Retained fetal programme — the infant-versus-adult stem cell difference, and the reason KMT2A::AFF1 needs no second mutation
also called fetal transcriptional signature · fetal programme retention · LIN28B-high fetal state · infant versus adult HSC state
A blood stem cell in a fetus is not the same as one in an adult: it grows faster, renews more readily and runs a set of genes that are switched off after birth. Infant leukaemia blasts keep that fetal setting, and new work shows that one fetal gene in particular has to be switched on for the fusion to cause leukaemia at all.
The same thing, for a scientist
Fetal and adult haematopoietic stem and progenitor cells differ transcriptionally in ways that are conserved between human and mouse, and infant KMT2A-rearranged ALL blasts retain the fetal signature. Two studies make this a state rather than a description. First, comparing human fetal and adult HSPC transcriptomes yielded a fetal-specific gene signature whose top hits included the fetal oncogene **LIN28B** and its downstream effectors, also expressed in infant ALL; in a human fetal-liver CRISPR-Cas9 KMT2A::AFF1 model, **FL CD34+ cells fail to transform upon induction of the translocation in the absence of LIN28B**, and LIN28B-expressing leukaemias were more proliferative in vitro and in vivo with that advantage lost on knockdown [E1]. Second, RNA sequencing of fetal, neonatal and adult HSPCs from human and mouse defined a conserved fetal signature of prominently proliferative and oncogenic character that **persists in infant ALL blasts**, from which PLK1 and ELOVL1 emerged as survival requirements of MLL-AF4-positive ALL cells, implicating altered fatty acid metabolism [E2]. The state is therefore permissive, targetable in principle, and — because it is a normal fetal programme — absent from the adult tissue a drug would also reach.
Fetal blood-making is a different system, not a younger one
It is tempting to think of a fetal blood stem cell as an adult one that has not grown up. It is not. The fetal system is built for rapid expansion: the cells cycle more, self-renew more readily, and run a distinct set of transcription factors and RNA-binding proteins, of which LIN28B is the best known. After birth these genes are progressively switched off, and the adult system settles into a slower, quiescence-weighted mode.
The human blood hierarchy is also structurally different early on: fetal liver contains many oligopotent progenitors with intermingled myeloid, erythroid and megakaryocytic fates while adult marrow contains almost none 3. So the same sorting gates do not capture the same cells across ontogeny, and a fetal progenitor is not simply an adult progenitor at an earlier point in its life.
This matters because infant KMT2A-rearranged leukaemia is initiated in utero, in the fetal system, and then presents in a baby who is already transitioning out of it.
LIN28B is required for KMT2A::AFF1 to transform a human fetal cell
This is the strongest statement in this area and it is recent. Comparing human fetal with adult HSPC transcriptomes produced a fetal-specific gene signature whose top hits included LIN28B and its downstream effectors, and those genes were also expressed in infant ALL. Functionally, LIN28B was required in human fetal liver CD34+ cells to maintain proliferation and stem-like properties and to support B- and NK-lymphopoiesis.
The decisive experiment used a human fetal-liver-derived CRISPR-Cas9 KMT2A::AFF1 model, in which the translocation is induced in primary human fetal cells. In that model, FL CD34+ cells fail to transform upon induction of the KMT2A::AFF1 translocation in the absence of LIN28B. LIN28B-expressing CRISPR-KMT2A::AFF1 leukaemias were more proliferative in vitro and in vivo, and that advantage was lost on LIN28B knockdown. Mechanistically LIN28B stabilises key early B-lymphoid genes, epigenetic regulators, and cell-cycle and anti-apoptotic genes 1.
That is a clean statement of the graph's long-standing puzzle. Infant ALL carries a very poor prognosis despite lacking the additional oncogenic driver mutations common in childhood ALL. If the fetal programme itself supplies what a second mutation would otherwise have to, then the missing co-operating lesion is not missing — it is the developmental context, and it is not encoded in the genome.
The programme persists in the blasts, and it is druggable in principle
Multiple RNA-sequencing experiments on fetal, neonatal and adult haematopoietic stem and progenitor cells from human and mouse defined a conserved fetal transcriptional signature characterised by a prominent proliferative and oncogenic nature, and that signature persists in infant ALL blasts. Functional validation from within the signature identified genes critical for survival of MLL-AF4-positive ALL cells, of particular interest PLK1 — for which inhibitors already exist — and ELOVL1, which highlights altered fatty acid metabolism as a feature of infant ALL 2.
The therapeutic logic here is unusually favourable and deserves to be stated plainly: a gene that is fetal-specific is, by construction, off in the tissues of the patient being treated. A drug against a retained fetal programme has a natural therapeutic window that a drug against a housekeeping pathway does not. This is the opposite situation from the haem-biosynthesis dependency discussed in cell-type-monocyte, where the target's normal function is running in every monocyte in the body.
The qualification is that infants are not fetuses. A neonate still has residual fetal-programme activity in normal haematopoiesis, and the window narrows the younger the patient — which is exactly the population this disease affects.
| state definition | Retention, in infant leukaemia blasts, of the transcriptional programme characteristic of fetal rather than adult haematopoietic stem and progenitor cells |
|---|---|
| location | Fetal liver and fetal bone marrow; retained in infant leukaemia blasts after birth |
markers
[ "LIN28B", "HMGA2 (LIN28B downstream)", "low let-7 family", "IGF2BP family", "conserved fetal signature score" ]
kmt2a relevance
{
"initiation_requirement": "human fetal liver CD34+ cells fail to transform upon CRISPR-induced KMT2A::AFF1 translocation in the absence of LIN28B",
"explains": "why KMT2A::AFF1 infant ALL is aggressive despite lacking the additional oncogenic driver mutations common in childhood ALL",
"derived_targets": [
"LIN28B",
"PLK1",
"ELOVL1"
],
"therapeutic_window": "a fetal-specific gene is off in the treated patient's normal tissue, which is a more favourable window than a housekeeping target; the window narrows in the youngest patients, who retain residual fetal-programme activity"
}infant vs adult
{
"fetal": "higher proliferation, readier self-renewal, LIN28B-high, many oligopotent progenitors with intermingled myeloid/erythroid/megakaryocytic fates",
"adult": "quiescence-weighted, LIN28B-low, few oligopotent progenitors, unilineage cells emerging more directly from the stem compartment"
}atlas refs — the numbers behind the chart
[
{
"label": "Fetal-specific signature and LIN28B requirement for KMT2A::AFF1 transformation",
"pmid": "42213620",
"title": "The fetal specific gene LIN28B is essential for human fetal B-lymphopoiesis and initiation of KMT2A::AFF1 infant leukemia.",
"authors": "Ling RE, Welsh AM, Hamer L, et al.",
"venue": "Blood",
"year": 2026,
"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": "Conserved fetal signature persisting in infant ALL blasts",
"pmid": "34706236",
"title": "Defining the fetal origin of MLL-AF4 infant leukemia highlights specific fatty acid requirements.",
"authors": "Symeonidou V, Jakobczyk H, Bashanfer S, et al.",
"venue": "Cell Rep",
"year": 2021,
"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": "Ontogeny-dependent reshaping of the human blood hierarchy",
"pmid": "26541609",
"title": "Distinct routes of lineage development reshape the human blood hierarchy across ontogeny.",
"authors": "Notta F, Zandi S, Takayama N, et al.",
"venue": "Science",
"year": 2016,
"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."
}
]- The infants who need therapy most are the ones whose normal haematopoiesis still runs the fetal programme, so the therapeutic window against a fetal-specific target is narrowest exactly where the drug would be used.
- If the fetal context substitutes for cooperating mutations, then genomic risk stratification will systematically under-classify infant disease, because the thing making it aggressive is not in the genome.
- LIN28B supports normal fetal B- and NK-lymphopoiesis, so inhibiting it in a neonate would be expected to impair the development of the lymphoid compartment the child needs.
- A drug that works by exploiting the fetal state loses its target as the child ages, which inverts the usual assumption that a biomarker measured at diagnosis remains valid through treatment.
- How fast does the fetal programme decay after birth, and does the rate predict which infants respond? If the programme is required for maintenance and not only for initiation, a baby whose normal haematopoiesis has already switched over may have a leukaemia that is running out of its own support.
- Is LIN28B required for MAINTENANCE of an established KMT2A::AFF1 leukaemia, or only for initiation? The knockdown result shows a lost proliferative advantage, which is suggestive but is not the same as a maintenance requirement in an established patient-derived model.
- Does the LIN28B requirement extend to KMT2A fusions other than AFF1, and to myeloid infant disease, or is it specific to fetal B-lymphopoiesis?
- Is the fetal programme upstream or downstream of the fusion? If the fusion imposes it, it is a consequence; if the cell brings it, it is a permissive context. The CRISPR fetal-liver model is the system in which that could be separated.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Comparison of human fetal and adult HSPC transcriptomes yielded a fetal-specific gene signature whose top hits included the fetal oncogene LIN28B and its downstream effectors, also expressed in infant ALL; LIN28B was essential in human fetal liver CD34+ cells to maintain proliferation and stem-like properties and to support B- and NK-lymphopoiesis; in a human fetal-liver-derived CRISPR-Cas9 KMT2A::AFF1 model, FL CD34+ cells fail to transform upon induction of the translocation in the absence of LIN28B, and LIN28B-expressing leukaemias were more proliferative in vitro and in vivo with that advantage lost on LIN28B knockdown; mechanistically LIN28B stabilises key early B-lymphoid genes, epigenetic regulators, and cell-cycle and anti-apoptotic genes.
The fetal specific gene LIN28B is essential for human fetal B-lymphopoiesis and initiation of KMT2A::AFF1 infant leukemia · Ling RE, Welsh AM, Hamer L, Jackson T, Neil E, Elliott N, et al., Roy A · Blood 2026What this evidence cannot show
A CRISPR-engineered translocation in primary human fetal liver cells is the closest available model to the real initiating event but remains a model: the translocation is induced acutely at a chosen time in cells taken out of their in vivo niche, and 'fail to transform' is a readout of a transformation assay rather than of leukaemia in a patient. LIN28B-null fetal cells also have impaired normal B-lymphopoiesis, so failure to transform may partly reflect loss of the target cell rather than loss of a transformation requirement. Read: abstract only.
Evidence that argues against this (1)
- The premise that infant KMT2A-rearranged ALL lacks additional oncogenic drivers is not uniformly true and its exceptions are the worst-prognosis cases. Multi-omics profiling of 84 infants with KMT2A-rearranged leukaemia found that the number of RAS pathway mutations predicts prognosis and that the most refractory subgroup, the undifferentiated IRX-type IC2, has 100 percent frequency and the heaviest burden of RAS pathway mutations. If a fetal programme were sufficient in place of cooperating mutations, the cases carrying the most cooperating mutations should not be the most refractory. Read: abstract only. Isobe T et al. 2022
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Multiple RNA-sequencing experiments on fetal, neonatal and adult haematopoietic stem and progenitor cells from human and mouse defined a conserved fetal transcriptional signature of prominent proliferative and oncogenic character that persists in infant ALL blasts; functional validation identified genes critical for survival of MLL-AF4-positive ALL cells, notably PLK1, for which an inhibitor is readily available, and ELOVL1, implicating altered fatty acid metabolism as a feature of infant ALL.
Defining the fetal origin of MLL-AF4 infant leukemia highlights specific fatty acid requirements · Symeonidou V, Jakobczyk H, Bashanfer S, Malouf C, Fotopoulou F, Kotecha RS, et al., Ottersbach K · Cell Reports 2021What this evidence cannot show
The fetal signature is derived from bulk RNA-seq of sorted populations across species, so the signature is conserved at the level of gene sets rather than of individual cell states; functional validation is in MLL-AF4-positive cell lines, and PLK1 is a broadly required mitotic kinase whose apparent selectivity may reflect proliferation rate. Read: abstract only.
Evidence that argues against this (1)
- Persistence of a fetal transcriptional signature in blasts is a statement about state, not about origin, and single-cell work in this disease warns explicitly against the inference. Mutation sharing across an infant KMT2A-rearranged B-ALL and its AML relapse in the same child placed the rearrangement before haematopoietic specification, with the authors stating that cell of origin cannot be inferred from transcriptional state. A retained fetal signature may be what the fusion imposes rather than what the cell brought with it. Read: abstract only. Khabirova E et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The human blood hierarchy is reshaped across ontogeny: fetal liver contains many oligopotent progenitors with intermingled myeloid, erythroid and megakaryocytic fates whereas adult bone marrow contains almost none, so fetal and adult progenitors captured by the same surface-marker gates are not equivalent cells.
Distinct routes of lineage development reshape the human blood hierarchy across ontogeny · Notta F, Zandi S, Takayama N, et al., Dror Y, Dick JE · Science 2016What this evidence cannot show
Functional single-cell and index-sorting assays; absence of a fate in culture is weaker evidence than its presence, and the fetal tissue studied is fetal liver rather than the fetal bone marrow that dominates later gestation. Read: abstract only.
Evidence that argues against this (1)
- Refining the picture further, integrative single-cell RNA-seq and ATAC-seq of human developmental haematopoiesis found extensive EPIGENETIC but not transcriptional priming in HSC/MPP, and required a refined sorting strategy to reach roughly 90 percent enrichment in fetal liver and bone marrow. A fetal-versus-adult difference defined transcriptionally may therefore miss the chromatin-level differences that actually determine what a fetal cell can become, and a 'fetal signature' scored on transcriptome alone is an incomplete description of the state. Read: abstract only. Ranzoni AM et al. 2021
the lymphoid lipid requirement is ELOVL1, a fetal-signature saturated/monounsaturated elongase — a different enzyme class from the myeloid ACSL4/PUFA requirement
the state is defined against fetal liver HSPC and retained by blasts
- supports →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
LIN28B is required for KMT2A::AFF1 to transform human fetal liver CD34+ cells
proposed explanation for aggressiveness despite few cooperating mutations
the adult counterpart from which the fetal signature was derived by contrast
- related to →The stem-like self-renewal state — a limited stem-cell module grafted onto a committed cell
fetal HSPC self-renew more readily, which the retained programme carries into the blast
- ← related toThe neonatal window — the state in which an MLL fusion can transform, bounded at both ends
bounds the fetal programme rather than assuming it is permissive: fetal context conveys MLL3-dependent heritable PROTECTION
- ← related toThe CD24-positive pre-leukaemic state — what the index fusion makes before it makes leukaemia
the background programme this state arises on
an oncofetal marker of the retained fetal programme - class 1 evidence; no experiment shows it MAINTAINS the state
SKIDA1 sustains MLL::ENL-expressing progenitors during neonatal stages and promotes B-lineage priming
competing account of the same age-dependent lineage phenotype, located inside the cell
identified from within the conserved fetal transcriptional signature
rests on the retained fetal programme of infant KMT2A::AFF1 ALL
top hit of the fetal-versus-adult HSPC signature; the fetal state is permissive for KMT2A::AFF1 transformation through it
the fetal state is permissive rather than incidental to transformation
let-7 accumulation is the clock that ends the fetal output state; inhibiting it in adults restores fetal-type haematopoiesis
the normal state whose retention in blasts the other entry describes
ID1 high in fetal-origin infant MLL-FP leukaemia, low in postnatal
fetal expression retained in infant but not childhood KMT2A::AFF1 ALL is the graph's existing support for the permissive reading
enforced IGF2BP3 skews murine haematopoietic development to the B-cell/myeloid lineage, so a B-cell programme found alongside IGF2BP3 may be its CONSEQUENCE rather than evidence of cell of origin - the network analysis cannot orient that arrow
HEXIM1 shifts RNAPII and GATA1 from beta- to gamma-globin, raising FETAL globin and BGLT3 in human erythroid cells (PMID 37738561)
Rybp loss in ADULT mice reverts B-cell progenitors toward the innate, predominantly FETAL B-1 programme - a fetal-retention state produced in an animal with no fusion, which is the lineage explanation in its cleanest form (murine; PMID 26711264)
MECHANISTIC PRECEDENT ONLY, not evidence about the leukaemia: HEXIM1 overexpression flips one locus from its adult to its fetal output by pause-release control, which is a second system in which fetal-versus-adult programme choice runs through P-TEFb regulation
ELOVL1 is one of two genes functionally validated out of the conserved fetal HSPC signature that persists in infant MLL-AF4 blasts (PMID 34706236); that cell state is why the gene is in this graph (Q89 2026-09-21)
The EUCHROMATIC isoform is the stem-like one: DNMT3A2 is predominant in embryonic stem and embryonal carcinoma cells and detectable in thymus and spleen, while full-length pericentric DNMT3A is low-level ubiquitous (PMID 12138111). Which form an infant KMT2A::AFF1 blast expresses is unmeasured and answerable from deposited RNA-seq, because the two use different promoters (Q93 2026-09-21)