Oncofusion

start / graph / localization-nucleolus-kmt2a-fusion

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q4b-localization

The nucleolus — the one compartment with direct published images of a KMT2A fusion inside it

also called MLL-AF4 nucleolus · nucleolar localization MLL · pSer SL1 RNA polymerase I · nucleolar stress MLL-AF4

Of all the sub-compartments inside the nucleus, the nucleolus — the factory that builds ribosomes — is the only one where somebody has actually photographed a KMT2A fusion protein. Two independent groups, twenty-two years apart, saw it there, and the second found the fusion switches on a nucleolar stress response.

The same thing, for a scientist

Immunofluorescence, confocal and immuno-electron microscopy of conditionally expressed MLL N-terminal fragments and of MLL-AF9 and MLL-AF4 chimeras in U937 and HeLa placed the proteins in punctate nucleoplasm plus the nucleolar periphery, and identified the MLL amino terminus as the targeting element common to essentially all chimeras [E1]. Twenty-two years later a Frankfurt group reported that human MLL-AF4 is predominantly nucleolar, interferes with Pol I transcription and ribosome biogenesis, and turns on the nucleolar stress pathway, while a murinised pSer variant sits more in the nucleoplasm [E2]. The mechanistic bridge is that the AF4/AFF1 pSer domain associates with SL1, the Pol I pre-initiation core, and uses it to load TBP at Pol II promoters [E3]. Both imaging studies used exogenously expressed constructs, so the honest reading is that nucleolar targeting is real and reproducible in overexpression systems and has never been confirmed at endogenous levels in a t(4;11) blast.

Go deeper
What was actually seen, and how1 deeper

Caslini et al. expressed a 50 kDa MLL N-terminal fragment spanning the AT-hooks (MLL3AT) conditionally in U937 and HeLa, and looked by indirect immunofluorescence, confocal laser scanning microscopy and immuno-electron microscopy. MLL3AT and endogenous MLL both gave a punctate nucleoplasmic pattern including regions attached to the nuclear envelope and the periphery of the nucleolus, and both were recovered in interphase nuclear matrices and on mitotic chromosomal scaffolds with topoisomerase II. Critically, conditionally expressed MLL-AF9 and MLL-AF4 chimeras showed the same nucleoplasm-plus-nucleolus pattern, which the authors read as a common address conferred by the MLL amino terminus 1.

Siemund et al. compared human MLL-AF4 with a murinised pSer variant (hMLL-mAf4, the construct that is leukemogenic in mice, versus hMLL-hAF4, which is not). Gene-expression profiling plus intracellular localisation studies put human MLL-AF4 predominantly inside the nucleolus, where it interferes with Pol I transcription and ribosome biogenesis and activates the nucleolar stress pathway; the murinised pSer version localises more to the nucleoplasm 2.

Note the discrepancy the graph should not paper over: Caslini saw the periphery of the nucleolus, Siemund reports predominantly inside it. Different constructs, different cells, twenty-two years apart, both overexpressed.

Why this is not the compartment people expected

The default mental picture of a KMT2A fusion is a Pol II elongation machine parked on CpG-island promoters. A nucleolar address puts it at a Pol I organelle instead. The connection is not arbitrary. Okuda et al. showed that the pSER domain of AF4-family proteins associates with SL1 — selectivity factor 1, a core component of the RNA polymerase I pre-initiation complex — on chromatin, and loads TBP onto the promoter to initiate Pol II-dependent transcription 3. So AEP and MLL-AEP fusions borrow a Pol I machine to start Pol II genes. If the fusion recruits SL1, some residence in the compartment where SL1 normally lives is not surprising.

How much weight this can carry

Both imaging papers rest on exogenously expressed constructs in non-leukemic or heterologous cell lines (U937, HeLa, transfected mammalian cells), not on endogenously tagged fusion in a t(4;11) line and not on fixed patient marrow. Overexpressed nucleolar accumulation is a recognised artefact class: basic, arginine-rich sequences partition into the nucleolus above a concentration threshold, and both the MLL AT-hook region and the AFF1 pSer region are exactly that kind of sequence.

The cheap decisive experiment is unchanged and undone: endogenously tag KMT2A::AFF1 at the native locus in SEM or RS4;11, co-stain fibrillarin or nucleophosmin, and count. Until that exists, the claim should be stated as the fusion goes to the nucleolus when overexpressed, twice, independently — not as the fusion is in the nucleolus.

The species difference is the interesting part

Siemund's framing matters for every mouse model in this programme. The human pSer domain recruits SL1 efficiently; the murine counterpart apparently does not, and the murinised protein sits more in the nucleoplasm 23. If localisation and SL1 recruitment track with species, then hMLL-hAF4 (not leukemogenic in mice) and hMLL-mAf4 (leukemogenic) differ in where the protein goes, not only in what it binds. That is a testable explanation for one of the most awkward facts in the KMT2A::AFF1 modelling literature, and it cuts against reading murine pSer-variant models as faithful reporters of human t(4;11) biology.

Facts and figures
compartmentnucleolus
evidence class for kmt2adirect imaging of the fusion, but only in overexpression systems
endogenous tag datafalse
patient material datafalse
marker proteins
[
  "fibrillarin (FBL)",
  "nucleophosmin (NPM1)",
  "UBF",
  "SL1 (TAF1A-D)"
]
cell systems used
[
  "U937",
  "HeLa",
  "transfected mammalian cells"
]
methods used
[
  "indirect immunofluorescence",
  "confocal laser scanning microscopy",
  "immuno-electron microscopy",
  "gene expression profiling"
]
How this could bite us back
  • If part of the fusion's function runs through Pol I/SL1 in the nucleolus, a menin inhibitor that only strips it off Pol II promoters leaves that arm untouched, and the residual activity would look like an incomplete response rather than resistance.
  • Targeting ribosome biogenesis to exploit a nucleolar dependency hits every proliferating normal cell; nucleolar stress agents have a poor therapeutic index in infants.
Still open
  1. Is endogenous KMT2A::AFF1 nucleolar in a t(4;11) blast? A knock-in tag in SEM or RS4;11 plus a fibrillarin co-stain settles it in one experiment and nobody has done it.
  2. Does the nucleolar stress pathway activation reported for MLL-AF4 contribute to transformation, or is it a toxicity of overexpression? A degron time-course reading RPL5/RPL11-MDM2-p53 output would separate them.
  3. Does the human-versus-murine pSer difference in localisation explain why hMLL-hAF4 is not leukemogenic in mice while hMLL-mAf4 is? If so, every murine pSer model reports a different subnuclear biology from the human disease.
  4. Does menin inhibition change the nucleolar fraction of the fusion? If the fusion is evicted from CpG-island chromatin, a nucleolar reservoir is one place it could go.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Conditionally expressed MLL N-terminal fragment, endogenous MLL, and conditionally expressed MLL-AF9 and MLL-AF4 chimeras all localise to punctate nucleoplasm including regions attached to the nuclear envelope and the periphery of the nucleolus, and are present in interphase nuclear matrices and colocalise with topoisomerase II on mitotic chromosomal scaffolds.

    PMID 11069025DOI
    What this evidence cannot show

    Conditionally expressed (exogenous) constructs in U937 and HeLa, an overexpression system in a monoblastic line and a cervical carcinoma line, neither carrying a KMT2A rearrangement. Immunofluorescence and immuno-EM on fixed cells; no live-cell or endogenous-tag data. The nucleolar signal is described as peripheral, not intranucleolar.

    Evidence that argues against this (1)
    • Every localisation in this item is a fixed-cell measurement of an overexpressed construct, and both halves of that are now known to distort the readout. Irgen-Gioro et al. imaged cells before and after paraformaldehyde fixation and found that fixation can both enhance and diminish apparent punctate/droplet behaviour, that for some proteins puncta appear in fixed cells that have none when live, and that the direction of the artefact depends on the relative rates of fixation and of the protein's own interaction dynamics - with glycine co-fixation reversing it. Conditionally expressed MLL fragments and wild-type MLL have different partners and therefore different interaction kinetics, so a comparison of their punctate patterns in fixed U937 and HeLa cells is not controlled for the thing being measured. The nucleolar-periphery signal in particular is a statement about where fixed material accumulated. Irgen-Gioro S et al. 2022
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Human MLL-AF4 localises predominantly inside the nucleolus, interfering with Pol I transcription and ribosome biogenesis and turning on the nucleolar stress pathway, whereas a murinised pSer variant of the fusion localises more to the nucleoplasm.

    MLL-AF4 and a murinized pSer-variant thereof are turning on the nucleolar stress pathway · Siemund AL, Hanewald T, Kowarz E, Marschalek R · Cell & Bioscience 2022
    PMID 35468859DOI
    What this evidence cannot show

    Transfected mammalian cells expressing the fusion from a construct, not endogenous tagging in a t(4;11) line and not patient material; basic arginine-rich sequences partition into nucleoli above a concentration threshold, which is precisely the artefact this design cannot exclude. Localisation is reported qualitatively, with no quantified nucleolar fraction in the abstract.

    Evidence that argues against this (1)
    • The earlier and methodologically broader imaging study (immuno-EM as well as confocal) placed MLL and its chimeras at the nucleolar periphery with a predominantly nucleoplasmic punctate distribution, not inside the nucleolus. Caslini C et al. 2000
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The pSER domain of AF4-family proteins associates on chromatin with SL1, a core component of the RNA polymerase I pre-initiation complex, and loads TATA-binding protein onto the promoter to initiate RNA polymerase II-dependent transcription, a mechanism used by AEP and by MLL-AEP fusion proteins.

    AF4 uses the SL1 components of RNAP1 machinery to initiate MLL fusion- and AEP-dependent transcription · Okuda H, Kanai A, Ito S, Matsui H, Yokoyama A · Nature Communications 2015
    PMID 26593443DOI
    What this evidence cannot show

    Biochemistry and chromatin association, not imaging: this paper shows SL1 is recruited, not that the fusion resides in the nucleolus. Using it to explain nucleolar localisation is an inference the paper itself does not make.

    Evidence that argues against this (1)
    • The same laboratory lineage publishes a different account of how AF4-family proteins activate MLL-fusion targets, and it does not run through Pol I machinery. Yokoyama et al. showed that AF4-family and ENL-family proteins form a higher-order complex with P-TEFb (AEP) that is normally recruited to MLL-target chromatin to facilitate transcription, that MLL oncoproteins fused to AEP components form constitutive MLL/AEP hybrid complexes causing sustained target expression and progenitor transformation, and that even MLL-AF6 recruits AEP by an alternative route - concluding that bypassing normal AEP recruitment is the strategy most MLL oncoproteins use. On that account the pSER domain's job is P-TEFb-dependent elongation control, not TBP loading via SL1. Using an SL1 interaction to explain a nucleolar address therefore chooses one of two competing mechanistic accounts without saying so, and the imaging it is meant to explain is independent of both. Yokoyama A et al. 2010
Connected to