The phosphatase side: SET binds both wild-type KMT2A and the fusion, and inhibiting it changes four kinases at once
also called SET PP2A KMT2A · FTY720 KMT2A-r · phosphatase arm
Phosphates get added by kinases and removed by phosphatases. A protein called SET blocks the main phosphatase PP2A, and SET physically sits on both the normal and the fused KMT2A protein. Blocking SET with a drug stops leukaemia cells growing and changes the activity of several kinases at once.
The same thing, for a scientist
Kinase-centred accounts of KMT2A-rearranged leukaemia have a matching blind spot: the phosphatases. Di Mambro et al. report that **SET, the endogenous inhibitor of PP2A, interacts with both KMT2A wild-type and fusion proteins and is recruited to the HOXA10 promoter**; SET is overexpressed in AML, correlates with poor prognosis and with MEIS/HOXA expression, and SET silencing abolishes clonogenic ability of KMT2A-rearranged cells and HOXA9/HOXA10 transcription. **FTY720 disrupts the SET-PP2A interaction**, causing cell-cycle arrest and increased chemosensitivity. Phosphoproteomics after FTY720 showed **reduced activity of PP2A-regulated kinases ERK1, GSK3-beta, Aurora B and PLK1**, with MYC suppression. **A SET contact on the fusion is a non-menin protein contact, it is phosphorylation-relevant by definition, and its position relative to the breakpoint is not reported** - which is the single most consequential gap in this entry.
Why a phosphatase inhibitor belongs in a phosphorylation beat
Every phospho-gated contact in this beat has an implicit second determinant: how fast the phosphate comes off. A site with a named kinase and an unnamed phosphatase is half-characterised.
SET is the point where that becomes concrete in this disease. It is a PP2A inhibitor; it is overexpressed in AML; it binds the fusion; and a drug that displaces it is already in clinical use for another indication. The phosphoproteomic readout after FTY720 - four kinases moving together plus MYC suppression - is what a phosphatase release looks like, and it is not a clean single-target result 1.
The consequence for the rest of this beat: if PP2A is held off by SET at KMT2A-fusion-occupied loci, then phospho-marks written there - including, in principle, the CK2 marks on the retained fragment - would persist longer than elsewhere in the nucleus. That is a testable prediction nobody has made, and it would explain why a constitutive kinase like CK2 produces locus-specific consequences.
What is not known, and it is the important part
Where on KMT2A does SET bind? The abstract says SET interacts with both wild-type and fusion proteins. If the contact were mapped, and if it lay N-terminal to the breakpoint, it would be another non-menin contact on the retained fragment. The paper does not say, and the full text was not read in this session.
That makes SET the highest-value unmapped contact this beat found: binding to the fusion is already demonstrated, so the only missing step is a truncation series, which is a week of work.
Also unknown: whether the SET-fusion interaction is direct or through the complex; whether it is phosphorylation-dependent; and whether the FTY720 phenotype runs through SET at all, since FTY720 is a sphingosine analogue with a well-documented S1P-receptor pharmacology of its own.
| phosphatase | PP2A |
|---|---|
| inhibitor protein | SET |
| drug | FTY720 (fingolimod) |
| contact | SET interacts with both KMT2A wild-type and KMT2A fusion proteins; recruited to the HOXA10 promoter |
| interface mapped | false |
| fate in fusion | unknown - the interaction with the fusion is demonstrated but the binding region on KMT2A is not reported |
| tested on fusion | Yes for the interaction; no for a mapped interface |
kinases affected
[ "ERK1", "GSK3B", "Aurora B", "PLK1" ]
- FTY720 is immunosuppressive at therapeutic doses through S1P receptor modulation; in a leukaemia patient with an already compromised immune system that is a serious liability independent of any anti-leukaemic effect.
- Reactivating PP2A broadly reverses phosphorylation everywhere, including the CK2 marks this beat argues are load-bearing for the fusion's LEDGF grip and the marks that hold normal progenitors in their state. Direction of benefit is not predictable from the target alone.
- Four kinases moving together after one drug is a systems response, not a mechanism; attributing the anti-leukaemic effect to any one of them would be unsupported.
- Where on KMT2A does SET bind? A truncation series would say whether this is another contact on the retained fragment, and the interaction with the fusion is already demonstrated so the experiment is short.
- Is PP2A activity locally suppressed at KMT2A-fusion-occupied loci? If so, phospho-marks written there persist longer, which would explain locus-specific consequences of a constitutive kinase like CK2.
- Does FTY720's effect in these cells survive SET knockdown? Without that control the phenotype cannot be attributed to the SET-PP2A axis.
- Which phosphatase removes the CK2 marks at S136/S142? Nobody has asked, and it determines whether a CK2 inhibitor produces a fast or a slow effect on the LEDGF grip.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
SET, the endogenous inhibitor of PP2A, interacts with both KMT2A wild-type and fusion proteins and is recruited to the HOXA10 promoter; SET is overexpressed in AML with elevated expression correlating with poor prognosis and with MEIS and HOXA expression; SET silencing abolishes clonogenic ability of KMT2A-rearranged cells and HOXA9/HOXA10 transcription; pharmacological SET inhibition by FTY720 disrupts SET-PP2A, causes cell-cycle arrest and increases chemosensitivity, and phosphoproteomics after FTY720 shows reduced activity of ERK1, GSK3-beta, Aurora B and PLK1 with MYC suppression.
SET-PP2A complex as a new therapeutic target in KMT2A (MLL) rearranged AML. · Di Mambro A, et al. · Oncogene 2023What this evidence cannot show
Abstract read 2026-09-19; full text not retrieved. The SET-KMT2A interaction is asserted without a mapped interface, without a stated assay format in the abstract, and without a retained-versus-lost call. FTY720 (fingolimod) is a sphingosine-1-phosphate receptor modulator with substantial pharmacology unrelated to SET, so its phenotype is not evidence that the SET-PP2A axis is the mediator. The kinase-activity inferences are from phosphoproteomic substrate signatures, not direct kinase assays.
Evidence that argues against this (2)
- A phosphoproteomic study of KMT2A-rearranged AML in 74 patients found that the phosphorylation signature splits the disease into two biologically distinct groups with different drug-response phenotypes, one of them marked by elevated CDK1 activity. A single-mechanism phosphatase account of the disease has to explain why the phospho-landscape is not one landscape, and the SET study does not address that heterogeneity. Casado P et al. 2023
- PP2A is a tumour suppressor whose reactivation is proposed as a therapy across many cancers, so a PP2A-reactivating phenotype in KMT2A-rearranged cells may reflect a general dependency of transformed cells rather than anything specific to the fusion; the study's own comparator is KMT2A-rearranged versus other AML rather than versus non-malignant progenitors. Lei Y et al. 2026
SET interacts with both wild-type KMT2A and the fusion; interface unmapped
SET overexpression correlates with poor prognosis and HOXA/MEIS expression
the phosphatase arm that would set the lifetime of the CK2 marks
GSK3-beta is one of four kinases moved by PP2A release in this disease