Hypothesis: KMT2A::AFF1 mislocalises to or disrupts nuclear speckles, and paraspeckle biology (NEAT1) is a fusion-repressed differentiation brake
also called nuclear speckle KMT2A-AFF1 · paraspeckle NEAT1 KMT2A-r · subnuclear localisation of the fusion · SON speckle association
Cell nuclei contain droplet-like compartments that concentrate the machinery for processing RNA. The partner half of this fusion belongs to a protein family that lives in those droplets and can disturb them, and in a different fusion leukemia a related compartment controls whether cells mature. Neither has been checked here.
The same thing, for a scientist
Two subnuclear compartments are implicated by strong analogy and untested in KMT2A-r. **Nuclear speckles:** AFF2, AFF3 and AFF4 localise to nuclear speckles and their overexpression interferes with speckle organisation or biogenesis; enlarged speckles are observed in FRAXE fibroblasts lacking AFF2 [E1]. KMT2A::AFF1 retains the AFF1 module and is expressed constitutively from a chromatin-tethered position, so speckle association or disruption is a direct prediction. **Paraspeckles:** in a different fusion leukemia, PML-RARα **represses** the paraspeckle-essential lncRNA NEAT1; NEAT1 is significantly lower in de novo APL than in healthy donors, rises during ATRA-induced differentiation of NB4 cells, and siRNA knockdown of NEAT1 blocks ATRA-induced differentiation [E2]. That is a fusion oncoprotein using a nuclear-body lncRNA as a differentiation brake — precisely the kind of mechanism KMT2A::AFF1, whose defining phenotype is differentiation block, might share. We could find no NEAT1 or paraspeckle study in KMT2A-rearranged leukemia, and no imaging study of where the KMT2A::AFF1 protein sits relative to speckles.
The speckle arm1 deeper
Nuclear speckles are not decorative: speckle-associated genes are more efficiently spliced and exported, and speckle proximity is now understood as a quantitative amplifier of expression. If KMT2A::AFF1 either (a) drags target loci toward speckles or (b) disturbs speckle integrity as overexpressed AFF family members do 1, then a subnuclear-positioning mechanism sits underneath the transcriptional output that everyone measures by RNA-seq.
The experiment is imaging, not sequencing, and it is cheap: immunofluorescence for SON or SRRM2 (speckle markers) with fusion staining or a tagged knock-in, plus DNA-FISH or TSA-seq for HOXA-locus speckle distance in KMT2A::AFF1 versus KMT2A-wild-type cells. The mechanism-kmt2a-subnuclear-localization node in another beat covers wild-type KMT2A subnuclear behaviour; what is missing is the fusion-specific speckle question.
Why the condensate literature does not already answer this
The KMT2A-r field has a condensate story (SEC/AFF4 phase behaviour, covered in mechanism-condensate-transcriptional-rewiring), but condensates formed by the fusion complex are a different object from the pre-existing nuclear bodies the cell already maintains. Showing that AFF4 can phase-separate does not tell you whether the fusion associates with, competes with, or dismantles speckles and paraspeckles. Conflating the two is the error to avoid here.
The paraspeckle arm, and why the APL analogy is the strongest lead in this node
The NEAT1/APL result has the structure of a template 2: a fusion oncoprotein represses a nuclear-body lncRNA; the lncRNA is required for differentiation; removing it blocks differentiation even when a differentiating agent is given.
Map that onto this disease and it makes a concrete, high-stakes prediction: if KMT2A::AFF1 also represses NEAT1, then residual NEAT1 level could determine whether a menin inhibitor produces differentiation or merely growth arrest. Menin inhibitors are differentiating agents whose clinical benefit and whose signature toxicity (differentiation syndrome) both scale with differentiation. A paraspeckle requirement would be a mechanistic predictor of response that nobody is measuring, and it is measurable in banked samples by qPCR and RNA-FISH.
We stress the honest status: the APL finding is a single 2014 study in one cell line plus a small patient series, from one group, and NEAT1's role in cancer is inconsistent across tissues (up in many solid tumours, down in APL).
| statement | KMT2A::AFF1 interacts with pre-existing nuclear bodies rather than only forming its own condensates: it associates with or disrupts nuclear speckles via the retained AFF1 module, and it represses the paraspeckle lncRNA NEAT1 as PML-RARa does, making paraspeckle status a determinant of whether differentiation therapy (menin inhibition) yields differentiation or only arrest. |
|---|---|
| origin | analogy |
| prior probability | 0.3 |
| planned experiments note | Moved out of a reference field on 2026-09-19. It named a node that does not exist, which made the pointer read as a real cross-reference and, for hypotheses, silently satisfied the 'needs at least one experiment or analysis' rule. The intent is preserved here; the node still needs writing. |
predictions
[ "Immunofluorescence of KMT2A::AFF1 with SON/SRRM2 shows speckle association or altered speckle number and size relative to KMT2A-wild-type leukemia cells", "TSA-seq or DNA-FISH shows the HOXA locus is closer to nuclear speckles in KMT2A::AFF1 cells than in matched controls", "NEAT1 is lower in KMT2A::AFF1 blasts than in normal pro-B progenitors and rises on fusion knockdown or menin inhibition", "NEAT1 knockdown blocks revumenib-induced differentiation in KMT2A::AFF1 cells, mirroring the ATRA/APL result", "Baseline NEAT1 level in banked trial samples correlates with differentiation response to menin inhibition" ]
analyses
[ "ana-noncoding-gap-ledger-kmt2a-r" ]
falsifiers
[ "The fusion shows no speckle association and no effect on speckle number or morphology by quantitative imaging", "NEAT1 levels in KMT2A::AFF1 blasts are indistinguishable from stage-matched normal progenitors", "NEAT1 knockdown has no effect on menin-inhibitor-induced differentiation", "HOXA-speckle distance is unchanged between KMT2A::AFF1 and control cells, removing the positioning mechanism" ]
evidence grade by class
{
"direct_kmt2a_r": "none",
"retained_partner_family_speckle_biology": "strong but on AFF2/3/4",
"other_fusion_family_paraspeckle_precedent": "moderate (PML-RARa/APL)"
}planned experiments
[ "exp-noncoding-dependency-crispr-screen-kmt2a-aff1" ]
- Nuclear speckles and paraspeckles are required in normal cells; targeting nuclear-body composition is close to targeting basal RNA processing, with a predictably poor therapeutic index.
- If NEAT1 is needed for differentiation, then any therapy that lowers NEAT1 would antagonise menin inhibitors; some chemotherapy backbones may do this unknowingly.
- The whole paraspeckle arm rests on a single APL study whose direction of effect is opposite to NEAT1 biology in most other cancers, so the analogy could simply be wrong.
- Speckle-morphology phenotypes are notoriously sensitive to overexpression and fixation artefacts; a positive imaging result would need knock-in-level rigour before it is believed.
- Where does the KMT2A::AFF1 protein actually sit in the nucleus relative to speckles and paraspeckles? No published image answers this, which is startling for a fusion first cloned in 1992.
- Is NEAT1 expression altered in any KMT2A-rearranged leukemia dataset? This is answerable today from public RNA-seq with no new experiment.
- Does differentiation syndrome on menin inhibitors correlate with paraspeckle or NEAT1 status, which would give the first mechanistic predictor of that toxicity?
- Is the AFF1 speckle function separable from its SEC-recruitment function? If the same protein surface does both, a drug against one hits the other.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
AFF2/FMR2, AFF3/LAF4 and AFF4/AF5q31 localise to nuclear speckles and bind RNA with high apparent affinity for G-quadruplex structures; overexpression of AFF2/3/4 interferes with nuclear speckle organisation or biogenesis, and enlarged nuclear speckles are present in FRAXE fibroblasts.
Functional characterization of the AFF (AF4/FMR2) family of RNA-binding proteins: insights into the molecular pathology of FRAXE intellectual disability · Melko M, Douguet D, Bensaid M, Zongaro S, Verheggen C, Gecz J, Bardoni B · Human Molecular Genetics 2011What this evidence cannot show
AFF1 itself was not the member tested for speckle localisation; the inference to KMT2A::AFF1 is ours. Non-haematopoietic cells, overexpression-based, and speckle morphology is sensitive to expression level, so the disruption phenotype may be an artefact of overexpression rather than a physiological function.
Evidence that argues against this (1)
- Every element of this is about the WILD-TYPE AFF proteins, and the entry is a hypothesis about the fusion. Overexpression disrupting speckle organisation is the standard outcome of overexpressing any speckle-resident protein, G-quadruplex affinity measured in vitro does not establish a nuclear target, and the enlarged speckles are in FRAXE fibroblasts — a different disease in a different cell type. AFF1 in KMT2A::AFF1 retains only part of the protein, so speckle behaviour of the full-length protein may not transfer at all. Ensembl, EMBL-EBI 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
NEAT1, the lncRNA essential for paraspeckle formation, is significantly repressed in de novo acute promyelocytic leukemia compared with healthy donors, is repressed by PML-RARα, is upregulated during ATRA-induced NB4 differentiation, and its siRNA knockdown blocks ATRA-induced differentiation.
Inhibition of long non-coding RNA NEAT1 impairs myeloid differentiation in acute promyelocytic leukemia cells · Zeng C, Xu Y, Xu L, Yu X, Cheng J, Yang L, Chen S, Li Y · BMC Cancer 2014What this evidence cannot show
Single group, one cell line (NB4) for the functional work plus a modest patient series; qRT-PCR-based expression comparisons in leukemia versus healthy donors are sensitive to cell-composition differences. NEAT1's direction of change in cancer is tissue-dependent, so the APL result does not generalise automatically.
Evidence that argues against this (1)
- NEAT1 is aberrantly OVER-expressed in many solid tumours where high levels associate with poor prognosis, the opposite direction to APL, so 'NEAT1 loss drives leukemia' cannot be assumed as a general rule. Yu X et al. 2017
Correction and counter (Ramilo Amor 2025, PMID 39793727): a transcriptome-wide splice-variant analysis of KMT2A-r leukaemia does exist (171 paediatric T-ALL) and attributes the KMT2A-r signature to the splicing regulators ESRP1 and MBNL3 and to CD44 promoter methylation, not to the fusion partner - so a KMT2A-r splicing signature can arise with no AFF1 involvement
Named by this hypothesis as the analysis that tests it; the reference was dangling until now.
grades the paraspeckle limb at weak-by-analogy and names the cheapest test
a speckle-resident protein in the pan-MLL-fusion core gives that hypothesis a concrete, testable partner
speckle- and paraspeckle-associated loci as a library element class