UBTF tandem duplications INVENT an export signal, and that invented signal is what puts the protein on chromatin - the cleanest proof the XPO1 address is real
also called UBTF-TD XPO1 · UBTF tandem duplication nuclear export signal · de novo NES creates a chromatin address
In a high-risk childhood leukaemia, a short stretch of the UBTF gene is duplicated, and the duplication accidentally creates a new export tag. That new tag grabs XPO1 - and instead of exporting the protein, it glues it onto the wrong genes. Delete the tag and the disease programme does not start.
The same thing, for a scientist
**UBTF tandem duplications (UBTF-TD)** define a high-risk AML subtype in children and young adults, and are a genotype in which **menin inhibitors show therapeutic promise with acquired resistance as the problem**. Barajas et al. report that the duplications cause **structural destabilisation and create nuclear export signal (NES) motifs that interact directly with XPO1**. In **cord-blood CD34+ UBTF-TD models** those interactions **drive aberrant chromatin binding and transcriptional activation** of the UBTF-TD gene signature; **mutagenesis of the NES motifs abolishes chromatin localisation, transcriptional dysregulation, proliferation and differentiation phenotypes**; and in preclinical human UBTF-TD models, **XPO1 inhibition disrupts UBTF-TD chromatin localisation, reduces tumour burden and promotes differentiation** [E1]. **Why this is the strongest single instance of the whole export axis.** Every other lesion on the XPO1 address *inherited* its grip module - the FG repeats were already in NUP214 and NUP98, the NES was already in CALM, the NES arrived with the NPM1 frameshift as part of a well-known oncogene. **Here the NES is created de novo by the duplication, and deleting just the NES is sufficient to abolish the chromatin phenotype.** That is a necessity experiment on the grip module itself, in a human haematopoietic cell, in the lesion's own genotype. **And it carries the menin question with it.** UBTF-TD AML is a menin-inhibitor-responsive subtype with acquired resistance, and this paper's proposal is XPO1 inhibition as a therapy for the same subtype. That is the second independent 2026 source pairing a menin-inhibitor genotype with an XPO1 dependency (the first being the KMT2A-r/NPM1-m ziftomenib-plus-selinexor preprint). **System frame:** human cord-blood CD34+ models and preclinical human UBTF-TD leukaemia models. **Paediatric and AYA myeloid disease**, not infant B-lineage KMT2A::AFF1.
What a de novo NES tells you that an inherited one cannot
Consider the alternative explanations for the XPO1 address across the other lesions:
- The FG repeats bind lots of things, XPO1 among them - available for NUP98 and NUP214 fusions.
- NPM1c's NES is doing transport and the chromatin signal is downstream - available for NPM1c, and in fact actively argued by Brunetti et al. (see
mechanism-npm1c-crm1-hoxa-recruitment). - The CALM NES is an export signal and export is the mechanism - the unresolved CALM-AF10 contradiction.
None of these survives UBTF-TD. The duplication has no pre-existing transport role to fall back on, and the reported experiment is exactly the right one: mutate the NES, lose the chromatin binding. If the NES were working as an export signal, mutating it would make the protein more nuclear, not less chromatin-bound.
That is the clearest available argument that in this family of lesions the NES-XPO1 contact is a chromatin-docking interaction, not a transport interaction - which is the exact question this graph left unresolved for CALM-AF10.
What it does not settle
- It is one paper, one group, abstract-level in this beat, published January 2026.
- UBTF-TD is not a fusion, and UBTF is a nucleolar transcription factor for RNA polymerase I. Whether its chromatin targets overlap the HOX clusters is not stated in the abstract; the abstract says 'genes dysregulated in UBTF-TD tumors', and the published UBTF-TD signature is HOXA/MEIS1-high, but that link was not verified in this beat and must not be asserted.
- 'XPO1 inhibition reduces tumour burden' in a preclinical model of a lesion whose own dependency you have just demonstrated is expected; the discriminating part is the chromatin-localisation disruption, not the survival.
- Nothing here is KMT2A-rearranged, lymphoid or infant.
| mechanism class | chromatin targeting by receptor capture |
|---|---|
| lesion | UBTF tandem duplication (UBTF-TD), a high-risk AML subtype of children and young adults |
| grip module | nuclear export signal motifs CREATED DE NOVO by the duplication |
| receptor | gene-xpo1 |
| necessity experiment | NES mutagenesis abolishes chromatin localisation, transcriptional dysregulation, proliferation and differentiation phenotypes |
| drug effect | XPO1 inhibition disrupts UBTF-TD chromatin localisation, reduces tumour burden and promotes differentiation in preclinical human models |
| menin context | UBTF-TD AML is a menin-inhibitor-responsive subtype in which acquired resistance is the stated problem; XPO1 inhibition is proposed for the same subtype |
| target loci | not named in the abstract - deliberately not asserted here |
| why it matters most | the only lesion on this axis whose grip module did not pre-exist, so 'the NES was already doing transport' is not available as an alternative explanation |
| not present | no KMT2A-rearranged, lymphoid or infant content |
systems
[ "cord-blood CD34+ UBTF-TD models", "preclinical human UBTF-TD leukaemia models" ]
- XPO1 is common-essential; a therapy proposed for UBTF-TD on this basis inherits the whole SINE toxicity profile, and UBTF-TD is a paediatric disease where selinexor's dose ceiling is set by cerebellar toxicity.
- The lesion creates the NES, so a resistant subclone could in principle lose or alter the duplication rather than the drug target - a resistance route with no precedent and no assay.
- If XPO1 inhibition is proposed as the answer to menin-inhibitor resistance in this subtype, the sequencing question matters: giving both may be better than giving one after the other, and nothing published addresses it.
- Does UBTF-TD occupy the HOXA cluster, and is chromatin-bound XPO1 already there? If yes, this is the seventh lesion on the same address rather than a parallel case; the abstract does not say.
- Does XPO1 inhibition restore menin-inhibitor sensitivity in menin-inhibitor-resistant UBTF-TD? Both halves are in this paper's framing and the combination experiment is not reported in the abstract.
- Could a de novo NES be created by other duplications and insertions in leukaemia genomes? Nobody appears to have scanned for NES-creating in-frame duplications systematically, and this paper makes that a well-posed question.
- Would NES mutagenesis do the same thing to KMT2A::AFF1? It has no NES, which is the point - but the negative control has never been run alongside a positive one in the same experiment.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
UBTF tandem duplications result in structural destabilisation and create nuclear export signal motifs that mediate direct interactions with XPO1; in cord-blood CD34+ UBTF-TD models these interactions drive aberrant chromatin binding and transcriptional activation of genes dysregulated in UBTF-TD tumours; mutagenesis shows the NES motifs are critical for chromatin localisation of UBTF-TD proteins, for transcriptional dysregulation and for proliferation and differentiation; and in preclinical human UBTF-TD leukaemia models XPO1 inhibition disrupts UBTF-TD chromatin localisation, reduces tumour burden and promotes differentiation.
Tandem Duplications in UBTF Create XPO1-Dependent Nuclear Export Signals that Reveal a Leukemic Therapeutic Dependency. · Barajas JM, Phillips AH, Wang J, et al. · Blood Cancer Discovery 2026What this evidence cannot show
Abstract-level for this beat; the mutagenesis and chromatin figures were not opened, and rule 80 applies to anything the abstract does not name. One group, published 2026-01. Cord-blood CD34+ models are human and primary-derived but are engineered expression systems, not patient blasts. The abstract does not name the target loci, so any statement that UBTF-TD shares the HOX address with the other lesions on this axis would be an inference and is deliberately not made here. Paediatric and AYA myeloid disease only. XPO1 is common-essential, so 'XPO1 inhibition reduces tumour burden' is not by itself evidence of a private dependency - the chromatin-localisation disruption is the load-bearing result.
Evidence that argues against this (1)
- XPO1 is common-essential, so 'XPO1 inhibition reduces tumour burden' in a UBTF-TD model is close to the null expectation for any leukaemia model, and the abstract names no fusion-negative lineage-matched comparator. The load-bearing result is the NES mutagenesis, which is genetic and clean; the pharmacological arm adds less than it appears to. A 2026 field review of XPO1 targeting in AML lists NPM1-mutated, DEK::NUP214-positive and SF3B1-mutated neoplasms as the responsive genotypes and does not yet list UBTF-TD. Liu Y et al. 2026
- E2weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
A commentary accompanies this paper in the same issue (Adams, p. 15), indicating the finding was judged notable by the journal.
Tandem Duplications in UBTF Create XPO1-Dependent Nuclear Export Signals that Reveal a Leukemic Therapeutic Dependency. · Barajas JM, Phillips AH, Wang J, et al. · Blood Cancer Discovery 2026What this evidence cannot show
This is a statement about editorial attention, taken from the source abstract's own closing line, and carries no evidential weight about the biology. Recorded only so a future reader knows a commentary exists and can find it; the commentary itself was not retrieved or read.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT on what the editorial attention is attention to, since notability and generalisability point in opposite directions here. Kaburagi et al. found UBTF-ITD in 6 of 503 Japanese paediatric de novo AML, 0 of 175 adult AML, and 0 of 65 cell lines including 39 acute lymphoblastic leukaemia lines; five of the six cases carried FLT3-ITD and all six had high PRDM16 expression, with gene-expression patterns resembling NUP98::NSD1 and FUS::ERG. A commentary marks a finding as interesting to a journal's readership; it is not a replication, and the lesion it concerns is rare, myeloid-restricted in every series and cell line tested so far, and confounded with FLT3-ITD in five of six index cases. This entry's title calls the result 'the cleanest proof the XPO1 address is real' - it is the cleanest proof that a created NES can put ONE protein on chromatin in ONE lineage. Kaburagi T et al. 2023
UBTF-TD instance - a de novo NES, the necessity experiment
NES motifs created de novo by the duplication bind XPO1
the necessity experiment on the grip module itself
UBTF-TD AML is a menin-inhibitor-responsive subtype with acquired resistance as the problem
UBTF-ITD was found in 6 of 503 Japanese paediatric AML, 0 of 175 adults and 0 of 65 cell lines including 39 ACUTE LYMPHOBLASTIC LEUKAEMIA lines, with FLT3-ITD in 5 of 6 cases (PMID 36448876) - the cleanest proof of a created NES is myeloid-restricted and confounded, and cannot reach infant B-lineage KMT2A::AFF1