Settle the monocytic confound isogenically: degrade the fusion, read haem-gene dependency before lineage has time to move
also called isogenic haem dependency test · monocytic confound resolution experiment
The project found that fusion-carrying leukaemia cells need certain iron-and-haem genes more than other cells do, but almost all those cells are monocytes, whose normal job is handling iron. The clean way to tell the two explanations apart is to switch the fusion off inside one and the same cell and see whether the iron dependency goes away before the cell has had time to change into a different kind of cell.
The same thing, for a scientist
Design to resolve the confound that `ana-depmap-haem-lineage-control` left open and that `ana-kmt2a-r-line-panel-monocytic-audit` shows cannot be resolved by adding comparison lines, because the required control set is effectively empty (one clean line). The between-line contrast is replaced by a within-line one: an acutely degradable KMT2A fusion in a monocytic KMT2A-rearranged line (MV4-11, THP-1 or MOLM-13), with the four haem genes (ALAD, IREB2, UROS, ALAS1) read out as dependency and as flux at timepoints short enough that monocytic identity has not yet shifted. The critical design constraint is that fusion loss itself moves lineage: oncoprotein level determines which lineage programme the fusion activates, and reduced levels shift cells toward a GMP-like programme [E4]. So the readout must be paired at every timepoint with a lineage-state measurement, and any haem effect appearing only after lineage has moved is uninterpretable as a fusion effect.
Design
Arm 1 — acute fusion loss, lineage held fixed by time. Knock a degron tag into the endogenous 3' partner allele of MV4-11 (KMT2A-AFF1) and THP-1 (KMT2A-MLLT3), so the fusion protein but not wild-type KMT2A is degraded. Sample at 2, 4, 8, 12 and 24 h. At each timepoint measure: (i) haem-gene transcript and protein; (ii) haem flux by stable-isotope labelling from glycine or from labelled haemin, not transcript alone; (iii) lineage state by a monocytic-signature score plus surface CD14/CD11b/CD34.
Arm 2 — dependency, not expression. Run CRISPR knockouts of ALAD, IREB2, UROS and ALAS1 in each line with the degron engaged versus disengaged, in a short competitive-growth format. This is the arm that matches the DepMap readout being explained.
Arm 3 — the lineage-only control that the panel cannot supply. Drive Mono-Mac-1 (the single clean monocytic non-rearranged, non-HOXA-axis line identified by the audit) through the same knockouts. One line is not a comparison group, but it is the only available direct measurement of whether the dependency exists without a fusion.
Arm 4 — the differentiation control. Induce monocytic differentiation of THP-1 with PMA and measure whether haem-gene dependency rises with monocytic maturation independently of the fusion. If it does, the lineage explanation gains direct support; if it does not, the lineage explanation weakens without needing more lines.
Controls throughout: GAPDH and a pan-essential gene for screen performance, MEN1 and a canonical fusion target (MEIS1, HOXA9) to confirm the degron is working, and a non-targeting guide.
What each outcome would mean
| result | reading |
|---|---|
| Haem dependency falls within 24 h of fusion degradation, before the monocytic score moves | the dependency is fusion-driven; the confound is resolved against lineage |
| Haem dependency is unchanged by fusion degradation but present in Mono-Mac-1 and rises with PMA differentiation | the dependency is monocytic physiology; the DepMap hit is a lineage artefact and the target is dropped |
| Haem dependency falls only after the monocytic score has moved | uninterpretable by this design; the fusion changed the cell type and the two explanations are not separable in this system |
| Haem dependency unchanged, absent in Mono-Mac-1, not induced by PMA | neither explanation holds; look for a shared genetic background in the panel |
The third row is the likely one and is the reason for the short timepoints. It is also why transcript alone is insufficient: a transcriptional signature can move within hours, whereas metabolic flux and dependency change more slowly, and the discrepancy between them is itself informative.
| design | Isogenic degron-tagged KMT2A fusion in monocytic KMT2A-rearranged lines (MV4-11, THP-1, MOLM-13) with four arms: (1) acute fusion degradation with paired haem and lineage readouts at 2-24 h; (2) CRISPR knockout of ALAD/IREB2/UROS/ALAS1 with degron engaged versus disengaged in short competitive growth; (3) the same knockouts in Mono-Mac-1, the only clean monocytic non-rearranged non-HOXA-axis comparator identified by the panel audit; (4) PMA-induced monocytic differentiation of THP-1 to test whether haem dependency scales with monocytic maturation independently of the fusion. |
|---|---|
| readout | Haem-gene dependency (competitive growth log fold-change), haem flux by stable-isotope labelling from glycine or labelled haemin, haem-gene transcript and protein, and a paired lineage-state readout (monocytic signature score plus surface CD14/CD11b/CD34) at every timepoint. |
| controls | GAPDH and a pan-essential gene for screen performance; MEN1, MEIS1 and HOXA9 to confirm degron function; non-targeting guides; oxidative-stress markers (NFE2L2 target panel) to exclude the stress-response explanation raised in E2 counter-evidence. |
| cost tier | med |
| timeline | 9-12 months: 3-4 months degron knock-in and validation, 2 months arms 1 and 4, 3 months arms 2 and 3, 2 months flux measurement and analysis |
| resolves | ana-depmap-haem-lineage-control |
| why not more cell lines | ana-kmt2a-r-line-panel-monocytic-audit shows the required control set contains one usable line, so adding comparison lines within DepMap cannot reach adequate power |
| primary falsifier | Haem-gene dependency unchanged by acute fusion degradation while present in Mono-Mac-1 and inducible by PMA differentiation would establish the dependency as monocytic lineage physiology and retire the haem target. |
- If the experiment shows the haem dependency is fusion-driven, the therapeutic consequence is an inhibitor of a housekeeping biosynthetic pathway shared with erythroid precursors and hepatocytes, so a positive result creates a toxicity problem rather than solving one.
- Degrading the fusion changes lineage, so a negative result at late timepoints cannot be distinguished from a lineage confound reappearing inside the isogenic system, which is the same failure mode the design was built to escape.
- Arm 3 rests on a single cell line, so a null there is uninformative and a positive there is not replicable within the available panel.
- How fast does monocytic identity actually drift after acute KMT2A-fusion degradation? Without that number the timepoint choice is guesswork, and measuring it is a prerequisite rather than a result.
- Is a degron knock-in feasible in MV4-11 and THP-1 at the endogenous fusion allele without also tagging wild-type KMT2A, given that the tag must sit on the 3' partner?
- Does haem flux in a leukaemic monocytic blast resemble a monocyte's or a proliferating progenitor's? If the latter, the lineage explanation may be weaker than the transcriptional signature suggests.
- Should the oxidative-stress explanation be elevated to a third arm rather than a control panel, given that HMOX1 induction is NFE2L2-driven and not lineage-restricted?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Monocytic AML is metabolically and apoptotically distinct from primitive AML: it loses BCL2, relies on MCL1 to sustain oxidative phosphorylation, and is therefore resistant to venetoclax-azacitidine, with monocytic subclones selected at relapse.
Monocytic Subclones Confer Resistance to Venetoclax-Based Therapy in Patients with Acute Myeloid Leukemia · Pei S, Pollyea DA, Gustafson A, Stevens BM, Minhajuddin M, Fu R, et al., Jordan CT · Cancer Discovery 2020What this evidence cannot show
Establishes that monocytic identity carries a distinct metabolic programme in AML generally; it does not measure haem biosynthesis or iron-regulatory gene dependency, so it motivates the experiment rather than predicting its result. Read: abstract only.
Evidence that argues against this (1)
- In an independent cohort, monocytic differentiation was not clinically predictive of azacitidine/venetoclax response even though cultured monocytic cells were resistant in vitro; response tracked the BCL2/BCL-xL/MCL1 ratio in leukaemic stem cells. If the in vitro monocytic metabolic phenotype is partly a culture artefact, then arm 4 of this design (PMA-differentiated THP-1) may report a culture effect rather than a lineage effect. Read: abstract only. Waclawiczek A et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Haem-iron handling is a defining lineage function of the monocyte-macrophage compartment: macrophages with both Nramp1/SLC11A1 and DMT1/SLC11A2 disabled show severely compromised recycling of haemoglobin-derived iron and blunted induction of ferroportin and haem oxygenase 1 after erythrophagocytosis.
Both Nramp1 and DMT1 are necessary for efficient macrophage iron recycling · Soe-Lin S, Apte SS, Mikhael MR, Kayembe LK, Nie G, Ponka P · Experimental Hematology 2010What this evidence cannot show
Mouse macrophage line, iron transport rather than haem biosynthesis. The four DepMap hits (ALAD, UROS, ALAS1) are biosynthetic and IREB2 is regulatory, so this evidence establishes that the lineage is iron-active, not that it is haem-biosynthesis-dependent. Read: abstract only.
Evidence that argues against this (1)
- HMOX1 and the oxidative-stress arm of haem handling are inducible in non-monocytic AML via NFE2L2/Nrf2 independently of TP53 status, so an apparently lineage-characteristic haem phenotype can be a general stress response. This weakens the premise that a haem dependency in a monocytic panel must be either fusion-driven or lineage-driven; a third, stress-driven explanation exists and this design does not distinguish it unless oxidative stress is measured alongside. Read: abstract only. Ali D et al. 2016
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
PICALM-MLLT10 (CALM-AF10), carried by U-937, is DOT1L-dependent for transformation in the same way MLL-AF10 is, and DOT1L inhibition suppresses Hoxa cluster genes and Meis1 in both, so U-937 cannot serve as a pathway-neutral non-KMT2A control.
Abrogation of MLL-AF10 and CALM-AF10-mediated transformation through genetic inactivation or pharmacological inhibition of the H3K79 methyltransferase Dot1l · Chen L, Deshpande AJ, Banka D, Bernt KM, Dias S, Buske C, et al., Armstrong SA · Leukemia 2013What this evidence cannot show
Mouse transformation models; shows a shared DOT1L requirement, not a shared haem requirement. Cited here only to justify excluding U-937 from arm 3. Read: abstract only.
Evidence that argues against this (1)
- The shared axis did not produce shared drug behaviour in patients: single-agent pinometostat reduced H3K79 methylation but gave complete remission in only 2 of 51 adults, both t(11;19). A shared genetic dependency in mice is a weak basis for excluding a line from a control group, and excluding U-937 costs arm 3 its only additional monocytic comparator. Read: abstract only. Stein EM et al. 2018
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MLL oncoprotein level, not only its presence, determines which lineage programme it activates: at high levels the oncoprotein preferentially activates either a pro-B-enriched ALL programme or an HSC-gene-enriched AML programme, and in lineage-switching samples oncoprotein levels are reduced and the oncoprotein preferentially activates granulocyte-monocyte progenitor genes.
MLL oncoprotein levels influence leukemia lineage identities · Janssens DH, Duran M, Otto DJ, Wu W, Xu Y, Kirkey D, et al., Henikoff S · Nature Communications 2024What this evidence cannot show
CUT&Tag oncoprotein-target profiling across 36 samples including 3 lineage-switching cases; the level-to-programme relationship is correlative across samples rather than demonstrated by titration within one cell. It nonetheless establishes the hazard this design must avoid: degrading the fusion is not a lineage-neutral perturbation. Read: abstract only.
Evidence that argues against this (1)
- Only three of the 36 profiled samples had lineage-switched, so the inference that reduced oncoprotein level causes a shift to a GMP-like programme rests on a very small subgroup, and the direction of causation between reduced oncoprotein level and lineage switch is not established. If lineage does not in fact move on the timescale of an acute degron, the short-timepoint design is more conservative than it needs to be, and longer, better-powered timepoints would be available. Read: abstract only (same paper). Janssens DH et al. 2024
replaces the between-line contrast the audit showed is unavailable
settles the monocytic alternative that analysis left open
tests whether haem dependency is monocytic physiology
degron-tagged MV4-11 KMT2A-AFF1 as arm 1 model
the isogenic design that could settle it