The decisive experiment: acute degron removal of IGF2BP3, IGF2BP2 and both, read on self-renewal and xenograft initiation rather than bulk dropout
also called IGF2BP degron experiment · IGF2BP double-degron
Instead of switching off IGF2BP3 for two weeks and counting cells, this experiment removes the protein within hours and asks whether the leukaemia can still start fresh disease in a mouse. It also removes the sibling protein, alone and together, because a gene that has a backup will look unimportant when tested by itself.
The same thing, for a scientist
A 2x2 acute-degradation design in human KMT2A::AFF1 lines (SEM, RS4;11, SEMK2) with endogenous degron tags on IGF2BP3 and IGF2BP2: vehicle, IGF2BP3 degraded, IGF2BP2 degraded, both degraded. Primary readouts are **serial replating** and **limiting-dilution xenograft initiation**, with bulk proliferation as a secondary readout kept explicitly to reproduce the negative that motivated the design. The experiment is built to answer two questions at once that the existing literature answers separately and inconsistently: whether a two-week CRISPR dropout is the wrong assay for an amplifier of leukaemia-initiating-cell function [E1][E4], and whether the family buffers itself [E2][E5].
Why not CRISPR, and why not shRNA
The gap this experiment closes is between a mouse deletion that greatly attenuates MLL-Af4 disease 1 and a human Chronos mean of -0.137 with 1.6% of lines depleted 5. Three differences could produce that gap, and the design controls all three.
Timescale and selection. A CRISPR dropout runs for two to three weeks in lines already selected for growth in culture - which is selection against dependence on a self-renewal amplifier. A degron removes the protein in hours, before the population can adapt, so the measured phenotype is the protein's acute requirement rather than the survivors' tolerance of its absence.
Readout. Bulk proliferation is the wrong endpoint for a leukaemia-initiating-cell phenotype. Serial replating and limiting-dilution transplantation measure the quantity the mouse genetics actually reported.
Gene number. Both single arms and the double arm are run in the same experiment on the same cells, so redundancy is measured rather than invoked. This is the falsifier for hyp-igf2bp-family-redundancy-kmt2a.
Design1 deeper
Cells. SEM, RS4;11 and SEMK2 (KMT2A::AFF1) as index lines; HB11;19 (KMT2A::MLLT1, B-lymphoid) as a KMT2A-rearranged non-AFF1 comparator; NALM-6 or a comparable non-KMT2A B-ALL line as a fusion-negative comparator. Three index lines because the entire lymphoid arm of this project's DepMap scan rests on four cell lines and a single line's idiosyncrasy moves every number in it 5.
Tagging. Degron knock-in at the endogenous locus of IGF2BP3 and of IGF2BP2, singly and in the same clone for the double arm. Endogenous tagging, not over-expression, because the whole objection to the RNA-binding-protein literature is that abundance tracks biosynthetic demand 6 and an over-expressed construct makes that objection unanswerable.
Arms. vehicle / degrade IGF2BP3 / degrade IGF2BP2 / degrade both, each in each line.
Primary readouts. (1) Serial replating in methylcellulose, three rounds, colony number per round. (2) Limiting-dilution xenograft in NSG mice with degradation maintained from before transplant, scoring engraftment frequency by Poisson statistics. Engraftment frequency, not survival, because an amplifier of initiation should change the frequency of cells able to start disease.
Secondary readouts. Bulk proliferation over 14 days, to reproduce the CRISPR result and show directly whether the assays diverge on the same cells. Differentiation markers, because both Igf2bp3 depletion and menin-MLL inhibition increase differentiation 3. Target-transcript panel covering HOXA-locus and RAS-pathway messages 1.
Controls. Degron-tagged but untreated clones, to exclude tag effects. A tagged pan-essential gene as an assay-sensitivity positive control. Parental untagged lines. Critically, an arm measuring IGF2BP2 protein after IGF2BP3 degradation and the reverse - the compensation measurement, which is the cheapest falsifier and can be run first as a standalone western blot.
Cost and timeline. High tier; roughly 9-12 months, dominated by the limiting-dilution xenograft.
Order of operations if the budget is smaller
The compensation western blot is a week of work on material that already exists in the laboratories that published 1 and 3, and it is a genuine falsifier: no compensation, no buffering mechanism, and the redundancy hypothesis is finished without any degron being built.
Second cheapest is serial replating on existing CRISPR-knockout cells. If IGF2BP3 knockout already shows a replating defect that the two-week dropout missed, the assay-mismatch explanation is supported and the degron becomes a refinement rather than a necessity.
The xenograft arm is the expensive part and should be run only after one of those two returns positive.
What each outcome would mean
| result | reading |
|---|---|
| Single IGF2BP3 degradation impairs replating and xenograft initiation but not bulk growth | The assay was wrong, not the biology. Reconciles mouse and DepMap without redundancy. |
| Neither single arm does anything, the double arm does | Redundancy confirmed; single-gene CRISPR under-reported both; a selective inhibitor is the wrong molecule. |
| Nothing in any arm | The human t(4;11) dependency does not exist and the mouse result is species- or construct-specific. The graph should say so. |
| Bulk growth falls as much as replating | The initiation framing is wrong and the CRISPR result was simply a small effect. |
The fourth row is the one that would most change this beat's entries, and it is not the expected outcome, which is why it is worth stating in advance.
| design | 2x2 acute degron removal of IGF2BP3 and IGF2BP2, singly and together, in three human KMT2A::AFF1 lines (SEM, RS4;11, SEMK2) plus a KMT2A-rearranged non-AFF1 line (HB11;19) and a fusion-negative B-ALL comparator; endogenous knock-in tagging, no over-expression |
|---|---|
| readout | primary: serial replating colony number over three rounds, and limiting-dilution xenograft engraftment frequency in NSG mice by Poisson statistics. secondary: 14-day bulk proliferation (to reproduce the CRISPR negative on the same cells), differentiation markers, HOXA-locus and RAS-pathway target transcripts |
| controls | degron-tagged untreated clones; parental untagged lines; a tagged pan-essential gene as assay-sensitivity control; fusion-negative comparator line; and a standalone compensation arm measuring IGF2BP2 protein after IGF2BP3 degradation and the reverse |
| cost tier | high |
| timeline | 9-12 months, dominated by the limiting-dilution xenograft; the compensation western blot is one week and is a standalone falsifier that should run first |
| cheapest falsifier first | IGF2BP2 protein does not rise after IGF2BP3 degradation - kills the redundancy hypothesis for the price of a western blot on existing material |
tests hypotheses
[ "hyp-igf2bp-family-redundancy-kmt2a" ]
- A degron shows what acute removal does, which is not what a drug does. Partial reversible occupancy of an RNA-binding protein may never reach the threshold a degron crosses.
- If the double arm is required to see an effect, the therapeutic implication is that both paralogues must be inhibited - and IGF2BP2 is high in normal marrow, so the window may close as soon as the biology opens.
- Xenograft engraftment frequency measures the ability to colonise an immunodeficient mouse, not leukaemia-initiating-cell frequency in a child, and the two have repeatedly diverged.
- The experiment is designed to find a phenotype the CRISPR data do not show. Designing an assay until it agrees with the mouse is a real risk here, and the bulk-proliferation secondary readout is kept specifically so the disagreement stays visible.
- Can IGF2BP2 and IGF2BP3 be degron-tagged in the same clone without disturbing either locus's regulation, given they are co-regulated in this disease?
- Should the xenograft arm maintain degradation throughout or model intermittent exposure, which is what a drug would actually deliver?
- Does the fusion-negative comparator need to be lineage- and maturation-matched to t(4;11) blasts for the biosynthetic-demand control to work?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Igf2bp3 deletion greatly attenuates MLL-Af4-driven leukaemia in mice with minimal impact on baseline haematopoiesis, MLL-Af4 leukaemia-initiating cells require Igf2bp3, and IGF2BP3 targets include HOXA-locus and RAS-pathway transcripts with effects on steady-state mRNA level and pre-mRNA splicing.
The RNA-binding protein IGF2BP3 is critical for MLL-AF4-mediated leukemogenesis · Tran TM, et al. · Leukemia 2022What this evidence cannot show
Abstract read; full text not read. Murine, chimeric construct, deletion before disease. This is the result the experiment is designed to test in human cells, not a result the experiment assumes.
Evidence that argues against this (1)
- Minimal impact on baseline haematopoiesis was measured in an unstressed mouse. This graph's menin work found preserved steady-state progenitors with a long-term HSC defect revealed only under competitive transplantation, so the safety half of this result is the one least likely to hold, and the proposed experiment does not test normal-tissue toxicity at all. Oncofusion knowledge graph (internal analysis) 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MLL/AF4-positive B-ALL over-expresses IGF2BP2 and IGF2BP3 together, which is the co-expression a buffering model requires and which motivates running both single arms and the double arm in the same experiment.
Identification of characteristic IGF2BP expression patterns in distinct B-ALL entities · Stoskus M, et al. · Blood Cells Mol Dis 2011What this evidence cannot show
Transcript-level, 2011, single centre. Co-expression motivates the double arm; it does not predict its outcome.
Evidence that argues against this (1)
- MLL-AF4 induces RNA-binding protein genes as a counter-regulatory response, so co-expression of two paralogues may simply reflect a shared upstream driver with no functional relationship - in which case the double arm returns nothing beyond the singles and the expense was unnecessary. Hanewald T et al. 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Depletion of Igf2bp3 sensitises MLL-AF4 leukaemia to menin-MLL inhibition on cell growth and on leukaemic initiating cells in vitro; both manipulations increase differentiation; IGF2BP3 knockdown had a greater effect on increasing survival and attenuating disease than the menin-MLL inhibitor MI-503 alone, with enhanced antileukaemic effects in combination.
Targeting IGF2BP3 enhances antileukemic effects of menin-MLL inhibition in MLL-AF4 leukemia · Lin TL, et al. · Blood Adv 2024What this evidence cannot show
Abstract read; full text not read. Murine MLL-AF4 models, genetic depletion of Igf2bp3 rather than pharmacological. MI-503 is a tool compound rather than a clinical menin inhibitor, so the comparison is not against revumenib as dosed in patients.
Evidence that argues against this (1)
- A genetic knockdown outperforming a small molecule is not a like-for-like comparison: complete protein removal beats partial reversible occupancy for almost any target, and the result says more about the assay than about the relative merit of the two targets. Oncofusion knowledge graph (internal analysis) 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MSI2 and SYNCRIP were identified in MLL-driven leukaemia through leukaemia-stem-cell readouts - loss of the self-renewal programme, delayed leukaemogenesis, increased differentiation and apoptosis - rather than through bulk proliferation, establishing that the phenotype class this experiment targets is the one this field's positive results come from.
Functional screen of MSI2 interactors identifies an essential role for SYNCRIP in myeloid leukemia stem cells · Vu LP, et al. · Nat Genet 2017What this evidence cannot show
Abstract read; full text not read. Murine, myeloid MLL leukaemia. Cited to justify the readout choice, not to transfer the biology to B-lymphoid disease.
Evidence that argues against this (1)
- Leukaemia-stem-cell assays are notoriously assay-dependent, and a field whose positive results come predominantly from one assay class should be read with that in mind rather than treated as proof the assay is correct. Oncofusion knowledge graph (internal analysis) 2026
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis
DepMap gives IGF2BP3 a Chronos mean of -0.137 with 1.6% of 1,208 lines depleted, and the three IMP paralogues show no co-dependency - pairwise correlations 0.066-0.105 across all lines, negative within 54 leukaemia lines, 4 lines depleted for two or more against 2.52 expected under independence. The entire B-lymphoid index arm of that scan is four cell lines.
The RNA layer in KMT2A-rearranged leukaemia lines: panel-wide selectivity scan and IGF2BP family co-dependency test · Oncofusion knowledge graph (internal analysis) 2026What this evidence cannot show
The scan cannot detect buffering, which is why an experiment rather than another analysis is proposed. Four lymphoid lines is also why the design uses three index lines rather than one.
Evidence that argues against this (1)
- If the flat DepMap row is simply correct and IGF2BP3 is not required in human KMT2A::AFF1 cells, this experiment is an expensive way to confirm a negative. That outcome is listed in the design's outcome table precisely so it cannot be quietly discarded. Oncofusion knowledge graph (internal analysis) 2026
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
IGF2BP3 associates with a proliferative phenotype in B-ALL, the confounder that forces endogenous degron tagging rather than over-expression in this design.
IGF2BP3 Associates with Proliferative Phenotype and Prognostic Features in B-Cell Acute Lymphoblastic Leukemia · Mäkinen A, et al. · Cancers (Basel) 2021What this evidence cannot show
Abstract-level, unstratified B-ALL. Cited as a design constraint rather than as support for the biology.
Evidence that argues against this (1)
- Endogenous tagging removes the over-expression confounder but not the demand confounder: a cell that stops proliferating for any reason will lower its RNA-binding protein load, so a degron experiment still cannot distinguish a specific requirement from a general biosynthetic one without the fusion-negative comparator line the design includes. Oncofusion knowledge graph (internal analysis) 2026
The double-degron arm is the falsifying experiment
Acute degradation on self-renewal and xenograft-initiation readouts
Second single arm and the double arm
Three KMT2A::AFF1 index lines plus a KMT2A-rearranged non-AFF1 comparator