PROPOSED: measure let-7 and restore it in the human fetal-liver KMT2A::AFF1 model - an axis measured in mouse MLL-fusion AML but never in the index fusion
also called let-7 phenocopy experiment · does let-7 restoration phenocopy LIN28B loss
LIN28B is famous for holding down a growth brake called let-7. If that is how it lets infant leukaemia start, then putting the brake back should stop the leukaemia just as removing LIN28B does. In mouse MLL-fusion leukaemia that is exactly what putting the brake back does - and nobody has tried it in the human infant fusion.
The same thing, for a scientist
Two steps, and the first is not optional. **Step 1 is a measurement that does not exist in this system**: mature let-7 levels in human fetal liver CD34+ cells, LIN28B-null versus intact, and in CRISPR-KMT2A::AFF1 derivatives. Live PubMed esearch on 2026-09-23 returns **zero** title/abstract records for `let-7[tiab] AND KMT2A[tiab]`, for `"MLL-AF4"` and for `"t(4;11)"`, with clean positive controls [E1]. **CORRECTED SCOPE:** the operand `MLL` returns three, and Chen et al. 2016 measured let-7g across mouse MLL-fusion AML models and showed forced let-7g differentiates blasts [E4] - so the axis is not untouched in MLL-rearranged disease, only in the index fusion and in human infant ALL. That makes step 1 a first measurement **in this system** rather than the first anywhere, and it makes step 2 better motivated, because forced let-7 is already known to be differentiating in a KMT2A fusion. **Step 2 is the epistasis test**: restore let-7 in LIN28B-intact cells and ask whether it phenocopies LIN28B loss. If it does, the druggable node moves from an RNA-binding protein to the let-7 axis; if it does not, LIN28B is acting through direct mRNA stabilisation, which is the only mechanism actually reported in the index fusion [E2] and the only one reported in any KMT2A fusion [E3]. **This experiment answers an open question already posed in `gene-lin28b` and left unresolved.**
Design
| step | what | why it is needed |
|---|---|---|
| 1. baseline | small-RNA qPCR or small-RNA-seq for the mature let-7 family (all 12 loci, and note that MIR98 is a family member whose name contains no 'let-7') in human FL CD34+ LIN28B-null vs intact, and in CRISPR-KMT2A::AFF1 derivatives | no let-7 measurement exists in KMT2A::AFF1, in t(4;11), or in any human infant ALL (it does exist in mouse MLL-fusion AML). Without it, step 2 is uninterpretable and the axis remains assumed |
| 2. epistasis | let-7 mimic or a let-7 sponge/decoy in LIN28B-INTACT cells; ask whether transformation frequency falls to the LIN28B-null level | tests whether let-7 de-repression is the mechanism or a bystander |
| 3. rescue | let-7 sponge in LIN28B-NULL cells; ask whether transformation is restored | the necessity arm. If blocking let-7 rescues transformation in LIN28B-null cells, route 1 is confirmed. If it does not, route 1 is excluded |
| controls | scrambled mimic; a let-7-insensitive HMGA2 3'UTR-truncated reporter to confirm mimic activity; LIN28A status (LIN28A uses a different, cytoplasmic mechanism and its presence would confound) | |
| readout | transformation frequency, plus HMGA2 and IGF2BP1 protein as let-7 target reporters |
Step 3 is the one that discriminates, and it is the one usually skipped. A mimic that blocks transformation proves let-7 is sufficient to block it, which almost any growth brake would be. Only the sponge-rescue arm shows that let-7 de-repression is what LIN28B is for in this cell.
Why the prior is against route 1 in the index fusion - but less strongly than this entry first claimed
CORRECTED IN-SESSION 2026-09-23. An earlier draft of this entry said let-7 had never been measured in any KMT2A rearrangement. That was a NAME-LIST ARTEFACT: the operand MLL was not tried and returns three records, one of which (Chen et al., Oncotarget 2016, PMID 27007052) measures let-7g across MLL-fusion AML models and shows forced let-7g differentiates blasts. The surviving, narrower absence is: no let-7 measurement in KMT2A::AFF1, in t(4;11), or in human infant ALL. See noncoding-let-7-microrna-family for the full query table.
Both LIN28B mechanisms ever measured in a KMT2A fusion are direct mRNA effects, not let-7 effects: MYBBP1A stabilisation in mouse MLL-ENL 3, and stabilisation of early B-lymphoid, epigenetic, cell-cycle and anti-apoptotic transcripts in human KMT2A::AFF1 2. So this beat's expectation is that let-7 restoration will NOT fully phenocopy LIN28B loss.
That is a reason to run it, not to skip it. A clean negative would:
- retire a mechanism this graph and the wider field have been drawing for years without a measurement,
- redirect any target-discovery effort from let-7 mimetics (a crowded and repeatedly disappointing modality) to LIN28B's RNA-binding surface or to its specific stabilised transcripts,
- and produce the first let-7 dataset in the INDEX fusion and in a human fetal B-lymphoid system, which has value independent of the outcome.
The cheapest useful version is step 1 alone, on material that already exists: a small-RNA panel costs almost nothing and would be the first such measurement in the literature.
| status | planned |
|---|---|
| design | Step 1: small-RNA qPCR/seq for the mature let-7 family across LIN28B-null and LIN28B-intact human fetal liver CD34+ cells and CRISPR-KMT2A::AFF1 derivatives. Step 2: let-7 mimic in LIN28B-intact cells, transformation readout. Step 3: let-7 sponge/decoy in LIN28B-null cells, transformation rescue readout. |
| readout | Mature let-7 family levels (all 12 loci including MIR98); transformation frequency; HMGA2 and IGF2BP1 protein as let-7 target reporters. |
| controls | Scrambled mimic; a let-7-insensitive 3'UTR-truncated HMGA2 reporter to confirm mimic activity; LIN28A expression status, since LIN28A uses a distinct cytoplasmic ZCCHC11-dependent mechanism and would confound. |
| cost tier | low for step 1, med for steps 2-3 |
| timeline | step 1: weeks, on existing material. Steps 2-3: 4-8 months. |
| decision rule | Route 1 is confirmed only if the SPONGE RESCUE arm works: blocking let-7 in LIN28B-null cells restores transformation frequency to within 2-fold of LIN28B-intact controls. A mimic that blocks transformation in step 2 alone is not sufficient evidence, because almost any growth brake would do that. |
| expected outcome | This beat expects let-7 restoration NOT to FULLY phenocopy LIN28B loss in the index fusion, because both mechanisms traced to a named target inside a KMT2A fusion are direct mRNA effects. But the prior is less one-sided than the first draft of this entry said: forced let-7g already differentiates blasts in mouse MLL-fusion AML (PMID 27007052), so a partial phenocopy is a live possibility. |
| cheapest useful version | Step 1 alone - the first let-7 measurement in KMT2A::AFF1 and in a human fetal B-lymphoid system, on material that already exists. (Not the first in MLL-rearranged leukaemia: that exists in mouse AML.) |
answers open question in
[ "gene-lin28b", "noncoding-let-7-microrna-family", "mechanism-lin28b-has-two-routes-and-the-leukaemia-evidence-is-all-on-the-non-let-7-one" ]
- let-7 mimetics are a modality with a long record of preclinical promise and delivery failure. A positive result here would point at a target class that has repeatedly not reached patients.
- let-7 restoration is a broad growth brake with hundreds of targets. In an infant, 'advance the developmental clock' is not a precisely dosable intervention.
- A negative result removes the mechanism but leaves the requirement, which is the worst position for a drug-discovery programme: a validated dependency with no tractable mechanism.
- Do human fetal liver CD34+ cells express LIN28A as well as LIN28B? If so, LIN28B knockout would not remove let-7 repression and every interpretation above changes.
- Which let-7 family members are expressed in fetal B progenitors? Family-level assays would hide a single-locus signal, and the 12 loci are not co-regulated.
- Does KMT2A::AFF1 itself change let-7 levels, independent of LIN28B? The fusion is a transcriptional regulator and several let-7 loci are intronic or in regulated clusters.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Live PubMed esearch on 2026-09-23 returns 0 records for `let-7[tiab] AND KMT2A[tiab]`, `let-7[tiab] AND "MLL-AF4"[tiab]` and `let-7[tiab] AND "t(4;11)"[tiab]`, each with a non-empty querytranslation, against positive controls of 1 for `LIN28B[tiab] AND KMT2A[tiab]` and 3 for `LIN28B[tiab] AND lymphoblastic[tiab]` in the same session.
What this evidence cannot show
Rule 80: bounded by titles and abstracts only. The paper most likely to hold a let-7 panel in a supporting figure - Ling et al. 2026 - is paywalled and absent from PMC, so its figures could not be searched. Rule 72: only the string `let-7` was tried on the microRNA side; `lethal-7`, `MIRLET7*` and `miR-98` were not. The KMT2A side used three names and did not try `MLL` (too polysemous to bound). **SCOPE CORRECTED IN-SESSION:** the operand `MLL` was subsequently tried and returns 3 records (1 for `"MLL-rearranged"`), so these zeros bound the index fusion and the modern symbol, NOT the field. See E4 and noncoding-let-7-microrna-family.
Evidence that argues against this (1)
- let-7 HAS been measured in AML more broadly: Zhou et al. report that LIN28B knockdown in AML cells acts through let-7a de-repression with IGF2BP1 downstream. So the axis is not untested in myeloid leukaemia generally - only in KMT2A-rearranged disease. J Hematol Oncol 2017
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In the human fetal-liver CRISPR KMT2A::AFF1 model, LIN28B is reported to act by stabilising key early B-lymphoid genes, epigenetic regulators, and cell-cycle and anti-apoptotic genes - a direct transcript-level mechanism, with no let-7 result in the abstract.
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 2026What this evidence cannot show
ABSTRACT ONLY and paywalled. Absence of a let-7 result in an abstract is not absence of one in the paper. If the full text contains a let-7 panel, step 1 of this experiment is already done.
Evidence that argues against this (1)
- Absence of a let-7 result in an abstract is not absence of one in the paper, and this paper's figures cannot be searched because it is paywalled and absent from PMC. If Ling et al. already measured let-7, step 1 of this experiment is redundant and the entry's premise is wrong. source 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In mouse MLL-ENL AML, LIN28B acts through a let-7-independent route: it binds MYBBP1A mRNA directly to raise the MYB co-repressor MYBBP1A, and MYBBP1A overexpression phenocopies while MYBBP1A perturbation abolishes the LIN28B effect.
A fetal tumor suppressor axis abrogates MLL-fusion-driven acute myeloid leukemia. · Eldeeb M, Yuan O, Guzzi N, Thi Ngoc PC, Konturek-Ciesla A, Kristiansen TA, Muthukumar S, Magee J, Bellodi C, Yuan J, Bryder D · Cell Rep 2023What this evidence cannot show
Abstract read; full text not opened. Mouse, myeloid, transgene, opposite phenotypic direction to the human lymphoid result. It establishes that a let-7-independent route exists in a KMT2A fusion, not that it is the route in KMT2A::AFF1.
Evidence that argues against this (1)
- A let-7-independent route existing in one KMT2A fusion does not make route 1 inoperative in another, and forced let-7g IS differentiating in mouse MLL-fusion AML - so the prior against route 1 in the index fusion is weaker than the MYBBP1A result alone suggests. Oncotarget 2016
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Across mouse AML models transformed with different MLL fusion proteins, Lin28B was differentially expressed in agreement with c-Myc, let-7g was significantly increased in the leukaemias with the longest disease latency, and forced let-7g expression induced differentiation of leukaemic blasts; the authors conclude the c-Myc/Lin28/let-7g programme is differentially regulated by different MLL fusion partners and relates to disease latency and BET-inhibitor resistance.
Differential regulation of the c-Myc/Lin28 axis discriminates subclasses of rearranged MLL leukemia. · Chen L, Sun Y, Wang J, Jiang H, Muntean AG · Oncotarget 2016What this evidence cannot show
Abstract read; full text not opened, so WHICH MLL fusion partners were compared is unknown - and whether AFF1/AF4 was among them is the single fact this project most needs from the paper. Mouse, myeloid, retroviral MLL-FP models. Oncotarget's editorial standards over this period have been questioned; treated as a strong pointer rather than as established. THIS ITEM WAS ADDED AS A SELF-CORRECTION - it refutes an earlier draft of this entry.
Evidence that argues against this (1)
- In human paediatric AML material the let-7/LIN28B/HMGA2 axis was reported NOT to be operative in MLL-rearranged blasts, in contrast to t(8;21) blasts - so the mouse MLL-FP result may not transfer. Leukemia 2014
step 1 would be the first let-7 measurement in any KMT2A-rearranged system
epistasis: does let-7 restoration phenocopy LIN28B loss
route 1 versus route 2 in the index fusion