DOT1L and H3K79 in KMT2A::MLLT3: recruitment is a graded dial that sets methylation and transformation - and the fusion it was discovered in produced no remission in the DOT1L trial
also called MLL-AF9 DOT1L · H3K79 MLL-AF9
The amount of DOT1L this fusion pulls in is not on or off: weaken the grip a little and you get a little less chemical mark and a little less cancer. That graded behaviour is exactly the shape that makes a drug hard, and the one DOT1L drug tested in patients helped nobody with this particular fusion.
The same thing, for a scientist
The DOT1L axis in KMT2A-rearranged leukaemia was discovered in, and is best characterised for, KMT2A::MLLT3. The retained AF9 AHD binds DOT1L directly at three motifs, and structure-guided AF9 point mutants give **graded** rather than binary loss of recruitment. ChIP-seq of H3K79me2 and H3K79me3 across that gradient shows **differential loss of the two marks at MLL-AF9 target genes**, and serial colony formation shows a **gradient of transforming capacity** tracking recruitment [E1]. So for this fusion the relationship is dose-like: recruitment sets methylation sets transformation. That is a quantitatively different claim from the KMT2A::MLLT1 position this graph records, where the interesting finding is that ENL's DOT1L contact and its AF4-family contact are **mutually exclusive**, making the ENL/DOT1L species a separate entity from the elongation complex. Both partners contact DOT1L through homologous AHDs with similar affinities [E1], so the partner difference is not in DOT1L affinity. Clinically, the only trial of a DOT1L inhibitor in this disease - pinometostat in 51 adults - produced **two complete remissions, both in t(11;19) patients**, and the report does not state how many t(9;11) patients were enrolled [E2]. The fusion that generated the entire DOT1L hypothesis therefore has **no reported complete remission on a DOT1L inhibitor, and no published denominator** against which to judge that.
The gradient result, which is the partner-specific mechanism
Kuntimaddi et al. built a series of AF9 and DOT1L point mutants of graded severity and asked what each did 1:
- Recruitment: AF9 D546R abolishes binding to both high-affinity DOT1L sites; D544R hits Site 3 hard and Site 2 lightly. So the mutants span a range rather than switching the contact off.
- Chromatin: ChIP-seq of H3K79me2 and H3K79me3 across that range shows differential loss of me2 and me3 at MLL-AF9 target genes - the two methylation states do not fall together.
- Phenotype: serial replating of murine bone marrow progenitors transduced with wild-type MLL-AF9, MLL-AF9 (D544R) and MLL-AF9 (D546R) shows colony-forming ability 'between' the arms, i.e. a gradient.
The authors connect these with a mechanistic argument worth recording: DOT1L is a distributive, non-processive enzyme, so at most one methyl group is added per enzyme-substrate encounter. Residence time therefore converts directly into methylation state, and me3 - requiring three encounters - is more sensitive to reduced recruitment than me2. That is a mechanism for why a partial inhibitor gives a partial mark rather than a cliff.
How this differs from the MLLT1 route, and how it does not
This graph records for KMT2A::MLLT1 that the ENL AHD contacts DOT1L directly and that this contact is mutually exclusive with ENL's AF4-family contact, so the ENL/DOT1L species is a distinct assembly from the AF4/ENL/P-TEFb elongation complex.
What this beat can add is that the affinity is not what separates the two partners: 'we tested binding of each of these DOT1L peptides with ENL which yielded similar binding affinities for ENL as seen for AF9' 1. Both AHDs bind DOT1L comparably.
So if MLL-AF9 and MLL-ENL differ in H3K79 output - and this graph records ChIP-qPCR showing higher relative H3K79me2 intensity at MLL-ENL than MLL-AF4 target loci - the difference is not a difference in the DOT1L contact itself. The remaining candidates are the other things the two fusions do not share: the ENL YEATS domain, which the AF9 fusion does not have, and the different lineage contexts the two fusions occupy.
The clinical negative, stated with its denominator missing
Pinometostat, a selective DOT1L inhibitor, was given to 51 adults with MLL-rearranged leukaemia. Two complete remissions, both t(11;19), one described by the authors as 'DOT1L-ELL' and one as 'DOT1L-ENL' - a labelling error for MLL-ELL and MLL-ENL that this graph already records 2. Efficacy overall was modest, the maximum tolerated dose was not reached, and H3K79 methylation did fall, so the pharmacodynamics worked 2.
The MLLT3-specific point is what the paper does not report. NGS 'enabled full identification of the specific fusion partners in all t(11:19) patients' 2 - the t(11;19) patients are partner-resolved and the rest are not. The words 'MLLT3', 'AF9' (outside the reference list) and 't(9;11)' appear nowhere in the report's own text. The number of t(9;11) patients exposed to pinometostat is not published.
So the honest statement is not 'DOT1L inhibition failed in MLL-AF9'. It is: the fusion in which the DOT1L dependency was discovered has no published response data in the only clinical trial of a DOT1L inhibitor in this disease. The authors themselves raise the possibility that different MLL fusion proteins have differential sensitivity to DOT1L inhibition and note the CRs were both t(11;19) 2; they could not test it, because the partner assignment was incomplete.
| fusion | fusion-kmt2a-mllt3 |
|---|---|
| partner | MLLT3 |
| dot1l contact | direct, via the retained AF9 AHD, at three DOT1L motifs |
| contact geometry | graded: AF9 D546R abolishes both high-affinity sites, D544R affects Site 3 more than Site 2 |
| enzyme kinetics argument | DOT1L is distributive/non-processive, so residence time maps onto methylation state and me3 is more recruitment-sensitive than me2 |
| paralogue affinity difference | none detected - ENL binds the same DOT1L peptides with similar affinity |
| trial | NCT01684150 |
| trial n | 51 |
| complete remissions | 2 |
| t 9 11 patients in trial note | not reported anywhere in PMID 29724899; the strings MLLT3, t(9;11) and AF9 (outside the reference list) do not appear in the report text |
| mllt3 complete remissions on dot1l inhibitor | 0 |
| mllt3 complete remissions denominator known | false |
readouts across the gradient
[ "H3K79me2 and H3K79me3 ChIP-seq at MLL-AF9 targets, differentially lost", "serial colony formation, graded" ]
cr partners
[ "KMT2A::ELL", "KMT2A::MLLT1" ]
- A graded recruitment-to-phenotype relationship predicts a shallow, partial clinical response to partial target occupancy, which is what pinometostat produced; deepening it requires occupancy levels that also strip H3K79 from the normal genome.
- DOT1L inhibition in a KMT2A::MLLT3 cell acts on the fusion, on wild-type AF9 and on wild-type ENL simultaneously, and ENL is required for AML maintenance - so part of any anti-leukaemic effect may not be fusion-directed at all.
- Because the MLL-AF9 mouse is the field's default assay and MLL-AF9 is the fusion with no published DOT1L trial data, the preclinical evidence base and the clinical evidence base for DOT1L inhibition are built on different fusions.
- How many t(9;11) patients received pinometostat, and what happened to them? The sponsor holds the partner assignments for the non-t(11;19) patients and has never published them.
- Does the graded recruitment-to-transformation relationship hold in vivo, in transplantation rather than serial replating? Every point on the published gradient is a replating number.
- If AHD affinity for DOT1L is the same for AF9 and ENL, what accounts for the reported difference in H3K79me2 intensity between MLL-ENL and MLL-AF4 target loci - the missing YEATS domain, the lineage, or the comparator being AF4 rather than AF9?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Structure-guided point mutations give graded effects on recruitment of DOT1L to MLL-AF9; graded reduction of the DOT1L interaction results in differential loss of H3K79me2 and H3K79me3 at MLL-AF9 target genes by ChIP-seq; the degree of DOT1L recruitment is linked to the level of MLL-AF9 haematopoietic transformation in serial colony formation; and ENL binds the same DOT1L peptides with affinities similar to AF9.
Degree of recruitment of DOT1L to MLL-AF9 defines level of H3K79 Di- and tri-methylation on target genes and transformation potential. · Kuntimaddi A, Achille NJ, Thorpe J, et al. · Cell Rep 2015What this evidence cannot show
Retroviral MLL-AF9 in murine bone marrow progenitors, plus in vitro peptide binding - a species, lineage and expression-level import. The transformation readout is serial replating, not transplantation. The authors note that DOT1L mutations affect binding to wild-type AF9 and ENL as well as to MLL-AF9, so phenotypes in the DOT1L-mutant arms are the sum of effects on three targets.
Evidence that argues against this (1)
- In the same figure series, mutation of DOT1L sites 1+2 produced significantly HIGHER colony numbers than wild-type DOT1L add-back, which is not what a simple monotonic recruitment-to-transformation relationship predicts and which the authors attribute to effects on wild-type AF9 and ENL rather than on the fusion. Kuntimaddi A et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the 51-patient adult phase 1 of the DOT1L inhibitor pinometostat, two patients achieved complete remission and both had t(11;19); the report states that next-generation sequencing enabled full identification of the specific fusion partners in all t(11;19) patients, and it does not report the number of t(9;11) or MLLT3-rearranged patients enrolled or their responses.
The DOT1L inhibitor pinometostat reduces H3K79 methylation and has modest clinical activity in adult acute leukemia. · Stein EM, Garcia-Manero G, Rizzieri DA, et al. · Blood 2018What this evidence cannot show
Read in full text from the PMC HTML page (PMC6265654, 51,473 characters of stripped text) on 2026-09-20. This is a dose-escalation phase 1 with no efficacy power and no partner-stratified prespecification; the absence of t(9;11) reporting is an absence of reporting, not evidence that no t(9;11) patients were enrolled or that none responded. The paper also mislabels the two responders' fusions as 'DOT1L-ELL' and 'DOT1L-ENL'.
Evidence that argues against this (1)
- The authors explicitly raise the alternative explanation for modest activity: that the schedule may have been inadequate rather than the target wrong, since the maximum tolerated dose was not reached and a continuous infusion schedule was used. A negative partner signal from an underdosed trial is not a partner result. Stein EM et al. 2018
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
ENL, but not its paralogue AF9, is required for disease maintenance in acute myeloid leukaemia, so the wild-type requirement runs opposite to the fusion architecture: the partner whose fusion discards its reader is the partner whose wild-type protein is dispensable in these models.
ENL links histone acetylation to oncogenic gene expression in acute myeloid leukaemia. · Wan L, Wen H, Li Y, et al. · Nature 2017What this evidence cannot show
CRISPR depletion in AML lines and mouse models; a negative for AF9 is a statement about those lines and guide sets. Several of the lines used are themselves KMT2A::MLLT3, in which an AF9 knockout also removes one AHD-bearing species that is not the fusion.
Evidence that argues against this (1)
- MLLT3 is required for human haematopoietic stem-cell self-renewal and engraftment, so AF9 dispensability is specific to established leukaemic maintenance and does not generalise to normal or regenerating haematopoiesis. Calvanese V et al. 2019
the partner-specific epigenetic route
graded recruitment via the retained AHD
me2 and me3 respond differentially to reduced recruitment
same contact, same affinity, different mutual-exclusivity result
the structural basis of the gradient
- related to →Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
no published t(9;11) denominator in the only DOT1L trial