PUBLISHED, DONE: human fetal liver CD34+ cells fail to transform on KMT2A::AFF1 induction without LIN28B (Ling et al., Blood 2026) - a dedicated human index-fusion experiment, read from the abstract only
also called Ling 2026 LIN28B requirement · CRISPR-KMT2A::AFF1 LIN28B knockout · the LIN28B initiation requirement
Scientists put the exact infant-leukaemia chromosome swap into human fetal liver blood stem cells with gene editing, and the cells only turned cancerous when a fetal-only gene called LIN28B was present. Switching LIN28B off in leukaemias that had already formed took away their growth advantage, in dishes and in mice.
The same thing, for a scientist
Ling et al. (*Blood*, 2026; PMID 42213620) compared human fetal and adult HSPC transcriptomes to derive a fetal-specific signature, found LIN28B and its downstream effectors among the top hits, and confirmed their expression in infant ALL. In a human fetal-liver-derived CRISPR-Cas9 KMT2A::AFF1 model, **human 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 the authors report LIN28B stabilising early B-lymphoid genes, epigenetic regulators, and cell-cycle and anti-apoptotic genes [E1]. **This entry exists because of what class of experiment this is.** This project measured that **0 of 24 sampled load-bearing experiments rest on a dedicated human KMT2A::AFF1 experiment** (rule of three: true rate under 12.5%). This is one, in the correct human cell of origin, with the endogenous translocation. It is also **abstract-only here**: Europe PMC returns `isOpenAccess: N`, `inEPMC: N`, `inPMC: N` for DOI 10.1182/blood.2025028877, verified live 2026-09-23 [E2]. No figure, no n, no effect size, no knockout efficiency and no control description was read.
What was done, as far as the abstract states it
| element | what the abstract says | what is NOT stated |
|---|---|---|
| perturbation | LIN28B absence (knockout) before translocation induction; LIN28B knockdown in established leukaemias | which nuclease/guide, knockout efficiency, whether knockout is clonal or bulk |
| cells | human fetal liver (FL) CD34+ HSPC | gestational age, donor number, purification scheme |
| lesion | CRISPR-Cas9-induced endogenous KMT2A::AFF1 translocation, inducible | induction efficiency, breakpoint, whether the reciprocal AFF1::KMT2A is formed |
| readout, arm 1 | transformation (yes/no) | the operational definition of transformation - colony serial replating? immortalisation? xenograft? |
| readout, arm 2 | proliferation in vitro and in vivo | doubling times, xenograft engraftment %, survival curves, n of mice |
| controls | none described in the abstract | this is the single most important gap; see caveats |
| mechanism | LIN28B stabilises early B-lymphoid genes, epigenetic regulators, cell-cycle and anti-apoptotic genes | which transcripts, by what assay (CLIP? RNA stability? ribosome profiling?), and whether any are fusion targets |
The normal-biology arm matters as much as the leukaemia arm and is easy to skip past. The same abstract states LIN28B was essential in human FL CD34+ cells to maintain proliferation and stem-like properties, and support B- and NK-lymphopoiesis. So the perturbation that abolishes transformation also damages the normal cell it is applied to. Nothing in the abstract separates "LIN28B is required for transformation" from "LIN28B is required for the target cell to be a competent, proliferating progenitor at all", and that distinction is the whole therapeutic argument. See risk-lin28b-target-overlaps-normal-fetal-lymphopoiesis.
Why this is a rare object in this graph
This project's partner-provenance audit found 0 of 24 sampled load-bearing experiments rest on a dedicated human KMT2A::AFF1 experiment - almost everything in the mechanism graph is imported from MLL-AF9 mouse, from a cell line, or from wild-type KMT2A. This experiment is:
- human (not mouse),
- in the index fusion KMT2A::AFF1 (not a partner-imported analogue),
- in the candidate cell of origin (fetal liver CD34+, the compartment
cell-type-fetal-liver-hscmodels), - with the endogenous translocation rather than a cDNA transgene,
- and it is an initiation experiment, which is the phase almost nothing in the literature can address because human infant ALL is diagnosed long after initiation.
The model system itself is already in the graph as model-crispr-t411-human-fetal-liver; this entry is the LIN28B perturbation run in it, not the model.
The counter-evidence, and it is not a footnote
Two mouse studies from the Magee and Bryder groups point the other way in the myeloid lineage, and one of them is a direct test of the fetal-permissiveness premise.
Okeyo-Owuor et al., Blood Adv 2019 3. An inducible transgenic MLL-ENL mouse. Induction in fetal or neonatal mice produced AML in almost all animals; induction in adults left most alive >1 year despite sustained expression. But initiation was most efficient NEONATALLY, not fetally, and - the part that bites - "Since fetal/neonatal progenitors were highly competent to initiate MLL-ENL-driven AML, we tested whether LIN28B, a fetal master regulator, could accelerate leukemogenesis. Surprisingly, LIN28B suppressed AML initiation rather than accelerating it." The authors propose this as the explanation for why MLL rearrangements arise before birth but transformation usually waits until after birth, when LIN28B declines.
Li et al., Leukemia 2024 4 (full text read from PMC10912017, not abstract-only). Conditional Lin28b deletion in fetal progenitors carrying Tet-OFF MLL::ENL did not accelerate AML initiation - survival curves were indistinguishable between genotypes - so LIN28B is not the effector of fetal protection. And enforced LIN28B in fully transformed MLL::AF9 AML significantly extended survival, with 80% of mice surviving >4 months, caused partial monocytic differentiation by day 6, and severely depleted colony-forming potential in 5 biologically independent AMLs without raising apoptosis.
How much of Ling et al. this actually threatens. Less than it first looks, because every axis differs - species (human vs mouse), lineage (B vs myeloid), partner (AFF1 vs MLLT1/MLLT3), lesion (endogenous translocation vs transgene) and direction of perturbation (loss vs enforced gain). But it does refute the general proposition that LIN28B is a pro-leukaemic fetal factor in KMT2A-rearranged disease, and it means the lineage and the species are doing load-bearing work in the interpretation. See hyp-fetal-state-is-permissive-for-lymphoid-and-protective-for-myeloid-kmt2a-transformation.
| status | PUBLISHED AND DONE - this is not a proposal. It is recorded as an experiment node so the graph can hold its design, readout and controls separately from the gene entry. |
|---|---|
| design | Human fetal liver CD34+ HSPC; CRISPR-Cas9-induced endogenous KMT2A::AFF1 translocation, inducible; LIN28B absent versus present at the point of induction. Second arm: shRNA/knockdown of LIN28B in already-established CRISPR-KMT2A::AFF1 leukaemias, assayed in vitro and in vivo. |
| readout | Arm 1: transformation (binary, operational definition not stated in the abstract). Arm 2: proliferation in vitro and in vivo. Mechanistic arm: transcript stabilisation of early B-lymphoid genes, epigenetic regulators, cell-cycle and anti-apoptotic genes. |
| controls | NOT DESCRIBED IN THE ABSTRACT. The paper is paywalled and the control arms were not read. This field is deliberately left as an absence rather than filled by inference. |
| cost tier | high |
| timeline | published 2026-05-29 |
| species | human |
| lineage | B-lymphoid (and NK, in the normal-biology arm) |
| fusion | fusion-kmt2a-aff1 |
| lesion type | endogenous CRISPR-engineered translocation, not a cDNA transgene |
| cell type | cell-type-fetal-liver-hsc |
| model | model-crispr-t411-human-fetal-liver |
| phase of disease addressed | INITIATION (arm 1) and maintenance/proliferative advantage (arm 2). The initiation arm is the rarer and more novel one; the maintenance arm is the therapeutically relevant one. |
| read depth | abstract only - full text paywalled, verified against Europe PMC 2026-09-23 |
| shown for fusion | YES, KMT2A::AFF1, in human fetal cells with the endogenous translocation - the strongest fusion-specific and cell-of-origin-specific claim this beat found. NOT shown for KMT2A::MLLT3, KMT2A::MLLT1, ETV6::RUNX1 or any other fusion; in mouse MLL-ENL and MLL-AF9 the sign is reversed. |
| why this experiment class is rare | 0 of 24 sampled load-bearing experiments in this graph rest on a dedicated human KMT2A::AFF1 experiment (rule of three: true rate under 12.5% with 95% confidence). This is one. |
- The perturbation that abolishes transformation also abolishes normal human fetal B- and NK-lymphopoiesis, in the same paper. In an infant patient those are the same cells.
- An initiation requirement is the least therapeutically actionable kind of requirement, because the patient presents after initiation. Only the knockdown-in-established-leukaemia arm speaks to treatment, and it reports a lost growth advantage rather than regression.
- If LIN28B's role is lineage-dependent - required in fetal B, tumour-suppressive in myeloid - then a LIN28B inhibitor given to a KMT2A-rearranged patient whose disease has switched lineage (a documented phenomenon in this disease under blinatumomab and CAR-T) could accelerate the myeloid clone.
- Because the full text is unreadable here, the graph cannot check the controls of the experiment it is leaning on most heavily for the fetal-permissiveness argument. That is a delivery-of-evidence risk, not a biology risk, and it is the honest state.
- What is the operational definition of 'transformation' in arm 1 - serial replating, immortalisation, or xenograft initiation? The strength of a negative result depends entirely on which, and the abstract does not say.
- Is the failure to transform specific to the fusion, or is it that a LIN28B-null fetal CD34+ cell is no longer a proliferating progenitor competent to be transformed by anything? The paper's own normal-biology arm - LIN28B is essential to maintain proliferation and stem-like properties - makes this the first control a reader wants, and it is invisible from the abstract. A second oncogene (e.g. a KMT2A::MLLT3 arm, or a non-KMT2A driver) in the same LIN28B-null cells would settle it.
- How many fetal donors, and at what gestational ages? Human fetal liver material is scarce and donor-to-donor variation in a transformation assay is large.
- Does the knockdown arm in established leukaemias produce regression or only a reduced proliferative rate? The abstract says 'advantage being lost', which is the weaker of the two and is the one that decides whether LIN28B is a therapeutic target or only an initiation factor.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In a human fetal-liver-derived CRISPR-Cas9 KMT2A::AFF1 model, human 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, with that advantage lost upon LIN28B knockdown; and LIN28B was separately essential in human FL CD34+ cells to maintain proliferation and stem-like properties and to support B- and NK-lymphopoiesis.
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, Cross JW, Wuppalapati AS, Smith AL, Chahrour C, Sevim O, Palmer DA, Bozhilov Y, Brown EJ, Olender L, Iskander D, Wang G, Rice S, O'Byrne S, Harman J, Ancliff P, Psaila B, Wilkinson AC, Morgan RG, Roberts IA, Milne TA, Roy A · Blood 2026What this evidence cannot show
ABSTRACT ONLY, AND THE PAPER IS PAYWALLED. Verified live 2026-09-23 against Europe PMC: isOpenAccess N, inEPMC N, inPMC N, the only fullTextUrl being the subscription DOI [E2]. Therefore this entry knows NO effect size, NO donor number, NO in vivo n, NO knockout efficiency, NO definition of 'transformation', and NO control arms. The quoted claim is the authors' own summary sentence, reproduced rather than assessed. A CRISPR-induced translocation in primary human fetal liver cells is a low-efficiency system, and 'fail to transform' in such a system is exactly the kind of negative that depends on the paper's own positive controls, which were not read. Identifier provenance: PMID, DOI, journal, year, full author list and abstract were fetched live from PubMed on 2026-09-23 and the PMID resolves to this exact title.
Evidence that argues against this (1)
- In mouse myeloid MLL-ENL, LIN28B SUPPRESSED AML initiation rather than accelerating it, and AML initiation peaked neonatally rather than fetally - the opposite sign for both the factor and the developmental window. Okeyo-Owuor T et al. 2019
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
As of 2026-09-23 the Ling et al. Blood 2026 paper is not open access and is not in PubMed Central or Europe PMC: the record returns isOpenAccess 'N', inEPMC 'N', inPMC 'N', no pmcid, and a single fullTextUrl of availability 'Subscription required' pointing at the DOI.
What this evidence cannot show
Fetched live 2026-09-23. This is a statement about availability on that date only; Blood deposits to PMC after an embargo, so a later reader should re-check before repeating 'abstract-only'. Rule 80 applies with unusual force here: because the full text cannot be reached at all, no supporting figure in this paper has been searched for anything, and the graph's bounded absences about this paper's content are bounded by ONE abstract.
Evidence that argues against this (1)
- Availability is a moving target, not a fact about the paper. Blood deposits to PMC after an embargo, so this record's 'not open access' status will become false and any bounded statement resting on 'the full text could not be read' silently expires. A later reader must re-check rather than inherit this. source 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In an inducible transgenic MLL-ENL mouse, AML initiation efficiency changed with developmental stage and peaked neonatally rather than fetally, and enforced LIN28B - tested specifically as a candidate accelerator because fetal/neonatal progenitors were so competent - suppressed AML initiation instead; the authors propose declining postnatal LIN28B as the explanation for why MLL rearrangements arise before birth but transformation usually follows birth.
The efficiency of murine MLL-ENL-driven leukemia initiation changes with age and peaks during neonatal development. · Okeyo-Owuor T, Li Y, Patel RM, Yang W, Casey EB, Cluster AS, Porter SN, Bryder D, Magee JA · Blood Adv 2019What this evidence cannot show
Abstract read; full text not opened. Mouse, transgenic MLL-ENL (KMT2A::MLLT1 analogue), MYELOID disease. A transgene is not an endogenous translocation, and the LIN28B arm is enforced expression, not loss - the converse perturbation to Ling et al. The abstract also notes that transplantation stress mitigated the developmental barrier, which means the readout is sensitive to handling.
Evidence that argues against this (1)
- The human, B-lineage, endogenous-translocation result in the index fusion points the other way, and none of species, lineage, partner or perturbation direction is shared between the two studies. Blood 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Conditional Lin28b deletion in fetal progenitors of Tet-OFF MLL::ENL mice did not accelerate AML initiation or shorten survival relative to Lin28b-intact controls, so LIN28B is not the sole effector of fetal leukaemia suppression; and enforced LIN28B in established MLL::AF9 AML significantly extended survival with 80% of secondary recipients alive beyond 4 months, induced partial monocytic differentiation by 6 days, and severely depleted colony-forming potential in 5 biologically independent AMLs without increasing programmed cell death.
LIN28B promotes differentiation of fully transformed AML cells but is dispensable for fetal leukemia suppression. · Li Y, Mendoza-Castrejon J, Patel RM, Casey EB, Denby E, Bryder D, Magee JA · Leukemia 2024What this evidence cannot show
FULL TEXT READ (PMC10912017) - this is not a title-only or abstract-only citation, unlike the entry's primary source. It is a Letter, so the experiments are compact. Mouse, myeloid, transgenic fusions (MLL::ENL and MLL::AF9), not the index fusion. The authors themselves flag two limits: LIN28B levels in the transgenic models exceeded physiological levels, and the fetal-deletion experiment was terminated at ~6 months so a later effect is not excluded (they argue against one on the grounds that adult progenitors and MLL::ENL AML cells do not express Lin28b transcript).
Evidence that argues against this (1)
- A third sign exists in the same disease: Zhou et al. report that INHIBITING LIN28B in AML cells causes cell-cycle arrest and loss of proliferation and colony formation via let-7a de-repression, and that LIN28B overexpression increases tumorigenicity in vivo - i.e. LIN28B as a myeloid oncogene, the opposite of the Magee/Bryder tumour-suppressor reading. Zhou J et al. 2017
LIN28B knockout before translocation induction, and knockdown in established leukaemia
endogenous CRISPR-engineered KMT2A::AFF1 in human fetal liver CD34+
the LIN28B perturbation run in this model
- expressed in →Fetal liver haematopoietic stem and progenitor cell — the leading candidate cell of origin for t(4;11) infant ALL
human fetal liver CD34+ HSPC - the cells the experiment was done in
the fetal state is permissive rather than incidental to transformation
an initiation experiment for the disease's defining in-utero origin
the index-fusion experiment reports a route-2 mechanism (transcript stabilisation)
the same perturbation abolishes normal progenitor competence, so the transformation block may not be fusion-specific
the lymphoid arm of the contradiction
the missing specificity control for the published result
most relevant system, least readable; the design-quality gradient runs against it
what the experiment shows versus what it is quoted as showing
liability 3 - the requirement demonstrated is for initiation, not maintenance