KMT2A-rearranged AML is two diseases on the phosphorylation layer, and DOT1L phosphorylation is one of the things that separates them
also called MLLGA MLLGB · Casado phosphoproteomics · KMT2A-r phospho subgroups
When researchers measured which proteins carried phosphate in 74 poor-prognosis leukaemia samples, the KMT2A-rearranged cases fell into two groups with different drug sensitivities. One group had more phosphate on DOT1L, a protein this whole field is trying to drug.
The same thing, for a scientist
Casado et al. profiled 74 poor-prognosis AML patients from two centres at the proteomic, phosphoproteomic and drug-response levels, with transcriptomics on 39. Integration produced **a phosphoproteomic signature defining two biologically distinct groups of KMT2A-rearranged leukaemia, MLLGA and MLLGB**. MLLGA showed **increased DOT1L phosphorylation**, higher HOXA expression, higher CDK1 activity and more phosphorylation of proteins in RNA metabolism, replication and DNA damage; it was **selectively sensitive to 15 compounds including genotoxics, mitotic-kinase inhibitors and IMPDH inhibitors**. Intermediate-risk KMT2A-MLLT3 cases sat mainly in a third group closer to MLLGA. **For this beat the operational point is that the phospho-layer stratifies the disease and the graph does not currently represent that anywhere** - and that DOT1L, the target of pinometostat, carries a phosphorylation state nobody has followed up.
What separates the groups
MLLGA versus MLLGB and versus non-KMT2A-rearranged cases 1:
- increased DOT1L phosphorylation in MLLGA
- higher HOXA gene expression
- higher CDK1 activity
- more phosphorylation of proteins in RNA metabolism, replication and DNA damage
- sensitivity to 15 compounds: genotoxic drugs, mitotic kinase inhibitors, IMPDH inhibitors
- higher IMPDH2 and nucleolar protein expression, correlating with IMPDH-inhibitor response
KMT2A-MLLT3, the intermediate-risk fusion, mostly forms a third group nearer MLLGA than MLLGB - so the split is not simply partner identity.
Why the DOT1L phosphorylation line should not be left where it is
DOT1L is recruited by several KMT2A fusion partners and is the target of pinometostat, a drug this graph already carries. A phosphorylation state on DOT1L that differs between two subgroups of the disease raises questions nobody has followed:
- Which residues? The abstract names none.
- Which kinase? None named.
- Does it change DOT1L activity or its recruitment by the fusion? Not addressed.
- Does phospho-DOT1L status predict pinometostat response? Pinometostat's clinical failure in KMT2A-rearranged leukaemia is one of this field's unexplained results, and a stratifying phospho-mark on the target itself is an obvious hypothesis that appears never to have been tested.
That last question is the most valuable thing in this entry, and it is entirely open.
How to read the sensitivity claim
'Sensitive to 15 compounds relative to other cases' is an index-versus-comparator statement, and this graph has a standing rule that such statements must be split before they are believed - the KMT2A-rearranged arm in most datasets is myeloid-enriched and lineage confounds fusion status. Here the cohort is AML throughout, which removes the lymphoid confound but not the others: two centres, 74 patients, and the subgroup sizes are not stated in the abstract.
So the honest reading is: the phospho-layer splits the disease (strong, and the paper's central claim), and MLLGA is more drug-sensitive in 15 assays (a hypothesis-generating result from a cohort of this size).
| status | run |
|---|---|
| result summary | KMT2A-rearranged AML splits into MLLGA and MLLGB on a phosphoproteomic signature; MLLGA carries increased DOT1L phosphorylation, higher CDK1 activity and selective sensitivity to 15 compounds; KMT2A-MLLT3 mostly forms a third group closer to MLLGA. |
| cohort | 74 poor-prognosis AML patients, two centres (UK and Finland), 39 with transcriptomics |
| key phospho observation | increased DOT1L phosphorylation in MLLGA; residues and kinase not named in the abstract |
layers
[ "proteomics", "phosphoproteomics", "transcriptomics", "ex vivo drug response" ]
- Which DOT1L residues are phosphorylated, and by what? The supplementary tables of this study would answer the first and nothing in the literature answers the second.
- Does phospho-DOT1L status predict pinometostat response? Pinometostat's clinical underperformance in KMT2A-rearranged leukaemia is unexplained and a stratifying mark on the target itself has never been tested.
- Do the MLLGA/MLLGB groups correspond to anything the graph already tracks - breakpoint subregion, partner identity, HOXA status, age?
- Does the same split exist in KMT2A-rearranged ALL, where most infants sit? This cohort is AML only.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Integrative proteomic, phosphoproteomic, transcriptomic and drug-response profiling of 74 poor-prognosis AML patients from two centres identified a phosphoproteomic signature defining two biologically distinct groups of KMT2A-rearranged leukaemia; the MLLGA group showed increased DOT1L phosphorylation, higher HOXA expression, higher CDK1 activity and increased phosphorylation of proteins in RNA metabolism, replication and DNA damage, and was selectively sensitive to 15 compounds including genotoxic drugs and inhibitors of mitotic kinases and IMPDH.
Integrative phosphoproteomics defines two biologically distinct groups of KMT2A rearranged acute myeloid leukaemia with different drug response phenotypes. · Casado P, et al. · Signal Transduct Target Ther 2023What this evidence cannot show
Abstract read 2026-09-19; full text and supplementary tables not retrieved, so the DOT1L phosphosites, the group sizes and the compound list are not recorded here. 74 patients across two centres with 39 transcriptomes; the KMT2A-rearranged subset is a fraction of that and the abstract does not state its size. The drug-response data are ex vivo. 'CDK1 activity' is inferred from substrate phosphorylation signatures, not measured directly.
Evidence that argues against this (2)
- A within-disease molecular split derived and validated in the same cohort is the setting where subgroup structure is most likely to be over-fitted, and the paper's own framing is that these data provide a rationale for developing specific therapies rather than that the split is established. No independent cohort is described in the abstract. Di Mambro A et al. 2023
- A second, independent objection to acting on the drug-sensitivity half: a comparable exercise in T-ALL found that menin-inhibitor sensitivity tracked a single phosphosite on a single transcription factor rather than a multi-compound sensitivity profile, which suggests that broad ex vivo sensitivity panels and mechanistic phospho-markers do not necessarily identify the same patients. Shimamoto K et al. 2026
- ← related toQ25 enumeration: what was searched, found, rejected and could not be read across the five 3' partners
same patient cohort, partner-side sites
phosphoproteomic signature splits the disease into MLLGA and MLLGB
increased DOT1L phosphorylation in MLLGA; residues and kinase unnamed
- related to →KMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)
intermediate-risk KMT2A-MLLT3 forms a third group closer to MLLGA
the disease-level phospho-layer the enumeration sits under