Oncofusion

start / graph / localization-nuclear-lamina-lads-kmt2a-r

Mechanismweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q4b-localization

The nuclear lamina and lamina-associated domains - evidence grade for KMT2A: weak and indirect, not none

also called LADs · lamin B1 DamID · nuclear periphery · lamina-associated domain leukemia

The inner surface of the nuclear envelope is a silencing shelf, and genes parked against it are switched off. One 2000 paper put the KMT2A amino terminus near that surface, but it never looked at the lamina itself, so the question is barely opened rather than untouched.

The same thing, for a scientist

Lamina-associated domains are a well-characterised repressive compartment: DamID in human fibroblasts identified more than 1,300 discrete LADs of 0.1-10 Mb, gene-poor and transcriptionally repressive, sharply bordered by CTCF sites, outward-facing promoters and CpG islands [E1]. LADs are cumulatively remodelled during lineage commitment, with genes leaving the lamina often activated and others becoming unlocked for later activation [E2]. In the lymphoid system, Lamin B1 DamID plus FISH in Jurkat T cells showed that activation releases T-cell-function loci from the lamina, that TADs reposition as units, and that released loci join A2 subcompartments [E3]. **As of 2026-09-19, PubMed returns 0 records for `("nuclear lamina" OR "lamina-associated domain" OR "lamina-associated domains" OR "lamin B1")[Title/Abstract] AND (HOXA OR KMT2A OR "MLL" OR "mixed lineage leukemia")[Title/Abstract]`, and 5 records for the same lamina terms against `(Hox OR HOXA OR HOXB)[Title/Abstract]`, none of them mammalian and none about a leukaemia; so no study links KMT2A, any KMT2A fusion, or the human HOXA cluster to the nuclear lamina or to LADs** — not direct imaging, not DamID, not a partner-mediated inference. The nearest positive result anywhere is in Drosophila, where the Hox protein Ultrabithorax represses autophagy genes by tethering their loci to the nuclear periphery through Lamin-C [E4] - a Hox protein acting as a lamina tether, in a fly, on non-Hox targets. The honest grade for KMT2A remains **none**, and the entry exists to make the absence visible and to state what would make it worth filling. Two features make this more than a routine gap. First, LAD borders are defined by CpG islands and CTCF, which are exactly the features the fusion's CXXC domain and the HOXA TAD boundary in this graph are built on. Second, the disease's defining phenotype is a differentiation block, and LAD remodelling is one of the principal genome-scale correlates of differentiation.

Go deeper
What is actually known about LADs, and why the border feature matters here1 deeper

Guelen et al. mapped nuclear-lamina interactions genome-wide by DamID in human fibroblasts: >1,300 discrete LADs, 0.1-10 Mb, gene-poor, transcriptionally repressive, and bordered by CTCF binding sites, outward-facing promoters, or CpG islands 1. The border composition is the interesting coincidence for this graph. The fusion's DNA-reading module is a non-methylated CpG-island binder, and the graph already carries genome-ctcf-cohesin-hoxa-tad-boundary. If the HOXA cluster sits at or near a LAD border in the relevant progenitor, then the fusion's target set and the lamina's edge are the same genomic feature seen from two directions — an entirely testable proposition that has never been stated.

Peric-Hupkes et al. showed LADs are remodelled cumulatively across a differentiation series, with locus release from the lamina correlating with activation, and some released loci merely 'unlocked' for later activation rather than immediately expressed 2. That two-step release-then-activate structure is a good formal match for the bivalent/primed chromatin states this graph already records under cell-state-primed-bivalent-chromatin.

The nearest haematopoietic dataset, and why it does not cover the question

Robson et al. did Lamin B1 DamID plus FISH in Jurkat T cells during activation and found constrained release of lamina-associated enhancers and genes, with TADs repositioning as units and released loci associating in chromosome compartments 3. This is the only lymphoid lamina dataset that could be verified. It is T-cell activation in a mature T-ALL line, not HSPC or pro-B differentiation, and it says nothing about HOXA or KMT2A. It establishes that the assay works in a haematopoietic cell and that lineage signals move loci off the lamina — nothing more that transfers.

What was searched, so the negative is auditable

Queries returning zero relevant primary studies: KMT2A AND nuclear lamina; MLL AND nuclear lamina; lamina-associated domains[Title] AND leukemia; HOXA[Title] AND nuclear lamina; lamin B1[Title] AND acute myeloid leukemia. No partner-mediated indirect route was found either: neither AFF1, AFF4, ENL, AF9 nor menin has a reported lamina association.

Grade recorded in data.evidence_class_for_kmt2a: none. Not weak, not indirect. None.

Bounded, dated re-run (2026-09-19, raw eutils esearch, counts included so the negative is reproducible):

Query (all [Title/Abstract])Records
("nuclear lamina" OR "lamina-associated domain" OR "lamina-associated domains" OR "lamin B1") AND (HOXA OR KMT2A OR "MLL" OR "mixed lineage leukemia")0
("nuclear lamina" OR "lamina-associated") AND (Hox OR HOXA OR HOXB)5 (PMIDs 42679029, 35177098, 25825760, 25151160, 23087710)

All five of the second set were read. Four are Drosophila or review/method papers about Hox topology, and the fifth (PMID 25825760) places mammalian Hox clusters with other H3K27me3 targets in an active nuclear domain, which is the opposite compartment. The one substantive positive is PMID 42679029 4. The negative therefore survives the re-run, but it is now a counted negative with a date rather than an unbounded 'nobody has looked'.

The one experiment that would be worth doing, and the reason it probably is not first in the queue

Lamin B1 DamID or CUT&RUN in a t(4;11) line with and without fusion degron depletion, asked as: does the fusion's presence change which domains touch the lamina, and does HOXA sit at a LAD border? The reagents exist and the graph already has the degron system under exp-degron-washout-recovery-kmt2a-aff1.

The honest counter-argument against prioritising it: the KMT2A fusion's target set is CpG-island promoters, which by definition are the gene-rich, non-LAD fraction of the genome. LADs are gene-poor. A prior that the fusion operates largely outside LADs is reasonable, and the interesting question collapses to the narrower one of whether borders matter. That narrower question can be answered by reanalysis of published Lamin B1 DamID from any haematopoietic cell against published KMT2A::AFF1 occupancy, with no new experiment at all — which is what ana-lad-border-overlap-kmt2a-targets proposes.

Correction: there is one piece of indirect evidence and this entry said there was none

This entry was graded "none, direct or indirect". That is wrong.

PMID 11069025 (Caslini et al., Leukemia 2000) reports, by immunofluorescence, confocal laser scanning and immunoelectron microscopy, that a conditionally expressed 50 kDa KMT2A amino-terminal fragment (MLL3AT) and endogenous MLL localise "in the nucleoplasm in a punctate pattern of distribution, including regions attached to the nuclear envelope and the periphery of the nucleolus". The full text adds the image-level detail: "the presence of label at the edge of the nuclear envelope (small arrows)".

The paper is already cited in five other entries in this graph. The gap was in this one.

Three limits, and the first is the one that keeps the grade weak.

  1. Envelope is not lamina. The word lamin appears zero times in the full text. Proximity to the envelope is compatible with lamina association and with several other things, and nothing here measures a lamina contact or a lamina-associated domain.
  2. The construct is conditionally over-expressed in U937 and HeLa - neither is KMT2A-rearranged - so this is attributable to the construct, not to an endogenous fusion.
  3. The same paper reports nucleolar periphery localisation for MLL-AF9 and MLL-AF4 too, which is the observation this graph uses elsewhere; the envelope observation travels with it and has had none of the same attention.

Retrieval note, because it matters for the rule. This paper was logged in the Q28 beat as having no PMCID and being readable only at abstract level. It has PMCID PMC7543881: fullTextXML returns HTTP 500, but the PMC HTML page returns 63,706 characters of text. The four-route test works here and route 4 is what works - the same route that has now rescued four beats.

Facts and figures
compartmentnuclear lamina / lamina-associated domains
evidence class for kmt2anone
direct evidencefalse
indirect evidence via partnerfalse
nearest haematopoietic datasetLamin B1 DamID in activated Jurkat T cells (PMID 28424353)
searches returning zero
("nuclear lamina" OR "lam…0.000("nuclear lamina" OR "lamina-associated domain" OR "lamina-associated domains" OR "lamin B1")[Title/Abstract] AND (HOXA OR KMT2A OR "MLL" OR "mixed lineage leukemia")[Title/Abstract]: 0KMT2A AND nuclear lamina …0.000KMT2A AND nuclear lamina (informal, cycle 1): 0MLL AND nuclear lamina (i…0.000MLL AND nuclear lamina (informal, cycle 1): 0
marker proteins
[
  "LMNB1",
  "LMNA/C",
  "LBR",
  "emerin"
]
searches returning zero — the numbers behind the chart
[
  {
    "query": "(\"nuclear lamina\" OR \"lamina-associated domain\" OR \"lamina-associated domains\" OR \"lamin B1\")[Title/Abstract] AND (HOXA OR KMT2A OR \"MLL\" OR \"mixed lineage leukemia\")[Title/Abstract]",
    "records": 0,
    "date": "2026-09-19",
    "source": "PubMed eutils esearch"
  },
  {
    "query": "KMT2A AND nuclear lamina (informal, cycle 1)",
    "records": 0,
    "date": "2026-09-18",
    "source": "PubMed"
  },
  {
    "query": "MLL AND nuclear lamina (informal, cycle 1)",
    "records": 0,
    "date": "2026-09-18",
    "source": "PubMed"
  }
]
lad border features
[
  "CTCF sites",
  "outward-facing promoters",
  "CpG islands"
]
bounded searches
[
  {
    "query": "(\"nuclear lamina\" OR \"lamina-associated\")[Title/Abstract] AND (Hox OR HOXA OR HOXB)[Title/Abstract]",
    "records": 5,
    "date": "2026-09-19",
    "relevant": 1,
    "note": "only PMID 42679029 is substantive, and it is Drosophila"
  }
]
How this could bite us back
  • Lamina biology is not currently druggable; even a positive result here would produce a mechanism with no intervention attached, which is a reason to keep the question as a cheap reanalysis rather than a wet-lab programme.
  • If LAD release turns out to be required for menin-inhibitor-induced differentiation, then cells with lamin mutations or altered lamin B1 levels — common in aggressive leukemia — would be intrinsically resistant, and nobody is stratifying for that.
Still open
  1. Does KMT2A or its fusion contact the lamina, as opposed to sitting near the envelope? Lamin B1 DamID in a KMT2A-rearranged line answers it and has never been run; the 2000 imaging cannot.
  2. Does the HOXA cluster sit at or near a LAD border in fetal-liver HSC or LMPP, the cells of origin this graph identifies? Published Lamin B1 DamID plus published KMT2A::AFF1 occupancy answers this by overlay, with no new experiment.
  3. Does fusion degron depletion change lamina contacts anywhere in the genome? If LAD architecture is unchanged, the compartment can be formally excluded from the mechanism and the beat can stop asking.
  4. Is the differentiation block accompanied by a failure of the normal LAD release that accompanies lineage commitment? That would make the lamina a read-out of the block rather than a cause, but a quantitative one.
  5. Do LAD borders, defined by CpG islands, overlap the fusion's CpG-island target set more than chance? A trivially runnable permutation test that nobody has run.
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    DamID mapping of nuclear-lamina interactions in human fibroblasts identified more than 1,300 discrete lamina-associated domains of 0.1-10 Mb that are gene-poor and transcriptionally repressive and are sharply bordered by CTCF binding sites, outward-facing promoters or CpG islands.

    Domain organization of human chromosomes revealed by mapping of nuclear lamina interactions · Guelen L, Pagie L, Brasset E, Meuleman W, Faza MB, Talhout W, Eussen BH, de Klein A, Wessels L, de Laat W, van Steensel B · Nature 2008
    PMID 18463634DOI
    What this evidence cannot show

    Human lung fibroblasts only; LAD maps are cell-type specific, so these coordinates do not transfer to haematopoietic progenitors. DamID reports contact frequency in a population, not the position of a locus in any single nucleus.

    Evidence that argues against this (1)
    • The founding DamID LAD map is a population average and single-cell measurement shows it does not describe any individual nucleus. Kind et al. report that only about 30 per cent of LADs are positioned at the nuclear periphery in a given nucleus, that contacts are intermittent rather than stable, and that after mitosis LAD positioning is stochastically reshuffled rather than inherited - though contact is linked to repression and H3K9me2 in single cells. For this beat the implication is specific: 'gene X is in a LAD' is a probability statement about a cell population, so a claim that a KMT2A target is peripheral (or that the fusion must overcome peripheral sequestration) cannot be made from a consensus map alone. Read: abstract only. Kind J et al. 2013
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Lamin B1 DamID across a mouse embryonic-stem-cell to neural-precursor to astrocyte differentiation series shows lamina-associated domains are cumulatively remodelled at hundreds of sites during lineage commitment, with genes leaving the lamina often activated and others becoming unlocked for later activation.

    Molecular maps of the reorganization of genome-nuclear lamina interactions during differentiation · Peric-Hupkes D, Meuleman W, Pagie L, Bruggeman SW, Solovei I, Brugman W, Graf S, Flicek P, Kerkhoven RM, van Lohuizen M, Reinders M, Wessels L, van Steensel B · Molecular Cell 2010
    PMID 20513434DOI
    What this evidence cannot show

    Mouse, and a neural lineage. Transfer to haematopoiesis is an assumption; no haematopoietic differentiation series of comparable depth has been published.

    Evidence that argues against this (1)
    • Population-level remodelling maps hide the fact that lamina association is stochastic in single cells and not heritable. Kind et al. showed with single-cell contact-memory tracking that only about 30% of LADs occupy the nuclear periphery in any given nucleus, that individual LADs make intermittent molecular contact with the lamina while remaining constrained peripherally, that LAD positioning is not detectably inherited through mitosis but is stochastically reshuffled, and that contact correlates with repression and H3K9me2 cell by cell, with G9a regulating it. 'Cumulative remodelling at hundreds of sites during lineage commitment' is therefore a change in the PROBABILITY of peripheral contact across a population, and a model in which a gene is 'released from the lamina and activated' during differentiation is an averaged description of a per-cell coin flip. Kind J et al. 2013
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Lamin B1 DamID with FISH in Jurkat T cells shows that T-cell activation releases lamina-associated enhancers and genes from the nuclear envelope in a constrained manner, with TADs repositioning as units and released loci associating in A2 chromosome subcompartments.

    PMID 28424353DOI
    What this evidence cannot show

    A mature T-cell leukemia line undergoing activation, not a haematopoietic stem or progenitor cell undergoing differentiation, and not a KMT2A-rearranged context. Included as the only verified haematopoietic lamina dataset, i.e. as proof the assay is feasible, not as evidence about KMT2A.

    Evidence that argues against this (1)
    • Searched and empty, recorded explicitly. No published study maps lamina association in a KMT2A-rearranged leukemia cell line or primary blast, by DamID, Lamin B1 ChIP or FISH. PubMed searches on 2026-09-18 combining KMT2A/MLL/MLL-AF4/infant ALL with lamina-associated domain, DamID, lamin B1 and nuclear periphery returned nothing in the disease system. Every lamina statement in this entry is therefore transferred from fibroblasts, mouse neural differentiation and Jurkat T cells; the `weak` grade is right and the gap is a genuine experiment, not a literature-search failure. No citation is attached because there is nothing to cite: this note is a bounded absence recording searches that returned nothing, not a sourced fact. Uncited and unciteable look identical in this field, so it is said here explicitly (Q42 citation-integrity repair, 2026-09-20).
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In Drosophila, the Hox protein Ultrabithorax represses autophagy-related (atg) genes by tethering their loci to the nuclear periphery, and this repression depends on interaction with the nuclear lamina component Lamin-C.

    A Hox-dependent anchoring mechanism mediates transcriptional repression of autophagy-related genes at the nuclear periphery · Vanderperre S, Ajuria L, Duffraisse M, Da Silva T, Vonau S, Brocard J, Forcet C, Merabet S · Proceedings of the National Academy of Sciences USA 2026
    PMID 42679029DOI
    What this evidence cannot show

    Drosophila, Lamin-C, and a Hox *protein* acting as a tether for non-Hox target loci - not the mammalian HOXA cluster, not a leukaemia, and not KMT2A. Cited only to show that a Hox-lamina axis exists somewhere, so the mammalian question is not absurd; it does not raise the evidence grade for KMT2A above none.

    Evidence that argues against this (1)
    • A deterministic tethering mechanism is not the mammalian default, which weakens the analogy this item is included to license. Kind et al. found that in human cells lamina contact is intermittent at the molecular level, that only about 30% of LADs are peripheral in any one nucleus, and - the decisive point for a tethering model - that LAD positioning is NOT detectably inherited through mitosis but is stochastically reshuffled each cycle. A Hox protein anchoring specific loci to the periphery through Lamin-C in Drosophila therefore describes a targeted, stable interaction of a kind that mammalian genome-lamina contacts have not been shown to make. The mammalian question is not absurd, as this entry says, but the Drosophila mechanism does not transfer as a mechanism - only as a motivation. Kind J et al. 2013
  5. E-envelopeweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.journal-article

    A conditionally expressed KMT2A amino-terminal fragment and endogenous MLL localise in a punctate nucleoplasmic pattern that includes regions attached to the nuclear envelope and the nucleolar periphery, by immunofluorescence, confocal and immunoelectron microscopy. The same pattern is reported for conditionally expressed MLL-AF9 and MLL-AF4.

    PMID 11069025DOI
    What this evidence cannot show

    Read in full text 2026-09-20 via the PMC HTML route (PMC7543881, 63,706 characters) after `fullTextXML` returned HTTP 500. **The word 'lamin' does not appear in the paper**, so this is envelope proximity and not a lamina claim - the distinction is the whole reason the grade stays weak. Everything here is conditional over-expression in U937 and HeLa, neither of which carries a KMT2A rearrangement, and immunofluorescence has no size dimension, so nothing in it is attributable to an endogenous fusion.

    Evidence that argues against this (1)
    • Envelope proximity in an over-expression system is very thin support for a lamina hypothesis, and reading it as more would repeat the error this correction fixes in the other direction. **The honest grade is: one paper, one construct, two non-rearranged cell lines, and no lamina measurement anywhere in this literature.** What changes is that the entry can no longer say nobody has produced anything adjacent. Caslini C et al. 2000
Connected to