Oncofusion

start / graph / concept-fetal-restricted-permissive-factors-across-fusion-families

Explainermoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q133-lin28b-fetal-ontogeny

Fetal-restricted permissive factors across fusion families - the class has exactly ONE fully worked member, and for ETV6::RUNX1 the fetal advantage is real but no factor is named

also called the analogous-biology axis for LIN28B · is there a LIN28B for other fusions · ontogeny-conditional transformation across fusion families

Several childhood leukaemias start before birth, so it is natural to ask whether each of them needs something that only exists in an unborn baby. Surprisingly, only one such thing has ever been pinned down to a named gene - LIN28B, for one MLL fusion. For the commonest childhood leukaemia fusion, the before-birth advantage is measurable but nobody knows what molecule supplies it.

The same thing, for a scientist

The PI's standing question - **is the analogous biology known in other fusion families?** - has a sharp answer here, and it is mostly an absence. Surveying the literature for a **fetal-restricted factor whose presence is required for a fusion to transform**, this beat found: **(1) LIN28B / KMT2A::AFF1**, human, endogenous translocation, the only fully worked instance [E1]. **(2) ID1 / KMT2A::MLLT3**, mouse, where the *requirement* rather than the *factor* is ontogeny-conditional and reverses sign between fetal liver and bone marrow, with human ID1 expression high in fetal-origin infant MLL-FP AML [E2]. **(3) ETV6::RUNX1**, where ontogeny demonstrably shapes the phenotype - fetal liver cells tolerate the fusion far better than adult bone marrow cells, and poly(I:C) exposure drives fetal E/R pro-B and pre-B compartments to ~5-fold and ~2.5-fold above wild type - but **no fetal-restricted factor is identified**; the mechanism offered is immune-evasion gene induction and an infection trigger [E3]. **(4) A developmentally restricted CELL rather than a factor**: a CD19-negative IL-7R-positive progenitor unique to first-trimester human embryos, expanded by endogenous-locus ETV6-RUNX1 in an hPSC model [E4]. **(5) Nothing at all for CBFA2T3::GLIS2 or NUP98 fusions on the LIN28B axis** - zero title/abstract records [E5]. **The class is therefore not a general principle with many instances. It has one worked example, one partial analogue, and for the best-studied in-utero fusion the causal molecule is unknown.**

Go deeper
The survey, laid out
fusion familyis ontogeny shown to matter?is a FACTOR named?systemsource
KMT2A::AFF1yes - human FL CD34+ fail to transform without LIN28BYES - LIN28B, fetal-restrictedhuman, endogenous CRISPR translocation1
KMT2A::MLLT3 (MLL-AF9)yes - Id1 loss prolongs survival from fetal liver, accelerates from bone marrowpartly - ID1, but ID1 is not itself fetal-restricted; the dependency ismouse, retroviral; human expression correlate2
KMT2A::MLLT1 (MLL-ENL)yes, and in the OPPOSITE direction - initiation peaks neonatally, not fetally; LIN28B suppresses itLIN28B, as a protective factormouse, transgenicsee hyp-fetal-state-is-permissive-for-lymphoid-and-protective-for-myeloid-kmt2a-transformation
ETV6::RUNX1yes - fetal liver cells far less impaired by the fusion than adult BM cells; poly(I:C) expands fetal E/R pro-B ~5x and pre-B ~2.5x over WTNO factor named. The mechanism offered is immune evasion (PD-L1/CD274, CD200R induction, much higher in fetal E/R cells) plus an infection triggermouse, inducible transgenic3
ETV6::RUNX1yes, as a restricted CELL - a CD19- IL-7R+ progenitor present in first-trimester human embryos that transitions myeloid-to-lymphoid during ontogeny; E/R expands it and partially blocks B commitmenta restricted compartment, not a moleculehuman embryo + hPSC, endogenous ETV6 locus4
CBFA2T3::GLIS2not assessed hereno - zero LIN28B co-mentions-5
NUP98 fusionsnot assessed hereno - zero LIN28B co-mentions-5

Read the second ETV6::RUNX1 row against the first. Eldeeb et al. and Böiers et al. reach compatible conclusions by different routes: the fetal window is favourable for E/R, and the favourable thing is a cell with properties adults do not have. Neither names a molecule. That is the honest shape of the field: ontogeny-conditional transformation is well established across fusion families; ontogeny-conditional transformation attributable to a named fetal-restricted gene product is essentially a single result.

Three distinctions this survey forces, and the graph should keep them apart
  1. A fetal-restricted FACTOR - a gene product present before birth and absent after, whose presence the fusion needs. LIN28B is the only clean case. It is the most therapeutically attractive shape, because the target is intrinsically absent from the healthy older patient.
  2. An ontogeny-conditional DEPENDENCY - a factor present at all ages whose requirement differs by cell of origin. ID1 is this, and its sign actually reverses. Much less selective, and potentially dangerous in the wrong age group.
  3. A developmentally restricted CELL - a compartment that exists only in fetal life and is uniquely transformable. Böiers' CD19- IL-7R+ progenitor is this, and so, arguably, is the whole cell-type-fetal-liver-hsc argument. No molecule is targetable in this shape, only the state, which is why this project's fetal entries have been hard to convert into interventions.

All three are routinely described as 'the fetal origin of infant leukaemia' and they are not the same claim. The first is druggable in principle, the second is a stratification problem, the third is not a drug target at all.

The bounded absences, with translations

Live PubMed esearch, 2026-09-23, one [tiab] tag per term, cytogenetics quoted, run through scripts/pubmed_safe.py:

querycount
LIN28B[tiab] AND "ETV6-RUNX1"[tiab]0
LIN28B[tiab] AND "TEL-AML1"[tiab]0
LIN28B[tiab] AND "CBFA2T3-GLIS2"[tiab]0
LIN28B[tiab] AND NUP98[tiab]0
LIN28B[tiab] AND "B-ALL"[tiab]0
LIN28B[tiab] AND "t(4;11)"[tiab]0
LIN28B[tiab] AND "MLL-AF4"[tiab]0
LIN28B[tiab] AND "KMT2A::AFF1"[tiab]1 (the Ling paper)
LIN28B[tiab] AND "infant leukemia"[tiab]2
LIN28B[tiab] AND "MLL-ENL"[tiab]2
LIN28B[tiab] AND lymphoblastic[tiab]3
ID1[tiab] AND "MLL-AF9"[tiab]2
ID1[tiab] AND KMT2A[tiab]0

The KMT2A-side lesson learned the hard way in this same beat (see noncoding-let-7-microrna-family): a zero against the modern symbol KMT2A systematically misses the pre-2018 literature, which for microRNA and developmental work is most of it. The zeros above are LIN28B-side zeros against fusion names, which are less exposed to that failure, because fusion names like ETV6-RUNX1 and CBFA2T3-GLIS2 have been stable. But TEL-AML1 was tried for ETV6::RUNX1 for exactly this reason, and no old-name synonym was tried for CBFA2T3::GLIS2 or for the NUP98 fusions.

Facts and figures
questionIs a fetal-restricted permissive factor known for fusion families other than KMT2A::AFF1?
answerEssentially no. One fully worked instance (LIN28B / KMT2A::AFF1), one ontogeny-conditional dependency that is not a fetal-restricted factor (ID1 / KMT2A::MLLT3), one family where ontogeny clearly matters but the causal molecule is unknown (ETV6::RUNX1), one developmentally restricted CELL rather than a factor (the first-trimester CD19- IL-7R+ progenitor), and nothing on this axis for CBFA2T3::GLIS2 or NUP98 fusions.
etv6 runx1 mechanism offeredimmune evasion - PD-L1/CD274 and CD200R induction much stronger in fetal than adult E/R cells - plus an infection trigger, with poly(I:C) expanding fetal E/R pro-B ~5x and pre-B ~2.5x over wild type. Anti-PD-L1 treatment reduced phenotypic HSC frequency and increased MPP Ly in the E/R setting.
shown for fusionKMT2A::AFF1 (LIN28B, human, factor named); KMT2A::MLLT3 (ID1, mouse, conditional dependency); KMT2A::MLLT1 (LIN28B, mouse, OPPOSITE sign); ETV6::RUNX1 (ontogeny matters, no factor named); CBFA2T3::GLIS2 and NUP98 fusions (nobody has looked on this axis).
three distinct shapes
{
  "fetal_restricted_factor": {
    "definition": "a gene product present before birth and absent after, whose presence the fusion requires",
    "instances": [
      "LIN28B / KMT2A::AFF1 (human)"
    ],
    "druggable_in_principle": true
  },
  "ontogeny_conditional_dependency": {
    "definition": "a factor present at all ages whose requirement differs by cell of origin, and here reverses sign",
    "instances": [
      "ID1 / KMT2A::MLLT3 (mouse, with a human expression correlate)"
    ],
    "druggable_in_principle": "only with age or origin stratification, and dangerous without it"
  },
  "developmentally_restricted_cell": {
    "definition": "a compartment existing only in fetal life that is uniquely transformable",
    "instances": [
      "first-trimester human CD19- IL-7R+ progenitor / ETV6-RUNX1",
      "human fetal liver CD34+ HSPC / KMT2A::AFF1 (this graph's cell-type-fetal-liver-hsc)"
    ],
    "druggable_in_principle": false
  }
}
bounded absences
{
  "LIN28B x ETV6-RUNX1": 0,
  "LIN28B x TEL-AML1": 0,
  "LIN28B x CBFA2T3-GLIS2": 0,
  "LIN28B x NUP98": 0,
  "LIN28B x B-ALL": 0,
  "ID1 x KMT2A": 0,
  "positive_controls": "LIN28B x KMT2A::AFF1 = 1, x infant leukemia = 2, x MLL-ENL = 2, x lymphoblastic = 3, ID1 x MLL-AF9 = 2",
  "verified": "2026-09-23"
}
How this could bite us back
  • If the fetal-restricted-factor class really has one member, then a therapeutic strategy built on 'target the fetal programme' generalises to almost nothing beyond KMT2A::AFF1, and the platform argument for it is weaker than the single result suggests.
  • The ETV6::RUNX1 work points at immune checkpoints as the ontogeny-linked handle. Checkpoint modulation in a fetus or infant is a very different safety proposition from inhibiting an RNA-binding protein, and the two 'fetal origin' stories therefore lead to incompatible therapeutic programmes.
  • ID1's sign reversal is a warning for the whole class: a factor that helps the fusion in a fetal cell may restrain it in an adult one, so any agent developed on fetal-origin biology needs an explicit answer for the postnatal patient it will actually be given to.
Still open
  1. What supplies the fetal advantage for ETV6::RUNX1? Eldeeb et al. show it exists and offer immune evasion; Böiers et al. show a restricted cell. Neither names a required molecule, and a LIN28B-style screen in the first-trimester CD19- IL-7R+ compartment has never been run.
  2. Is LIN28B required for ETV6::RUNX1 pre-leukaemia in the Böiers hPSC system? Zero papers address it, both systems are human, and it is the single most direct import of this beat's result into another fusion family.
  3. Does CBFA2T3::GLIS2 - an infant AMKL fusion with an even more age-restricted epidemiology than KMT2A::AFF1 - have a fetal-restricted requirement? Nothing on this axis exists for it, and its age restriction is the most extreme in paediatric leukaemia, which makes it the highest-prior untested case.
  4. Is the poly(I:C)/infection trigger relevant to KMT2A::AFF1 too? An existing Mll-AF4 fetal-liver mouse study found the cell of origin responds directly to poly(I:C) but that a single maternal immune activation did not produce leukaemia - so the two fusions have been probed with the same reagent and only E/R showed a progenitor expansion.
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Human fetal liver CD34+ cells fail to transform upon induction of the KMT2A::AFF1 translocation in the absence of LIN28B, a fetal-specific gene identified among the top hits of a fetal-versus-adult human HSPC signature and expressed in infant ALL.

    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. · Blood 2026
    PMID 42213620DOI
    What this evidence cannot show

    ABSTRACT ONLY and paywalled (Europe PMC isOpenAccess N, inEPMC N, inPMC N, verified 2026-09-23). The claim that this is the ONLY fully worked fetal-restricted permissive factor is a claim about a literature survey bounded by the queries in this entry, not an exhaustive review.

    Evidence that argues against this (1)
    • The claim that this is the only fully worked fetal-restricted permissive factor is bounded by this entry's own queries, and this same beat had a bounded absence collapse when one extra operand was added (see noncoding-let-7-microrna-family). A survey absence in this project has a demonstrated failure rate and should be treated as provisional. Oncotarget 2016
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MLL-AF9 expressed in mouse fetal liver versus bone marrow cells gives opposite Id1 dependencies - Id1 loss prolongs survival in fetal-liver recipients and accelerates leukaemogenesis in bone-marrow recipients, both rescued by Cdkn1a deletion - and human ID1 expression is elevated in fetal-origin (infant) MLL-fusion-protein leukaemia and reduced in postnatal MLL-FP AML.

    Differential role of Id1 in MLL-AF9-driven leukemia based on cell of origin. · Man N, Sun XJ, Tan Y, García-Cao M, Liu F, Cheng G, et al. · Blood 2016
    PMID 26944543DOI
    What this evidence cannot show

    Abstract read; full text not opened. Mouse, retroviral MLL-AF9, myeloid. ID1 is NOT a fetal-restricted gene - it is expressed in adults - so this is an ontogeny-conditional dependency, not a fetal-restricted factor, and the entry keeps the distinction.

    Evidence that argues against this (1)
    • ID1 is not a fetal-restricted factor at all - it is expressed in adults - so including it in a survey of fetal-restricted permissive factors already requires relaxing the category. If the category is relaxed that far, other ontogeny-conditional dependencies may qualify and the class would be larger than this entry says. Blood 2016
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In an inducible transgenic ETV6::RUNX1 mouse, E/R expression compromised haematopoietic reconstitution and lymphoid differentiation with an almost complete absence of B-cell progenitors from adult BM cells, while the reconstitution ability of fetal liver cells was less affected; E/R-dysregulated genes were enriched for inflammatory pathways, MHC class I antigen presentation, and PD-L1/CD274 and CD200R expression, much more strongly in fetal than adult E/R cells; and poly(I:C) administration increased the pro-B and pre-B compartments of fetal E/R cells to approximately 5-fold and 2.5-fold above wild-type levels, effects not seen in adult E/R cells.

    Ontogeny shapes the ability of ETV6::RUNX1 to enhance hematopoietic stem cell self-renewal and disrupt early lymphopoiesis. · Eldeeb M, Konturek-Ciesla A, Zhang Q, Kharazi S, Tingvall-Gustafsson J, Ungerbäck J, Sigvardsson M, Bryder D · Leukemia 2024
    PMID 38243088DOI
    What this evidence cannot show

    FULL TEXT READ (PMC10844086); it is a Letter, so compact, and most quantitative detail sits in supplementary figures that were not inspected. MOUSE, transgenic E/R rather than an endogenous translocation, and the fold-changes quoted are the authors' own text description of a figure, not values read from a table. Critically for this entry's argument: the paper demonstrates that ontogeny matters and offers immune evasion plus an infection trigger as the mechanism - it does NOT name a fetal-restricted factor, and the authors themselves phrase the leukaemia link as speculation ('we speculate that these cells might subsequently be amenable to leukemic transformation').

    Evidence that argues against this (1)
    • The same paper's headline mechanism is arguably NOT about the fetal state at all: E/R HSCs are outcompeted by wild-type HSCs and removing E/R restores normal differentiation, so the framing is about preleukaemic clone persistence and immune checkpoints rather than about a fetal permissive factor. Citing it as analogous to the LIN28B result requires reading it selectively, and this entry flags that rather than hiding it. Leukemia 2024
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Characterising the first emerging B cells in first-trimester human embryos identified a developmentally restricted CD19-negative IL-7R-positive progenitor compartment that transitions from a myeloid to a lymphoid programme during ontogeny; the series is recapitulated in differentiating human pluripotent stem cells, and genome-engineered hPSCs expressing ETV6-RUNX1 from the endogenous ETV6 locus show expansion of that compartment, a partial block in B-lineage commitment, and proB cells with aberrant myeloid gene expression signatures and potential.

    A Human IPS Model Implicates Embryonic B-Myeloid Fate Restriction as Developmental Susceptibility to B Acute Lymphoblastic Leukemia-Associated ETV6-RUNX1. · Böiers C, Richardson SE, Laycock E, Zriwil A, Turati VA, Brown J, Wray JP, Wang D, James C, Herrero J, Sitnicka E, Karlsson S, Smith AJH, Jacobsen SEW, Enver T · Dev Cell 2017
    PMID 29290585DOI
    What this evidence cannot show

    Abstract read; full text not opened. HUMAN, and the ETV6-RUNX1 is expressed from the ENDOGENOUS ETV6 locus, which is a strong design - but the readout is a pre-leukaemic state, not transformation, and the authors describe the features as 'collectively consistent with a pre-leukemic state' rather than as leukaemia. The restricted compartment is a CELL, not a factor, which is the whole point of citing it here.

    Evidence that argues against this (1)
    • This is a PRE-LEUKAEMIC readout, not transformation - the authors describe the features as 'collectively consistent with a pre-leukemic state'. A restricted compartment that supports a pre-leukaemic expansion is not demonstrated to be required for leukaemia, so its inclusion alongside the KMT2A::AFF1 transformation result compares unlike things. Leukemia 2024
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Live PubMed esearch on 2026-09-23 returns 0 title/abstract records for LIN28B against each of `"ETV6-RUNX1"`, `"TEL-AML1"`, `"CBFA2T3-GLIS2"`, `NUP98`, `"B-ALL"`, `"t(4;11)"` and `"MLL-AF4"`, and 0 for `ID1[tiab] AND KMT2A[tiab]`, against positive controls of 1 for `LIN28B[tiab] AND "KMT2A::AFF1"[tiab]`, 2 for `"infant leukemia"`, 2 for `"MLL-ENL"`, 3 for `lymphoblastic` and 2 for `ID1[tiab] AND "MLL-AF9"[tiab]`.

    What this evidence cannot show

    Rule 80: titles and abstracts only. Rule 72: for ETV6::RUNX1 both the modern and the legacy name (`TEL-AML1`) were tried, but NO legacy synonym was tried for CBFA2T3::GLIS2 or for the NUP98 fusions, and the NUP98 query used the bare gene name rather than any specific fusion. **A beat in this same session was burned by exactly this**: a zero against `KMT2A` collapsed once `MLL` was added (see noncoding-let-7-microrna-family), so these zeros should be read as provisional and re-run with wider name lists before anything is built on them.

    Evidence that argues against this (1)
    • These zeros were generated by the same method that produced a false headline elsewhere in this beat: a KMT2A-side zero that collapsed once the legacy symbol MLL was added. No legacy synonym was tried for CBFA2T3::GLIS2 or the NUP98 fusions, so those two zeros in particular are the least trustworthy rows in the table. Oncotarget 2016
Connected to