CBFB::MYH11 (CBFbeta-SMMHC), inv(16)(p13;q22)

This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.

• Restoring RUNX1 activity is the therapeutic mechanism [E4], but RUNX1 and CBFB are required for definitive haematopoiesis [E3] and the pathway is dose-sensitive in both directions - germline RUNX1 loss causes familial platelet disorder with predisposition to AML. An agent that perturbs CBF stoichiometry could plausibly cause cytopenias even while correcting the fusion.
• AI-10-49 delayed leukaemia in mice; it did not cure it [E4]. A PPI inhibitor that releases a sequestered transcription factor removes the fusion's dominant-negative effect but does nothing about cooperating lesions such as KIT mutations, so monotherapy resistance should be expected and the realistic use is in combination.
• The clinical development bar is brutally high here precisely BECAUSE the disease is favourable-risk. Showing that a novel agent improves on high-dose cytarabine plus gemtuzumab in a subgroup with already-good outcomes requires very large numbers, and the subgroup is small - a structural disincentive that has nothing to do with the biology.
• The inv(16) name collision with CBFA2T3::GLIS2 is a live hazard in both directions: calling a CBFA2T3::GLIS2 infant 'favourable-risk inv(16)' under-treats an aggressive leukaemia, and mislabelling a CBFB::MYH11 patient could trigger unnecessary transplant. This is a reporting and database-schema risk, not a biological one, and it is entirely preventable.
• KIT inhibition, the available co-mutation pharmacology, carries the usual on-target resistance mutations plus KIT's essential roles in normal haematopoiesis, mast cells and germ cells.
• The captured RUNX1 is intact, which is what makes release therapeutically attractive - but it also means the leukaemia has never had to evolve a RUNX1-independent programme. Sustained pharmacological release could select for exactly that, and a RUNX1-independent inv(16) blast would be far harder to treat than the original.
• If inv(16) cells need native RUNX1, an interface drug that frees RUNX1 is adjusting a ratio rather than lifting a blockade, and the therapeutic window is bounded on both sides -- too little and nothing happens, too much and you are supplying a factor the leukaemia requires.

1. Why has AI-10-49 not advanced clinically in the decade since publication [E4]? Is it pharmacokinetics, a therapeutic window problem, the difficulty of showing benefit on top of already-favourable chemotherapy outcomes, or commercial disinterest in a small favourable-risk subgroup? The answer matters far beyond this fusion, because it is the test case for whether transcription-factor PPI inhibition is a viable modality at all.
2. There is no structure of the CBFB::MYH11 fusion, and none of AI-10-49 bound to it [E5][E6]. What is the fusion-specific conformation that gives the drug its selectivity? A structure would enable second-generation design and would test the general rule that fusion-specific conformational change is what makes an inherited interface druggable.
3. Is the increased Runt-domain affinity [E2] sufficient on its own, or is coiled-coil multimerisation independently required? An affinity-increased but multimerisation-dead mutant, and the converse, would separate capture from sequestration - and would tell us whether AI-10-49's mechanism is the right one or whether blocking multimerisation would be better.
4. RUNX1::RUNX1T1 and CBFB::MYH11 attack opposite subunits of the same heterodimer and both give favourable-risk AML with a KIT-mutant adverse subset. Why do they produce different morphologies (M2 with Auer rods versus M4Eo with abnormal eosinophils) from what appears to be the same stem-progenitor cell of origin? That is a clean question about how the mode of CBF corruption - sequestration versus corepressor delivery - shapes lineage output.
5. Cbfb-MYH11 knock-in embryos die of failed definitive haematopoiesis rather than developing AML [E3]. What postnatal cooperating events convert the dominant-negative block into leukaemia, and does AI-10-49 act on the fusion's contribution only, leaving the cooperating lesions intact?
6. Why has no CBFbeta-SMMHC:RUNX1 interface inhibitor been registered in eleven years? The answer determines whether this beat's central hypothesis is limited by biology or by chemistry and funding, and it is answerable by asking the AI-10-49 authors rather than by reading more papers.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   inv(16)(p13q22) breaks near the end of the CBFB coding region and interrupts MYH11, producing in six of six patient samples an in-frame fusion mRNA joining the first 165 amino acids of CBFbeta to the tail region of smooth-muscle myosin heavy chain; the repeated SMMHC coiled coil was predicted to dimerise the fusion and thereby alter transcriptional regulation.

   Fusion between transcription factor CBF beta/PEBP2 beta and a myosin heavy chain in acute myeloid leukemia · Liu P, Tarle SA, Hajra A, Claxton DF, Marlton P, Freedman M, Siciliano MJ, Collins FS · Science 1993
   PMID 8351518DOI
   ▸What this evidence cannot show

   Six patient samples; the dimerisation mechanism was a sequence-based prediction in this paper, later confirmed functionally and structurally by others. CBFB was described here as regulating genes expressed in T cells, which understates its now-established role across haematopoiesis.

   ▸Evidence that argues against this (1)

   • Six of six is a small denominator, and the uniformity does not hold at scale. Across 1629 consecutive CBFB break-apart FISH tests, 262 (16.1%) were positive and 271 CBFB rearrangements were identified in total, but five cases were discordant between FISH and RT-PCR because the partner was a NEW gene (n=3), because the rearrangement was an insertion (n=1), or because the CBFB-MYH11 variant was rare (n=1); eight further cases carried a 3-prime CBFB deletion, and 133 (8.2%) showed other atypical CBFB signal patterns including copy-number changes. 'inv(16) produces the in-frame CBFB-MYH11 mRNA joining CBFbeta residue 165 to the SMMHC tail' is true of the common case and wrong often enough that a pipeline resting on one RT-PCR primer set mis-assigns a measurable fraction of patients - the same failure mode as M-bcr-only assays in BCR::ABL1. Yang RK et al. 2021
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Isothermal titration calorimetry shows CBFbeta-SMMHC binds the Runx1 Runt domain with HIGHER affinity than native CBFbeta, and NMR identifies contributing interactions in both the CBFbeta portion and the SMMHC coiled-coil domain; this higher affinity explains the dominant-negative phenotype of the CBFB-MYH11 knock-in.

   Altered affinity of CBF beta-SMMHC for Runx1 explains its role in leukemogenesis · Lukasik SM, Zhang L, Corpora T, Tomanicek S, Li Y, Kundu M, Hartman K, Liu PP, Laue TM, Biltonen RL, Speck NA, Bushweller JH · Nat Struct Biol 2002
   PMID 12172539DOI
   ▸What this evidence cannot show

   In vitro biophysics on purified recombinant fragments, not full-length proteins in cells; the abstract reports higher affinity without our having verified the exact Kd values, so the graph states the direction of the change rather than a number. Whether the affinity difference alone is sufficient to explain sequestration in vivo, as opposed to acting together with coiled-coil-driven multimerisation, is not separable from these data.

   ▸Evidence that argues against this (1)

   • 'Higher affinity therefore dominant-negative' is the wrong end state. inv(16) ME-1 cells REQUIRE wild-type RUNX1 to survive: RUNX1 knockdown attenuates the mitotic checkpoint and causes apoptosis, and an active RUNX1 allele is usually preserved in inv(16) and t(8;21) patients while RUNX1 is frequently inactivated in other leukaemias. Higher-affinity capture sets a balance the leukaemia depends on; it does not abolish native RUNX1 function. Ben-Ami O et al. 2013
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Mouse embryos heterozygous for a Cbfb-MYH11 knock-in lack definitive haematopoiesis, phenocopying Cbfb-null embryos; CBFB is expressed in midgestation HSC/progenitor-enriched populations and that population is absent in Cbfb-MYH11 heterozygotes, showing the fusion blocks haematopoiesis at the stem-progenitor level; inv(16) is found in almost all AML M4Eo, and CBFB or CBFalpha2 rearrangements occur in more than 30% of AML.

   Role of Cbfb in hematopoiesis and perturbations resulting from expression of the leukemogenic fusion gene Cbfb-MYH11 · Kundu M, Chen A, Anderson S, Kirby M, Xu L, Castilla LH, Bodine D, Liu PP · Blood 2002
   PMID 12239155DOI
   ▸What this evidence cannot show

   Mouse embryonic phenotypes; the embryonic block is a developmental readout of dominant-negative activity, not a model of the human leukaemia, which requires additional postnatal cooperating events (the knock-in embryos die rather than develop AML). The >30% figure covers both CBF fusions combined and is quoted from the introduction rather than measured in this study.

   ▸Evidence that argues against this (1)

   • The same knock-in does not produce leukaemia by itself, which bounds what the embryonic phenotype licenses. Castilla et al. showed that chimeric mice expressing Cbfb-MYH11 have blocked myeloid differentiation and are PREDISPOSED to acute myelomonocytic leukaemia - but leukaemia required ENU mutagenesis. A complete block of definitive haematopoiesis in heterozygous embryos and a requirement for a second mutagenic hit in adult mice are different dose-and-context regimes, and the embryonic result therefore does not support the inference that the fusion alone drives human AML. It also cuts against reading the fusion as a straightforward dominant-negative of CBFB: a true phenocopy of Cbfb-null in the embryo coexists with a merely predisposing, second-hit-dependent phenotype in the adult marrow, and the entry's claim reports only the first. Castilla LH et al. 1999
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   CBFbeta-SMMHC outcompetes wild-type CBFbeta for RUNX1 binding, deregulates RUNX1 activity and induces AML; the protein-protein interaction inhibitor AI-10-49 selectively binds CBFbeta-SMMHC, disrupts its binding to RUNX1, restores RUNX1 transcriptional activity, has favourable pharmacokinetics, delays leukaemia progression in mice and triggers selective cell death in primary inv(16) AML patient blasts; current cytotoxic therapy gives a good initial response but limited long-term survival.

   Chemical biology. A small-molecule inhibitor of the aberrant transcription factor CBFbeta-SMMHC delays leukemia in mice · Illendula A, Pulikkan JA, Zong H, Grembecka J, Xue L, Sen S, Zhou Y, Boulton A, Kuntimaddi A, Gao Y, Rajewski RA, Guzman ML, Castilla LH, Bushweller JH · Science 2015
   PMID 25678665DOI
   ▸What this evidence cannot show

   Entirely preclinical: mouse survival was DELAYED, not cured, and the patient-blast data are ex vivo. AI-10-49 has not become an approved drug in the decade since publication, which is itself evidence about the difficulty of developing PPI inhibitors. The 'limited long-term survival' framing is the authors' motivation statement and is a harsher characterisation of favourable-risk inv(16) AML than most clinical series would give.

   ▸Evidence that argues against this (2)

   • Eleven years after the Science paper, a ClinicalTrials.gov API query for CBFB-SMMHC and for AI-10-49 returns no registered interventional study of any CBFbeta-SMMHC:RUNX1 interface inhibitor. This is the single most important number for the beat that uses inv(16) as its proof that obligate partner interfaces are druggable: validated biology, good chemistry, favourable pharmacokinetics, zero trials. US National Library of Medicine 2026
   • The therapeutic logic also needs restating. 'Restores RUNX1 transcriptional activity' reads as relieving a block, but RUNX1 is required for the survival of AML1-ETO and MLL-AF9 cells through BCL2, RUNX2 can compensate for its loss, and the authors conclude that inhibiting rather than enhancing RUNX1 activity may be the tractable direction in fusion-driven AML. Goyama S et al. 2013
5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   The RUNX1 Runt domain and the CBFB core domain form a DNA-induced molecular clamp in which CBFB makes no DNA contact but binds a Runt-domain face opposite the DNA-contacting loops (PDB 1H9D, 1E50; 1CMO is a solution-NMR structure of the isolated Runt domain, not of the heterodimer), and the heterodimer can act combinatorially with a C/EBPbeta bZIP homodimer at the CSF1R promoter (PDB 1IO4).

   PDB entries 1H9D (Aml1/cbf-beta/dna complex - 'The Leukemia-Associated AML1 (RUNX1)-CBFbeta Complex Functions as a DNA-Induced Molecular Clamp'), 1E50 (AML1/CBFbeta complex), 1CMO (Runt-domain immunoglobulin motif DNA recognition and heterodimerization - SOLUTION NMR, Runt domain alone, no CBFB and no DNA) and 1IO4 (RUNX1 Runt-CBFbeta heterodimer with C/EBPbeta bZIP on the CSF1R promoter) · RCSB Protein Data Bank · RCSB PDB 2001
   ▸What this evidence cannot show

   Verified by fetching each RCSB entry record. All entries are of WILD-TYPE CBFB with the Runt domain, not of the CBFB::MYH11 fusion - there is no structure of the fusion, and no structure of AI-10-49 bound to it among these entries. The description of the interface geometry relative to the DNA-contacting surface is drawn from the entry titles and the standard structural literature on this complex.

   ▸Evidence that argues against this (2)

   • One of the three listed clamp entries is not a clamp structure. Fetched from RCSB on 2026-09-19, 1CMO is a SOLUTION NMR structure with a single 127-residue polymer entity mapping to UniProt Q01196 (RUNX1) and zero non-polymer entities - no CBFB chain, no DNA. It shows the isolated Runt domain in solution and cannot evidence a heterodimer in which 'CBFB makes no DNA contact', because neither CBFB nor DNA is present. The claim survives on 1H9D and 1E50 (and, for the combinatorial C/EBPbeta complex, 1IO4), which are genuine crystal structures of the assembly; the claim text and the PDB note have been corrected. The identical error in fusion-runx1-runx1t1#E3 is corrected there. RCSB Protein Data Bank 1999
   • The clamp metaphor itself has been superseded, and the replacement matters for this node's affinity argument. Yan et al. showed by 15N backbone relaxation that the Runt domain's S-switch region undergoes microsecond-to-millisecond conformational exchange in the Runt-DNA complex, that the DNA does too, and that CBFB binding QUENCHES that exchange in several regions of the Runt domain making energetically important DNA contacts. CBFB therefore raises DNA affinity by shifting a pre-existing dynamic conformational equilibrium, not by mechanically clamping a rigid object around DNA. That reframing is load-bearing here: if the interface is an equilibrium rather than a geometry, then CBFbeta-SMMHC's measured affinity increase [E2] and AI-10-49's ability to reverse it [E4] are thermodynamic events at a dynamic surface, which is a different design problem from occluding a fixed contact patch. Yan J et al. 2004
6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   Structural coverage of the retained MYH11 module is limited to the myosin motor domain and related smooth-muscle myosin structures, not the coiled-coil tail that the fusion actually keeps: an RCSB search on MYH11 (P35749) returns a single entry, 9FU2, 'Smooth Muscle Myosin II in complex with MT-228'.

   RCSB PDB: search by UniProt accession P35749 (MYH11) returns one entry, 9FU2 (Smooth Muscle Myosin II in complex with MT-228); Q13951 (CBFB) structures appear within the RUNX1 heterodimer entries · RCSB Protein Data Bank · RCSB PDB 2026
   ▸What this evidence cannot show

   Verified by querying the RCSB search API on the reference-sequence accession field and fetching the entry record. 9FU2 covers the MOTOR domain region, which the fusion DELETES, so it is listed as a negative result: the coiled-coil tail that the fusion retains and that drives its multimerisation has no dedicated experimental structure in this search. Coiled coils are structurally well understood generically, so this is a gap in specific coordinates rather than in understanding.

   ▸Evidence that argues against this (1)

   • The absence is real, and it is accession-specific rather than fold-specific - which is worth stating because the entry's phrasing invites the stronger reading. A UniProt cross-reference query on P35749 (MYH11) run on 2026-09-19 returns exactly one PDB entry, 9FU2 ('Smooth Muscle Myosin II in complex with MT-228', X-ray, 2.58 A), confirming that the retained coiled-coil tail has no coordinates under this accession. What would falsify the stronger reading is a myosin-rod structure, and those exist under other accessions: Blankenfeldt et al. solved the 126 N-terminal residues of human cardiac beta-myosin II subfragment 2, wild-type and the disease-associated E924K mutant, as a straight parallel dimeric coiled coil, and described local asymmetry, axial stagger from bulky non-canonical a/d side chains, and three rings of negative surface potential. Nobody has solved the SMMHC segment the fusion actually keeps, but the fold class is characterised and a homology model has a real template. Blankenfeldt W et al. 2006
7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   inv(16) and t(8;21) leukaemia cells are addicted to native RUNX1, and no registered interventional trial of any CBFbeta-SMMHC:RUNX1 interface inhibitor exists eleven years after AI-10-49.

   Addiction of t(8;21) and inv(16) acute myeloid leukemia to native RUNX1. · Ben-Ami O, Friedman D, Leshkowitz D, et al. · Cell Reports 2013
   PMID 24055056DOI
   ▸What this evidence cannot show

   Abstract only for PMID 24055056 (two cell lines plus an allele-preservation inference). The trial-absence half is a ClinicalTrials.gov v2 API query run 2026-09-19 (query.term=CBFB-SMMHC, query.term=AI-10-49) and shows no registered trial, not the absence of any programme.

   ▸Evidence that argues against this (1)

   • The absence of a trial is real; the preclinical reason for it is visible in the chemistry's own follow-up, and 'nobody developed it' is the wrong explanation. Pulikkan et al. showed that pharmacological disruption of the CBFbeta-SMMHC/RUNX1 interaction works by letting RUNX1 rebind three MYC distal enhancers, where it replaces the SWI/SNF component BRG1 with the Polycomb component RING1B, represses MYC and triggers apoptosis - a clean, on-mechanism account. But ELIMINATING inv(16) leukaemia in human cells and in a mouse model required COMBINING the interface inhibitor with the BET inhibitor JQ1. Single-agent interface disruption was mechanistically correct and therapeutically insufficient. That is a direct warning for the lab's menin-interface programme: an obligate, correctly targeted partner interface can still need a chromatin-directed partner drug before it clears disease, and a registry search for single-agent interface trials will under-report the field's actual position. Pulikkan JA et al. 2018

• ← refutesCBFB — domain architecture and residue coordinates

  Counter-evidence added cycle 41: the four-transcript picture is an undercount. 162 patients gave ten CBFB-MYH11 transcript types, 79.0% type A and 21.0% across nine rare types enriched for therapy-related AML with atypical cytomorphology and low white cell count (Schnittger 2007, PMID 17287858). Also, PDB 1CL3 - called a crystal structure in this entry until 2026-09-19 - is a 25-model solution NMR ensemble.

• ← related toMYH11 — domain architecture and residue coordinates

  Counter-evidence added cycle 41: the multimerisation-causes-inhibition chain is confounded with nuclear localisation. ACD-deficient CBFbeta-SMMHC variants are predominantly CYTOPLASMIC, and forcing deltaACD into the nucleus with an SV40 NLS still only mildly inhibited CBF activities (Kummalue 2002, PMID 12417730). Items E4 and E5 downgraded strong -> moderate.

• ← related toCBFB — domain architecture and residue coordinates

  Coordinate correction (PDB 8CX0 and 6P59 observed residues): CBFB is ordered to residue 157-160 in deposited complexes and Pfam PF02312 models the family domain to 166, against the stored fold end of 141. The CBFB moiety retained by the fusion (1-165) is therefore essentially the entire ordered protein, not a fold plus a long disordered tail.

• ← part ofCBFB — domain architecture and residue coordinates

  5-prime partner domain architecture; the Runt-binding core is retained, the last 17 residues replaced

• ← part ofMYH11 — domain architecture and residue coordinates

  3-prime partner domain architecture; the coiled-coil rod is the retained multimerisation element

• ← related toHypothesis — the menin lesson generalises: every non-enzymatic fusion has an obligate partner interface that is more druggable than the fusion

  Translational-ceiling measurement (ClinicalTrials.gov API, queried 2026-09-19): no registered interventional study of any CBFbeta-SMMHC:RUNX1 interface inhibitor exists, eleven years after AI-10-49. The closest replica of the menin story has produced zero trials.

• ← related toMechanism class 2 — transcription-factor corruption (RUNX1::RUNX1T1, PML::RARA, ETV6::RUNX1, CBFB::MYH11) — a balance, not a pure dominant negative

  Same correction applied to the inv(16) row, and to the therapeutic logic behind AI-10-49: ME-1 cells die on RUNX1 knockdown, so 'restoring RUNX1 activity' is tuning a set point in a system that needs native RUNX1, not relieving a blockade.

• related to →Hypothesis — the menin lesson generalises: every non-enzymatic fusion has an obligate partner interface that is more druggable than the fusion

  Translational-ceiling measurement plus mechanism correction: no registered trial of any CBFbeta-SMMHC:RUNX1 inhibitor exists 11 years after AI-10-49 (ClinicalTrials.gov, 2026-09-19), and inv(16) cells REQUIRE native RUNX1 (Ben-Ami 2013, PMID 24055056), so 'restoring RUNX1' is tuning a ratio rather than lifting a block.

• ← related toAnalysis (planned) — comparative druggability scan of every obligate fusion-partner interface in the atlas

  Positive control never reached the clinic (NCT05821738 check, 2026-09-19): no CBFB-SMMHC interface study is registered, while a phase 2 of avapritinib against the acquired KIT co-lesion has run since 2022-06-01.

• related to →Step 5: molecular confirmation and naming the partner

  Assay-uniformity bound (Yang 2021, PMID 34771519): among 1629 CBFB break-apart FISH tests, 271 CBFB rearrangements included three with a NEW partner gene, one insertion, one rare variant and eight 3-prime CBFB deletions, so a single RT-PCR primer set mis-assigns a measurable fraction.

• related to →Which step is rate-limiting, and what compensates when you block it — the bite-back map for the causal chain

  Sufficiency bound (Castilla 1999, PMID 10508507): Cbfb-MYH11 knock-in chimeric mice are only PREDISPOSED to acute myelomonocytic leukaemia and required ENU mutagenesis to develop it, so the embryonic block does not show the fusion alone drives disease.

• related to →Dataset: PDB and AlphaFold - the structural census, including the two absences that shape every plan

  Structural correction (RCSB 1CMO, fetched 2026-09-19): 1CMO is a SOLUTION NMR structure of a single 127-residue RUNX1 Runt domain with no CBFB and no DNA, so it cannot evidence the clamp. Entry corrected.

• related to →Menin is the best handle on KMT2A::AFF1 because of where the break falls, not because menin is biologically special

  Mechanism reframing (Yan 2004, PMID 15322525): CBFbeta raises Runt-domain DNA affinity by quenching a pre-existing microsecond-to-millisecond conformational exchange, so the CBF interface is a dynamic equilibrium rather than a rigid clamp geometry.

• related to →49 of 91 proteins in this graph have never had a drug-like molecule bound to them

  Absence bound (Blankenfeldt 2006, PMID 17095604): MYH11 (P35749) maps to exactly one PDB entry, 9FU2, but human cardiac beta-myosin II S2 coiled-coil structures exist under other accessions, so the retained SMMHC rod is unsolved yet modellable.

• related to →Hypothesis: the fusion output program fails at a threshold on total cut-set suppression, so three nodes at 40% beat one node at 90%

  Sufficiency bound on interface disruption (Pulikkan 2018, PMID 29958106): eliminating inv(16) leukaemia required combining the CBFbeta-SMMHC/RUNX1 interface inhibitor with the BET inhibitor JQ1; single-agent interface disruption was on-mechanism but insufficient.

• ← analogous toRUNX1::RUNX1T1 (AML1-ETO), t(8;21)(q22;q22)

  THE TWO CORE-BINDING-FACTOR LEUKAEMIAS, disrupting the same transcription factor complex from its two subunits. The canonical example in this graph of one complex broken two ways giving one disease class, which is the comparison the KMT2A partner families are measured against.

• depends on →Mechanism class 2 — transcription-factor corruption (RUNX1::RUNX1T1, PML::RARA, ETV6::RUNX1, CBFB::MYH11) — a balance, not a pure dominant negative
• ← analogous toBCAT1 in core-binding-factor AML - the closest analogue in another fusion family, where TWO different fusions converge on one metabolic enzyme by rewiring the chromatin at its locus

  CBFB::MYH11 directly activates BCAT1 by rewiring chromatin at its locus

• ← analogous toIn inv(16), disrupting the fusion interface swaps BRG1 for RING1B at three MYC enhancers - the cleanest version anywhere of the experiment nobody has run on menin

  the source system: AI-10-49 interface disruption releases RUNX1 onto three MYC enhancers

• ← related toThe fusion-oncoprotein TurboID experiment this graph proposes has already been published - for PML::RARA, RUNX1::RUNX1T1, CBFB::MYH11, NUP98::HOXA9 and SET::NUP214, and not for any KMT2A fusion

  CBFB::MYH11-TurboID recovered a largely cytoplasmic neighbourhood and cytoplasmic RUNX1 sequestration

• ← analogous toRUNX2 - a direct KMT2A-chimera target and dependency in T-ALL, a patient-specific enhancer target in KMT2A::AFF1 B-ALL, and the gene a menin inhibitor represses to restore steroid sensitivity in t(4;11) infants

  the RUNT-CBFB heterodimerisation interface targeted pharmacologically in KMT2A-R T-ALL is the interface that CBFB::MYH11 subverts

• ← analogous toCBFA2T3::GLIS2, inv(16)(p13.3q24.3)

  AXIS: SHARED NOMENCLATURE AND NOTHING ELSE - and that is the point. Both lesions are written inv(16); the breakpoints are p13.3q24.3 against p13.1q22, the genes are unrelated, and the risk assignments are opposite - CBFB::MYH11 is favourable risk in ELN 2022, CBFA2T3::GLIS2 reports 15-30% overall survival. The failure mode is not misreading one inversion as the other but that CBFA2T3::GLIS2 is CRYPTIC and invisible to karyotyping, so the only inv(16) string in the record belongs to the favourable lesion, and the direction of the error is de-escalation of therapy in infant AMKL. A nomenclatural analogy predicts nothing about biology and everything about data handling.