Oncofusion

start / graph / concept-an-export-receptor-used-as-a-postal-address

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q61-export-axis

How a lift operator became a postal address: the idea behind the XPO1 axis, in plain language

also called what chromatin-bound CRM1 means · addressing by receptor capture explained

Cells have a protein whose job is to carry things out of the nucleus, grabbing them by a short tag. In several leukaemias, some of that carrier is parked on specific genes instead of carrying anything - so a mutant protein with the right tag gets stuck there rather than being carried out, and switches those genes on.

The same thing, for a scientist

A layman-first explainer for the mechanism this beat is about. The technical entries are `mechanism-nes-crm1-chromatin-addressing`, `gene-xpo1` and `localization-xpo1-chromatin-bound-versus-soluble-pool`.

Go deeper
The normal job

Think of the nucleus as a building with one kind of doorway. XPO1 (also called CRM1, also called exportin-1) is the porter who takes things out. It recognises its passengers by a short greasy patch on their surface called a nuclear export signal - an NES. Hundreds of different proteins carry one.

The porter only picks a passenger up inside the building, because it needs a second thing in its other hand: a small protein called RAN loaded with GTP, which is only available inside. Out in the corridor, RAN's GTP is spent, the porter's grip relaxes, and the passenger is released. That one-way chemistry is the whole reason export goes outward. 1

The twist

Some of the porter is not walking about. A fraction of XPO1 is stuck to the DNA itself, at a small number of places - in particular the HOX gene clusters, which control how a blood cell decides what to become. 2

Now imagine a mutant protein that happens to carry the right greasy patch. It walks up to the porter expecting a lift. But this porter is bolted to the floor at the HOX genes. So instead of leaving, the mutant protein gets parked at the HOX genes - which is exactly where a leukaemia wants to be, because switching those genes on freezes the cell in an immature, dividing state. 2

The grip can be a greasy NES patch or a different sticky surface called FG repeats that some nuclear-pore proteins carry. Either works. Seven different leukaemia lesions have independently found this parking spot, including one - UBTF tandem duplication - where the greasy patch did not exist before and the mutation accidentally created it. 3

Why it is hard to turn into a treatment

There are drugs that jam the porter - selinexor is approved for two adult cancers and eltanexor is in trials. Jamming it does knock the mutant protein off the HOX genes and does switch those genes down.

Three problems:

  1. Every cell needs the porter. In a survey of 1,208 human cancer cell lines, removing XPO1 kills essentially all of them. There is no safety margin built into the target.
  2. The drug jams both jobs at once. It binds the same pocket the greasy patch binds, so there is no way to stop the parking without also stopping the export. Nobody has designed a molecule that could.
  3. Leukaemia cells answer back. Blocking XPO1 in AML switches on a survival signal through a receptor called P2RY2, immediately, in every cell.

So the mechanism is elegant and the intervention is blunt. This graph records both.

The bit that is still argued about

Two respectable laboratories disagree about which half of the story is doing the work in one of these diseases. For mutant NPM1, one group says the mutant has to be out in the cytoplasm and the drug works by dragging it back in; the other says the mutant has to be in, parked on the HOX genes, and the drug works by knocking it off. Same drug, same result, opposite explanations. See mechanism-npm1c-crm1-hoxa-recruitment.

That disagreement is not a detail. If you were designing a treatment you would want to know whether you were trying to move a protein in or out.

Facts and figures
audiencelayman first; a bright secondary-school reader should follow every section
metaphor health warningthe porter metaphor breaks down in one important place - a real porter that is bolted down would simply be idle, whereas chromatin-bound XPO1 appears to do positive work holding an assembly together. The metaphor explains the trap; it does not explain the scaffold.
technical entries
[
  "gene-xpo1",
  "mechanism-nes-crm1-chromatin-addressing",
  "localization-xpo1-chromatin-bound-versus-soluble-pool",
  "hyp-xpo1-is-a-shared-address-for-hoxa-driven-leukaemia"
]
How this could bite us back
  • The metaphor makes the intervention sound simple - jam the porter. The target is needed by every cell in the body, the only approved agent failed a randomised trial in AML, and in children it caused reversible brain toxicity at the higher dose tested.
  • An explainer that makes a mechanism feel inevitable is a risk in itself: this one is built on cell lines and mouse models, with a single study using primary patient cells.
Still open
  1. Is the parking spot normal or created by the disease? The honest answer is that nobody has looked in a healthy blood progenitor, and the metaphor reads very differently depending on the answer.
  2. If seven lesions found the same parking spot independently, how many more are there? Nobody has scanned leukaemia genomes for mutations that create a new NES.
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    XPO1 mediates nuclear export of cargoes bearing a leucine-rich NES, binding cooperatively to the NES and to RAN-GTP in the nucleus and releasing cargo in the cytoplasm when RANBP1 and RANGAP1 drive Ran-GTP hydrolysis; the asymmetric Ran-GTP/GDP distribution confers directionality.

    UniProtKB - O14980 (XPO1_HUMAN), Exportin-1 · UniProt Consortium · UniProtKB 2026
    What this evidence cannot show

    Curated database synthesis fetched live 2026-09-20; the porter metaphor is this entry's, not the database's.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT against the completeness of the lift-operator picture the curated record supports. Chen et al. show that XPO1 directly binds chromatin, co-occupies hundreds of loci with p300, and that degrading XPO1 diminishes p300's chromatin occupancy and phenocopies p300 inhibition - a transport-independent structural function that no part of the UniProt export annotation carries. The metaphor this entry is built on ('a lift operator that became a postal address') therefore has a third reading the entry does not offer: XPO1 may simply be a widespread chromatin cofactor that also exports things, in which case being found at a locus is unremarkable. Chen YF et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Chromatin-bound CRM1 is present at HOX cluster regions in leukaemia cells and recruits SET-Nup214 and NPM1c onto those clusters, with CRM1 upstream of the fusion since fusion knockdown does not reduce chromatin-bound CRM1.

    PMID 31755865DOI
    What this evidence cannot show

    Human leukaemia cell lines only. The same paper finds CRM1 at HOX in a line carrying none of these lesions, so 'the parking spot' may be a normal feature the lesions exploit rather than something they create.

    Evidence that argues against this (1)
    • The same paper that establishes the parking spot also finds CRM1 at HOX regions in HEL cells, which carry none of these lesions, and states the significance is undetermined. The metaphor of a lesion 'finding' a parking spot may have the causality backwards: the spot may be a normal feature of HOX regulation, in which case blocking it is not a tumour-selective act. Oka M et al. 2019
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    UBTF tandem duplications create nuclear export signal motifs de novo that interact with XPO1 and drive aberrant chromatin binding, and mutating those created motifs abolishes chromatin localisation and the downstream phenotypes.

    PMID 41182175DOI
    What this evidence cannot show

    Abstract-level; human cord-blood CD34+ models; paediatric and AYA myeloid disease. Used here as the example that makes the concept vivid, because the parking tag did not exist before the mutation.

    Evidence that argues against this (1)
    • NAMED METHODOLOGICAL DEFEATER for 'creates NES motifs de novo'. The leucine-rich NES consensus is degenerate: Kosugi et al. screened a random peptide library in a nuclear-export-dependent yeast selection and recovered 101 distinct CRM1-dependent NESs falling into six different patterns of conserved hydrophobic spacing, and state explicitly that the traditional consensus is too poorly defined to identify functional NESs. A tandem duplication that juxtaposes hydrophobic residues will generate sequences matching that consensus by chance. The UBTF-TD claim is only as strong as its functional test - direct XPO1 interaction plus loss of chromatin localisation on mutagenesis - and not at all strong as a motif-creation claim, which is how this concept entry paraphrases it ('the parking tag did not exist before the mutation'). Kosugi S et al. 2008
Connected to