Oncofusion

start / graph / target-hox-pbx-interface-peptide-hxr9

Targetweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q56-homeodomain

The HOX-PBX interface, and its one peptide: HXR9/HTL001 has been tried in eight tumour types and in KMT2A-rearranged leukaemia exactly once

also called HXR9 · HTL001 · CXR9 (control peptide) · HOX/PBX dimer antagonist · hexapeptide antagonist

The one drug-like thing ever aimed at this biology is a short peptide that copies the piece of HOXA9 that plugs into its partner PBX, so it competes the two apart. It kills cancer cells in melanoma, ovarian, breast, mesothelioma, oral and B-cell tumours - and in KMT2A-fusion leukemia it has been tested in a single published paper. That breadth is the reason to be interested and also the reason to be careful.

The same thing, for a scientist

HXR9 is a cell-penetrating peptide built on the conserved HOX hexapeptide and designed as a competitive antagonist of HOX/PBX binding; HTL001 is its clinical-stage successor [E1]. In KMT2A-rearranged leukaemia the entire published basis is **one paper**: `(HXR9 OR HTL001) AND (MLL OR KMT2A)` returns **1** PubMed record, PMID 23264595, where HXR9 selectively kills leukaemic cells overexpressing HOXA/PBX3 genes and is explicitly framed by its authors as proof of concept [E2][E3]. Set against that, the same peptide has reported activity in melanoma, ovarian, breast, meningioma, mesothelioma, oral potentially-malignant and squamous carcinoma, malignant B cells including myeloma, and lymphangioleiomyomatosis [E4]. **Three specific mechanistic problems sit on top of the breadth**: the interface is redundant on the HOXA9 side, so competing at the hexapeptide pocket need not dissociate the complex [E5]; the pairing the fusion transcriptome actually tracks is MEIS1-PBX3, not HOX-PBX [E6]; and HXR9 is a polyarginine-tailed cationic peptide, a class whose membrane activity is a known confounder. No HXR9 or HTL001 trial was found on ClinicalTrials.gov.

Go deeper
The bounded statement that should govern how much weight this carries

Measured on PubMed via E-utilities esearch on 2026-09-20, with query translations read back:

queryrecords
HXR9[tiab]18
HTL001[tiab]4
HTL001 (all fields)4
(HXR9[tiab] OR HTL001[tiab] OR CXR9[tiab]) AND (leukemia[tiab] OR leukaemia[tiab])4
(HXR9 OR HTL001) AND (MLL OR KMT2A)1
"HOX/PBX"[tiab] (positive control for the concept)71
hexapeptide[tiab] AND PBX[tiab] (positive control for the motif)17

The single record is PMID 23264595, title-checked. The controls return 71 and 17, so the instrument can return a one and the count of one is a measurement rather than an artefact (rule 67). Name list tried: HXR9, HTL001, CXR9, HOX/PBX, hexapeptide - and not any compound code this beat has not seen, which is the residual risk under rule 72.

So: one paper, in this disease, from the group that also proposed PBX3 as the critical cofactor. Everything else about this peptide is from other tumour types.

The case for it, made as strongly as it can honestly be made

The rationale is not naive and the group states it plainly: because of functional redundancy amongst HOX proteins, knockdown or knockout of individual HOX genes often produces no phenotype, so the productive move is to target the interaction they all share rather than any one gene 1.

This project's own data support that premise exactly. Live DepMap gives HOXA9 a mean gene effect of -0.223 across 57 leukaemia lines and HOXA10 -0.0997; neither is a dependency, and the HOXA cluster is a textbook redundant family. A paralogue-blind agent is the rational response to a redundant family, and the fact that single-gene knockouts look empty is the prediction of the hypothesis rather than evidence against it.

In the one leukaemia paper, HXR9 selectively kills cells overexpressing HOXA/PBX3 genes 2 - selectivity being the claim that matters, and the claim that is hardest to establish for a cationic CPP.

The case against, in three separable parts

1. Breadth without a unifying dependency. Reported activity spans melanoma, ovarian, breast, meningioma, mesothelioma, oral keratinocyte-derived lesions, malignant B cells including myeloma, and LAM 4. In the B-cell panel, HXR9 induced significant cytotoxicity in the entire panel of six lines despite the authors finding different levels of expression of HOX family members across them 3. A compound whose potency does not track its stated target's expression across a panel is a compound whose mechanism is unproven. Against that, one study specifically reports sparing of normal oral keratinocytes while killing malignant and potentially-malignant cells 4 - which is the single best specificity datum in the peptide's record and is not from a leukaemia system.

2. The interface is redundant on the HOXA9 side. Human HOXA9's TALE binding rests on the hexapeptide and two paralog-specific homeodomain residues, either of which suffices; and MEIS presence remodels the interaction so the hexapeptide becomes globally dispensable for trimer formation in most HOX proteins 5. MEIS1 is abundant in exactly this disease. A competitor at the hexapeptide pocket is therefore competing at the route this literature says is the more dispensable one, in the cell type where it is most dispensable.

3. It may be the wrong interface entirely. The combination that reproduces the MLL-fusion core transcriptome in murine HSPCs is PBX3 + MEIS1 with no ectopic HOX gene; HOXA9/PBX3 does not 6. If the load-bearing contact is MEIS-PBX, then HOX-PBX antagonism is aimed at a contact the disease can run without.

Regulatory and chemical status, and what is genuinely unknown
  • No trial. A ClinicalTrials.gov API query for HTL001 returned one unrelated study (NCT06952894, an ultrasound feasibility study) and a query for HXR9 returned nothing relevant, on 2026-09-20. No registered HXR9 or HTL001 trial was located. Bounded by those two search terms only; a sponsor-name search was not run.
  • Modality. A cell-penetrating peptide is a delivery problem before it is a pharmacology problem: serum stability, renal clearance, tissue distribution and immunogenicity all apply, and none was read for this entry.
  • Paediatric data. None found, for either peptide, in any indication.
  • The control peptide. CXR9 is the published inactive control. Nothing read this cycle reports a CXR9 arm in a leukaemia experiment, so the CPP-backbone contribution to the leukaemia result is not bounded here.

What would change the assessment cheaply. A dose-response of HXR9 against CXR9 in KMT2A::AFF1 lines (SEM, RS4;11) versus KMT2A-germline B-ALL lines, with a HOXA/PBX3 expression panel and an LDH-release readout to separate apoptosis from membrane lysis. That is one experiment and it addresses parts 1 and the CPP confound together.

Facts and figures
target interfaceHOX hexapeptide binding the PBC-class TALE pocket
modalitycell-penetrating cationic peptide (hexapeptide sequence with a polyarginine delivery tail)
control peptideCXR9
approval statusnone; no registered trial located in ClinicalTrials.gov for either agent on 2026-09-20
adjacent chemical matterMEIS inhibitors (MEISi) exist and reduce MEIS1 transcriptional activity and LSC survival in primary leukaemia cells (PMID 37902585), but they are described as inhibitors of MEIS1 transcriptional activity rather than of the MEIS1-PBX3 dimerisation interface, and none has been tested in KMT2A-rearranged disease in anything read this cycle.
decisive cheap experimentHXR9 versus CXR9 dose-response in KMT2A::AFF1 lines (SEM, RS4;11) against KMT2A-germline B-ALL lines, with a HOXA/PBX3 expression panel and an LDH-release readout alongside apoptosis, to separate on-mechanism killing from cationic membrane activity
target genes
[
  "gene-hoxa9",
  "gene-hoxa10",
  "gene-pbx3"
]
agents
[
  "HXR9",
  "HTL001"
]
trials
[]
kmt2a evidence base
{
  "published_records": 1,
  "pmid": "23264595",
  "query": "(HXR9 OR HTL001) AND (MLL OR KMT2A)",
  "date": "2026-09-20",
  "tool": "PubMed E-utilities esearch",
  "positive_controls": {
    "\"HOX/PBX\"[tiab]": 71,
    "hexapeptide[tiab] AND PBX[tiab]": 17,
    "HXR9[tiab]": 18
  },
  "name_list_tried": [
    "HXR9",
    "HTL001",
    "CXR9",
    "HOX/PBX",
    "hexapeptide"
  ],
  "residual_risk": "any compound code for this peptide series that this beat has not seen would not have been searched (brief rule 72)",
  "title": "PBX3 is an important cofactor of HOXA9 in leukemogenesis.",
  "authors": "Li Z, Zhang Z, Li Y, et al.",
  "venue": "Blood",
  "year": 2013,
  "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
indications with reported activity
[
  "melanoma",
  "ovarian",
  "breast",
  "meningioma",
  "mesothelioma",
  "oral potentially-malignant and squamous carcinoma",
  "malignant B cells incl. myeloma",
  "lymphangioleiomyomatosis",
  "MLL-rearranged / HOXA-PBX3-high leukaemia"
]
mechanistic objections
[
  "the HOXA9-TALE interaction is redundant between the hexapeptide and two paralog-specific homeodomain residues, and MEIS presence makes the hexapeptide the more dispensable route",
  "the combination that reproduces the MLL-fusion core transcriptome is PBX3+MEIS1, not HOXA9+PBX3",
  "potency did not track HOX expression across the one B-cell panel where both were measured",
  "no control-peptide arm was read in any leukaemia experiment, so the cationic-CPP backbone contribution is unbounded here"
]
How this could bite us back
  • Paralogue-blind by design means toxicity-blind by design. The hexapeptide is shared across the HOX family, so this agent cannot distinguish leukaemic HOXA9 from the HOX proteins running normal haematopoiesis, gut and limb patterning, or endometrial receptivity - in patients who are infants.
  • Polyarginine cell-penetrating tails are membrane-active. Killing in every cell line of a panel regardless of target expression [E3] is the signature this class produces when the backbone rather than the payload is doing the work, and no leukaemia experiment read here includes the published CXR9 control.
  • It may be the wrong interface. The contact whose disruption abolishes the fusion transcriptome in the best combinatorial experiment is MEIS1-PBX3, not HOX-PBX [E6], and the HOX-PBX contact is redundant on the HOXA9 side [E5].
  • One paper. The entire KMT2A evidence base is a single proof-of-concept arm, from the group that developed the peptide, with no independent replication in 13 years - a bounded count of 1 against controls of 71 and 17.
  • Peptide delivery in a marrow-resident disease. Serum stability, renal clearance, marrow penetration and immunogenicity are all unaddressed in anything read here, and a CPP that works in a flank xenograft has not been shown to reach a blast in a bone.
  • Resistance would be cheap. If the complex can form through homeodomain residues when the hexapeptide pocket is occupied [E5], resistance requires no mutation at all - only the MEIS1 the cell already has.
Still open
  1. Does HXR9 kill KMT2A-rearranged B-lineage cells more than KMT2A-germline B-lineage cells, with a CXR9 control in the same plate? The one leukaemia paper does not answer this from its abstract, and it is the experiment that decides whether the whole approach is worth anything here.
  2. Is the cell death on-mechanism? An LDH-release or membrane-integrity readout alongside apoptosis, in the same experiment, would separate target engagement from cationic-peptide lysis - and no leukaemia study read this cycle reports one.
  3. Does HTL001 exist as a clinical agent, and under what registry or IND? The developing group describes it as clinical-stage; no trial was found under either compound code.
  4. Would an agent aimed at the MEIS-PBX interface instead do better? That interface is the one the transcriptome data implicate and the one with no chemical matter against it; see hyp-meis1-pbx3-interface-is-the-unattempted-target.
Every source behind this entry (7)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review

    Because of functional redundancy amongst HOX proteins, knockdown or knockout studies of individual HOX genes often do not result in an altered phenotype; an alternative approach is to target the HOX-PBX interaction, which depends on a conserved hexapeptide region within HOX, and several peptides have been developed on the hexapeptide sequence that act as competitive antagonists of HOX/PBX binding, including HXR9 and HTL001.

    Using the Conserved Hexapeptide in HOX Proteins as an Antagonist of HOX/PBX Interactions · Morgan R, Smith C, Pandha H · Methods in Molecular Biology 2889:139-153 2025
    PMID 39745610DOI
    What this evidence cannot show

    A methods review written by the group that developed both peptides, so it is authoritative on the approach and interested in its success. The exact peptide sequences were NOT read and are not asserted here. Used as a map, per the standing preference for primary sources.

    Evidence that argues against this (1)
    • The redundancy premise does not hold for HOXA9 in this disease: HOXA9 knockdown alone kills MLL-rearranged and not MLL-germline human cells, so at least one paralogue has a non-redundant phenotype and the argument for a paralogue-blind agent is weaker here than in general. Faber J et al. 2009
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    As a proof of concept, the small peptide HXR9, developed to specifically disrupt the interactions between HOX and PBX proteins, can selectively kill leukaemic cells with overexpression of HOXA/PBX3 genes, in a study whose other arms show PBX3 (but not PBX1 or PBX2) is co-expressed with HOXA9 in MLL-rearranged AML and that Pbx3 depletion inhibits MLL-fusion-mediated transformation.

    PBX3 is an important cofactor of HOXA9 in leukemogenesis · Li Z, Zhang Z, Li Y, Arnovitz S, Chen P, Huang H, Jiang X, Hong GM, Kunjamma RB, Ren H, He C, Wang CZ, Elkahloun AG, Valk PJ, Dohner K, Neilly MB, Bullinger L, Delwel R, Lowenberg B, Liu PP, Morgan R, Rowley JD, Yuan CS, Chen J · Blood 121:1422-1431 2013
    PMID 23264595DOI
    What this evidence cannot show

    Abstract read this cycle; full text not retrieved. Graded MODERATE and not strong because: the authors themselves call it proof of concept; no cell lines, concentrations, control-peptide arm or normal-cell comparator were read; 'selectively' is unquantified in the abstract; and R Morgan, who developed HXR9, is a co-author, so this is not an independent test of the peptide. This is the ONLY published HXR9 or HTL001 result in KMT2A-rearranged or MLL leukaemia - a bounded count of 1, controls 71 and 17, measured 2026-09-20.

    Evidence that argues against this (1)
    • In a six-line malignant B-cell panel with different levels of HOX family expression, HXR9 induced significant cytotoxicity in the ENTIRE panel - activity that does not track target expression, which is what an off-mechanism cationic-peptide effect would also look like. Daniels TR et al. 2010
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a panel of six malignant B-cell lines including multiple myeloma cells, expression profiling of all 39 HOX genes found different levels of expression of HOX family members between lines, and HXR9 nonetheless induced significant cytotoxicity in the entire panel of cell lines; the cytotoxic effect could be enhanced by combination with an antibody-avidin fusion protein against transferrin receptor 1 (CD71), with iron starvation contributing through a partly caspase-independent pathway.

    Disruption of HOX activity leads to cell death that can be enhanced by the interference of iron uptake in malignant B cells · Daniels TR, Neacato II, Rodriguez JA, Pandha HS, Morgan R, Penichet ML · Leukemia 24:1555-1565 2010
    PMID 20574452DOI
    What this evidence cannot show

    Abstract read this cycle; the author surname Rodriguez is rendered without its accent. Malignant B-cell lines including myeloma, NOT KMT2A-rearranged B-ALL. The abstract does not report a control-peptide arm or a normal-B-cell comparator, so 'cytotoxicity in the entire panel' cannot be separated from a backbone effect from what was read. Cited here as the best available evidence about the peptide's BREADTH in B-lineage cells, which is the property this entry is assessing.

    Evidence that argues against this (1)
    • One study reports the opposite property - death in oral potentially-malignant and squamous carcinoma cells but NOT in normal oral keratinocytes - which is the strongest specificity datum in the peptide's record, though it is not a haematopoietic system. Platais C et al. 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    HXR9 or the HOX-PBX antagonism approach has reported anti-tumour activity across many unrelated tumour types: melanoma (blocking in vivo proliferation of flank tumours), ovarian cancer, breast cancer, meningioma, malignant mesothelioma, oral potentially-malignant and squamous carcinoma cells with sparing of normal oral keratinocytes, malignant B cells including myeloma, and lymphangioleiomyomatosis.

    Antagonism of HOX/PBX dimer formation blocks the in vivo proliferation of melanoma · Morgan R, Pirard PM, Shears L, Sohal J, Pettengell R, Pandha HS · Cancer Research 67:5806-5813 2007
    PMID 17575148DOI
    What this evidence cannot show

    The melanoma paper is the founding report and its abstract was read this cycle. The OTHER indications in this claim were established by a PubMed title survey of the 18 HXR9[tiab] records on 2026-09-20 - TITLES ONLY, abstracts not read for most of them, so the breadth is documented at title level and the strength of each individual result is not assessed here. The claim is about the RANGE of reported activity, which titles can support, and not about the size or quality of any one result, which they cannot.

    Evidence that argues against this (1)
    • Breadth is not by itself evidence against a mechanism: HOX genes are dysregulated across many solid tumours, so a genuine pan-HOX agent would be expected to show pan-tumour activity. The concern is specifically that potency did not track HOX expression in the one panel where both were measured. Daniels TR et al. 2010
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The human HOXA9-TALE interaction relies on the redundant activity of the hexapeptide motif and two paralog-specific residues of the HOXA9 homeodomain, and the additional presence of MEIS induces a remodelling of the HOX-PBX interaction leading to global dispensability of the hexapeptide motif for trimeric complex formation in the large majority of HOX proteins.

    The human HOXA9 protein uses paralog-specific residues of the homeodomain to interact with TALE-class cofactors · Dard A, Jia Y, Reboulet J, Bleicher F, Lavau C, Merabet S · Scientific Reports 9:5664 2019
    PMID 30952900DOI
    What this evidence cannot show

    Abstract read this cycle. In vitro DNA-binding assays and BiFC in heterologous cell lines, not leukaemia cells and not a functional leukaemia readout. It is a statement about complex FORMATION, and a complex that forms by the alternative route is not shown to be transcriptionally competent. It is an argument that the peptide's design premise is incomplete, not a demonstration that the peptide fails.

    Evidence that argues against this (1)
    • HXR9 demonstrably kills leukaemic cells overexpressing HOXA/PBX3 genes, so whatever the binding-assay redundancy shows, competing at this motif produces a cellular effect - though the effect may not be on-mechanism. Li Z et al. 2013
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In murine haematopoietic stem/progenitor cells, PBX3 plus MEIS1 without an ectopic HOX gene is sufficient to transform cells and cause AML, and uniquely among the combinations tested recapitulates the MLL-fusion core transcriptome; HOXA9 plus PBX3 does not. Disrupting the MEIS1-PBX3 binding diminishes transformation and HOX upregulation.

    PBX3 and MEIS1 Cooperate in Hematopoietic Cells to Drive Acute Myeloid Leukemias Characterized by a Core Transcriptome of the MLL-Rearranged Disease · Li Z, Chen P, Su R, Hu C, Li Y, Elkahloun AG, Zuo Z, Gurbuxani S, Arnovitz S, Weng H, Wang Y, Li S, Huang H, Neilly MB, Wang GG, Jiang X, Liu PP, Jin J, Chen J · Cancer Research 76:619-629 2016
    PMID 26747896DOI
    What this evidence cannot show

    Abstract read this cycle. Murine HSPCs, myeloid readout, retroviral overexpression. Cited here for the negative half - that HOXA9/PBX3 does NOT reproduce the fusion transcriptome - which is what bears on whether the HOX-PBX contact is the right one to antagonise. Sufficiency of PBX3/MEIS1 is not necessity, and the same group authored E2.

    Evidence that argues against this (1)
    • The same group published both this and the HXR9 leukaemia result, and continued to describe the HOX/PBX interaction as a feasible therapeutic strategy, so they evidently do not read their own 2016 result as invalidating the 2013 one. Li Z et al. 2013
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.trial_registry

    A ClinicalTrials.gov API v2 query for the term HTL001 on 2026-09-20 returned no study of that agent (the single returned record, NCT06952894, is an unrelated ultrasound feasibility study), and a query for HXR9 returned no relevant study. No registered clinical trial of either peptide was located.

    ClinicalTrials.gov API v2 free-text query for HTL001 and for HXR9 · ClinicalTrials.gov (NIH/NLM); query run by this project 2026-09-20 · Trial registry 2026
    What this evidence cannot show

    A bounded absence with a date and a query, bounded ONLY by the two compound codes searched. No sponsor-name search, no search of EudraCT, ISRCTN, ANZCTR, ChiCTR or the WHO ICTRP, and no search of any alternative compound code this beat has not seen - which is the exact residual risk brief rule 72 names. Graded moderate rather than established for that reason. Note also brief rule 75: a drug's absence from one registry can be regulatory or geographic rather than real.

    Evidence that argues against this (1)
    • The developing group describes HTL001 as a clinical-stage successor to HXR9 in a 2025 methods review, so a trial or a trial application may exist under a name or registry this query did not reach. Morgan R et al. 2025
Connected to