Oncofusion

start / graph / target-lsd1-kdm1a-kmt2a-r

Targetmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6b-therapeutics2

LSD1/KDM1A inhibition — a differentiation therapy with real clinical pharmacodynamics and no randomised proof

also called KDM1A · iadademstat · ORY-1001 · bomedemstat · tranylcypromine · lysine-specific demethylase 1

LSD1 is an enzyme that keeps leukemia cells locked in an immature state. Blocking it makes the cells grow up and stop dividing, and in patients it visibly forces the blasts to differentiate, especially in KMT2A-rearranged disease.

The same thing, for a scientist

LSD1/KDM1A demethylates H3K4me1/2 and acts as a scaffold in CoREST and GFI1-containing repressive complexes that enforce the stem-like state in KMT2A-rearranged leukemia. Iadademstat (ORY-1001) inhibits both the enzymatic and the scaffolding function. In the first-in-human phase 1, 27 patients with relapsed/refractory AML were treated; the extension cohort of 14 was deliberately enriched for MLL/KMT2A-rearranged AML, and reductions in blasts with induction of blast differentiation were observed, particularly in patients with MLL translocations. One CR with incomplete count recovery was reported. In combination with azacitidine in newly diagnosed unfit AML (ALICE, phase 2a), the recommended dose was established. There is no randomised evidence and no paediatric data.

Go deeper
What was actually observed in patients

Phase 1 monotherapy (dose escalation plus extension) 1:

  • 27 patients with relapsed/refractory AML, iadademstat 5-220 µg/m²/day, days 1-5 weekly in 28-day cycles.
  • Recommended single-agent dose 140 µg/m²/day; the 14-patient extension cohort was enriched for MLL/KMT2A-rearranged AML.
  • Dose-dependent plasma exposure and potent time- and exposure-dependent induction of differentiation biomarkers — i.e. confirmed pharmacodynamics in humans, not just in cells.
  • Reductions in blood and bone marrow blast percentages with induction of blast differentiation, in particular in patients with MLL translocations.
  • Adverse events were as expected for R/R AML: myelosuppression, infections, asthenia, mucositis, diarrhoea.
  • One complete remission with incomplete count recovery is the extent of the formal response data in the abstract.

ALICE (phase 2a, EudraCT 2018-000482-36) combined iadademstat with azacitidine in newly diagnosed AML unfit for intensive chemotherapy, across six Spanish hospitals, establishing a recommended phase 2 dose 2. The trial is completed. We do not state its response rate here because we did not verify it from the full text.

Why LSD1 is a better-shaped target than BET or CDK9

LSD1's role in KMT2A-rearranged leukemia is to enforce a state, not to run a housekeeping process. Inhibiting it produces differentiation rather than cytotoxicity, which means:

  • the therapeutic index argument is about state-specific dependency rather than proliferation, so it is not automatically defeated by essentiality in normal tissue;
  • the toxicity profile is dominated by on-mechanism differentiation, including the thrombocytopenia that follows from LSD1's role in megakaryopoiesis, rather than by broad cytotoxicity;
  • it is mechanistically parallel to menin inhibition. Both are differentiation agents acting on the same locked stem program from different directions. That makes them either a rational combination or a redundant one, and nobody has published which.

The honest summary for the lab: iadademstat has the cleanest demonstration of in-human, KMT2A-enriched, on-target differentiation pharmacodynamics of any non-menin agent in this beat. It also has, after fifteen years, no randomised efficacy data and no paediatric development. That combination — good pharmacology, no trial — is what a stalled programme looks like, and it is worth asking whether it stalled for a scientific reason or a commercial one.

Facts and figures
modalitysmall-molecule inhibitor of both enzymatic and scaffolding function; irreversible (tranylcypromine-derived) chemotypes predominate
clinical status in kmt2a rPhase 1/2a only. No approval in any leukemia. No randomised efficacy data. No paediatric trial in KMT2A-rearranged leukemia identified.
recommended single agent dose140 µg/m²/day, days 1-5 weekly, 28-day cycles
in human pharmacodynamicsconfirmed time- and exposure-dependent induction of differentiation biomarkers; blast differentiation particularly in MLL-translocated patients
eudract alice2018-000482-36
mechanism classdifferentiation therapy (parallel to menin inhibition)
pediatric eligiblefalse
target genes
[
  "KDM1A"
]
agents
[
  "iadademstat (ORY-1001)",
  "bomedemstat",
  "seclidemstat",
  "tranylcypromine (with ATRA)"
]
How this could bite us back
  • On-mechanism differentiation toxicity: LSD1 inhibition can produce differentiation syndrome, the same potentially fatal toxicity that scales with efficacy in menin inhibitors and in ATRA.
  • Essential-function toxicity: LSD1 is required for normal megakaryopoiesis and erythropoiesis; thrombocytopenia is the predictable class effect and it is additive with leukemic marrow failure.
  • Lineage plasticity: pushing a lineage-ambiguous KMT2A-rearranged blast to differentiate may drive it toward the myeloid compartment, which is the same direction as immunotherapy-driven lineage switch. Differentiation therapy and CD19-directed therapy may therefore be in tension.
  • Redundancy risk: if LSD1 and menin inhibition act on the same state lock, combining them buys nothing and doubles the differentiation-syndrome risk.
  • Neurological class risk: tranylcypromine-derived chemotypes are monoamine oxidase-related, and MAO inhibition in an infant on multiple supportive-care drugs is a drug-interaction hazard that has never been characterised.
Still open
  1. Is LSD1 inhibition additive to menin inhibition, or redundant with it? Both force differentiation of the same locked program; a simple combination matrix in KMT2A-rearranged primary cells would answer this and has not been published.
  2. Does LSD1 inhibition work in KMT2A-rearranged B-ALL, or only in myeloid disease? Every clinical observation above is in AML, and KMT2A-rearranged ALL responds to differentiation pressure differently.
  3. Why has no paediatric LSD1 trial in KMT2A-rearranged leukemia opened, given that the in-human pharmacodynamics are among the cleanest in the field? Determining whether this is a scientific or a commercial gap is a cheap and decision-relevant question for the lab.
  4. Does LSD1 inhibition alter surface CD19 or myeloid antigen expression? If differentiation therapy changes the immunophenotype, it directly affects whether CD19-directed therapy can follow it.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the first-in-human phase 1 of iadademstat, 27 patients with relapsed/refractory AML were treated at 5-220 µg/m²/day with a recommended single-agent dose of 140 µg/m²/day and a 14-patient extension cohort enriched for MLL/KMT2A-rearranged AML; pharmacodynamic data confirmed time- and exposure-dependent induction of differentiation biomarkers, and reductions in blood and bone marrow blasts with blast differentiation were observed particularly in patients with MLL translocations, with one complete remission with incomplete count recovery.

    PMID 33052756DOI
    What this evidence cannot show

    Phase 1, n=27, non-randomised, relapsed/refractory adults. 'Reductions in blast percentages with differentiation' is a pharmacodynamic observation, not a response rate; only one CRi is reported. The MLL-translocation observation comes from a small enriched extension cohort and is descriptive. No paediatric patients, no infant ALL data, and myelosuppression is a class effect that will be harder to tolerate in infants.

    Evidence that argues against this (1)
    • A different LSD1 inhibitor in the same indication was stopped before it could be tested for efficacy. The GlaxoSmithKline phase 1 of GSK2879552, alone or with ATRA, in relapsed/refractory AML (NCT02177812, n=41, 2014-2017) is registered as TERMINATED, and the registry records against every Part 2 outcome measure - objective response rate, duration of response, progression-free survival, safety - that the analysis 'was planned but not performed for Part 2 as the study was terminated early during Part 1'. Pharmacodynamic differentiation alone has therefore not been enough to carry an LSD1 inhibitor through phase 1 in AML. GlaxoSmithKline 2017
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The ALICE open-label phase 2a dose-finding study combined iadademstat (starting 90 µg/m²/day, 5 days on 2 days off weekly, range 60-140) with azacitidine 75 mg/m² subcutaneously for 7 of 28 days in adults with newly diagnosed AML ineligible for intensive chemotherapy, across six hospitals in Spain, with safety and recommended phase 2 dose as primary objectives; the study is completed.

    PMID 38824932DOI
    What this evidence cannot show

    Phase 2a dose-finding in older, unfit adults — a population and a question far removed from infant KMT2A-rearranged ALL. We did not verify the response rates from this trial and therefore do not state them. No randomised comparison exists for iadademstat in any setting.

    Evidence that argues against this (1)
    • ALICE is a phase 2a dose-finding study whose primary objectives were safety and the recommended phase 2 dose; it has no randomised azacitidine-alone comparator, so nothing in it separates iadademstat's contribution from the known single-agent activity of azacitidine in this population. The class precedent argues for caution: GSK's LSD1 inhibitor GSK2879552 was terminated during Part 1 of its relapsed/refractory AML phase 1 (NCT02177812) and its planned efficacy analyses were never performed. GlaxoSmithKline 2017
Connected to