Blot all three PBX proteins in KMT2A::AFF1 B-lineage cells, then delete them singly and in combination - the family has never been tested together in any human leukaemia line
also called PBX1 · PBX2 · PBX3 · PBX family knockout · SEM · RS4;11
Three near-identical proteins can each partner the HOX factors that drive this leukaemia, and every experiment so far has removed only one of them at a time. This proposes measuring all three in the right cells and then deleting them in every combination.
The same thing, for a scientist
Two linked experiments, both cheap, neither previously done in this disease. **Part A - a three-lane blot.** Measure PBX1, PBX2 and PBX3 protein in human KMT2A::AFF1 B-lineage lines (SEM, RS4;11) and, if available, primary infant KMT2A-rearranged blasts, with paralogue-discriminating antibodies validated against single-gene knockouts. **This measurement does not exist in this disease in either species.** The only protein-level PBX measurement in haematopoietic tissue is Selleri et al. 2004's immunoblot of murine postnatal bone marrow and thymus, which found **PBX2** to be the predominant high-molecular-weight PBX isoform - the opposite of what this graph's transcript-derived PBX3 position would lead a reader to expect. **Part B - combinatorial CRISPR deletion.** Single, double and triple knockouts of PBX1/2/3 in the same lines, reading out proliferation, HOXA9/MEIS1 target expression, MEIS1 protein stability and colony formation. `gene-pbx3`'s own open questions ask for this and note that the family *"has never been tested combinatorially in a human leukaemia line"*. **The reason it matters is a specific, published conflict.** Li et al. 2013 excluded PBX1 and PBX2 on transcript coexpression in myeloid CA-AML; Selleri et al. 2004 put PBX2 at the top of the protein ladder in bone marrow; and DepMap shows PBX3 is not a dependency in any human leukaemia line. Those three facts are consistent only if the family is redundant at the protein level, which is precisely what a single-gene experiment cannot see.
Design
Part A - paralogue-resolved protein panel
- Cells: SEM and RS4;11 (human KMT2A::AFF1 B-ALL); THP-1 or MOLM-13 as a myeloid KMT2A-rearranged comparator; a non-KMT2A B-ALL line (REH or NALM-6) as lineage control.
- Readout: immunoblot for PBX1, PBX2, PBX3 with the same loading, plus a high- versus low-molecular-weight isoform distinction, since Selleri's result is specifically about the HMW isoform class.
- Antibody validation is the whole experiment. PBX1 and PBX2 are both 430 aa and PBX3 is 434 aa, the three accessions are consecutive (P40424/P40425/P40426), and the homeodomain is near-identical. An antibody that cross-reacts produces a confident wrong answer with no signature. Validate each antibody against a single-gene knockout of its own target and against knockouts of the other two.
- Orthogonal readout: targeted mass spectrometry on paralogue-unique peptides, which sidesteps the antibody problem entirely and should be run if any antibody fails validation.
Part B - combinatorial deletion
- Seven arms: three singles, three doubles, one triple, plus non-targeting control. Inducible Cas9 with paired guides, or degron tagging if a knockout proves lethal.
- Readouts: growth over 21 days; RNA-seq for the HOXA9/MEIS1 output programme this graph already carries as
mechanism-hoxa9-meis1-output-program; MEIS1 protein by blot and by cycloheximide chase, because Pbx3 is reported to stabilise Meis1 against ubiquitination; colony-forming ability. - Critical control: run the same arms in the non-KMT2A B-ALL line. A pan-PBX requirement in every B-ALL line is a lineage property, not a fusion dependency, and this project has repeatedly found that distinction to be where the claim lives.
What each result would mean, decided before the experiment
| result | reading |
|---|---|
| PBX2 is the dominant PBX protein in KMT2A::AFF1 cells | Li 2013's transcript-level exclusion does not hold at the protein level in this lineage, and gene-pbx3's position needs a lineage caveat in its claim, not just its caveat field |
| PBX3 dominant, PBX2 low | the graph's current position survives and gains its first protein-level support in the index disease |
| all three present, no single knockout has an effect, triple knockout does | family redundancy - the DepMap zero for PBX3 is explained, and the target becomes the shared HOX-PBX interface rather than any one paralogue |
| triple knockout has no effect either | the PBX requirement is myeloid and murine, and this graph should stop building on it in B-lineage disease |
The fourth row is the one worth running the experiment for. It is a real possibility - every positive PBX result in this field is murine and most of it myeloid - and nothing currently in the graph could produce it.
| design | Part A: paralogue-resolved immunoblot of PBX1/PBX2/PBX3 in SEM and RS4;11 (KMT2A::AFF1 B-ALL), a myeloid KMT2A-rearranged comparator and a non-KMT2A B-ALL control, with antibodies validated against single-gene knockouts of all three paralogues and a targeted-MS fallback on paralogue-unique peptides. Part B: inducible-Cas9 single, double and triple knockout of PBX1/2/3 in the same lines, seven arms plus non-targeting control. |
|---|---|
| readout | PBX1/2/3 protein level and high- versus low-molecular-weight isoform distribution; 21-day growth; RNA-seq of the HOXA9/MEIS1 output programme; MEIS1 protein by blot and cycloheximide chase; colony-forming ability. |
| controls | Non-KMT2A B-ALL line in every arm, to separate a fusion dependency from a B-lineage property. Antibody validation against knockouts of all three paralogues, not only the cognate one. Non-targeting guide arm. Targeted MS as an orthogonal quantification if any antibody fails. |
| cost tier | low |
| timeline | Part A 4-6 weeks including antibody validation; Part B 3-4 months for seven arms with the RNA-seq readout. |
| why nobody has done it | Every published PBX experiment in this field removes one paralogue. The combinatorial arm is the only design that can distinguish family redundancy from dispensability, and redundancy is the hypothesis that reconciles Li 2013's transcript exclusion, Selleri 2004's protein ladder and DepMap's zero. |
| tests hypothesis | hyp-the-pbx-requirement-has-only-ever-been-tested-one-paralogue-at-a-time |
decision table
{
"PBX2_dominant_protein": "Li 2013's transcript exclusion does not hold at protein level in B-lineage; gene-pbx3's claim needs a lineage clause",
"PBX3_dominant": "current graph position survives and gains its first protein-level support in the index disease",
"no_single_effect_triple_effect": "family redundancy; target moves to the shared HOX-PBX interface",
"no_triple_effect": "the PBX requirement is murine and myeloid and this graph should stop building on it in B-lineage disease"
}- If the triple knockout kills every line including the non-KMT2A control, the result is a general requirement for TALE cofactors in B-lineage cells and offers no therapeutic window - which is the most likely outcome and should be stated before the experiment rather than after it.
- A positive combinatorial result argues for a pan-PBX agent, and PBX2 is the dominant high-molecular-weight PBX in normal bone marrow, so the on-target normal-marrow cost of such an agent is larger than any single-paralogue rationale implies.
- Are paralogue-specific PBX1/PBX2/PBX3 antibodies that pass knockout validation commercially available, or does Part A require the targeted-MS route from the start? The answer determines whether this is a four-week experiment or a three-month one.
- Should the panel include PREP1/PKNOX1 and MEIS1 as the MEINOX partners that determine PBX nuclear retention? This graph carries mechanism-pbx-nuclear-retention-requires-a-meinox-partner, and a PBX protein panel without its partners measures abundance rather than available dimer.
- Is primary infant KMT2A::AFF1 material obtainable for Part A? Cell lines are adult- and paediatric-derived and this graph repeatedly finds that the infant disease is the least-measured system in its own literature.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In mice, Pbx2 is the predominant high-molecular-weight-isoform Pbx protein detectable by immunoblot in postnatal bone marrow and thymus, yet Pbx2-null animals show no measurable effect on steady-state haematopoiesis, which the authors attribute to compensation by low-molecular-weight Pbx isoforms.
The TALE homeodomain protein Pbx2 is not essential for development and long-term survival. · Selleri L, DiMartino J, van Deursen J, et al., Cleary ML · Mol Cell Biol 2004What this evidence cannot show
Murine, germline, steady-state, normal tissue, read as abstract rather than full text. Nothing in it concerns leukaemia or KMT2A. It is cited here as the motivation for measuring the protein panel, not as evidence about this disease.
Evidence that argues against this (1)
- Li et al. 2013 report that it is PBX3, not PBX1 or PBX2, that is consistently coexpressed with HOXA9 in MLL-rearranged AML - so the transcript evidence in leukaemia points the other way from the protein evidence in normal marrow. That conflict is the reason for the experiment and also the reason its result is genuinely uncertain. Li Z et al. 2013
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PBX1, PBX2 and PBX3 are UniProt P40424 (430 aa), P40425 (430 aa) and P40426 (434 aa) - three consecutive accessions, two of them identical in length - so neither sequence length nor accession proximity can discriminate them and antibody cross-reactivity cannot be caught by a length check.
UniProtKB entries P40424, P40425 and P40426 (PBX1_HUMAN, PBX2_HUMAN, PBX3_HUMAN) · UniProtKB/Swiss-Prot 2026What this evidence cannot show
Retrieved 2026-09-21. Length identity is a statement about the canonical sequences; the paralogues also have alternative isoforms of differing length that were not enumerated here, and the high- versus low-molecular-weight isoform distinction Selleri's result turns on is an isoform question this lookup does not resolve.
Evidence that argues against this (1)
- Commercial antibodies to PBX family members are widely sold as paralogue-specific and many are. The design point is not that they cannot work, it is that a failure produces a clean-looking blot, so validation against knockouts is not optional here in the way it often is. Li Z et al. 2013
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.analysis
This graph's existing PBX3 entry records that the PBX family has never been tested combinatorially in a human leukaemia line, that PBX3 is not a dependency in any human leukaemia line in DepMap, and that every positive PBX3 cell-state result in the corpus is murine and mostly myeloid.
PBX3 - the TALE cofactor the field says is non-redundant, and the experiment that says HOXA9 is the redundant one · Oncofusion knowledge graph, gene-pbx3 2026What this evidence cannot show
An internal entry of this project, not an external source; its own caveats about the DepMap scan apply, including that a near-zero Chronos effect in a paralogue-redundant family is uninterpretable rather than negative - which is itself part of the argument for the combinatorial arm.
Evidence that argues against this (1)
- A triple knockout of a transcription-factor family is a demanding experiment with a real chance of a non-specific fitness effect that cannot be attributed to the HOXA programme. The non-KMT2A B-ALL control arm is what separates those, and if it moves too the experiment has measured lineage biology rather than a fusion dependency. Selleri L et al. 2004
DEK recruits GABPA to upregulate PBX3 - an upstream handle this panel does not include
the combinatorial arm with a non-KMT2A B-ALL lineage control separates family redundancy from a myeloid-only requirement
Part A measures PBX2 protein in KMT2A::AFF1 cells for the first time in this disease
- tests →Step 5 — The output program: HOXA9/HOXA10/MEIS1/PBX3 plus MEF2C, MYB, BCL2, FLT3 and PROM1/CD133
RNA-seq readout is the HOXA9/MEIS1 output programme this graph already carries