Step 4 — The histone mark: DOT1L and H3K79me2/3 at fusion target genes (and why global expression barely moves)
also called H3K79 methylation · DOT1L recruitment · aberrant H3K79me2 · ectopic H3K79 methylation
The fusion brings along an enzyme called DOT1L that stamps a chemical tag (H3K79 methylation) on the genes it sits on, and that tag keeps them switched on. Remarkably, removing the enzyme barely changes the rest of the cell's genes — only the fusion's own targets collapse.
The same thing, for a scientist
DOT1L, the sole H3K79 methyltransferase, is recruited to KMT2A-fusion-bound loci (most directly via AF9/ENL, which bind DOT1L in a complex distinct from the AF4-P-TEFb SEC) and deposits ectopic H3K79me2/3 across the bodies of fusion target genes. The mark is the field's best mechanistic biomarker: H3K79me2 profiles distinguish KMT2A-rearranged ALL from other ALLs, and Dot1l inactivation downregulates direct fusion targets and the translocation-associated signature while leaving global gene expression largely unaffected. That selectivity is the single most important property of this step — and the reason it produced both the field's cleanest biology and its most disappointing clinical result.
The precise claim, stated the way the authors stated it3 deeper
Two papers carry this step, and both are routinely over-claimed, so the wording matters.
Bernt & Armstrong, 2011. Epigenetic profiling of H3K79me2, H3K4me3, H3K27me3 and H3K36me3 in haematopoietic progenitors and leukemia stem cells found abnormal profiles only for H3K79me2 at MLL-AF9 fusion target loci. Inactivating Dot1l downregulated direct MLL-AF9 targets and the MLL-translocation-associated signature, whereas global gene expression remained largely unaffected, and the in-vivo dependence of MLL-AF9 leukemia on Dot1l tracked that suppression 1.
Krivtsov & Armstrong, 2008. A conditional Mll-AF4 mouse develops B-precursor ALL or AML; the ALL profile significantly overlaps human KMT2A-rearranged ALL; H3K79 methylation profiles correlated with fusion-associated expression; and human KMT2A-rearranged ALLs could be distinguished from other ALLs by their H3K79 profiles, with shRNA suppression of DOT1L inhibiting critical MLL-AF4 target genes 2.
Note the two hedges the authors themselves put in: 'correlated with', and DOT1L suppression by shRNA rather than by inhibitor. Neither paper demonstrates that H3K79me2 deposition is the causal step rather than a co-travelling consequence of elongation.
How selective is 'largely unaffected'? The number, and its status
H3K79me2 is associated with more than 5,000 genes, yet on Dot1l inactivation only a small subset changed expression at p<0.01 — on the order of ~137 genes (≈75 probe sets up, ≈92 down) — and the down-regulated set was highly enriched for direct MLL-AF9 targets. Expression of the majority of H3K79me2-marked genes did not significantly change after loss of the mark 1.
Provenance warning: the qualitative statement ('global gene expression remained largely unaffected') is in the peer-reviewed abstract and is safe. The specific gene counts and the '>5,000' figure come from the full text as extracted in this cycle and have not been re-read directly in the PDF, so they are carried at moderate confidence. Any percentage derived from them (e.g. '~2-3% of marked genes') is our arithmetic, not a published figure, and must not be attributed to the authors.
The deeper implication is a mechanistic puzzle the field has never resolved: if H3K79me2 is required for expression, why do most marked genes not care when it is removed? The most parsimonious reading is that H3K79me2 is permissive and redundant at most loci and rate-limiting only where it is ectopically broad — which is exactly what the spreading and enhancer work predicts.
Where the mark actually matters: spreading regions and KEEs
The selectivity resolves into two concrete, KMT2A::AFF1-specific findings from the Milne group.
Spreading. At a subset of targets, KMT2A::AFF1 binding spreads into the gene body with menin, raising H3K79me2/3 and disrupting H3K36me3. Compared with other H3K79me2/3-marked genes, spreading-gene expression is preferentially downregulated by DOT1L inhibitors — so spreading, not the presence of the mark, predicts DOT1L sensitivity, and it mediates synergy with other targeted drugs 3.
KEEs. DOT1L inhibition defines a distinct class of H3K79me2/3-dependent enhancer elements (KEEs). At KEEs — but not at non-KEE enhancers — H3K79me2/3 is required to maintain chromatin accessibility, histone acetylation and transcription-factor binding, and it maintains enhancer-promoter interactions at a subset of them 4.
So H3K79me2 is not a general 'on' switch. It is a locus-specific architectural requirement at a minority of elements, and those elements are the fusion's own.
DOT1L is recruited by the partner, in a complex separate from SEC
The recruitment route is partner-dependent and, importantly, not via the same assembly that delivers P-TEFb. In reconstituted minimal complexes, DOT1 forms a stable complex with AF9 and does not interact with AF9-AF4-P-TEFb complexes 5. For KMT2A::MLLT3/MLLT1 the fusion partner itself is the DOT1L docking site. For KMT2A::AFF1 the route must be indirect — through endogenous ENL/AF9 recruited into the fusion complex — which is a real and under-appreciated asymmetry between fusion families and a plausible reason DOT1L dependence is not uniform across partners.
The clinical verdict, and what it teaches1 deeper
Preclinically DOT1L looked outstanding. EPZ-5676 (pinometostat) has Ki ≈ 80 pM and is ~37,000-fold selective over other methyltransferases, and continuous IV infusion produced complete tumour regressions sustained beyond the infusion period in a rat KMT2A-rearranged xenograft 6.
In adults it did not deliver. In 51 patients across dose escalation and expansion (54 and 90 mg/m²/day by continuous IV infusion, 28-day cycles), the MTD was not reached, H3K79 methylation was reduced as intended, and there were only 2 complete remissions — both in t(11;19). Single-agent efficacy was explicitly described as modest 7.
Three lessons for the lab, in order of importance:
- Kinetics, not potency, was the problem. A writer inhibitor acts only as fast as the mark dilutes through replication. That is days, and it requires continuous infusion — a pharmacological posture that cannot be sustained clinically. Menin inhibition, which evicts the complex, does not have this problem.
- The one responding genotype was t(11;19) — i.e. KMT2A::MLLT1, whose partner binds DOT1L directly. Small numbers, but exactly the direction the recruitment biology predicts.
- Acquired resistance to pinometostat does not go through the target. In five KMT2A-rearranged lines, treatment-emergent resistance arose without DOT1L mutation; KOPN-8 upregulated ABCB1/P-gp efflux, and NOMO-1's mechanism was efflux-independent and unresolved 8. A drug whose resistance route is transporter-mediated is a combination agent, not a backbone.
Where DOT1L is still the right idea
Combined DOT1L plus menin inhibition resembles outright fusion degradation, whereas either single agent has more modest effects on fusion occupancy and gene expression, and the combination releases the oncoprotein from chromatin globally 9. This is the strongest available argument that DOT1L belongs in a combination rather than being abandoned — and it is also, notably, in tension with the in-vitro finding that DOT1L and SEC assemblies are mutually exclusive 5. Reconciling those two results is a real open problem.
| step | 4 |
|---|---|
| chain | mechanism-kmt2a-fusion-causal-chain |
| enzyme | gene-dot1l |
| mark | H3K79me2/H3K79me3 |
| mark uniqueness | DOT1L is the sole H3K79 methyltransferase in mammals; no known H3K79 demethylase, so removal depends on replicative dilution |
| genes marked by h3k79me2 | >5000 |
| genes changed after dot1l loss p001 | 137 |
| selectivity statement | global gene expression largely unaffected; direct fusion targets and the translocation signature selectively lost |
| recruitment route | direct via AF9/ENL YEATS-containing partner; indirect for AFF1 fusions |
| dependent element class | KEEs (H3K79me2/3-dependent enhancers) and spreading gene bodies |
| drug | drug-pinometostat |
| clinical result | phase 1, n=51 adults, 2 CRs both t(11;19), modest single-agent activity |
| rate limiting | partially — kinetically slow because the mark is diluted rather than erased; best as a combination partner |
| resistance route | ABCB1/P-gp efflux and unresolved non-target mechanisms; not target mutation |
- Kinetic bite-back: with no H3K79 demethylase, the mark is removed only by replicative dilution, so a writer inhibitor is intrinsically slow and needed continuous infusion — the drug lost on pharmacology, not on biology.
- Resistance to DOT1L inhibition runs through ABCB1/P-gp efflux rather than target mutation, which means it would arise against any structurally similar agent and would not be predicted by target sequencing.
- DOT1L is required for normal haematopoiesis and erythroid development in mouse knockouts, so the mark is not leukemia-specific even though its aberrant breadth is.
- If H3K79me2 is permissive rather than instructive at most loci, blocking it may lower fusion-target output without collapsing the program — cytostasis and reversible differentiation rather than eradication.
- Why do the majority of H3K79me2-marked genes not change expression when the mark is removed? Until this is answered, 'H3K79me2 maintains expression' is a locus-specific statement masquerading as a general one.
- Is the t(11;19) pinometostat response signal real? KMT2A::MLLT1 is the partner that binds DOT1L directly, so the biology predicts it. With n=2 this is a hypothesis worth a partner-stratified retrospective analysis of the trial data.
- How do DOT1L-containing and SEC-containing assemblies coexist at the same locus, given they are mutually exclusive in vitro yet DOT1L+menin inhibition behaves like fusion degradation? Sequential ChIP or proximity ligation at HOXA9 would resolve it.
- What is the actual decay half-life of H3K79me2 at fusion targets after fusion removal? No paper reports it, and it is the number that determines whether a writer inhibitor can ever be fast enough.
- Was pinometostat ever tested in infants, where cell-cycle kinetics and fetal chromatin differ? No indexed pediatric report was found.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Abnormal chromatin profiles at MLL-AF9 target loci in leukemia stem cells were found only for H3K79me2 (not H3K4me3, H3K27me3 or H3K36me3); Dot1l inactivation downregulated direct MLL-AF9 targets and the MLL-translocation-associated signature while global gene expression remained largely unaffected, and MLL-AF9 leukemia was Dot1l-dependent in vivo. H3K79me2 marks >5,000 genes but only ~137 genes changed expression at p<0.01 after Dot1l loss.
MLL-rearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L · Bernt KM, Zhu N, Sinha AU, et al. (Armstrong SA, senior author) · Cancer Cell 20(1):66-78 2011What this evidence cannot show
The qualitative selectivity claim is from the abstract and is solid. The specific gene counts and the >5,000-gene figure were extracted from the full text during this cycle and have not been independently re-read in the source PDF — treat those numbers as moderate confidence. Any derived percentage is our arithmetic, not the authors'. Model is conditional Dot1l deletion in mouse MLL-AF9 leukemia; microarray-era expression platform; MLL-AF9, not KMT2A::AFF1.
Evidence that argues against this (1)
- Most H3K79me2-marked genes do not change expression when the mark is removed, which argues the mark is permissive/redundant at the majority of loci rather than instructive. This weakens any claim that H3K79me2 deposition is the causal switch, as opposed to a co-travelling mark of elongation. Godfrey L et al. 2019
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
A conditional Mll-AF4 mouse develops B-precursor ALL or AML whose expression profile overlaps human KMT2A-rearranged ALL; H3K79 methylation profiles correlated with Mll-AF4-associated expression, human KMT2A-rearranged ALLs could be distinguished from other ALLs by their H3K79 profiles, and shRNA suppression of DOT1L inhibited critical MLL-AF4 target genes.
H3K79 methylation profiles define murine and human MLL-AF4 leukemias · Krivtsov AV, Feng Z, Lemieux ME, et al. (Armstrong SA, senior author) · Cancer Cell 14(5):355-68 2008What this evidence cannot show
The authors' own wording is 'correlated with', not 'caused'. DOT1L suppression is by shRNA, not inhibitor. The profiling method is ChIP-chip at 2008 resolution, so the gene-body/enhancer distinctions later made with ChIP-seq are not resolvable here. The mouse is an Mll-AF4 knock-in, which recapitulates the disease imperfectly.
Evidence that argues against this (1)
- A methylation PROFILE defining a leukaemia is a correlation between a mark and a disease class, and the entry uses it as a step toward requirement. The conditional Mll-AF4 mouse develops B-precursor ALL or AML, which is itself informative: the same fusion gives different lineages in mice, so the profile is being defined across a phenotype that is not stable. Cancer Cell 2008
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
At a subset of KMT2A::AFF1 targets, fusion and menin binding spread into the gene body with elevated H3K79me2/3 and disrupted H3K36me3; compared with other H3K79me2/3-marked genes, spreading-gene expression is preferentially downregulated by DOT1L inhibition, and this sensitivity mediates synergy with additional targeted agents.
MLL-AF4 Spreading Identifies Binding Sites that Are Distinct from Super-Enhancers and that Govern Sensitivity to DOT1L Inhibition in Leukemia · Kerry J, Godfrey L, Repapi E, et al. (Milne TA, senior author) · Cell Reports 18(2):482-495 2017What this evidence cannot show
Small number of t(4;11) cell lines; 'spreading' is a threshold-dependent computational definition; the synergy claims are cell-line combination indices, not in-vivo.
Evidence that argues against this (1)
- Spreading with elevated H3K79me2/3 is occupancy measured in the steady state, so it cannot order cause and effect - the mark may follow the elongation it is taken to enable. The degron experiment in this same entry makes the point: removing the fusion changes elongation within 15 minutes and the chromatin landscape only afterwards. Molecular Cell 2022
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
DOT1L inhibition in MLL-AF4 cells defines H3K79me2/3-dependent enhancer elements (KEEs) where the mark is required to maintain chromatin accessibility, histone acetylation and transcription-factor binding, and maintains enhancer-promoter interactions at a subset; non-KEE enhancers are unaffected.
DOT1L inhibition reveals a distinct subset of enhancers dependent on H3K79 methylation · Godfrey L, Crump NT, Thorne R, et al. (Milne TA, senior author) · Nature Communications 10(1):2803 2019What this evidence cannot show
KEE is an operationally defined class from a small number of t(4;11) lines treated with EPZ-5676; the causal chain from mark loss to accessibility loss is inferred from co-kinetics rather than from a separation-of-function mutant of the mark itself.
Evidence that argues against this (1)
- KMT2A::AFF1-bound enhancers whose H3K79me2/3 is required are DEFINED by their response to DOT1L inhibition, which makes the category partly circular: the elements that lose the mark and lose activity are labelled H3K79-dependent. The finding is real; the inference that H3K79me2 maintains those enhancers needs a perturbation that does not also define the set. Nature Communications 2019
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In reconstituted minimal complexes DOT1 forms a stable complex with AF9 and does not detectably interact with AF9-AF4-P-TEFb complexes, indicating that DOT1L-containing and SEC/P-TEFb-containing assemblies are distinct rather than a single super-complex.
Function of leukemogenic mixed lineage leukemia 1 (MLL) fusion proteins through distinct partner protein complexes · Biswas D, Milne TA, Basrur V, Kim J, Elenitoba-Johnson KSJ, Allis CD, Roeder RG · Proceedings of the National Academy of Sciences USA 108(38):15751-6 2011What this evidence cannot show
In-vitro reconstitution with recombinant proteins; an undetected interaction under those conditions may still occur on chromatin with bridging factors. Directly in tension with the DOT1L+menin combination result [E9], which is an unresolved conflict.
Evidence that argues against this (1)
- This item argues AGAINST the unified SEC/DOT1L picture the rest of the entry builds, and it is filed as support. DOT1 forming a stable complex with AF9 while NOT detectably interacting with AF9-AF4-P-TEFb means the two activities may sit in mutually exclusive complexes rather than one machine - and reconstituted minimal complexes can equally miss an interaction that needs chromatin or a bridging factor. PNAS 2011
- E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
EPZ-5676 (pinometostat) inhibits DOT1L with Ki approximately 80 pM and about 37,000-fold selectivity over other methyltransferases; continuous IV infusion produced complete tumour regressions sustained beyond the infusion period in a rat KMT2A-rearranged xenograft.
Potent inhibition of DOT1L as treatment of MLL-fusion leukemia · Daigle SR, Olhava EJ, Therkelsen CA, et al. (Pollock RM, senior author) · Blood 122(6):1017-25 2013What this evidence cannot show
Rat xenograft of a single cell line with continuous infusion — a dosing regimen that does not translate to durable clinical use. Company-sponsored preclinical work.
Evidence that argues against this (1)
- Ki ~80 pM and 37,000-fold selectivity are enzymatic measurements, and DOT1L's problem was never potency. The same compound reduced H3K79 methylation demonstrably in patients and produced 2 complete remissions in 51 - so biochemical excellence and clinical failure coexisted here, which is the reason to state the two numbers together. Blood 2018
- E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In a phase 1 trial of pinometostat in 51 adults with acute leukemia (54 and 90 mg/m2/day continuous IV infusion, 28-day cycles) the MTD was not reached, H3K79 methylation was reduced, and there were only 2 complete remissions, both in t(11;19); single-agent efficacy was modest.
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. (Tallman MS, senior author) · Blood 131(24):2661-2669 2018What this evidence cannot show
Adults only; no PubMed-indexed pediatric pinometostat report was found in this cycle, so the infant/pediatric KMT2A-r question is formally untested in the indexed literature. n=2 responders cannot support a partner-specific efficacy claim, only a hypothesis.
Evidence that argues against this (1)
- The trial's own design limits what its failure proves: the maximum tolerated dose was NOT reached, dosing was continuous intravenous infusion over 28-day cycles, and a differentiation agent may need longer exposure than that to show benefit. Pinometostat's result bounds DOT1L as a single agent on that schedule rather than bounding H3K79 methylation as a target. Blood 2018
- E8moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Treatment-emergent resistance to pinometostat was generated in five KMT2A-rearranged lines (KOPN-8, MOLM-13, MV4-11, NOMO-1, SEM) without DOT1L target mutation; KOPN-8 resistance was mediated by upregulation of the ABCB1/P-glycoprotein efflux pump, and NOMO-1 resistance was efflux-independent with an unresolved mechanism.
Mechanisms of Pinometostat (EPZ-5676) Treatment-Emergent Resistance in MLL-Rearranged Leukemia · Campbell CT, Haladyna JN, Drubin DA, et al. (Bernt KM, Daigle SR, senior authors) · Molecular Cancer Therapeutics 16(8):1669-1679 2017What this evidence cannot show
Cell lines under prolonged in-vitro selection, which strongly favours efflux mechanisms over the biology that dominates in patients; no patient resistance data exist because the drug did not reach efficacy.
Evidence that argues against this (1)
- Treatment-emergent resistance was generated by continuous in vitro exposure in five lines, which selects for what survives culture under a drug rather than for what emerges in a patient. With only 2 clinical responses in 51 patients there was almost no clinical resistance to characterise, so this mechanism has never been checked against the setting it is meant to describe. Molecular Cancer Therapeutics 2017
- E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Combined DOT1L and menin inhibition resembles outright MLL::AF9 degradation and releases the oncoprotein from chromatin globally, whereas either single agent has more modest effects on fusion occupancy and gene expression.
MLL::AF9 degradation induces rapid changes in transcriptional elongation and subsequent loss of an active chromatin landscape · Olsen SN, Godfrey L, Healy JP, et al. (Armstrong SA, senior author) · Molecular Cell 82(6):1140-1155.e11 2022What this evidence cannot show
MLL::AF9 degron in cell lines; 'resembles degradation' is a similarity of transcriptional and occupancy profiles, not an equivalence of therapeutic effect. No KMT2A::AFF1 degron equivalent exists in the indexed literature.
Evidence that argues against this (1)
- 'Resembles outright degradation' is a similarity claim from chromatin and elongation readouts, not a demonstration of equivalent phenotype. Degrading MLL::AF9 also removes menin binding, LEDGF binding, DNA targeting and every other fusion function at once, so a combination reproducing the chromatin consequence need not reproduce the therapeutic one. Molecular Cell 2022
would produce the H3K79me2 decay half-life, currently absent from the literature
gives residue coordinates for the AF10 surface that recruits DOT1L and for the PZP reader that licenses H3K79me2
- ← depends onThe differentiation block — what it actually is, operationally, in KMT2A::AFF1 leukaemia
MLL-AF4 spreading targets carry elevated H3K79me2/3 and are preferentially DOT1L-dependent
Counter-evidence (Alexander 2019, PMID 31124784): live-cell imaging of Sox2 and its control region shows no enhanced enhancer-promoter proximity and no association between transcriptional bursting and proximity, so a fixed-cell contact-frequency increase is not evidence of causal looping
- ← related toThe trans-histone crosstalk H2Bub → DOT1L → H3K79me: modulatory for me1, effectively required for me2/me3
Counter-evidence (Darwanto 2010, PMID 20442396): the only human-leukaemia measurement of H2Bub and H3K79 methylation together - LC-MS/MS MRM in U937 plus or minus AF10 - found them inversely correlated, which is opposition to the licensing chain the graph's 35 H3K79me2 entries rest on
state-resolved granularity for the mark that entry treats as one thing
state-resolved granularity for the step-4 mark entry
state-resolved granularity for the step-4 mark entry
H2BK120ub licenses DOT1L-mediated H3K79 methylation; Rnf20 loss lowers H3K79me at Hoxa9 and Meis1
modulatory for H3K79me1, effectively required for H3K79me2/me3
required to maintain Dot1l-dependent H3K79 methylation at Hoxa9 and Meis1
- ← depends onStep 5 — The output program: HOXA9/HOXA10/MEIS1/PBX3 plus MEF2C, MYB, BCL2, FLT3 and PROM1/CD133
the mark sustains the program
the founding evidence that only H3K79me2 is abnormally distributed at MLL-fusion targets in leukemia stem cells
partner-resolved instance of the general H3K79 mechanism, with a cross-line intensity difference against KMT2A::AFF1
catalytic DOT1L inhibition raises the mark's mid-occupancy share to 1.89-2.01 while lowering it at the fusion's own sites to 0.13-0.15, i.e. the residual mark is relatively enriched in mid-occupancy chromatin
the compartmental signature of catalytic DOT1L inhibition: the mark's mid-occupancy share roughly doubles while its share at the fusion's own sites falls to 0.13-0.15, consistent with turnover-coupled deposition failing fastest where turnover is highest
BAG6 knockdown reduces the DOT1L-H3 interaction and H3K79me2
H3K79me2 at five HOXA promoters in LOUCY, partially lost on fusion knockdown
step 4 - H3K79me2/3; parallel to step 3
bounds the data field h3k79me2_decay_halflife recorded as unreported
- ← related toMLLT10 (AF10): the discarded PZP reader — the one partner where loss is proven to be the oncogenic event
PZP loss releases H3K79 methylation from H3K27-dependent targeting
- ← inhibitsPinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
reduces H3K79 methylation in patient marrow, confirmed pharmacodynamically