Oncofusion

start / graph / mechanism-dot1l-normal-b-lymphopoiesis

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q3b-dot1l-h3k79

H3K79me2 builds up as B cells mature - so in B-ALL the drug's target is what the rescued compartment runs on

also called DOT1L in B-cell development · H3K79me2 in B lymphopoiesis · Dot1l pre-B expansion

In normal mice, the H3K79 tag accumulates as B cells grow up, and without DOT1L the pre-B stage does not expand properly, the antibody repertoire narrows and a whole class of B cells fails to form. Since infant KMT2A leukaemia is usually a B-cell leukaemia, a DOT1L drug is attacking the same machinery the healthy B cells need.

The same thing, for a scientist

Conditional knockout and deep sequencing across murine B-cell development show that *Dot1l* is upregulated in **pro-B** cells, that H3K79me2 **accumulates during maturation**, and that the genomic distribution of H3K79me2 peaks implicates DOT1L in transcription and cell-cycle control across developmental stages [E1]. Functionally DOT1L is **essential for pre-B-cell expansion**: its absence significantly reduces pre-B numbers and **skews the B-cell receptor repertoire toward proximal VH usage**. DOT1L is also required to establish the marginal-zone B-cell programme, with attenuation of MZ B cells and a bottleneck at the pre-MZ stage correlating with H3K79me2 peaks at MZ-regulatory genes such as *Lfng* and *Dock10*; and *Dot1l* is re-expressed in germinal-centre B cells post-immunisation to deposit H3K79me2 at GC loci [E1]. **This is the bite-back that the myeloid-centred normal-tissue argument misses.** KMT2A-rearranged infant leukaemia is predominantly B-lineage, and the compartment a DOT1L inhibitor must spare is the one in which the mark is actively accumulating. A narrowed BCR repertoire is also an immune-reconstitution toxicity with no obvious biomarker and a long latency.

Go deeper
Why this changes the safety conversation and not the efficacy one

The published human safety data for pinometostat are reassuring and were generated in adults with relapsed/refractory acute leukaemia 2 - a population with baseline cytopenias, prior chemotherapy and no expectation of intact B-cell development. Nothing in that dataset could have detected a defect in pre-B expansion or a skewed BCR repertoire.

The population where this matters is the one the project actually cares about: infants and children with KMT2A-rearranged B-ALL, in whom B lymphopoiesis is active and repertoire formation is ongoing. The paediatric phase 1 of pinometostat (NCT02141828) completed and has never posted results, so the one dataset that could speak to this does not exist in the public record.

Note the asymmetry this creates with menin inhibitors. Both act on the same programme, but a menin inhibitor works by displacing the fusion from chromatin, whereas a DOT1L agent removes a mark that normal developing B cells are actively depositing. Those are different toxicity hypotheses and the graph should not treat 'epigenetic therapy in infant ALL' as one risk class.

The repertoire finding, and why it is the hardest kind of toxicity to see

Skewing toward proximal VH usage 1 is a change in the diversity of the antibody repertoire, not in a cell count. It would not appear in a blood count, would not be flagged as an adverse event, and would only become visible as impaired responses to new antigens months or years later - a latency that outlasts most phase 1 follow-up.

Marginal-zone B cells matter for the same reason: they carry much of the early response to encapsulated bacteria, and their loss is the immunological signature of functional asplenia. A drug that attenuates MZ B-cell generation in a young child is importing an infection risk that a leukaemia trial is not designed to measure.

This is a bite-back that belongs on any DOT1L intervention entry in this graph, and it is recorded here rather than asserted from analogy.

Facts and figures
mechanism classnormal-tissue requirement / bite-back
genegene-dot1l
markmark-h3k79me2
lineageB lymphoid
repertoire effectskewing toward proximal VH usage on DOT1L loss
relevanceKMT2A-rearranged infant leukaemia is predominantly B-lineage, so the host compartment a DOT1L agent must spare is the one actively depositing the mark
detectabilitya narrowed BCR repertoire and attenuated MZ B cells produce no change in blood counts and would not be recorded as adverse events in a phase 1
human data gapthe paediatric phase 1 of pinometostat (NCT02141828) completed and has never reported
stages affected
[
  "pro-B (Dot1l upregulated)",
  "pre-B (expansion requires DOT1L)",
  "marginal zone (programme requires DOT1L)",
  "germinal centre (Dot1l re-expressed)"
]
named mz target genes
[
  "Lfng",
  "Dock10"
]
How this could bite us back
  • Pre-B expansion requires DOT1L, so the drug suppresses the developmental step the patient needs to rebuild a normal B compartment after leukaemia is cleared.
  • Repertoire narrowing and marginal-zone attenuation are invisible to blood counts and to standard adverse-event reporting, so this toxicity could be real and unrecorded in every trial run so far.
  • Marginal-zone B-cell loss carries an infection risk pattern resembling functional asplenia, which in a young child is a long-term harm that a short leukaemia trial cannot detect.
  • Germinal-centre B cells re-express Dot1l, so ongoing DOT1L inhibition would be expected to impair responses to new antigens and to vaccination during and after therapy.
Still open
  1. Does pharmacological DOT1L inhibition at clinically achievable exposure narrow the BCR repertoire in a young animal? Nothing published tests the pharmacological version of the genetic phenotype.
  2. What happened in the paediatric phase 1 of pinometostat, NCT02141828? It completed in 2016, has never reported, and is the only human dataset that could speak to B-developmental toxicity of this target class.
  3. Are the H3K79me2-marked B-developmental loci the same loci a KMT2A::AFF1 fusion marks? If the normal and leukaemic H3K79me2 programmes overlap heavily in B cells, the therapeutic window in B-ALL is narrower than the myeloid data suggest.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In murine B-cell development, Dot1l is upregulated in pro-B cells and H3K79me2 accumulates during maturation, with peak distribution indicating DOT1L regulates transcription and the cell cycle across developmental stages; DOT1L is essential for pre-B-cell expansion, its absence significantly decreasing pre-B cells and skewing the B-cell receptor repertoire to favour proximal VH usage; DOT1L is required to establish the marginal-zone B-cell gene expression programme, with attenuated MZ B cells and a bottleneck at the pre-MZ stage correlating with H3K79me2 peaks at MZ-regulatory genes such as Lfng and Dock10; and Dot1l is re-expressed in germinal-centre B cells post-immunisation to deposit H3K79me2 at GC loci.

    PMID 40972012DOI
    What this evidence cannot show

    Abstract read from the eutils record; full text not read. Mouse conditional knockout - complete genetic loss in the B lineage, not pharmacological inhibition, and the Cre driver and its stage of action are not stated in the abstract. Published 2026, not independently replicated. The correlation between H3K79me2 peaks at Lfng/Dock10 and the MZ phenotype is associative; no rescue experiment is described in the abstract. **Resolved:** the driver is **Mb1-Cre (Cd79a-Cre), acting from the pro-B stage**; the genotypes are `Dot1l f/f Mb1Cre/+` against `Dot1l +/+ Mb1Cre/+`, across pro/pre-B, naive, follicular and germinal-centre B cells. GSE306634. Resolved 2026-09-19 by the Q18b unstated-systems sweep from the study's deposition, not from its text; bounded as a statement about the deposited data rather than about the whole study.

    Evidence that argues against this (2)
    • Genetic deletion of Dot1l in the B lineage is a far more complete perturbation than a catalytic inhibitor achieves. In the adult human trial the same target class was well tolerated at doses whose maximum was never reached, and in mice EPZ-5676 in vivo minimally perturbed erythropoiesis - so the developmental phenotype bounds the worst case rather than predicting the clinical one. Malcom CA et al. 2021
    • Deletion from the pro-B stage cannot speak to DOT1L's requirement before that point, which is where the cell of origin for infant KMT2A-rearranged ALL is thought to sit. A normal B-lymphopoiesis result conditioned on Mb1-Cre leaves the fetal and pre-pro-B window untested. NCBI Gene Expression Omnibus 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Pinometostat was studied in 51 adults with acute leukaemia, reduced H3K79 methylation and produced 2 complete remissions with modest overall clinical activity.

    PMID 29724899DOI
    What this evidence cannot show

    Abstract read from the eutils record; full text not read. Cited here for the population and the absence of any B-developmental readout. An adult relapsed/refractory population cannot report on pre-B expansion or repertoire formation, so this is an argument about what the trial could not have seen rather than about what it saw.

    Evidence that argues against this (1)
    • A paediatric phase 1 of pinometostat (NCT02141828) was conducted and completed, so a dataset in the population where this toxicity would appear does exist; it has never posted results or produced an indexed primary publication, which is an absence of reporting rather than an absence of data. ClinicalTrials.gov 2016
Connected to