Patricia Ernst (University of Colorado) — what normal KMT2A and menin actually do, which is the therapeutic-window question
also called Ernst Lab · University of Colorado Department of Pediatrics · MLL wild-type function
Ernst studies what the normal, unbroken versions of these genes do in healthy blood. That matters because a drug that blocks the fusion also hits the normal machinery, and her work is the main reason to think menin can be blocked without wrecking blood production.
The same thing, for a scientist
Patricia Ernst, in the Department of Pediatrics and the Pharmacology and Cancer Biology Program at the University of Colorado, works on the physiological roles of wild-type MLL1/KMT2A and menin in haematopoiesis. Her group showed that the MLL-dependent transcriptional network in haematopoietic stem cells extends well beyond the classical HOX targets, and that menin and MLL1 are genetically separable in normal haematopoiesis with no genetic interaction in steady-state or regenerative haematopoiesis or in B-cell differentiation. She now co-authors the field's mechanism-to-drug synthesis with Akihiko Yokoyama.
Why this is the therapeutic-window evidence base1 deeper
The obvious objection to menin inhibition is that wild-type KMT2A is essential for blood stem cells, so removing an essential KMT2A cofactor should destroy haematopoiesis. The reason the field does not believe that objection is largely Ernst's genetics: menin and MLL1 are separable in normal blood, with no genetic interaction in steady-state or regenerative haematopoiesis or in B-cell differentiation, and only a small shared target set 1. That result is already cited as E2 in this graph's gene-men1 dossier, so it is load-bearing for the lab's central hypothesis.
The complementary result cuts the other way and is equally important: the Mll-dependent transcriptional network in haematopoietic stem cells is much broader than HOX genes, comprising many characterised self-renewal regulators 2. So the normal function that a KMT2A-directed drug could disturb is wider than the field's HOXA9/MEIS1 shorthand implies.
Taken together: menin inhibition looks safe because it spares most of MLL1's normal network, not because that network is small.
Partnering angle
Ernst is the right collaborator for the bite-back question, which META_KING treats as a first-class deliverable. For any new obligate-partner interface the lab proposes to disrupt, the question 'what does this interaction do in normal haematopoiesis, and is it separable from the fusion's use of it?' is exactly the question her laboratory is built to answer, using conditional mouse genetics and competitive transplantation.
The menin case was answered because someone did that experiment. For every other candidate partner in the KMT2A::AFF1 interactome, it has not been. That is a well-defined, fundable joint programme and it directly serves the lab's differentiating claim (that the right target is dispensable in normal cells and indispensable to the fusion).
She is a collaborator, not a competitor: her programme is normal-function genetics, not drug discovery or interface chemistry. She is also now co-writing the field's synthesis with Yokoyama 3, which makes her a route into that line as well.
| group | Ernst Laboratory (Patricia Ernst) |
|---|---|
| institution | Department of Pediatrics and Pharmacology and Cancer Biology Program, University of Colorado, Aurora CO, USA (2025 affiliation as printed on PubMed); earlier work at the Geisel School of Medicine at Dartmouth |
| focus | Physiological function of wild-type MLL1/KMT2A and menin in haematopoietic stem cells and B-cell development; genetic separability of menin and MLL1; the normal-function basis of the therapeutic window |
| partnering angle | The right partner for the bite-back deliverable. For every new obligate-partner interface the lab nominates, her laboratory can run the experiment that made menin credible: conditional genetics plus competitive transplantation to establish whether the interaction is separable from normal haematopoietic function. That experiment has been done for menin and for essentially no other KMT2A::AFF1 partner. |
| relationship | natural collaborator, zero overlap, directly serves the lab's central claim |
key papers — the numbers behind the chart
[
{
"title": "Distinct pathways regulated by menin and by MLL1 in hematopoietic stem cells and developing B cells",
"year": 2013,
"venue": "Blood",
"pmid": "23908472",
"doi": "10.1182/blood-2013-03-486647"
},
{
"title": "An MLL-dependent network sustains hematopoiesis",
"year": 2013,
"venue": "PNAS",
"pmid": "23744037",
"doi": "10.1073/pnas.1301278110"
},
{
"title": "KMT2A-rearranged leukemia: from mechanism to drug development",
"year": 2025,
"venue": "Experimental Hematology",
"pmid": "40921261",
"doi": "10.1016/j.exphem.2025.105247"
}
]leverage they hold
[ "conditional mouse genetics and competitive transplantation for normal haematopoietic function", "the wild-type MLL1 and menin normal-function evidence base the therapeutic window rests on", "co-authorship axis with Yokoyama" ]
- Separability shown by complete genetic deletion in adult mouse marrow may not hold under partial pharmacological inhibition in a fetal-origin human progenitor, which is the actual clinical setting.
- Menin is still required for long-term HSC function under transplantation stress, so haematopoietic recovery after intensive infant therapy may be impaired in a way that is attributed to the chemotherapy instead.
- If the MLL-dependent normal network is broad, then any drug that perturbs KMT2A-complex assembly, not just menin, has a wider normal-function footprint than the HOXA9/MEIS1 framing suggests.
- The menin-MLL1 separability experiment has been done. Has it been done for PSIP1/LEDGF, AFF1, DOT1L, CDK9, WDR5, KAT7 or CBX8 in the same rigorous way - conditional deletion plus competitive transplantation? If not, every other 'druggable partner interface' in the graph lacks the evidence that made menin credible.
- Does menin-MLL1 separability hold in FETAL haematopoiesis, which is where the lab's disease originates, or only in adult marrow? The published work is adult mouse.
- Given that the MLL-dependent HSC network is much broader than HOX, which parts of it does a menin inhibitor actually leave alone in a treated child over months of exposure?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Ernst's group showed menin and MLL1 are genetically separable in normal haematopoiesis, with no genetic interaction in steady-state or regenerative haematopoiesis or B-cell differentiation, and only a small set of co-regulated genes (HOXA9, MEIS1, MECOM, EYA1).
Distinct pathways regulated by menin and by MLL1 in hematopoietic stem cells and developing B cells · Li BE, Gan T, Meyerson M, Rabbitts TH, Ernst P · Blood 2013What this evidence cannot show
Mouse genetics with complete conditional deletion, which models a drug only loosely: partial reversible pharmacological occupancy is a different perturbation, and species differences in haematopoietic tolerance are real. The reassuring result may not transfer to years-long dosing in a growing child.
Evidence that argues against this (1)
- Conditional menin deletion in mouse haematopoiesis preserves steady-state progenitor numbers but reveals a marked long-term HSC defect under competitive transplantation, so menin is not simply dispensable in normal blood. Maillard I et al. 2009
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The Mll-dependent transcriptional network in haematopoietic stem cells extends well beyond clustered homeobox genes and comprises many characterised regulators of self-renewal, mapped by conditional loss-of-function, genome-wide expression, ChIP and functional rescue.
An MLL-dependent network sustains hematopoiesis · Artinger EL, ... Ernst P · Proceedings of the National Academy of Sciences USA 2013What this evidence cannot show
Murine haematopoietic stem cells; the human and specifically the fetal human network may differ, which matters because the lab's disease originates in fetal haematopoiesis. Affiliation at the time was the Geisel School of Medicine at Dartmouth, before the move to Colorado.
Evidence that argues against this (1)
- The network was mapped by conditional loss-of-function in MOUSE haematopoietic stem cells, so it describes what wild-type Mll does in normal murine HSCs. Extending it to the fusion requires that the retained KMT2A portion behaves the same way, which the neomorph-versus-haploinsufficiency debate in this graph leaves open. Oncofusion knowledge graph (internal analysis) 2026
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review
Ernst, now in the Department of Pediatrics and the Pharmacology and Cancer Biology Program at the University of Colorado, co-authored the field's current mechanism-to-drug-development synthesis with Akihiko Yokoyama.
KMT2A-rearranged leukemia: from mechanism to drug development · Ernst P, ... Yokoyama A · Experimental Hematology 2025What this evidence cannot show
A review; establishes current affiliation and the Ernst-Yokoyama axis but contributes no primary data.
Evidence that argues against this (1)
- Co-authoring a synthesis places someone at the centre of how a field describes itself and says nothing about current experimental capacity, group size or willingness to collaborate. Departmental affiliation is also a fact with a short half-life. Blood Cells, Molecules and Diseases 2010
- ← proposed byKMT2A / MLL1 — the gene that breaks
what normal KMT2A and menin actually do, which is the safety argument
showed menin and MLL1 are genetically separable in normal haematopoiesis, the basis of the therapeutic window
- related to →KMT2A / MLL1 — the gene that breaks
mapped the MLL-dependent transcriptional network sustaining normal haematopoietic stem cells beyond HOX genes
co-author with Yokoyama on the field's current mechanism-to-drug synthesis
- supports →Analysis: cross-validating the single-group non-menin KMT2A-fusion dependencies in one framework
the menin-MLL1 separability experiment is the template; no other candidate partner interface has had it done
one of the groups positioned to say whether a replication was attempted