An anti-MLL-C antibody in a t(4;11) cell is looking at two proteins at once
also called MLL-C ambiguity · taspase1 cleavage confounds the fusion question · anti-MLL-C trap
KMT2A is cut in half inside the cell before it works. The mirror-image leukemia fusion gets cut too, and the piece it releases looks exactly like the normal protein's second half. So an antibody against that half cannot tell you which of the two proteins it just pulled down - and this is the antibody the field uses.
The same thing, for a scientist
Wild-type KMT2A is cleaved by taspase1 into MLL-N (p320) and MLL-C (p180), which then re-associate. The reciprocal t(4;11) fusion AFF1::KMT2A contains the KMT2A taspase1 sites and is likewise cleaved, releasing an MLL-C fragment that is, in sequence terms, the same species the wild-type allele produces. Consequently **an antibody raised against the KMT2A C-terminus cannot distinguish wild-type MLL-C from reciprocal-fusion-derived MLL-C** in any t(4;11) cell, by western blot, immunoprecipitation or ChIP. Every published WRAD-adjacent observation in a KMT2A::AFF1 cell rests on exactly this reagent class: the RUNX1 co-IP experiments in SEM and RS4;11, which recover 'MLL-C' alongside a band assigned to cleaved AF4-MLL by apparent molecular weight [E1]; and the ASH2L-knockdown ChIP in RS4;11, which reports 'MLL' occupancy at the HOXC8 promoter [E2]. The trap is symmetrical with the one this graph already records for MLL1-versus-fusion N-terminal antibodies: in the murine MLL-AF9 work the authors explicitly note that their MLL1 antibody recognises the 180 kDa C-terminal fragment and therefore **cannot** detect MLL-AF9 [E3], which is the correct use of the same property and the reason that study could separate the two proteins at all.
Why the ambiguity is not fixable by a better antibody
The problem is sequence identity, not reagent quality. After taspase1 cleavage, wild-type MLL-C and reciprocal-fusion MLL-C differ only in the N-terminal residues immediately following the cleavage site, which vary with the patient's breakpoint. No epitope on the retained C-terminal fragment distinguishes them. Three designs do:
- Tag the reciprocal allele. An endogenous knock-in tag on the der(4) product - the only construct in the cell that carries partner N-terminal sequence upstream of the KMT2A portion - is the clean answer, and CRISPR-engineered t(4;11) human HSPC models now exist that could carry it 4.
- Use a reciprocal-negative comparator. About half of t(4;11) cases and cell lines express no reciprocal product 5. A paired experiment in a reciprocal-positive and a reciprocal-negative line assigns any difference to the reciprocal fusion without needing to see it.
- Degron the wild-type allele. What remains after acute wild-type KMT2A degradation is, in a reciprocal-positive line, reciprocal-derived.
None of the three has been applied to a WRAD question.
What this invalidates and what it leaves standing
It does not invalidate the observation that WRAD subunits sit on chromatin at KMT2A-fusion target genes in t(4;11) cells - that is well supported and uncontroversial. It invalidates the attribution of that occupancy to any particular KMT2A-derived protein. Three sources are available in a reciprocal-positive t(4;11) cell and the published experiments cannot rank them: the wild-type allele, the reciprocal fusion, and the other five SET1/MLL enzymes that share the same WRAD module.
This is a bounded, transferable methodological point and it applies beyond WRAD to every 'MLL-C complex' claim made in a t(4;11) system.
| ambiguous reagent class | antibodies raised against the KMT2A C-terminus (e.g. mmC2.1) |
|---|---|
| none applied to a wrad question | true |
species confounded
[ "wild-type KMT2A MLL-C (p180) after taspase1 cleavage", "MLL-C fragment released from the reciprocal AFF1::KMT2A fusion" ]
affected published results
[ "PMID 23352661 RUNX1/MLL-C co-IP in SEM and RS4;11", "PMID 32021283 ASH2L knockdown ChIP in RS4;11" ]
resolving designs
[ "endogenous epitope tag on the der(4) product", "reciprocal-positive versus reciprocal-negative paired lines", "acute degron of the wild-type KMT2A allele" ]
- If chromatin-bound WRAD in t(4;11) cells is loaded on the wild-type allele rather than on any fusion, an agent aimed at it is an agent against normal KMT2A function in an infant, which is the least favourable therapeutic framing available.
- Any biomarker built on 'MLL-C occupancy' in a t(4;11) sample is measuring an unresolved mixture and would behave differently in reciprocal-positive and reciprocal-negative patients without anyone knowing why.
- Is MV4-11 reciprocal-positive? **Answered 2026-09-19.** RS4;11 and SEMK2-M1 (a SEM subline) **both express MLL-AF4 and AF4-MLL** - stated directly in the abstract of PMID 20869771, whose own experiment was to knock AF4-MLL down in them. So both standard lines are **reciprocal-positive**, and there is no reciprocal-negative t(4;11) line: no cell-based experiment in this disease has ever had a comparator arm. **MV4-11 remains open** - it is t(4;11) AML, and this beat did not find its status stated anywhere. Every WRAD-adjacent result in t(4;11) was generated in these lines, and because both known ones carry both products the antibody ambiguity applies to all of it.
- Is the reciprocal-derived MLL-C fragment detectable at the protein level in patient blasts at all, or does taspase1-dependent destabilisation remove it?
- How many 'MLL-C complex' claims in the t(4;11) literature would change if the reciprocal product were assigned separately?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In SEM and RS4;11, anti-RUNX1, anti-MLL-C and anti-AF4-N immunoprecipitations each recover a complex containing RUNX1, MLL-C, wild-type AF4 and a band assigned to the cleaved ~194 kDa AF4-MLL protein, with the assignment made on apparent molecular weight.
RUNX1 is a key target in t(4;11) leukemias that contributes to gene activation through an AF4-MLL complex interaction. · Wilkinson AC, Ballabio E, Geng H, et al. · Cell Rep 2013What this evidence cannot show
The species labelled 'MLL-C' is not resolved between wild-type and reciprocal-fusion origin by any experiment in the paper. The AF4-MLL band assignment rests on apparent molecular weight, which the authors themselves discuss as requiring explanation. Their AF4-MLL-specific siRNAs failed, so no loss-of-function control exists.
Evidence that argues against this (1)
- Taspase1-dependent cleavage of MLL1 destabilises it, and loss of taspase1 increases uncleaved MLL1 chromatin association and can displace MLL chimeras from chromatin - so the abundance of any cleaved MLL-C species in a leukaemia cell is itself a regulated variable rather than a fixed readout. Zhao Z et al. 2019
- E2weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
ASH2L knockdown in RS4;11 reduces occupancy of RBBP5, WDR5, MLL and BPTF at the HOXC8 promoter by ChIP, with 'MLL' detected by an antibody that does not distinguish the wild-type protein from fusion-derived species.
ASH2L-Promoted HOXC8 Gene Expression Plays a Role in Mixed Lineage Leukemia-Rearranged Acute Leukemia. · Wu YJ, Li LX, Liu L, et al. · Onco Targets Ther 2020What this evidence cannot show
siRNA, one target gene, no growth readout, and the antibody ambiguity described in this entry. The result is best read as evidence that a WRAD-loaded H3K4 methyltransferase complex occupies that promoter, with the identity of the catalytic subunit unassigned.
Evidence that argues against this (1)
- The wild-type KMT2A allele is required for KMT2A-fusion leukaemia maintenance and binds chromatin in these cells, which is a complete alternative explanation for the occupancy being measured. Thiel AT et al. 2010
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
An MLL1 antibody that specifically recognises the 180 kDa MLL1 C-terminal fragment cannot detect the MLL-AF9 protein, a property the authors relied on to separate wild-type MLL1 from the fusion in ChIP-seq.
MLL1 and MLL1 fusion proteins have distinct functions in regulating leukemic transcription program. · Xu J, Li L, Xiong J, et al. · Cell Discov 2016What this evidence cannot show
The separation works in MLL-AF9 because that fusion has no KMT2A C-terminus at all. It does NOT work in a reciprocal-positive t(4;11) cell, where a second protein carrying the KMT2A C-terminus is present - a case that study did not face.
Evidence that argues against this (1)
- A 2024 study using different antibodies reports that MLL1 and MLL-AF9 occupy chromatin with identical patterns, contrary to the separation reported here, so antibody-based species assignment in this field is not settled. Yang Z et al. 2024
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CRISPR/Cas9-engineered t(4;11) translocations can now be generated in primary human haematopoietic stem and precursor cells with both reciprocal products concomitantly present.
CRISPR/Cas9-mediated t(4;11) translocation in human hematopoietic stem/precursor cells demonstrates plasticity to differentiate into either the myeloid or lymphoid lineage. · Benz T, Larghero P, Meyer C, et al. · Leukemia 2026What this evidence cannot show
A recent model system; whether an endogenous epitope tag can be placed on the der(4) product without perturbing the biology has not been demonstrated.
Evidence that argues against this (1)
- Engineering the translocation in HSPCs does not by itself solve the attribution problem, because the resulting cells contain the same three sources of chromatin-bound WRAD; the tag, not the translocation, is what would resolve it. Wilhelm A et al. 2021
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The reciprocal AFF1::KMT2A transcript is expressed in only about 45% of t(4;11) cases, which supplies a natural reciprocal-negative comparator arm.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al. · Haematologica 2019What this evidence cannot show
Transcript-level detection in patient samples; whether the widely used t(4;11) cell lines SEM, RS4;11 and MV4-11 are reciprocal-positive or reciprocal-negative is stated inconsistently in the literature and was not resolved in this beat.
Evidence that argues against this (1)
- Reciprocal-positive patients have markedly better event-free survival, so a reciprocal-positive versus reciprocal-negative comparison confounds the reciprocal product with whatever biology drives that prognostic difference. Agraz-Doblas A et al. 2019
- E-linesstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.journal-article
The two standard t(4;11) cell lines used for essentially all mechanistic work in this disease, RS4;11 and SEMK2-M1 (a SEM subline), both express the reciprocal AF4-MLL transcript as well as MLL-AF4. There is therefore no reciprocal-negative t(4;11) line in routine use.
t(4;11) leukemias display addiction to MLL-AF4 but not to AF4-MLL · Leukemia Research 2011What this evidence cannot show
Abstract read verbatim 2026-09-19: *"In the human leukemia cell lines - RS4;11 and SEMK2-M1, both of which express MLL-AF4 and AF4-MLL, we knocked down the expression of AF4-MLL using siRNA."* This is an RT-PCR-level statement about transcript, not protein, and it is made in passing as a premise rather than tested. SEMK2-M1 is a subline, so it does not strictly settle SEM itself, and **MV4-11 is not addressed**.
Evidence that argues against this (1)
- The same paper reports the reciprocal product is **dispensable** in both lines - siRNA knockdown of AF4-MLL had no effect on growth, cell cycle or apoptosis, while MLL-AF4 knockdown arrested and killed them. So 'every line is reciprocal-positive' is a weaker complaint than it sounds: the one experiment that removed the reciprocal product from those lines found nothing changed. The unstratified preclinical package matters for initiation and for drug response, not obviously for maintenance. Leukemia Research 2011
- ← refutesThe protein nobody can attribute by antibody is the one mass spectrometry finds most easily
the reciprocal product's unmeasurability is a property of the antibody epitope, not of the molecule: by tryptic MS it is the easier of the two to identify
the mirror reagent: N-terminal antibodies pool wild-type with the direct fusion, C-terminal antibodies pool wild-type with the reciprocal product and KMT2A*
extends the two-source ambiguity to three sources and to every KMT2A fusion family
Both Taspase1 sites (2666/2667, 2718/2719) are retained, so the reciprocal releases a sequence-identical MLL-C fragment.
Identical cleavage sites are the mechanical reason the released MLL-C fragments cannot be told apart.
Extends the reagent problem from immunoprecipitation to the clinical assay layer: DNA, RNA and protein give three different answers.
Answers that entry's first open question: RS4;11 and SEMK2-M1 both express the reciprocal fusion, so the three-way attribution problem applies to every experiment in them.
the MLL-C180 control band is fusion-excluding but not wild-type-specific in a reciprocal-positive line
in a reciprocal-positive cell the free C180 pool has two sources and no reagent separates them
- depends on →Forty-five percent of t(4;11) patients express a KMT2A fusion transcript that keeps the whole WRAD interface
The ambiguity exists only because the reciprocal product is taspase1-cleaved to release an MLL-C fragment indistinguishable from the wild-type one.
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
Every published WRAD-adjacent observation in a KMT2A::AFF1 cell rests on anti-MLL-C or anti-MLL reagents that cannot assign the species they detect.
This is why the RBBP5 chromatin-occupancy data in SEM and RS4;11 cannot be attributed to any particular KMT2A-derived protein.
Same limitation applies to the ASH2L knockdown ChIP in RS4;11.
- ← depends onSettle the WRAD question: tagged co-IP plus degron-timed proximity across three fusion classes
The design exists because an anti-MLL-C antibody cannot assign the species it detects; endogenous tagging of the der(4) allele is the fix.