AF10 fusions recruit and stabilise JAK1 - a kinase bound to the fusion with no phosphosite on the fusion
also called JAK1 AF10 · CALM-AF10 JAK1 · MLL-AF10 JAK/STAT
The AF10 fusion protein physically grabs a signalling enzyme called JAK1 and protects it from being broken down, which switches on an inflammation programme. But nobody has shown that JAK1 puts a phosphate on the fusion itself - the kinase is a passenger, not an editor, as far as anyone has looked.
The same thing, for a scientist
Inducible mouse models of PICALM::MLLT10 and KMT2A::MLLT10 show that AF10 fusion proteins activate a JAK/STAT-mediated inflammatory cascade through **direct recruitment of JAK1**, confirmed by Flag immunoprecipitation of both fusions and by reciprocal co-IP of endogenous Jak1. Both fusions are associated with higher Jak1 protein without a change in JAK1 mRNA, which the authors interpret as fusion-mediated stabilisation; they explicitly attribute the increased JAK1 phosphorylation in fusion-on cells to higher total JAK1 rather than to increased specific activity. Genetic Jak1 deletion or pharmacological JAK/STAT inhibition is antioncogenic in mouse and human AF10-fusion models. **This is the closest thing to a kinase-gated contact that exists for MLLT10, and it is not one**: no phosphorylation site on AF10 or on either fusion is mapped anywhere in the paper or in this beat's searches. The entry is written to mark the distinction, because a fusion that binds a kinase and a fusion that is phosphorylated by a kinase are different objects and the literature is easy to read as if they were the same.
What is established and what is not
Established 1: AF10 fusions co-immunoprecipitate with JAK1; Jak1 protein is higher in fusion-on cells with unchanged mRNA; STAT3 phosphorylation is elevated; Jak1 deletion or JAK/STAT inhibition is antioncogenic; atovaquone reduces STAT3 phosphorylation and impairs proliferation of both CALM-AF10 and MLL-AF10 AML cells.
Not established, and not claimed by the authors: that JAK1 phosphorylates AF10 or either fusion; that any residue of AF10 is a JAK1 substrate; that the interaction requires a phosphorylation event on the fusion at all. The authors state that the increase in JAK1 phosphorylation in fusion-on cells is attributable to higher total JAK1.
The directionality is therefore fusion-to-kinase, not kinase-to-fusion. That is the opposite of every other entry in this beat.
Which AF10 region binds JAK1 is not resolved in the accessible record read here, so whether the contact is on retained or discarded sequence cannot be stated - although both PICALM::MLLT10 and KMT2A::MLLT10 do it, and the sequence those two fusions have in common is the retained AF10 portion, which makes a retained-region contact the more parsimonious reading. That is an inference, not a mapping.
Why it belongs in a phospho beat anyway
Three reasons.
First, it is the reason MLLT10's bounded absence is an absence of sites and not an absence of kinase biology. A reader who queries 'MLLT10 kinase' finds this paper and could reasonably conclude the phospho-regulation question is answered. It is not.
Second, it is the only KMT2A partner for which a kinase inhibitor is already an evidenced therapeutic proposition, and the evidence runs entirely through the kinase's downstream substrates rather than through the fusion.
Third, it creates a clean experiment. If JAK1 is held against the fusion, the simplest thing to ask is whether the fusion is a substrate. An in vitro kinase assay with recombinant JAK1 and the retained AF10 fragment, read out by mass spectrometry, would either open a new site or close the question. AF10 carries 12 S/T-Q motifs and 20 observed phosphosites with no kinase attached to any of them.
| partner a | gene-mllt10 |
|---|---|
| partner b | JAK1 |
| contact type | kinase recruited BY the fusion, not a kinase acting ON the fusion |
| interface region on af10 | not resolved in the accessible record |
| retained in fusion | not determinable - the interaction occurs in both PICALM::MLLT10 and KMT2A::MLLT10, whose common sequence is the retained AF10 portion, but no mapping is published |
| druggability | JAK inhibitors are approved drugs; the target is the kinase, not the interface, and the evidence runs through downstream STAT3 rather than through the fusion |
fusions tested
[ "PICALM::MLLT10 (CALM-AF10)", "KMT2A::MLLT10 (MLL-AF10)" ]
- JAK inhibitors are approved and immediately available, which makes this the most actionable kinase finding for any KMT2A partner - and the published human counter-case is a tofacitinib failure in an MLLT10-fusion leukaemia with activating JAK3 mutations, i.e. in the patient where JAK dependence should have been strongest.
- JAK1 inhibition suppresses normal cytokine signalling across haematopoiesis and immunity. In a paediatric patient receiving myelosuppressive chemotherapy, adding a JAK inhibitor has an infection-risk floor that the mouse models do not measure.
- Because the fusion stabilises JAK1 protein rather than activating it, a catalytic inhibitor leaves the stabilised pool in place. A degrader and an inhibitor would not be interchangeable here, and only the inhibitor has been tested.
- Is AF10, or the KMT2A::MLLT10 fusion, a JAK1 substrate? An in vitro kinase assay with recombinant JAK1 and the retained AF10 fragment read out by MS would settle it.
- Which region of AF10 binds JAK1, and is it retained? Both fusions do it, which points at the retained portion, but nobody has mapped it.
- MLLT10 has 20 observed phosphosites and 12 S/T-Q motifs and not one assigned kinase. Given that a PIKK acts on all three of the other partners, has anyone asked whether AF10 is phosphorylated after DNA damage?
- Does JAK inhibition change the fusion's chromatin occupancy, or only its downstream inflammatory output?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
AF10 fusion proteins activate a JAK/STAT-mediated inflammatory signalling cascade through direct recruitment of JAK1; Flag immunoprecipitation of CALM-AF10 or MLL-AF10 and reciprocal endogenous Jak1 co-IP confirm the interaction, fusion-on cells show higher Jak1 protein without a change in JAK1 mRNA, and genetic Jak1 deletion or pharmacological JAK/STAT inhibition is antioncogenic in mouse and human AF10-fusion AML models.
A JAK/STAT-mediated inflammatory signaling cascade drives oncogenesis in AF10-rearranged AML. · Chen BR, et al. · Blood 2021What this evidence cannot show
Full text read via Europe PMC on 2026-09-19. Inducible mouse models plus human lines; the interaction was first identified by proteomics in the CALM-AF10 dataset and then confirmed for both fusions. No phosphorylation site on AF10 or on either fusion is mapped, and the authors attribute the increase in JAK1 phosphorylation in fusion-on cells to higher total JAK1 rather than to increased activity. The region of AF10 that contacts JAK1 is not resolved in what was read here.
Evidence that argues against this (1)
- A clinical counter-case is on record in this graph's own literature: a DDX3X-MLLT10 T-lymphoblastic leukaemia with activating JAK3 mutations failed to achieve an objective response to tofacitinib. JAK-pathway engagement by an MLLT10 fusion is not the same as JAK-pathway dependence in a patient. Wong J et al. 2020
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
No phosphorylation site on MLLT10 or on any MLLT10 fusion has been mapped: PubMed returns 1 record for (MLLT10[tiab] OR AF10[tiab]) AND (phosphosite[tiab] OR "phosphorylation site"[tiab]), and MLLT10 carries 20 observed phosphosites in the PeptideAtlas re-analysis and 12 S/T-Q motifs, none with an assigned kinase.
PubMed esearch 2026-09-19 and Proteins API proteomics-ptm P55197; motif scan of the P55197 canonical FASTA · EMBL-EBI 2026What this evidence cannot show
A counted absence over a stated symbol set, subject to the usual limits: 'AF10' is a word-like token that also matches an unrelated B-cell line name in older immunology literature, and complex-level phrasing can make a subunit-keyed query miss a relevant paper entirely. See ana-partner-phospho-enumeration for the full query ledger.
Evidence that argues against this (1)
- This project has produced a confident wrong absence before by keying a query to current symbols. MLLT10 publishes as AF10, and the C. elegans orthologue literature uses ZFP-1, which no query here covered; a phosphorylation site mapped in the worm would be invisible to this search. Casado P et al. 2023
the gene entry behind the AF10-fusion kinase contact
AF10 fusions directly recruit JAK1
MLL-AF10 co-IPs endogenous JAK1
a kinase contact with no mapped site, beside two sites with no kinase
the one kinase relation runs the other way
kinase contact with no mapped site