p279 and p318: one infant, two diseases, and an oncoprotein target map that does not carry over
One baby's leukaemia started as a B-cell type, was treated, and came back as a myeloid type. Profiling the cancer protein in both shows the list of genes it controls changed as much as if it were a different child.
The same thing, for a scientist
p279 and p318 are two chromatin-profiled states of a single infant enrolled at Texas Children's Hospital, deposited in GSE252378: a KMT2A::AFF1 infant B-ALL and the AML it lineage-switched into. The paper states the switch happened in response to ALL-directed chemotherapy, and that in p318 the oncoprotein level is reduced. Re-measured on one promoter universe with one calling rule and an activity-stratified null, 54.6% of the B-ALL's oncoprotein-target promoters are still called in the AML and 40.7% of the AML's calls are new - an overlap that sits at the 37th percentile of 27 unrelated KMT2A::AFF1 sample pairs. This is a chromatin-level case, not a clinical narrative: the deposit records no age in months, no treatment detail, no outcome, and no consent-bearing clinical record. All of that is absent by design.
What the public record does and does not say about this child
Does say (GEO !Sample_description and the publication): infant B-ALL, KMT2A::AFF1, Texas Children's Hospital; paired with p318; p318 is the AML resulting from the lineage switch; the switch occurred in response to ALL-directed chemotherapy; oncoprotein levels are reduced after the switch, with a corresponding fall in the fusion transcript's expression; the paper names B-ALL-specific and GMP-like target gene sets that differ between the two samples.
Does not say: age in months, sex, exact regimen, time to switch, whether a menin inhibitor was ever given (it was not, in this pair - the menin-inhibitor switch is the separate 148752/152985 case), MRD trajectory, or outcome. No inference about this child's course beyond the above is supportable from the public record, and this entry makes none.
Why this one pair carries so much weight in this corpus
It is the only chemotherapy-driven lineage switch in the public record with paired KMT2A N-terminal and C-terminal oncoprotein profiling at both timepoints. The second switch pair (148752/152985) has a post-switch sample with undetectable oncoprotein because it was taken during revumenib, so it cannot contribute an overlap at all. Three beats of this corpus have now argued about a single patient, which is the strongest available argument for the serial replication experiment rather than a fourth re-analysis.
What a clinician could take from it, and what they could not
Could: after a lineage switch, roughly half the oncoprotein-target genes are different, so a target list from the diagnostic sample describes the new disease about as well as a list from another child with the same fusion would.
Could not: anything about this individual's prognosis, the timing of the switch, or whether the switch was preventable. And nothing about menin-inhibitor response, because this child was not on one.
| age at dx | infant (exact age not stated in the public record) |
|---|---|
| fusion | fusion-kmt2a-aff1 |
| course | Lineage switch to AML in response to ALL-directed chemotherapy (sample p318), with reduced oncoprotein score and reduced fusion transcript expression. No menin inhibitor in this case. |
| outcome | not stated in the public record |
| source type | peer-reviewed publication plus its public data deposit |
| consent note | public sources only; GEO records carry no PHI and none is reproduced here. No age in months, sex, regimen, timing or outcome is available and none is inferred. |
presentation
[ "infant B-ALL, KMT2A::AFF1, Texas Children's Hospital (sample p279)" ]
chromatin measurements — the numbers behind the chart
{
"retained_target_promoters_pct": 54.6,
"new_target_promoters_in_aml_pct": 40.7,
"percentile_among_unrelated_aff1_pairs": 37
}- The clinical instinct after a lineage switch is to re-target the new phenotype. If the oncoprotein target map is no more this child's than any other child's, then the mechanistic rationale for a patient-specific salvage target is weak, and a fusion-directed agent is the better-supported choice.
- How long after diagnosis did the switch occur? The deposit carries no dates and the paper does not state an interval, which matters because a switch at 4 weeks and one at 18 months are different biological events.
- Was this child's B-ALL already IRX1-positive and HOXA-low at diagnosis - the signature this corpus associates with switch-prone infant disease? An earlier beat found p279 positive for both HOXA and IRX1 occupancy, which is the double-positive state; whether that predicted the switch is untestable at n=1.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
p279 is a KMT2A::AFF1 infant B-ALL that lineage switched in response to ALL-directed chemotherapy, and p318 is the AML that resulted; in p318 the oncoprotein score and the fusion transcript's expression are reduced.
MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, Mullighan CG, Yi JS, Meshinchi S, Sarthy JF, Ahmad K, Henikoff S · Nature Communications 15:9341 2024What this evidence cannot show
Read from the article full text. The GEO sample descriptions call them 'lineage-switching' and record the pairing but do NOT mention chemotherapy - the chemotherapy attribution is the paper's, and an agent reading GEO alone would not find it.
Evidence that argues against this (1)
- The paper reports that oncoprotein scores are generally reduced following relapse as AML and cites a prior RNA-seq report finding the same in three of four MLL::AF4 samples - so the level fall is not unique to this child, and this pair's behaviour may be typical rather than individual. Janssens DH et al. 2024
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
The two samples are deposited as separate GSM records with reciprocal 'paired with' statements in their descriptions, at Texas Children's Hospital, with KMT2A N1/N2 and C1/C2 AutoCUT&RUN libraries for each.
GSE252378: AutoCUT&RUN of KMT2A-rearranged primary leukemia samples (KMT2A N- and C-terminal antibodies, ENL, DOT1L, H3K4me3, IgG) · Janssens DH, et al. · NCBI Gene Expression Omnibus 2024What this evidence cannot show
A 'paired with' string in a deposit is a second timepoint of one patient, not a second patient - a distinction this corpus has recorded being got wrong. No PHI is present in the deposit and none is reproduced here.
Evidence that argues against this (1)
- The deposit gives no collection dates, so the interval between the two samples is unknown and 'before and after chemotherapy' is the paper's framing rather than a dated observation. Janssens DH et al. 2024
the two samples whose overlap the headline measures
the 54.6% / 40.7% asymmetric retention
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
both states carry KMT2A::AFF1
the first disease is infant B-ALL
the second disease is the switched AML
p279 and p318 are the worked case of that rule
the same diagnostic sample's HOXA/IRX1 occupancy state
p279/p318 is the pair the peak classes were derived from
an 8.56x swing in programme balance across the same infant's two samples