Kathrin Bernt (University of Pennsylvania / Children's Hospital of Philadelphia) — DOT1L and H3K79 methylation in KMT2A-rearranged leukemia
also called Bernt Lab · CHOP Center for Childhood Cancer Research · Penn pediatric oncology
Bernt proved that KMT2A-rearranged leukemia depends on one specific chemical mark placed on DNA-packaging proteins by an enzyme called DOT1L. That discovery produced the first targeted drug tried in this disease, and its clinical disappointment is one of the most instructive failures in the field.
The same thing, for a scientist
Kathrin M. Bernt established the DOT1L dependency of KMT2A-rearranged leukemia while in the Armstrong group at Boston Children's and Dana-Farber, showing that among H3K79me2, H3K4me3, H3K27me3 and H3K36me3, only H3K79me2 is abnormally distributed at MLL-fusion target loci in leukemia stem cells, and that Dot1l inactivation collapses the MLL-translocation gene signature in vivo while leaving global gene expression largely intact. She subsequently ran her own programme through Colorado and is currently at the University of Pennsylvania with the Children's Hospital of Philadelphia Center for Childhood Cancer Research, working on H3K79 methylation patterns and oncogene cooperativity, and writes commentary on the menin-inhibitor field.
What she owns that the Iqbal Lab does not1 deeper
The H3K79 methylation expertise and the models built around it. The 2011 paper is the definitive statement of the DOT1L dependency, and it is unusual in the field for being a selectivity result: only H3K79me2 was abnormal, and Dot1l loss suppressed the translocation signature without globally disrupting transcription 1. That is the kind of result the lab needs for menin, and it has not been produced with the same rigour.
A cautionary dataset the lab should study before committing. DOT1L was the field's first great targeted hope and pinometostat did not deliver clinically. Bernt's later work adds nuance that makes the failure interpretable rather than merely disappointing: specific patterns of H3K79 methylation, not just its presence, govern how oncogenes interact, and the expected cooperation between IDH1/2 mutation and KMT2A alterations failed to appear experimentally despite both converging on H3K79me2 2. In other words the mark is real but the simple dose model of it is wrong.
Current visibility into the menin field from a paediatric vantage point. She writes editorial commentary on menin inhibitors and mutant MEN1 3, which means she reads the resistance literature as it appears and is positioned to see where the paediatric gaps are.
Partnering angle
Bernt is a collaborator, with a specific and unusual value: she is the person best placed to tell the lab why an epigenetic dependency that looks airtight in a model can fail as a drug target. That is worth more to an interface-disruption programme than another positive result.
The concrete joint project is comparative: run the same selectivity logic that established the DOT1L dependency on the menin dependency. Does menin inhibition collapse the fusion signature while leaving global transcription intact, in leukemia stem cells specifically, with the same marker-by-marker discrimination? The DOT1L answer exists; the equivalently rigorous menin answer, in paediatric leukemia stem cells, appears not to.
Overlap with the lab is low. She works on chromatin marks and combination biology, not on protein-protein interface chemistry.
| group | Bernt Laboratory (Kathrin M. Bernt) |
|---|---|
| institution | University of Pennsylvania, Philadelphia PA, USA (2026 affiliation as printed on PubMed); previously Center for Childhood Cancer Research / Division of Pediatric Oncology (CHOP-Penn) and University of Colorado; DOT1L work performed at Boston Children's Hospital and Dana-Farber Cancer Institute |
| focus | H3K79 methylation and DOT1L dependency in KMT2A-rearranged leukaemia; patterns of chromatin marks and oncogene cooperativity; paediatric perspective on menin inhibitors |
| partnering angle | Ask her the question the field avoids: why did the DOT1L dependency, which was as clean a target-validation result as exists, fail clinically, and which of those failure modes apply to menin? Concrete joint project: replicate the DOT1L selectivity logic (mark-by-mark discrimination in leukemia stem cells, signature collapse without global transcriptional disruption) for menin inhibition in a paediatric KMT2A::AFF1 context, where it has not been done with equivalent rigour. |
| relationship | natural collaborator, low overlap, high value as a sceptic |
key papers — the numbers behind the chart
[
{
"title": "MLL-rearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L",
"year": 2011,
"venue": "Cancer Cell",
"pmid": "21741597",
"doi": "10.1016/j.ccr.2011.06.010"
},
{
"title": "Specific patterns of H3K79 methylation influence genetic interaction of oncogenes in AML",
"year": 2020,
"venue": "Blood Advances",
"pmid": "32634241",
"doi": "10.1182/bloodadvances.2020001922"
},
{
"title": "Whack-a-menin: zifto's preclinical activity and mutant MEN1",
"year": 2026,
"venue": "Blood",
"pmid": "42560720",
"doi": "10.1182/blood.2026033836"
},
{
"title": "Targeting epigenetic programs in MLL-rearranged leukemias (review with Armstrong)",
"year": 2011,
"venue": "Hematology Am Soc Hematol Educ Program",
"pmid": "22160057",
"doi": "10.1182/asheducation-2011.1.354"
}
]leverage they hold
[ "H3K79 methylation profiling expertise and models", "institutional history inside the group that validated both DOT1L and menin", "CHOP/Penn paediatric leukaemia sample access" ]
- DOT1L is the field's worked example of a clean, replicated, mechanistically satisfying dependency that produced a clinically ineffective drug; the same can happen to menin and the lab should plan for it.
- Her IDH/KMT2A non-cooperation result shows that mechanistically obvious combination rationales fail more often than the literature admits, which should discipline any combination plan the lab proposes.
- If specific patterns rather than levels of a chromatin mark govern the phenotype, then simple occupancy or displacement readouts for an interface inhibitor may be measuring the wrong thing.
- Has anyone applied the 2011 DOT1L selectivity framework to menin inhibition in leukemia stem cells - mark-by-mark, signature collapse versus global transcription - in a paediatric KMT2A::AFF1 model? If not, that is a well-defined and relatively cheap gap.
- Is the DOT1L clinical failure pharmacological (exposure, schedule, delivery) or biological (redundancy, adaptation)? The answer determines whether menin inhibitors are the same story with better chemistry or a genuinely different one.
- KMT2A::AFF1 B-ALL versus KMT2A::MLLT3 AML: is the DOT1L dependency of equal magnitude? Most of the foundational work is in the AML fusion.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Bernt showed that among H3K79me2, H3K4me3, H3K27me3 and H3K36me3, only H3K79me2 is abnormally distributed at MLL-AF9 fusion target loci in leukemia stem cells, and that Dot1l inactivation downregulates direct MLL-AF9 targets and the MLL-translocation gene signature while leaving global gene expression largely unaffected, with corresponding in-vivo dependence.
MLL-rearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L · Bernt KM, Zhu N, Sinha AU, et al., Armstrong SA · Cancer Cell 2011What this evidence cannot show
A murine MLL-AF9 model, which is an AML rather than an infant B-ALL context; the KMT2A::AFF1 biology the lab cares about is not the same fusion and may not have the same DOT1L dependency profile. Published back-to-back with an industry paper on the DOT1L inhibitor EPZ004777 (Daigle et al., Cancer Cell 2011, PMID 21741596), so the two should be read as a coordinated target-validation package rather than independent confirmation.
Evidence that argues against this (1)
- A joint Oxford-Princess Maxima resistance model showed KMT2A::AFF1B ALL cells can become largely independent of DOT1L-mediated H3K79 methylation while still requiring DOT1L protein, the fusion and HOXA9, which limits how far catalytic H3K79 dependency can be read as an obligate requirement. Schneider P et al. 2023
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Her later work found that the expected cooperation between IDH1/2 mutation and KMT2A alterations, predicted because 2-hydroxyglutarate increases H3K79 methylation, failed to appear experimentally; specific patterns of H3K79 methylation rather than its overall level govern oncogene interaction in AML.
Specific patterns of H3K79 methylation influence genetic interaction of oncogenes in AML · Kingsley MC, ... Bernt KM · Blood Advances 2020What this evidence cannot show
An experimental model result contradicting a clinical correlation, which is the harder direction to trust; the model may simply lack a context present in patients. Affiliation on this paper spans University of Colorado and the Center for Childhood Cancer Research, reflecting a move in progress.
Evidence that argues against this (1)
- A failed prediction is informative and this one is bounded: absence of cooperation between IDH1/2 mutation and KMT2A alterations was established in particular model systems at particular 2-hydroxyglutarate concentrations. IDH mutation and KMT2A rearrangement also rarely co-occur in patients, so the experiment tests a combination the disease does not often present. eLife 2021
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Bernt is currently affiliated with the University of Pennsylvania and writes commentary on menin-inhibitor preclinical activity and mutant MEN1.
Whack-a-menin: zifto's preclinical activity and mutant MEN1 · Bernt KM · Blood 2026What this evidence cannot show
A commentary piece, so it establishes her current affiliation and engagement with the menin field but contributes no primary data. Institutional listing is 'University of Pennsylvania' without department detail; the associated Children's Hospital of Philadelphia appointment is inferred from her earlier Center for Childhood Cancer Research affiliation and was not separately verified.
Evidence that argues against this (1)
- Affiliation and commentary authorship describe where someone works and what they were invited to write about. Commentary is not a research programme, and "currently affiliated" is a fact with a short half-life that this graph cannot keep current. Blood Cells, Molecules and Diseases 2010
established the H3K79 methylation dependency
established the H3K79me2 selectivity of the DOT1L dependency in KMT2A-rearranged leukaemia
the founding evidence that only H3K79me2 is abnormally distributed at MLL-fusion targets in leukemia stem cells
- related to →Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
her dependency result is the target validation behind the DOT1L inhibitor programme and its clinical disappointment
Bernt's DOT1L dependency work was performed in the Armstrong group before she established her own programme