Romidepsin - the HDAC inhibitor with the best preclinical case in KMT2A-rearranged INFANT ALL specifically, and a mechanism that is not about histones at all
also called FK228 · depsipeptide · FR901228 · Istodax · NSC 630176
Romidepsin is an approved HDAC-blocking drug that came out top in an unbiased drug screen against cell lines made directly from infants with MLL-rearranged leukaemia. What it does there is unexpected: it makes an existing chemotherapy drug, cytarabine, work better, by switching off the enzyme that destroys cytarabine inside the cell.
The same thing, for a scientist
Romidepsin is a bicyclic depsipeptide **class I-selective HDAC inhibitor**, FDA-approved for cutaneous T-cell lymphoma. It has the **best disease-matched preclinical case of any HDAC inhibitor in this graph**, because it was identified by unbiased screening in a panel of cell lines generated from **primary infant ALL patient samples** rather than being carried in from adult AML. Cruickshank et al. built that panel, checked its fidelity to the primary disease by RNA-sequencing and cytogenetics, and screened FDA-approved chemotherapeutics: three drug classes showed consistent efficacy - **proteasome inhibitors, HDAC inhibitors and CDK inhibitors** - and romidepsin was the named HDAC inhibitor [E1]. **The mechanism they report is the interesting part and it is not a chromatin mechanism**: romidepsin **reduces expression of cytidine deaminase, the enzyme that deactivates cytarabine**, and enhances the DNA-damage response to cytarabine, enhancing cytarabine activity in vitro; combined romidepsin plus cytarabine in xenografted mice further reduced leukaemia burden [E1][E2]. That is a **pharmacological-potentiation mechanism** - the HDAC inhibitor is modifying the metabolism of a drug already in the infant ALL backbone. It predicts a combination, it explains why monotherapy trials of this class have failed, and it is directly testable in the regimens children actually receive. **The three drug classes that came out of that screen are the same three that reached the infant clinic**: the TINI trial gave a proteasome inhibitor (bortezomib) and an HDAC inhibitor (vorinostat) together on a chemotherapy backbone [E3] - convergent, and worth noticing, although romidepsin itself has never been given to an infant with ALL as far as this beat established.
Why the screen is better evidence than most of this beat
Almost every other result in the acetylation layer comes from a mouse fusion model (myeloid), an immortalised line of adult origin, or a xenograft of convenience. The Cruickshank panel is different in three ways that matter 1:
- Generated from primary patient samples of MLL-rearranged infant ALL - the index disease.
- Fidelity checked, by RNA-sequencing and cytogenetics against the matched primary specimens, including what the authors describe as the first systematic analysis of transcriptome reproducibility in vitro. Drug-target gene expression was highly reproducible between lines and matched primaries.
- Unbiased, a high-throughput screen of FDA-approved agents rather than a hypothesis-led test of a favoured class.
An unbiased screen returning HDAC inhibitors is a much stronger statement than an HDAC paper reporting that HDAC inhibitors work. It is also consistent with the independent connectivity-mapping result that nominated HDAC inhibitors for t(4;11) infant ALL from a hypomethylation signature - a completely different route to the same class 4.
What the screen does not establish: that romidepsin is better than the other HDAC inhibitors, that class I selectivity is why it works, or anything about a single HDAC enzyme. It is a class-level result with a named representative.
A mechanism that explains the clinical record
The cytidine deaminase finding deserves to be read as more than a detail 1.
Cytidine deaminase (CDA) deaminates cytarabine to the inactive ara-U. If romidepsin lowers CDA expression, then romidepsin raises intracellular active cytarabine. That is a pharmacokinetic/metabolic potentiation, not an epigenetic one, and it makes three predictions the graph can hold the field to:
- Monotherapy should disappoint. It has: the only efficacy-endpoint HDAC-inhibitor trial in acute leukaemia in this graph was terminated for futility, and it was panobinostat monotherapy.
- The benefit should be specific to cytarabine-containing regimens, and should be absent or smaller where cytarabine is not given.
- The effect should be measurable - CDA expression in blasts before and after dosing is a pharmacodynamic assay that exists, is cheap, and has never been run in any HDAC-inhibitor trial in this disease.
This is the clearest example in the beat of why "an HDAC inhibitor entry that names only histone acetylation is incomplete". Three distinct mechanisms are now on record for this drug class in this disease: bulk histone acetylation; suppression of the RNF20/RNF40/WAC ligase complex depleting H2Bub (panobinostat); acute release of P-TEFb from the 7SK snRNP (SAHA); and now a fourth, CDA-mediated potentiation of cytarabine. None of them is attributable to a single HDAC enzyme.
| target | gene-hdac1 |
|---|---|
| target note | Class I-selective HDAC inhibitor covering HDAC1, HDAC2 and HDAC3. gene-hdac1 is named in the schema-required target field as a representative class I enzyme; romidepsin is NOT HDAC1-selective and no romidepsin result attributes to an enzyme. |
| modality | bicyclic depsipeptide, prodrug reduced intracellularly to the active thiol, class I-selective HDAC inhibitor |
| status | FDA approved for cutaneous T-cell lymphoma; never given to an infant with ALL as far as this beat established |
| trials in this disease | none found; romidepsin's infant ALL evidence is entirely preclinical |
| discovery route | unbiased high-throughput screen of FDA-approved chemotherapeutics against cell lines generated from PRIMARY MLL-rearranged infant ALL samples, validated against matched primaries by RNA-seq and cytogenetics |
| reported mechanism | reduces cytidine deaminase expression, lowering cytarabine deactivation, and enhances the DNA damage response to cytarabine - a pharmacological-potentiation mechanism, not a chromatin one |
| combination partner | cytarabine |
| in vivo result | romidepsin plus cytarabine in xenografted mice further reduced leukaemia burden |
| independent route to the same class | connectivity mapping on the t(4;11) hypomethylation signature (PMID 22015773) |
| pharmacodynamic assay that exists and has never been run | cytidine deaminase expression in blasts before and after HDAC inhibitor dosing |
| attribution to an hdac enzyme | false |
trials
[]
classes that scored
[ "proteasome inhibitors", "HDAC inhibitors", "CDK inhibitors" ]
distinct mechanisms on record for this drug class in this disease
[ "bulk histone acetylation", "RNF20/RNF40/WAC suppression depleting H2B ubiquitination (panobinostat)", "acute P-TEFb release from the 7SK snRNP (SAHA/vorinostat)", "cytidine deaminase suppression potentiating cytarabine (romidepsin)" ]
key ae
[ "not enumerated in this entry; romidepsin's label carries warnings for QT prolongation, myelosuppression, infections and tumour lysis syndrome, recorded here as a pointer and not verified against the label in this beat" ]
resistance
[ "not characterised in KMT2A-rearranged leukaemia", "a cytidine-deaminase-mediated potentiation mechanism is defeated by any resistance route that does not run through cytarabine metabolism, and adds nothing where cytarabine is not given" ]
- Lowering cytidine deaminase raises active intracellular cytarabine everywhere, not only in blasts. Cytarabine's dose-limiting toxicities - myelosuppression, mucositis, cerebellar toxicity at high dose - would be potentiated along with its efficacy, and in an infant the cerebellar risk is the one with lifelong consequences.
- An unbiased screen returning proteasome, HDAC and CDK inhibitors has returned three of the most broadly cytotoxic targeted classes in existence. Without a lineage-matched non-KMT2A comparator arm, the screen cannot distinguish a KMT2A-rearranged vulnerability from general cytotoxicity, and it did not have one.
- A potentiation mechanism is invisible in a monotherapy trial and would be scored as drug failure. Conversely a combination trial that works cannot tell you the HDAC inhibitor contributed, unless CDA or another pharmacodynamic marker is measured - which no trial in this disease has done.
- Romidepsin carries QT prolongation warnings. Infants on this protocol already receive anthracyclines or mitoxantrone, azoles for fungal prophylaxis and ondansetron, all QT-active, and the TINI experience shows infection is what actually kills these children.
- The 2012 connectivity-mapping paper called for urgent clinical testing. Fourteen years later there is no efficacy-endpoint HDAC-inhibitor trial in this disease and the one in adjacent disease failed. Enthusiasm in this class has a track record and it should be priced in.
- Does cytidine deaminase knockdown phenocopy romidepsin's cytarabine potentiation? The abstract reports the expression change but not the causal test, and that one experiment would convert a correlation into a mechanism.
- Is the potentiation class-wide or romidepsin-specific? If vorinostat and panobinostat also lower CDA, the TINI trial may have delivered this mechanism without scheduling for it; if not, the wrong HDAC inhibitor reached the infant clinic.
- Romidepsin has never been given to an infant with ALL. Given that it has the best disease-matched preclinical case of any agent in this class, why is the agent in the infant trial vorinostat?
- Should an HDAC inhibitor be scheduled adjacent to cytarabine rather than in induction? The CDA mechanism predicts the timing matters and no trial in this disease has been designed around it.
- Would CDA expression in blasts serve as the pharmacodynamic biomarker this entire drug class lacks in this disease?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A panel of cell lines generated from primary MLL-rearranged infant ALL patient samples, validated against matched primary specimens by RNA-sequencing and cytogenetics, was screened against FDA-approved chemotherapeutics; high-throughput screening identified consistent efficacy of three novel drug classes - proteasome inhibitors, histone deacetylase inhibitors and cyclin-dependent kinase inhibitors; histone deacetylase inhibitors including romidepsin enhanced the activity of cytarabine in vitro and combined romidepsin plus cytarabine in xenografted mice further reduced leukaemia burden; and molecular studies showed romidepsin reduces expression of cytidine deaminase, the enzyme involved in cytarabine deactivation, and enhances the DNA damage response to cytarabine.
Systematic chemical and molecular profiling of MLL-rearranged infant acute lymphoblastic leukemia reveals efficacy of romidepsin. · Cruickshank MN, Ford J, Cheung LC, Heng J, Singh S, Wells J, et al. · Leukemia 31(1):40-50 2017What this evidence cannot show
Cell lines and xenografts, not patients. The screen was of FDA-approved chemotherapeutics, so the comparator set is what was already approved and not the full chemical space - a class absent from the library cannot be identified. Romidepsin is class I-selective but covers HDAC1, HDAC2 and HDAC3, so nothing here attributes to an enzyme. The number of distinct patients contributing lines is not given in the abstract; the full text was not retrieved in this beat (PMCID PMC5220136). The cytidine deaminase mechanism is a gene-expression observation and the abstract does not state that CDA knockdown phenocopies it, so the causal step in that chain was not confirmed here.
Evidence that argues against this (1)
- A high-throughput screen returning proteasome inhibitors, HDAC inhibitors and CDK inhibitors returns three of the most broadly cytotoxic targeted classes available. A separate high-throughput drug-library screen in PRIMARY KMT2A-rearranged infant ALL cells reports that the screen format itself 'favors the identification of drug candidates that activate P53 signaling' - i.e. that this class of screen has a characterisable bias toward a general death pathway rather than toward anything specific to the rearrangement. Neither screen included a lineage-matched non-KMT2A comparator arm, which is what would separate a KMT2A vulnerability from general cytotoxicity. Only the title of this counter-citation was read. Wander P et al. 2022
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Romidepsin enhances the efficacy of cytarabine in vivo, revealing histone deacetylase inhibition as a promising therapeutic strategy for KMT2A-rearranged infant acute lymphoblastic leukaemia.
Romidepsin enhances the efficacy of cytarabine in vivo, revealing histone deacetylase inhibition as a promising therapeutic strategy for KMT2A-rearranged infant acute lymphoblastic leukemia. · Cheung LC, Cruickshank MN, Hughes AM, Singh S, Chua GA, Ford J, et al. · Haematologica 104(7):e300-e303 2019What this evidence cannot show
This is a four-page Haematologica letter (e300-e303) and PubMed returns NO abstract for it, so this entry carries only the claim asserted by the title itself. The full text was not retrieved in this beat; Haematologica is open access and is a route-5 candidate via its DOI, and a future cycle should read it rather than relying on this block. Same laboratory as E1, so this is an extension rather than an independent replication. Confidence is 'moderate' precisely because it rests on a title.
Evidence that argues against this (1)
- Enhancing cytarabine in vivo is a combination result, and the mechanism proposed elsewhere in this graph - reduced cytidine deaminase - would make romidepsin a pharmacokinetic modifier of the partner drug rather than a therapy directed at the leukaemia's chromatin. The class IIA work in KMT2A-r ALL found in-vitro efficacy that did not survive the xenograft, which is the standard this claim should be held to. Verbeek TCAI et al. 2024
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The three drug classes identified as consistently efficacious in the infant ALL screen - proteasome inhibitors, HDAC inhibitors and CDK inhibitors - overlap with what reached the infant clinic: NCT02553460 incorporated the proteasome inhibitor bortezomib and the HDAC inhibitor vorinostat into a chemotherapy backbone for 50 newly diagnosed infants, 35 of them KMT2A-rearranged.
Bortezomib and vorinostat in combination with mitoxantrone, dexamethasone, and pegasparaginase during induction and reinduction for infants with acute lymphoblastic leukaemia: a multicentre single-arm phase 1/2 study. · Gruber TA, Jeha S, Deyell RJ, Lewis V, Chang BH, Lowe EJ, et al. · The Lancet Haematology 13(3):e144-e156 2026What this evidence cannot show
The overlap is an observation made by this entry, not a claim either paper makes; the TINI investigators cite their own preclinical work identifying bortezomib and vorinostat, not the Cruickshank screen. The trial used vorinostat, not romidepsin, and romidepsin has never been given to an infant with ALL as far as this beat established. The trial's endpoint was tolerability and it makes no efficacy claim. Cytarabine is in the TINI backbone but in later phases rather than alongside the HDAC inhibitor in induction, so the CDA-potentiation mechanism was not scheduled for in that protocol.
Evidence that argues against this (1)
- That three drug classes recur across infant ALL screens is a statement about what screens find, not about what works in patients: proteasome, HDAC and CDK inhibitors are broadly cytotoxic to proliferating cells and recur in screens across many diseases. The infant KMT2A-rearranged literature also contains a direct counter-direction, in which restoring HDAC7 rather than inhibiting HDACs is the proposed advantage. de Barrios O et al. 2025
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Connectivity mapping applied to the genes most significantly hypomethylated in t(4;11)-positive infant ALL relative to healthy bone marrow independently nominated HDAC inhibitors as suitable candidates; HDAC inhibitors induced leukaemic cell death in t(4;11)-positive primary infant ALL cells, accompanied by downregulation of MYC, SET, RUNX1, RAN and of the MLL-AF4 fusion product itself, and DNA methylation was restored after HDAC inhibitor exposure.
Connectivity mapping identifies HDAC inhibitors for the treatment of t(4;11)-positive infant acute lymphoblastic leukemia. · Stumpel DJ, Schneider P, Seslija L, Osaki H, Williams O, Pieters R, Stam RW · Leukemia 26(4):682-692 2012What this evidence cannot show
Cited here as an INDEPENDENT route to the same drug class in the same disease, not as romidepsin evidence - the agents used are not specified in the abstract and romidepsin is not named. Primary infant ALL cells in vitro; the authors themselves conclude the agents 'should urgently be tested in in vivo models and subsequent clinical trials', i.e. the in vivo work was not done here. Connectivity mapping is a signature-matching method and nominates classes, never enzymes. The reported downregulation of the MLL-AF4 fusion product itself is a striking claim that this entry carries at the level the abstract states it. No PMCID reported by efetch; full text not retrieved.
Evidence that argues against this (1)
- Fourteen years after this paper called for urgent clinical testing, the only HDAC inhibitor to reach newly diagnosed infants is vorinostat in a combination tolerability study with no efficacy endpoint, and the only efficacy-endpoint HDAC-inhibitor trial in acute leukaemia in this graph was terminated for futility. The preclinical enthusiasm and the clinical record have not converged. City of Hope Medical Center; National Cancer Institute 2010
identified by unbiased screening of primary MLL-rearranged infant ALL lines; preclinical only, never given to an infant
class I-selective, covering HDAC1, HDAC2 and HDAC3
- treated by →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
the connectivity-mapping route nominated this class for t(4;11) infant ALL and reported downregulation of the MLL-AF4 fusion product
romidepsin + low-dose cytarabine gave a profound survival increase in three KMT2A-r infant ALL xenografts
synergy in only 3 of 6 infant ALL cell lines, against 6 of 6 for daunorubicin
synergy in 4 of 6 infant ALL cell lines
synergy in 6 of 6 infant ALL cell lines, the highest score in that panel
ANTAGONISED in 4 of 6 infant ALL cell lines - the reason the HDAC scheduling proposal avoids MARMA
ANTAGONISED in 4 of 6 infant ALL cell lines
4-hydroperoxycyclophosphamide synergy in 6 of 6 infant ALL cell lines