Diagnosis versus relapse in the same child: relapse comes from a subclone that was already there and already quiescent, shown twice, seven years apart, by two unrelated methods
also called relapse subclone state · quiescent subclone at diagnosis · paired diagnosis-relapse state · subclonal dynamics in infant ALL
When infant KMT2A-rearranged leukaemia comes back, it does not usually come from cells that changed during treatment. It comes from a group of cells that was already present at diagnosis and was already sitting quietly, which is why treatment aimed at the dividing bulk misses it.
The same thing, for a scientist
The brief asked whether paired diagnosis-versus-relapse data exist in the same patients. They do, in exactly the disease of interest, and they agree with the single-cell persister work from a completely different direction. Bardini et al. analysed Ig/TCR rearrangements in **MLL-AF4-positive infant ALL at diagnosis** and in xenografts generated from those same diagnostic samples, and established that the disease is **already subclonally branched at diagnosis**, frequently driven by an Ig/TCR-rearranged founder clone. Some clones were **largely quiescent at diagnosis but reactivated and dominated on serial transplantation**, while clones dominant at diagnosis became more quiescent — a dynamic competition between proliferating and quiescent subclones. Investigation of **paired diagnostic and relapse samples suggested that relapses often occur from subclones already present but more quiescent at diagnosis**, with copy-number alterations identified at relapse possibly contributing to their activation; and each identified subclone could contribute to the phenotypic pool of leukaemia-propagating cells [E1]. Seven years later, single-cell transcriptomics of diagnostic samples from infants with MLL-rearranged ALL identified relapse-associated cells characterised by **a quiescent programme with stemness properties** whose abundance predicted relapse [E2]. Clonal genetics and single-cell transcriptomics, with no methodological overlap, converge on the same statement: **the relapse compartment is present and quiescent before treatment starts.**
Why the convergence is the finding
Bardini's instrument is clonal: Ig/TCR rearrangement marks lineage-related cells and xenografts test their functional capacity. It cannot see a transcriptional state.
Candelli's instrument is transcriptional: single-cell expression scored against a prednisone signature. It cannot see clonal relationships.
Each identified a quiescent minority at diagnosis that goes on to cause relapse. Neither could have found the other's evidence, and neither depends on the other's assumptions. In a field where most convergences are between studies sharing a method, this one is worth more than either result alone.
Bardini adds something Candelli cannot: the direction is reversible. Clones dominant at diagnosis became more quiescent while quiescent ones reactivated 1. So this is not a fixed subpopulation with a fixed identity; it is a competition whose winner changes when the selective pressure changes. That matches cell-state-quiescence's record that dormancy in ALL xenografts reverses on removal from the niche.
What changed at relapse, and what did not
Did not change: the founding lesion. This graph's mechanism-fusion-clonal-stability-diagnosis-to-relapse records that the fusion itself is stable from diagnosis to relapse, and Bardini's founder-clone structure is consistent with that.
Did change: copy-number alterations were identified at relapse and might contribute to the activation and expansion of previously quiescent subclones 1. Note the authors' own hedging — 'might contribute' — and note what it implies: the genetic change is a candidate activator of a pre-existing clone, not the origin of a new one.
Changed at the state level, in the broader disease: in paired diagnosis/residual/relapse single-cell profiling of paediatric B-ALL, leukaemic cells at relapse shifted toward poorly differentiated states, while the changes in residual cells were more complicated 3. That is a different observation from Bardini's — a state shift in the bulk rather than a clonal substitution — and the two are compatible if the surviving clone is also the less differentiated one.
The bounded absence: nobody has profiled paired infant KMT2A-rearranged samples at single-cell resolution
Searched 2026-09-20 on PubMed:
infant acute lymphoblastic leukemia AND (paired OR matched) AND relapse AND (gene expression OR transcriptomic OR single-cell)— 33 records, title-checked. The only one describing paired diagnosis and relapse material in MLL-rearranged infant ALL is Bardini 2015 1, which is Ig/TCR clonal analysis, not transcriptomics.(MLL OR KMT2A) AND infant leukemia AND minimal residual disease AND single-cell— 0 records.querytranslationread and confirmed sane: all three concept blocks expanded normally, so this is a real absence and not a parser artefact.infant acute lymphoblastic leukemia AND single-cell RNA sequencing— 25 records, title-checked; Candelli 2022 profiles diagnostic samples only, and no record profiles infant KMT2A-rearranged material at residual disease or at relapse.
So: paired-timepoint single-cell profiling in infant KMT2A-rearranged ALL does not exist. The paired single-cell work is paediatric B-ALL broadly 3; the infant-specific single-cell work is diagnosis-only 2; the infant-specific paired work is clonal rather than transcriptional 1. The three cover the space between them and none covers the centre of it.
| state definition | The relationship between the cell states present at diagnosis and those present at relapse in the same patient; in MLL-AF4-positive infant ALL the relapse population is a subclone already present at diagnosis and relatively quiescent there |
|---|---|
| what is stable | the fusion itself, recorded in mechanism-fusion-clonal-stability-diagnosis-to-relapse |
| what is acquired | copy-number alterations identified at relapse, proposed by the authors as possible activators of previously quiescent subclones rather than as the origin of new ones |
two independent lines
[
{
"method": "Ig/TCR clonal architecture plus serial xenotransplantation, paired diagnosis and relapse samples",
"conclusion": "relapse often arises from subclones present but more quiescent at diagnosis; competition between proliferating and quiescent subclones is bidirectional",
"pmid": "24798483",
"title": "Clonal variegation and dynamic competition of leukemia-initiating cells in infant acute lymphoblastic leukemia with MLL rearrangement.",
"authors": "Bardini M, Woll PS, Corral L, et al.",
"venue": "Leukemia",
"year": 2015,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"method": "single-cell transcriptomics of diagnostic samples, two platforms, external prednisone signature",
"conclusion": "relapse-associated cells at diagnosis are quiescent with stemness features and basal glucocorticoid-response activation",
"pmid": "34304246",
"title": "Identification and characterization of relapse-initiating cells in MLL-rearranged infant ALL by single-cell transcriptomics.",
"authors": "Candelli T, Schneider P, Garrido Castro P, et al.",
"venue": "Leukemia",
"year": 2022,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]bounded absence
{
"claim": "no published single-cell profiling of paired diagnosis and relapse (or residual disease) material in infant KMT2A-rearranged ALL",
"queries_run_2026_09_20": [
{
"query": "(MLL OR KMT2A) AND infant leukemia AND minimal residual disease AND single-cell",
"count": 0,
"querytranslation_checked": true
},
{
"query": "infant acute lymphoblastic leukemia AND (paired OR matched) AND relapse AND (gene expression OR transcriptomic OR single-cell)",
"count": 33,
"on_beat_after_title_check": 1
},
{
"query": "infant acute lymphoblastic leukemia AND single-cell RNA sequencing",
"count": 25,
"on_beat_after_title_check": 1
}
]
}atlas refs
[
{
"label": "Relapse from pre-existing quiescent subclones in MLL-AF4 infant ALL, paired samples",
"pmid": "24798483",
"title": "Clonal variegation and dynamic competition of leukemia-initiating cells in infant acute lymphoblastic leukemia with MLL rearrangement.",
"authors": "Bardini M, Woll PS, Corral L, et al.",
"venue": "Leukemia",
"year": 2015,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"label": "Quiescent stem-like relapse-associated cells at diagnosis, single-cell",
"pmid": "34304246",
"title": "Identification and characterization of relapse-initiating cells in MLL-rearranged infant ALL by single-cell transcriptomics.",
"authors": "Candelli T, Schneider P, Garrido Castro P, et al.",
"venue": "Leukemia",
"year": 2022,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]- If the relapse clone is present and quiescent at diagnosis, then intensifying induction targets the wrong compartment harder, and the historical response to poor early response in this disease has been to intensify.
- A therapy that kills the proliferating bulk creates the selective pressure under which quiescent subclones were shown to reactivate and dominate — so the more effective the induction, the cleaner the selection.
- Copy-number alterations appearing at relapse mean that a strategy of holding dormant cells dormant indefinitely is betting against an accumulating genetic process in a child with decades of life ahead.
- Predicting relapse from a diagnostic single-cell assay would change treatment allocation on the basis of a state that the clonal data show is reversible, so the prediction is of current occupancy rather than of fixed identity.
- Is the quiescent subclone of Bardini the quiescent transcriptional state of Candelli? A single experiment combining clonal tracking with single-cell transcriptomes on the same diagnostic sample would answer it, and both technologies are now routine.
- Is the copy-number alteration at relapse necessary, or only an accelerant of a clone that would have relapsed anyway? The distinction decides whether relapse prevention means eliminating dormant cells or preventing a genetic event.
- Does a menin inhibitor change the clonal competition? If quiescent and proliferating subclones trade places under pressure, a drug that kills the proliferating compartment could actively select the quiescent one, and nobody has tracked clonal composition through menin-inhibitor treatment.
- Why has nobody profiled infant KMT2A-rearranged ALL at single-cell resolution at residual disease or relapse? The samples are taken as a matter of routine care; the gap appears to be one of assembly rather than availability.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Ig/TCR rearrangement analysis of MLL-AF4-positive infant ALL at diagnosis and of xenograft leukaemias from mice transplanted with the same diagnostic samples established that the disease is composed of a branching subclonal architecture already at diagnosis, frequently driven by an Ig/TCR-rearranged founder clone; some clones largely quiescent at diagnosis reactivated and dominated on serial transplantation while clones dominant at diagnosis became more quiescent; paired diagnostic and relapse samples suggested that relapses often occur from subclones already present but more quiescent at diagnosis, with copy-number alterations identified at relapse possibly contributing to their activation and expansion; and each identified subclone was able to contribute to the phenotypic pool of MLL-AF4-positive leukaemia-propagating cells.
Clonal variegation and dynamic competition of leukemia-initiating cells in infant acute lymphoblastic leukemia with MLL rearrangement · Bardini M, Woll PS, Corral L, Luc S, Wittmann L, Ma Z, et al., Jacobsen SE · Leukemia 2015What this evidence cannot show
The paired diagnosis-relapse conclusion is stated as a suggestion by the authors, from a small number of paired samples whose count is not given in the abstract; relapse material in infant ALL is scarce. Quiescence here is inferred from clonal behaviour on serial transplantation, not measured by cycle status or label retention in the patient. Serial transplantation into immunodeficient mice imposes its own selection, and a clone that dominates in a mouse is not thereby shown to be the clone that would relapse in the child. Ig/TCR rearrangement is a clonal marker and does not report cell state. Read: abstract only.
Evidence that argues against this (1)
- The same study identifies copy-number alterations at relapse that 'might contribute to the activation and expansion of previously quiescent subclones', which admits an alternative reading of the whole result: relapse may require an acquired genetic event and the pre-existing quiescent subclone may be necessary but not sufficient. On that reading, preventing relapse would mean preventing the acquisition rather than eliminating the quiescent compartment, and therapies aimed at the dormant cells would not by themselves change outcome. The paper does not distinguish these. Read: abstract only (same paper). Bardini M et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Single-cell RNA sequencing on two platforms of diagnostic bone marrow and peripheral blood from infants with MLL-rearranged ALL, scored against an independently derived prednisone-dependent signature, identified leukaemic cells associated with high relapse risk characterised by basal activation of the glucocorticoid response, smaller size and a quiescent gene expression programme with cell stemness properties, and quantification of resistant versus sensitive cells predicted future relapse in individual patients.
Identification and characterization of relapse-initiating cells in MLL-rearranged infant ALL by single-cell transcriptomics · Candelli T, Schneider P, Garrido Castro P, Jones LA, Bodewes E, Rockx-Brouwer D, et al., Stam RW · Leukemia 2022What this evidence cannot show
Diagnostic samples only; no relapse or residual-disease material was profiled, so the link to relapse is a prediction from a diagnostic measurement and not an observation of the same cells at two timepoints. Cohort size is not stated in the abstract. The resistant population is defined by one externally derived prednisone signature, so 'relapse-initiating' names an inference rather than a demonstrated function. Read: abstract only.
Evidence that argues against this (1)
- A quiescent, stem-like minority at diagnosis is not by itself unusual, and the direction of clonal competition in this disease is documented to be reversible: clones dominant at diagnosis became more quiescent while quiescent clones reactivated on serial transplantation. If quiescence is a position clones move into and out of under changing pressure rather than a fixed property of particular cells, then a diagnostic snapshot measures the current occupancy of the state and not the identity of the cells that will occupy it under therapy - which is a weaker basis for prediction than the entry's framing implies. Read: abstract only. Bardini M et al. 2015
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 161,986 single-cell transcriptomes with paired B-cell-receptor sequencing from paediatric B-ALL sampled at diagnosis, residual disease and relapse, leukaemic cells at relapse tended to shift to poorly differentiated states relative to diagnosis, whereas the changes in the residual cells were more complicated, and hypoxia-pathway activation distinguished residual cells, resistant clones and MLL-rearranged B-ALL.
Elucidating minimal residual disease of paediatric B-cell acute lymphoblastic leukaemia by single-cell analysis · Zhang Y, Wang S, Zhang J, Liu C, Li X, Guo W, et al., Cheng T · Nat Cell Biol 2022What this evidence cannot show
Paediatric B-ALL broadly rather than infant KMT2A-rearranged disease; the MLL-rearranged subgroup size and age range are not stated. A shift toward 'poorly differentiated states' at relapse is a transcriptional description against a reference and could reflect clonal substitution, state change within a clone, or both - the paired B-cell-receptor sequencing could in principle distinguish these but the abstract does not report that it did. Read: abstract only.
Evidence that argues against this (1)
- A relapse shift toward poorly differentiated states is hard to interpret in KMT2A-rearranged infant disease because the diagnostic state is already extremely immature: single-cell projection against a developing marrow reference found this disease uniquely dominated by an early lymphocyte precursor state, with hybrid myeloid-lymphoid features, and the same study established that transcriptional state does not report cell of origin. If the blasts start at the earliest lymphoid position in the atlas, there is little room for a further shift in that direction, and a 'poorly differentiated' relapse signature derived across all paediatric B-ALL may not describe a change in this subtype at all. Read: abstract only. Khabirova E et al. 2022
clonal genetics and single-cell transcriptomics converge with no shared method: the relapse compartment is present and quiescent at diagnosis
- expressed in →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
paired diagnosis and relapse samples in MLL-AF4-positive infant ALL; branching subclonal architecture already present at diagnosis
the fusion is stable while the dominant subclone is not; what changes at relapse is which pre-existing clone is in front, plus acquired copy-number alterations
quiescence here is measured as clonal behaviour on serial transplantation rather than by label retention, and it is bidirectional