Oncofusion

start / graph / cell-type-fetal-liver-hsc

Cell typestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q4-cells

Fetal liver haematopoietic stem and progenitor cell — the leading candidate cell of origin for t(4;11) infant ALL

also called fetal liver HSPC · FL HSC · CD34+CD38- fetal liver stem cell · pre-VDJ fetal progenitor

Before birth, blood-forming stem cells live in the fetal liver, not the bone marrow. The gene activity of t(4;11) infant leukaemia looks most like these fetal liver cells, which is the strongest clue that the disease is born in the womb.

The same thing, for a scientist

The human fetal liver is the dominant site of definitive haematopoiesis from around the end of the first trimester until fetal bone marrow takes over, and its most immature CD34-positive CD38-negative lineage-negative haematopoietic stem and progenitor cells are the compartment whose transcriptome most closely matches t(4;11)/KMT2A::AFF1 infant B-ALL. Two independent lines of evidence support a fetal origin: unbiased transcriptomic matching of 124 uniformly treated infant B-ALL cases to fetal ontogeny references, and CRISPR-Cas9 engineering of a genuine t(4;11) translocation directly into primary human fetal liver haematopoietic cells, which reproduces both the clinical phenotype and an infant-ALL-specific, fetal-specific gene expression program that is absent from childhood KMT2A::AFF1 ALL.

Go deeper
Why fetal liver rather than bone marrow1 deeper

Two arguments, from opposite directions, converge.

From patients, looking backwards. A multi-layered genome-wide study of 124 infant B-ALL cases uniformly diagnosed and treated on Interfant-99/06 found the most silent mutational landscape reported for any sequenced paediatric cancer, with recurrent mutations confined to a small set of subclonal events that were frequently lost at relapse. B-cell receptor repertoire analysis found only minor, non-expanded B-cell clones in t(4;11) cases, unlike non-KMT2A-rearranged B-ALL, and RNA-sequencing placed the leukaemic transcriptome closest to the most immature human fetal liver haematopoietic stem and progenitor cells. The authors read this as a pre-VDJ fetal cellular origin 1. The silence of the genome matters as much as the transcriptome match: a cancer that needs essentially one hit is a cancer whose permissive context was supplied by the cell, not by additional mutations.

From the bench, looking forwards. CRISPR-Cas9 was used to create an authentic t(4;11)/MLL-AF4 translocation in primary human fetal liver haematopoietic cells. The resulting disease replicated the clinical features of infant-ALL and switched on both infant-ALL-specific and fetal-specific gene expression programs. Critically, the same study showed that fetal-specific programs are maintained in MLL-AF4 infant-ALL but not in MLL-AF4 childhood-ALL 2. That is the cleanest available explanation for the field's central clinical puzzle: identical driver, radically different prognosis by age.

The age paradox this resolves

KMT2A::AFF1 occurs in infants, older children and adults. Outcomes are worst in infants. If the fusion were the whole story, that gradient would make no sense. The fetal-program model says the fusion is only half the oncogene: the other half is the developmental state of the cell it lands in, and that state is available only in a window that closes around birth 2.

The practical corollary for a drug programme is uncomfortable. A therapy aimed purely at the fusion or at its obligate partner is aimed at the half of the lesion that is shared across ages. The half that makes infant disease distinctively lethal is a developmental transcriptional context, which is not a protein and has no pocket.

Surface marker definition and where these cells physically sit

The immunophenotype used to sort the relevant compartment is the standard human HSPC gate: CD34-positive, CD38-negative, lineage-negative, with CD90 (THY1) and CD45RA used to separate stem from multipotent progenitor fractions. Cell Ontology provides CL:0000037 (haematopoietic stem cell), the marker-explicit CL:0001024 (CD34-positive, CD38-negative haematopoietic stem cell) and CL:0002353 (fetal liver haematopoietic progenitor cell) 5.

Anatomically, human definitive haematopoiesis is a relay, not a single address. Lymphopoiesis starts in utero at about 6 weeks post-conception; CD10-negative PreProB and CD10-positive ProB progenitors appear early in the first trimester in embryonic liver, followed by a sustained second wave of B-progenitor development in fetal bone marrow, where B-progenitors together make up more than 40% of the total stem/progenitor pool 3. So 'fetal liver origin' does not require that the leukaemia initiated inside liver parenchyma: it means the cell was in a fetal-liver-type state. Fetal bone marrow in the second and third trimester still carries fetal, not adult, programs.

What a postnatal stem cell cannot do — and the study that disagrees

If the fetal state is the missing cofactor, then putting the fusion into a postnatal stem cell should fail. Two experiments say it largely does. Enforced MLL-AF4 expression in human cord blood CD34+ HSPCs enhanced repopulating and clonogenic potential, proliferation and survival, and improved multilineage engraftment and homing after intra-bone-marrow transplantation into NOD/SCID-IL2Rg-null mice, but was insufficient to initiate leukaemia 4. Adding activated KRAS, among the commonest cooperating lesions in real patients, promoted extramedullary engraftment and migration of cord blood CD34+ HSPCs but again did not initiate leukaemia 6.

The counter-argument is real and should not be buried. A retroviral MLL-Af4 construct (murine Af4, highly conserved with human AFF1) transduced into human CD34+ cells produced a model that fully recapitulated the immunophenotypic and molecular features of t(4;11) pro-B ALL 7. Either the murine Af4 moiety, the higher titre, or the specific CD34+ source made the difference. The honest reading is that a postnatal CD34+ cell is hard, not impossible to transform, and that the fetal advantage is quantitative rather than absolute.

Facts and figures
ontologyCL:0002353
locationHuman fetal liver from around the end of the first trimester; relays to fetal bone marrow in the second and third trimester. Not an adult bone marrow state.
developmental stagefetal, pre-VDJ
evidence for origintranscriptomic matching of patient blasts (n=124) plus an engineered translocation in primary human fetal tissue
evidence against originpostnatal cord blood CD34+ HSPCs can be made to model t(4;11) pro-B ALL with a high-titre retroviral MLL-Af4, so the fetal requirement is quantitative rather than absolute
atlas refs
Infant B-ALL genome-wide …2,019Infant B-ALL genome-wide cohort, n=124, Interfant-99/06: 2019Human fetal B-cell develo…2,019Human fetal B-cell developmental hierarchy, fetal liver and fetal bone marrow: 2019CRISPR t(4;11) in primary…2,021CRISPR t(4;11) in primary human fetal liver haematopoietic cells: 2021
ontology alternatives
[
  "CL:0000037",
  "CL:0001024"
]
markers
[
  "CD34+",
  "CD38-",
  "Lineage-",
  "CD90/THY1+",
  "CD45RA-",
  "CD19-",
  "CD10-"
]
atlas refs — the numbers behind the chart
[
  {
    "label": "Infant B-ALL genome-wide cohort, n=124, Interfant-99/06",
    "pmid": "30679323",
    "title": "Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis.",
    "authors": "Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al.",
    "venue": "Haematologica",
    "year": 2019,
    "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": "CRISPR t(4;11) in primary human fetal liver haematopoietic cells",
    "pmid": "34824279",
    "title": "A human fetal liver-derived infant MLL-AF4 acute lymphoblastic leukemia model reveals a distinct fetal gene expression program.",
    "authors": "Rice S, Jackson T, Crump NT, et al.",
    "venue": "Nat Commun",
    "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": "Human fetal B-cell developmental hierarchy, fetal liver and fetal bone marrow",
    "pmid": "31383639",
    "title": "Discovery of a CD10-negative B-progenitor in human fetal life identifies unique ontogeny-related developmental programs.",
    "authors": "O'Byrne S, Elliott N, Rice S, et al.",
    "venue": "Blood",
    "year": 2019,
    "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."
  }
]
How this could bite us back
  • If the fetal transcriptional context, not the fusion, is what makes infant disease lethal, then a perfect menin-interface inhibitor addresses only the half of the lesion that is shared with the more curable childhood cases.
  • Any therapy targeting a fetal-specific program risks hitting normal fetal or neonatal haematopoiesis in the very age group being treated.
  • Transcriptome matching to a bulk-sorted reference can make a heterogeneous leukaemia look like one clean origin, hiding a mixed or shifting cell of origin that would defeat a single-target strategy.
Still open
  1. Which specific fetal-restricted transcription or chromatin factors supply the permissive context, and is any of them druggable or at least a biomarker of infant-type biology?
  2. Can the fetal program be switched off in an established infant leukaemia, and does that convert it to the more curable childhood-type disease?
  3. Could single-cell lineage tracing or mitochondrial barcoding on neonatal blood spots identify the actual cell in which the translocation arose, rather than inferring it from transcriptome similarity?
  4. Why does a high-titre retroviral MLL-Af4 transform postnatal CD34+ cells when a lentiviral MLL-AF4 does not? Dose, the murine Af4 moiety, or the vector integration landscape?
Every source behind this entry (7)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 124 uniformly treated infant B-ALL cases, RNA-sequencing placed t(4;11) leukaemic transcriptomes closest to the most immature human fetal liver haematopoietic stem and progenitor cells, and BCR repertoire analysis showed only minor non-expanded B-cell clones, supporting a pre-VDJ fetal cellular origin; the mutational landscape was the most silent reported for any sequenced paediatric cancer.

    Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis · Agraz-Doblas A, Bueno C, Bashford-Rogers R, Roy A, Schneider P, Bardini M, Ballerini P, Cazzaniga G, Moreno T, Revilla C, Gut M, Valsecchi MG, Roberts I, Pieters R, De Lorenzo P, Varela I, Menendez P, Stam RW · Haematologica 2019
    PMID 30679323DOI
    What this evidence cannot show

    Transcriptomic similarity is a correlative argument about cell state, not lineage-tracing proof of which cell acquired the translocation. The fetal reference panels available were bulk-sorted populations, so 'most immature fetal liver HSPC' is a gate, not a single cell. The cohort was Interfant-treated and therefore enriched for the most aggressive presentations.

    Evidence that argues against this (4)
    • Transcriptome proximity is not ontogeny. A bulk RNA-seq profile placing t(4;11) blasts nearest the most immature fetal liver HSPC is consistent with the leukaemia ARISING there, and equally consistent with an oncoprotein imposing a fetal-like programme on a cell of different origin - which is what a KMT2A fusion is known to do to lineage programmes. The comparison also runs against a reference atlas whose own cell-type boundaries set the answer. Haematologica 2019
    • The fetal context in this disease is known to be functionally required rather than merely correlated, which cuts against reading the transcriptome result as evidence of origin specifically: LIN28B is required for KMT2A::AFF1 to transform human fetal liver CD34+ cells at all. A fetal-like transcriptome may therefore be the permissive state the fusion needs, not the address it started from. SOURCING CORRECTION 2026-09-20 (Q40 counter-evidence quality sweep): the LIN28B result is NOT in the paper cited on this note and this note should never have carried that paper's PMID. Its primary source, and a second study pointing the other way, are given in the two notes below. Haematologica 2019
    • INDEPENDENT SOURCE for the LIN28B sentence in the preceding note, added 2026-09-20 by the Q40 sweep after fetching the record. In a human fetal-liver-derived CRISPR-Cas9 KMT2A::AFF1 model, FL CD34+ cells FAIL TO TRANSFORM on induction of the translocation in the absence of LIN28B, and LIN28B-expressing leukaemias were more proliferative in vitro and in vivo with the advantage lost on LIN28B knockdown. This is a functional requirement for a fetal-restricted gene, and it is what makes the transcriptome-proximity argument for fetal ORIGIN weaker rather than stronger: a fetal programme that the fusion needs is not evidence of the address it started from. Ling RE et al. 2026
    • AND THE INDEPENDENT LITERATURE DISAGREES WITH ITSELF, which no same-source note in this entry could have shown. Added 2026-09-20 by the Q40 sweep. In a dual-transgenic mouse co-expressing MLL-ENL and LIN28B, LIN28B IMPEDES progression of MLL-rearranged AML through compromised leukaemia-initiating-cell activity and suppression of MYB signalling, acting by stabilising MYBBP1A mRNA; the authors propose developmentally restricted LIN28B as a layer of PROTECTION against MYB-dependent AML. So the fetal programme is required in one system and protective in another, and neither result licenses reading a fetal transcriptome as a statement about cell of origin. Eldeeb M et al. 2023
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CRISPR-Cas9 engineering of a t(4;11)/MLL-AF4 translocation in primary human fetal liver haematopoietic cells replicates the clinical features of infant-ALL and drives infant-ALL-specific and fetal-specific gene expression programs, and these fetal programs are maintained in MLL-AF4 infant-ALL but not in MLL-AF4 childhood-ALL.

    A human fetal liver-derived infant MLL-AF4 acute lymphoblastic leukemia model reveals a distinct fetal gene expression program · Rice S, Jackson T, Crump NT, Fordham N, Elliott N, O'Byrne S, Lara Fanego MDM, Addy D, Crabb T, Dryden C, Inglott S, Ladon D, Wright G, Bartram J, Ancliff P, Mead AJ, Halsey C, Roberts I, Milne TA, Roy A · Nature Communications 2021
    PMID 34824279DOI
    What this evidence cannot show

    Primary human fetal tissue is scarce, so the number of independent biological donors in such experiments is inherently small. A xenograft readout imposes a mouse microenvironment on a human fetal cell, which itself perturbs developmental programs. The study establishes that the fetal context is sufficient and permissive, not that it is uniquely necessary.

    Evidence that argues against this (1)
    • A retrovirally delivered MLL-Af4 fusion in human CD34+ cells produced a faithful model of t(4;11) pro-B ALL, showing that fetal liver tissue is not an absolute requirement for modelling the disease. Lin S et al. 2016
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Human lymphopoiesis begins in utero about 6 weeks post-conception, with CD10-negative PreProB and CD10-positive ProB progenitors appearing early in the first trimester in embryonic liver, followed by a sustained second wave of B-progenitor development in fetal bone marrow where B-progenitors together form more than 40% of the total haematopoietic stem cell/progenitor pool.

    Discovery of a CD10-negative B-progenitor in human fetal life identifies unique ontogeny-related developmental programs · O'Byrne S, Elliott N, Rice S, Buck G, Fordham N, Garnett C, Godfrey L, Crump NT, Wright G, Inglott S, Hua P, Psaila B, Povinelli B, Knapp DJHF, Agraz-Doblas A, Bueno C, Varela I, Bennett P, Koohy H, Watt SM, Karadimitris A, Mead AJ, Ancliff P, Vyas P, Menendez P, Milne TA, Roberts I, Roy A · Blood 2019
    PMID 31383639DOI
    What this evidence cannot show

    Fetal tissue availability restricts gestational sampling, so the developmental timeline is reconstructed from discrete windows rather than continuously observed. Frequencies depend on the sorting gate used and on tissue dissociation efficiency.

    Evidence that argues against this (1)
    • Normal fetal B-lymphopoiesis is the right background for this beat and it is background: nothing in the developmental timing establishes which of those progenitors acquires the translocation. The window argument is also loose — a first-trimester origin is inferred from twin concordance and backtracking of the fusion in neonatal blood spots, not from observing the cell. Blood Cells, Molecules and Diseases 2010
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Enforced MLL-AF4 expression in human cord blood CD34+ HSPCs enhances repopulating and clonogenic capacity, proliferation and survival, and improves multilineage engraftment and homing in NOD/SCID-IL2Rg-null mice, but is insufficient on its own to initiate leukaemia.

    Enforced expression of MLL-AF4 fusion in cord blood CD34+ cells enhances the hematopoietic repopulating cell function and clonogenic potential but is not sufficient to initiate leukemia · Montes R, Ayllon V, Gutierrez-Aranda I, Prat I, Hernandez-Lamas MC, Ponce L, Bresolin S, te Kronnie G, Greaves M, Bueno C, Menendez P · Blood 2011
    PMID 21389315DOI
    What this evidence cannot show

    Lentiviral enforced expression is not an endogenous translocation: it lacks the native promoter, the correct stoichiometry and the reciprocal AF4-MLL product, all of which the authors themselves flag as possible reasons for the negative result. A negative xenograft result also depends on follow-up duration and recipient strain.

    Evidence that argues against this (1)
    • Enhanced repopulation without leukaemia is the central negative of the cord-blood models and is consistently filed as a partial success. Cord blood is NEONATAL, and this beat's own evidence has an authentic CRISPR t(4;11) in fetal liver cells reproducing infant-ALL programmes. The failure is therefore attributable to the cell, which makes every conclusion drawn from cord-blood experiments in this beat provisional. Oncofusion knowledge graph (internal analysis) 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Cell Ontology provides resolvable identifiers for the relevant compartments: CL:0000037 haematopoietic stem cell, CL:0001024 CD34-positive CD38-negative haematopoietic stem cell, and CL:0002353 fetal liver haematopoietic progenitor cell.

    Cell Ontology (CL) term lookup via the EMBL-EBI Ontology Lookup Service OLS4 API · EMBL-EBI Ontology Lookup Service; Cell Ontology consortium · Database 2026
    What this evidence cannot show

    Cell Ontology marker definitions are frequently written from mouse phenotypes and are not always stage-specific for human fetal tissue, so a CL id should be treated as a stable label rather than an authoritative human marker panel.

    Evidence that argues against this (1)
    • Cell Ontology identifiers resolve and that is all they do. An ontology term records a community naming decision, its definitions are revised between releases, and the compartments this beat cares about — a fetal-liver progenitor at a specific developmental moment — are finer than any available term. A resolvable identifier is precision of reference, not precision of biology. OBO Foundry / EMBL-EBI OLS 2026
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Activated KRAS cooperates with MLL-AF4 to promote extramedullary engraftment and migration of human cord blood CD34+ HSPCs but is insufficient to initiate leukaemia.

    Activated KRAS Cooperates with MLL-AF4 to Promote Extramedullary Engraftment and Migration of Cord Blood CD34+ HSPC But Is Insufficient to Initiate Leukemia · Prieto C, Stam RW, Agraz-Doblas A, Ballerini P, Camos M, Castano J, Marschalek R, Bursen A, Varela I, Bueno C, Menendez P · Cancer Research 2016
    PMID 26837759DOI
    What this evidence cannot show

    Same enforced-expression limitation as E4, compounded by the fact that RAS pathway mutations in patients are often subclonal and frequently lost at relapse, so their role may be in progression rather than initiation.

    Evidence that argues against this (1)
    • Cooperation that enhances migration without initiating leukaemia is the same negative as the fusion alone, in the same neonatal cells. Read together with the fetal-liver result, the informative conclusion is that the missing ingredient is developmental context rather than a second lesion. Oncofusion knowledge graph (internal analysis) 2026
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A retroviral MLL-Af4 fusion using murine Af4 transduced into human CD34+ cells generates a model that fully recapitulates the immunophenotypic and molecular features of t(4;11) pro-B ALL.

    Instructive Role of MLL-Fusion Proteins Revealed by a Model of t(4;11) Pro-B Acute Lymphoblastic Leukemia · Lin S, Luo RT, Ptasinska A, Kerry J, Assi SA, Wunderlich M, Imamura T, Kaberlein JJ, Rayes A, Althoff MJ, Anastasi J, O'Brien MM, Meetei AR, Milne TA, Bonifer C, Mulloy JC, Thirman MJ · Cancer Cell 2016
    PMID 27846391DOI
    What this evidence cannot show

    Uses murine Af4 rather than human AFF1 and a retroviral vector at high titre, so it reports what a strongly expressed fusion can instruct rather than what the endogenous t(4;11) does at physiological dose. Species specificity of the oncogene is explicitly one of the paper's own themes.

    Evidence that argues against this (1)
    • A retroviral fusion using MURINE Af4 in human CD34+ cells recapitulates the phenotype, which is a striking result and a warning: if a cross-species chimeric fusion expressed retrovirally reproduces the disease's immunophenotype, then immunophenotypic recapitulation is not a stringent test of a model. Oncofusion knowledge graph (internal analysis) 2026
Connected to