage without imatinib, which is not the same as conferring a disadvantage. A resistance allele that is fitness-neutral off drug gives adaptive therapy no lever in the classical model — and yet, per Wodarz, benefit can still arise without a cost. The honest position is that the fitness-cost axis is neither established as present nor established as necessary, and an experiment designed to detect a cost should be powered and interpreted accordingly rather than being treated as pass/fail for the whole strategy. Wodarz D 2021
the stated prerequisite of this analogy — that adaptive therapy needs a fitness cost to resistance — is contradicted in the mathematical-oncology literature
- analogous to →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
lineage switch read as competitive release after removal of the dominant competitor
- analogous to →Lineage switch as antigen escape from CD19-directed therapy in KMT2A-rearranged leukemia
the clinical risk entry this analogy reinterprets
- related to →MEN1 drug-interface mutations — the first proof that a chromatin drug can drive Darwinian escape in patients
adaptive therapy requires a fitness cost that this resistance route probably lacks
generated by competitive release
generated by the fitness-cost prerequisite of adaptive therapy
- analogous to →Lineage switch in KMT2A-rearranged leukemia — the same clone changes what kind of blood cell it is
removing a dominant clone lets a suppressed one expand