Oncofusion

start / graph / fusion-bcr-abl1

Fusionestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q11-comparative

BCR::ABL1 (Philadelphia chromosome), t(9;22)(q34;q11)

also called BCR-ABL · BCR-ABL1 · Ph chromosome · p210 BCR-ABL · p190 BCR-ABL · Ph+ ALL · CML driver

A swap between chromosomes 9 and 22 welds a self-clumping piece of one protein onto an enzyme that normally tells blood cells to divide, and the weld jams the enzyme permanently on. Because that enzyme has a real pocket where a drug can sit, a pill called imatinib turned a fatal leukaemia into a manageable condition.

The same thing, for a scientist

t(9;22)(q34;q11) joins BCR (P11274) exon 1, 13/14 or 19 to ABL1 (P00519) exon 2, producing p190, p210 or p230 chimeras. The fusion always keeps BCR's N-terminal coiled-coil oligomerisation domain and always keeps the entire ABL1 tyrosine kinase domain, SH2 and SH3, while deleting ABL1 exon 1a/1b including the N-terminal cap and the myristoylation site that normally autoinhibits the kinase by docking into the C-lobe myristate pocket [E1][E2]. BCR-driven oligomerisation plus loss of the cap gives constitutive, ligand-independent kinase activity; p210 and p190 differ measurably in kinase output and transforming potency, and that difference tracks with their different clinical diseases [E2]. Because the driver is a **classical enzyme with a druggable ATP site that adopts a distinctive DFG-out inactive conformation**, imatinib achieves both potency and selectivity [E3], converting chronic-phase CML into a chronic illness (10-year OS 83.3% on IRIS) [E4] and lifting paediatric Ph+ ALL from ~20-30% cure to ~70% 5-year DFS with chemotherapy plus imatinib and no transplant advantage [E5].

What it looks like
loading 1K1F…
PDB 1K1F ↗
shape

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

I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
What is kept and what is lost

Kept from BCR (5-prime partner): the N-terminal coiled-coil oligomerisation domain (residues ~1-72; PDB 1K1F), which drives the fusion into dimers and tetramers 6. Longer breaks (b2a2/b3a2) additionally retain the BCR Ser/Thr kinase-like region, the RhoGEF/DH-PH module and the GRB2-binding phosphotyrosine 177.

Lost from BCR: the C-terminal GAP-related domain.

Kept from ABL1 (3-prime partner): the SH3 domain, SH2 domain and the entire tyrosine kinase (SH1) catalytic domain, plus the C-terminal actin-binding/last-exon region.

Lost from ABL1 — and this is the key deletion: ABL1 exon 1a/1b, encoding the N-terminal cap and, on the 1b isoform, the myristoylation site. In autoinhibited ABL1 the myristate inserts into a hydrophobic pocket in the kinase C-lobe and bends the alpha-I helix, clamping the SH3-SH2 unit against the kinase domain. Truncating exon 1 removes the myristate and dismantles that latch 12.

So BCR::ABL1 is a double activation: the brake (the cap/myristoyl latch) is deleted and an accelerator (forced oligomerisation, driving trans-autophosphorylation of the activation loop) is added. p210 has higher tyrosine kinase activity and transformation potency than p190 in matched assays, so the amount of BCR contributed is not cosmetic 2.

Mechanism: this is the constitutive-kinase archetype

The mechanistic class is constitutive kinase — the only one of the eleven comparative fusions in this beat whose driver is an enzyme with a small-molecule-accessible catalytic pocket (the NUP98::KDM5A demethylase and NUP98::NSD1 methyltransferase retain catalytic domains, but in both the leukaemogenic function runs through chromatin reading and condensate formation rather than through the catalytic site, and in NUP98::NSD1 the SET domain is required in a way that is still debated).

Downstream, constitutive ABL1 activity fires RAS/MAPK, PI3K/AKT, JAK/STAT5 and CRKL/paxillin cytoskeletal signalling. Y177 in the retained BCR region recruits GRB2 and is a major route into RAS. But the actionable insight was never the downstream network; it was the pocket.

Why this fusion is the comparative anchor for the whole beat: the graph's other ten fusions are transcription-factor corruptions, chromatin hijacks, condensate rewirings or enhancer hijacks. None has an enzyme active site whose occupancy shuts off the oncogenic program. BCR::ABL1 is the control case that tells you how much of the difference in therapeutic progress is about molecular class rather than about effort or funding.

WHY THIS ONE BECAME CURABLE (PI depth request)

Four structural facts, in order of importance:

1. The retained module is a kinase with an ATP pocket, i.e. a classical enzyme active site. Unlike a protein-protein interface, an ATP site is a deep, enclosed, largely hydrophobic cleft evolved to bind a small molecule with high affinity. Medicinal chemistry has a century of precedent for competing with a small natural ligand in a deep pocket; it has almost none for occluding a flat, extended protein-protein interface, which is exactly what ETV6::RUNX1, RUNX1::RUNX1T1 and CBFB::MYH11 present.

2. ABL1's inactive conformation is idiosyncratic. Imatinib does not bind active ABL1. The crystal structure of the ABL1 catalytic domain with an STI-571 variant showed that the kinase must first adopt an inactive conformation in which the activation loop is unphosphorylated and folded in a way distinct both from active kinases and from inactive Src — a DFG-out conformation with the Phe of the DFG motif flipped into the ATP site 3. Imatinib is a type II inhibitor that reaches past the gatekeeper Thr315 into the adjacent hydrophobic back pocket created by that flip. This is the single most important point in the whole node: selectivity came from conformation, not from the ATP site itself. The Schindler paper's own conclusion is the generalisable lesson — 'compounds that exploit the distinctive inactivation mechanisms of individual protein kinases can achieve both high affinity and high specificity' 3. Structures 2HYY (human ABL1) and 1IEP (murine c-Abl) show the bound complex 7.

3. The kinase activity is not just a marker, it is the requirement. Kinase-dead BCR-ABL does not transform; kinase output correlates quantitatively with transforming potency across p190/p210 2. So occupying the pocket is not a partial intervention — it removes the driver function.

4. The leukaemia is genetically simple at diagnosis, especially chronic-phase CML. A single driver in a disease with few cooperating lesions means there is little pre-existing machinery to route around the block. Ph+ ALL, which carries a heavier secondary lesion burden (IKZF1 deletions, CDKN2A loss), responds less completely — a useful internal control showing that even an ideal target is only as good as the genetic context.

What did NOT make it curable: the BCR coiled-coil oligomerisation interface (PDB 1K1F) was an obvious second target and has never yielded a drug 6. That is consistent with the rest of this beat: the interaction module resisted, the enzyme did not.

Comparison with PML::RARA, the other curable fusion: PML::RARA was tractable for a different structural reason — it retains a nuclear-receptor ligand-binding pocket (a natural small-molecule site, hence ATRA) and a PML RBCC/coiled-coil that forces the fusion into degradable multimers (hence arsenic-triggered degradation). BCR::ABL1 was inhibited; PML::RARA was degraded and differentiated. Two curable fusions, two distinct structural reasons, and in both cases the reason is the presence of a natural small-molecule-binding pocket on the retained oncogenic module. None of the other nine fusions in this beat has one.

Clinical phenotype, age and prognosis

Two diseases, set largely by the breakpoint:

  • p210 (e13a2/b2a2, e14a2/b3a2) → chronic myeloid leukaemia, a stem-cell disease with a chronic phase, accelerated phase and blast crisis. Rare in children (CML is <3% of paediatric leukaemia); median adult age ~55-65.
  • p190 (e1a2) → predominantly Ph+ B-cell precursor ALL, the form relevant to a paediatric fusion atlas. Ph+ ALL is ~3-5% of childhood ALL, rising with age to ~25% of adult ALL.

Prognosis, CML: on the IRIS trial with median 10.9 years of follow-up, estimated 10-year overall survival on first-line imatinib was 83.3%, with 82.8% achieving complete cytogenetic response and no unacceptable late cumulative toxicity 4.

Prognosis, paediatric Ph+ ALL: historically one of the worst paediatric ALL subtypes — only 20-30% cured with chemotherapy alone, 60% with matched allogeneic transplant 8. Adding continuous imatinib to intensive chemotherapy on COG AALL0031 gave 5-year disease-free survival of 70% +/- 12% on chemotherapy plus imatinib, statistically indistinguishable from sibling-donor (65% +/- 11%) or unrelated-donor (59% +/- 15%) transplant, i.e. no transplant advantage 5. Additional cytogenetic abnormalities worsened outcome (p = 0.05) 5.

That single result is the paediatric point of this node: a small molecule aimed at a fusion's active site replaced allogeneic transplantation in a childhood leukaemia. Nothing comparable has happened for any of the other ten fusions in this beat.

Druggability: clinically validated, five generations deep

Status: clinically validated. Imatinib (2001), then dasatinib, nilotinib, bosutinib, radotinib; ponatinib for the resistant gatekeeper mutant; asciminib as an allosteric agent binding the myristate pocket — the very pocket the fusion vacated by deleting ABL1 exon 1, which is an unusually elegant example of drugging a lesion's structural consequence rather than its retained function.

Resistance is overwhelmingly on-target kinase-domain mutation, and the archetype is T315I, the gatekeeper threonine whose side-chain hydroxyl hydrogen-bonds imatinib; substituting isoleucine both removes that bond and sterically blocks the back pocket. Ponatinib was designed around it: in the PACE phase 2 trial, 70% of T315I chronic-phase patients achieved a major cytogenetic response versus 51% of those resistant to dasatinib/nilotinib, and no single mutation conferring ponatinib resistance was found 9.

But the price is visible: serious arterial thrombotic events in 9% of ponatinib patients, treatment-related in 3% 9. The lesson for the rest of the beat is that even the best-validated fusion target ends up with a fifth-generation drug carrying vascular toxicity — target validation does not exempt you from tolerability.

Correction 2026-09-19: 1IEP is murine, not human

Evidence item E7 previously read: "Crystal structures of the human ABL1 kinase domain bound to imatinib exist (PDB 1IEP, 2HYY)". The word human is withdrawn for 1IEP.

Fetched from RCSB on 2026-09-19, PDB 1IEP has a single polymer entity described as PROTO-ONCOGENE TYROSINE-PROTEIN KINASE ABL, source organism Mus musculus, UniProt P00520, 293 residues, X-ray at 2.1 A. 2HYY is the human protein: Proto-oncogene tyrosine-protein kinase ABL1, Homo sapiens, UniProt P00519, 273 residues, 2.4 A.

What survives: the type-II, DFG-out binding mode and the reach past gatekeeper Thr315. The murine and human ABL kinase domains are effectively identical across the imatinib site, so no mechanistic conclusion in this node depends on the error.

What does not survive: the provenance statement. This is the same defect the graph has caught before in accession-assembled structure lists - a PDB id is collected because it maps to the right protein family, and the organism, the construct boundaries and the experimental method are never re-read. The correction is recorded here rather than only in counter_evidence so that the entry does not assert 'human 1IEP' in one place while denying it in another.

Analogy: the resistance archetype for menin inhibition, and the three places it stops being one

Closest analogue, and the axis. fusion-kmt2a-aff1, on the axis of shared clinical behaviour under a targeted drug - specifically, acquired resistance by missense mutation at the drug-binding interface that spares the pathogenic function. This is the comparison the menin-inhibitor field reaches for by default, and the point of writing it down as an analogy is to record where it holds and where it silently stops.

Where it holds, and it holds remarkably well. Imatinib resistance came back as substitution of the ABL1 threonine that forms the critical hydrogen bond with the drug, sufficient on its own in a reconstitution experiment, with BCR-ABL signalling reactivated in every resistant case 10. Revumenib resistance came back as somatic MEN1 mutations at the revumenib-menin interface, recurring in xenografts and in an unbiased base-editor screen, attenuating drug binding while leaving the natural menin-MLL1 interaction intact and restoring chromatin occupancy 11. Two different drug classes, two different kinds of target - an enzyme active site and a protein-protein interface - and the same escape geometry: mutate the drug contact, keep the biology. As a prediction generated by the analogy and then tested, this is about as clean as this beat gets, and it is the strongest available argument that menin inhibition is a genuine on-target occupancy drug rather than a pleiotropic chromatin effect.

Where it breaks. 1: target ownership. T315I is a mutation in the oncoprotein. MEN1 interface mutations are in a wild-type host gene the leukaemia did not create - a gene whose germline loss causes multiple endocrine neoplasia type 1 and which every normal cell needs. Two consequences follow that have no counterpart in CML. A next-generation compound must be designed against a mutant form of a normal protein, so the therapeutic-index arithmetic is different from designing against a mutant oncoprotein that exists only in the tumour. And the resistance allele is not a fusion-specific handle: nothing about a MEN1 escape mutation distinguishes a leukaemic cell from a parathyroid cell carrying the same allele. The drug target here is a cofactor, not the driver.

Where it breaks. 2: resistance that precedes the drug. Imatinib resistance is selected by imatinib. Menin-inhibitor resistance need not be: heavily pretreated KMT2A::AFF1 ALL samples were markedly less menin-sensitive than earlier samples from the same patients with no prior exposure to the class, and with on-target transcriptional activity intact in the resistant cells (counter-evidence on 11). That is a different clinical object - not escape under selection but a pre-existing state - and it argues for giving the class early, which is a recommendation the CML analogy would never produce.

Where it breaks. 3: there is no orthogonal site to retreat to. The field's actual escape from ATP-site resistance was not more chemotypes at the same site but a structurally distinct allosteric site on the same oncoprotein - the ABL1 myristoyl pocket, now asciminib. KMT2A::AFF1 contains no catalytic domain at all; the KMT2A SET methyltransferase is always lost at the breakpoint, so there is no second pocket on the fusion to design against. concept-why-bcr-abl1-became-curable works this through; the analogy section records it here because it is the reason the resistance parallel should not be read as a roadmap.

And the larger disanalogy this entry exists for. BCR::ABL1 is the archetype of the constitutive kinase class: a fusion whose driver is an enzyme with a deep, enclosed, conformationally exploitable active site, where the drug hits the driver directly. KMT2A fusions are the chromatin class: the driver has no active site, the drug hits a cofactor that tethers it, and the pathogenic function survives the drug's own resistance mutations. Treating BCR::ABL1 as the template for what a fusion oncoprotein is has been costly in this field, and the honest summary is that it transfers a method - define the non-negotiable molecular function, find a structured surface that executes it, design against that surface - and not a target class.

Facts and figures
gene5BCR
gene3ABL1
cytobandt(9;22)(q34;q11)
frequency notecml_pct is 100 by definition (BCR::ABL1 is the defining lesion of CML in current WHO criteria). The paediatric B-ALL figure of ~4% and the adult ~25% are standard textbook ranges consistent with [E8]; they are not single-cohort measurements, and the adult figure rises monotonically with age.
prognosisTwo very different numbers. CML: estimated 10-year overall survival 83.3% on first-line imatinib with 82.8% complete cytogenetic response (IRIS, median 10.9 y follow-up) [E4]. Paediatric Ph+ ALL: 5-year disease-free survival 70% +/- 12% with chemotherapy plus imatinib, versus a pre-TKI historical 20-30% with chemotherapy alone, and with no advantage for allogeneic transplant [E5][E8].
uniprot5P11274
uniprot3P00519
PDB noteNo structure of intact BCR::ABL1 exists. 1K1F is the isolated BCR N-terminal coiled-coil oligomerisation domain (the retained 5-prime module). 1IEP (murine c-Abl, UniProt P00520) and 2HYY (human ABL1, UniProt P00519) are the ABL1 kinase domain with imatinib and show the type II DFG-out binding mode that makes the drug selective. 1OPJ is autoinhibited c-Abl and is included because it shows the SH3-SH2-clamp and myristoyl latch that the fusion deletes — the structural basis for constitutive activation. Every entry is a fragment.
druggability classclinically-validated
stage of hematopoiesisBreakpoint-dependent. p210/CML arises in a multipotent bone-marrow haematopoietic stem cell (CD34+CD38-), which is why CML is a myeloproliferative stem-cell disease capable of lymphoid blast crisis and why TKIs suppress but rarely eradicate the leukaemic stem cell. p190/Ph+ ALL behaves as a committed B-lineage (pro-B/pre-B, CD19+CD34+) disease. This node ranks by the CML stem-cell origin because that is the more primitive of the two.
age peakCML: median 55-65 y in adults, <3% of pediatric leukemia. Pediatric Ph+ ALL: incidence rises with age across childhood, peaking in adolescence (10-18 y), unlike ETV6::RUNX1
developmental order rank2
frequency
pediatric_leukemia_cml_sh…3.000pediatric_leukemia_cml_share_pct: 3pediatric_b_all_pct4.000pediatric_b_all_pct: 4adult_b_all_pct25.0adult_b_all_pct: 25cml_pct100.0cml_pct: 100
breakpoints
bp-bcr-e1-abl1-exon2-p1901.000bp-bcr-e1-abl1-exon2-p190: 1bp-bcr-e13e14-abl1-exon2-…14.0bp-bcr-e13e14-abl1-exon2-p210: 14bp-bcr-e19-abl1-exon2-p23019.0bp-bcr-e19-abl1-exon2-p230: 19
diseases
[
  "chronic myeloid leukemia",
  "pediatric Ph-positive B-cell precursor ALL",
  "adult Ph-positive B-ALL",
  "Ph-positive AML (rare)",
  "CML blast crisis"
]
frequency — the numbers behind the chart
{
  "cml_pct": 100,
  "pediatric_b_all_pct": 4,
  "adult_b_all_pct": 25,
  "pediatric_leukemia_cml_share_pct": 3
}
breakpoints — the numbers behind the chart
[
  {
    "id": "bp-bcr-e13e14-abl1-exon2-p210",
    "gene5_exon_last": 14,
    "gene3_exon_first": 2,
    "frequency_pct": 95,
    "population": "chronic myeloid leukemia (M-bcr; e13a2/b2a2 and e14a2/b3a2 combined)",
    "domains_retained_5": [
      "N-terminal coiled-coil oligomerisation domain (aa 1-72)",
      "Ser/Thr kinase-like region",
      "GRB2-binding Y177",
      "RhoGEF DH-PH module"
    ],
    "domains_lost_5": [
      "C-terminal GAP-related domain"
    ],
    "domains_retained_3": [
      "SH3 domain",
      "SH2 domain",
      "tyrosine kinase (SH1) catalytic domain including the ATP pocket and gatekeeper Thr315",
      "C-terminal actin-binding region"
    ],
    "domains_lost_3": [
      "ABL1 exon 1a/1b N-terminal cap",
      "N-terminal myristoylation site (1b isoform) that autoinhibits by docking into the C-lobe myristate pocket"
    ],
    "phenotype_notes": "The p210 chimera; defines chronic myeloid leukaemia. Higher kinase activity and transformation potency than p190 in matched assays [E2]. The vacated myristate pocket is the binding site later exploited by the allosteric inhibitor asciminib."
  },
  {
    "id": "bp-bcr-e1-abl1-exon2-p190",
    "gene5_exon_last": 1,
    "gene3_exon_first": 2,
    "frequency_pct": 70,
    "population": "pediatric Ph-positive B-cell precursor ALL (m-bcr; e1a2)",
    "domains_retained_5": [
      "N-terminal coiled-coil oligomerisation domain (aa 1-72) only"
    ],
    "domains_lost_5": [
      "Ser/Thr kinase-like region",
      "GRB2-binding Y177",
      "RhoGEF DH-PH module",
      "GAP-related domain"
    ],
    "domains_retained_3": [
      "SH3 domain",
      "SH2 domain",
      "tyrosine kinase (SH1) catalytic domain",
      "C-terminal actin-binding region"
    ],
    "domains_lost_3": [
      "ABL1 exon 1a/1b N-terminal cap and myristoylation site"
    ],
    "phenotype_notes": "The p190 chimera; predominates in Ph+ ALL, especially paediatric. Lower kinase output than p210 but still fully transforming [E2]. The frequency_pct is the approximate share of transcript types within paediatric Ph+ ALL, not a share of all ALL; the p190/ALL mapping is a strong association rather than a rule."
  },
  {
    "id": "bp-bcr-e19-abl1-exon2-p230",
    "gene5_exon_last": 19,
    "gene3_exon_first": 2,
    "frequency_pct": null,
    "population": "chronic neutrophilic leukemia / indolent CML variant",
    "domains_retained_5": [
      "coiled-coil oligomerisation domain",
      "Ser/Thr kinase-like region",
      "Y177",
      "DH-PH module",
      "most of BCR"
    ],
    "domains_lost_5": [
      "extreme BCR C-terminus"
    ],
    "domains_retained_3": [
      "SH3",
      "SH2",
      "tyrosine kinase domain",
      "C-terminal actin-binding region"
    ],
    "domains_lost_3": [
      "ABL1 exon 1a/1b cap and myristoylation site"
    ],
    "phenotype_notes": "The mu-bcr p230 chimera, classically associated with a neutrophilic, thrombocytosis-predominant indolent phenotype. Rare; frequency left null because we did not find a cohort-level percentage we could verify.",
    "frequency_pct_note": "Left null deliberately: p230 is rare enough that published frequencies are case-series-derived and we could not verify a defensible number."
  }
]
PDB IDS
[
  "1K1F",
  "1IEP",
  "2HYY",
  "1OPJ"
]
mechanism class
[
  "constitutive-kinase"
]
How this could bite us back
  • On-target resistance is the rule, not the exception: kinase-domain point mutations, with T315I as the archetype, reliably emerge under single-agent pressure. A target being ideal does not prevent resistance; it just makes resistance predictable and sequenceable [E9].
  • Escalating potency escalates toxicity. Ponatinib solves T315I but caused serious arterial thrombotic events in 9% of patients at the original dose, treatment-related in 3%, with the rate rising on longer follow-up [E9]. Any fifth-generation agent for a fusion target should be assumed to carry an off-target vascular or cardiac liability until proven otherwise.
  • ABL1 has essential normal functions and the TKIs are not fusion-selective — they inhibit wild-type ABL1, ABL2, KIT, PDGFR and more. In a growing child this shows up as growth impairment and bone-mineral effects on long-term imatinib, a toxicity class with no adult analogue and one that matters for every paediatric fusion target.
  • Kinase inhibition is a suppressive, indefinite therapy, not a cure, for the stem-cell form of the disease; that means lifelong adherence, lifelong cost and lifelong late-effect accrual. Treatment-free remission succeeds in only a minority.
  • The uncomfortable comparative lesson: BCR::ABL1's success may say more about molecular class than about strategy. If the tractability came from having an enzyme active site plus a distinctive inactive conformation [E3], then the ten transcription-factor, chromatin and condensate fusions in this beat cannot be expected to follow the same route, and pursuing 'an imatinib for fusion X' where X has no pocket is a predictable dead end.
Still open
  1. TKIs suppress but do not eradicate the BCR::ABL1 leukaemic stem cell, and most patients relapse molecularly on discontinuation. What keeps that cell alive when its kinase is inhibited, and is the surviving state kinase-independent (i.e. would even a perfect inhibitor fail to cure)?
  2. Asciminib works by occupying the myristate pocket that the fusion vacated when ABL1 exon 1 was deleted. Is there a generalisable principle here — that a fusion's deletion creates an orphan pocket which is a better-than-average drug site precisely because the fusion no longer needs it? If so, which of the other ten fusions in this beat has an analogous vacated pocket?
  3. Why does p190 drive lymphoid and p210 myeloid disease when both retain the identical kinase? Is it kinase output, the extra BCR domains, or the cell of origin, and can this be settled with isogenic human progenitor models rather than retroviral overexpression?
  4. Paediatric Ph+ ALL still fails in ~30% of children despite an ideal target. How much of that residual failure is attributable to the cooperating-lesion burden (IKZF1, CDKN2A) rather than to inadequate kinase inhibition, and does that predict a ceiling for targeted therapy in genetically complex fusions?
  5. Does the analogy hold for the outcome as well as for the escape route? Imatinib converted a fatal disease into a chronic one because single-agent target occupancy is durable in chronic phase; against the same fusion in acute disease the same drug gave median survival of months. If menin inhibition behaves like the acute case rather than the chronic case, the T315I/MEN1 parallel is a parallel of mechanism and not of prognosis, and the discriminating measurement is duration of response on monotherapy in a treatment-naive KMT2A-rearranged patient - which no trial has yet reported because the class is given at relapse.
Every source behind this entry (11)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The Ph translocation places a variable number of BCR exons upstream of ABL1 in a chimaeric bcr/abl mRNA, with all Ph-positive patients breaking within a 5.8-kb breakpoint cluster region of BCR.

    Structural organization of the bcr gene and its role in the Ph' translocation · Heisterkamp N, Stam K, Groffen J, de Klein A, Grosveld G · Nature 1985
    PMID 2989703DOI
    What this evidence cannot show

    A 1985 restriction-mapping and cDNA-hybridisation study; the 'all patients examined to date' claim rested on a small series, and the subsequent recognition of the minor (e1a2/p190) and micro (e19a2/p230) breakpoint regions shows the original 5.8-kb M-bcr is not the only cluster.

    Evidence that argues against this (1)
    • The 'all Ph-positive patients' generalisation was falsified within two years, by the same Rotterdam laboratory. In Philadelphia-positive but bcr-negative ALL the chromosome-22 breakpoint lies well 5-prime of the 5.8-kb M-bcr, within the same BCR gene, and the chimaeric transcript is 7 kb rather than 8.5 kb, encoding a novel smaller fusion protein rather than the 210 kDa one. The M-bcr is therefore one of at least two breakpoint clusters, and which one is used tracks with a different disease. An assay or a frequency estimate built on M-bcr probes alone misses the entire p190 population, which is the majority of Philadelphia-positive ALL and the population this node's paediatric evidence [E5][E8] is actually about. Hermans A et al. 1987
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Variation in the BCR exon contribution produces a functional difference between p210 and p185/p190 BCR-ABL: foreign upstream BCR sequence deregulates ABL1 kinase activity, and the extent of deregulation correlates with transformation potency and with the pathological phenotype.

    Tyrosine kinase activity and transformation potency of bcr-abl oncogene products · Lugo TG, Pendergast AM, Muller AJ, Witte ON · Science 1990
    PMID 2408149DOI
    What this evidence cannot show

    Retroviral overexpression in rodent fibroblast and haematopoietic lines with a single-step transformation readout; the measured kinase-activity difference does not by itself prove that p190's ALL phenotype versus p210's CML phenotype is caused by kinase output rather than by the differing BCR domains retained or by cell-of-origin.

    Evidence that argues against this (1)
    • Breakpoint alone does not determine disease: p210 occurs in a substantial fraction of Ph+ ALL and p190 transcripts are detectable in some CML, so the p190/ALL and p210/CML mapping is a strong statistical association rather than a deterministic rule. Long-term clinical series such as the IRIS cohort treat both transcript types within CML. Hochhaus A et al. 2017
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Imatinib (STI-571) binds the ABL1 catalytic domain only when the kinase adopts an inactive conformation with an unphosphorylated activation loop folded distinctly from both active kinases and inactive Src, and this conformational selectivity is the source of the drug's combined high affinity and high specificity.

    Structural mechanism for STI-571 inhibition of abelson tyrosine kinase · Schindler T, Bornmann W, Pellicena P, Miller WT, Clarkson B, Kuriyan J · Science 2000
    PMID 10988075DOI
    What this evidence cannot show

    The crystal used a close variant of STI-571 and the isolated catalytic domain of murine Abl, not the full-length human BCR::ABL1 fusion; the structure therefore says nothing about how the BCR oligomerisation domain or the SH3-SH2 unit modulates drug binding in the intact fusion.

    Evidence that argues against this (1)
    • The conformation is not what discriminates. Seeliger et al. solved the c-Src kinase domain bound to imatinib and found it closely resembles the Abl-imatinib and Kit-imatinib complexes and differs sharply from the inactive 'Src/CDK' conformation of the Src-family kinases - that is, c-Src DOES adopt the supposedly distinctive inactive state, while being inhibited far more weakly. Swapping individual Abl residues into c-Src failed to transfer affinity, showing that the thermodynamic penalty is distributed over a broad region of the fold, whereas two mutations expected to destabilise the alternative inactive Src/CDK state raised drug sensitivity 15-fold. Selectivity is therefore a free-energy balance between competing inactive states, not the uniqueness of one conformation, and the entry's wording attributes to a structure what is a property of an equilibrium. Seeliger MA et al. 2007
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    With median 10.9 years of follow-up, estimated 10-year overall survival on first-line imatinib for chronic-phase CML was 83.3%, with 82.8% achieving complete cytogenetic response and no unacceptable late cumulative toxicity.

    Long-Term Outcomes of Imatinib Treatment for Chronic Myeloid Leukemia · Hochhaus A, Larson RA, Guilhot F, Radich JP, Branford S, Hughes TP, Baccarani M, Deininger MW, Cervantes F, Fujihara S, Ortmann CE, Menssen HD, Kantarjian H, O'Brien SG, Druker BJ · N Engl J Med 2017
    PMID 28273028DOI
    What this evidence cannot show

    Survival is not disease-specific and the randomised comparison collapsed because 65.6% of the interferon arm crossed over after a median of only 0.8 years, so this is effectively a single-arm long-term analysis of an industry-sponsored trial; 48.3% of the imatinib arm did not complete study treatment, and the surviving analysed population is therefore selected. Adults only.

    Evidence that argues against this (1)
    • IRIS is a trial cohort and 83.3% is not what a population sees. The Swedish CML registry captures more than 95% of national diagnoses: 779 patients, 2002-2010, median age 60. Relative survival at 5 years was close to 1.0 below age 60, 0.9 at ages 60-80 and only 0.6 above 80, and even in 2007-2009 only 79% of patients over 80 received a tyrosine kinase inhibitor as initial therapy against 94% of those under 70. IRIS enrolled a younger, fitter, newly diagnosed chronic-phase population; the registry shows the same drug in the same era leaving a large age-stratified residue that the trial figure conceals. 'Chronic illness' is accurate for the trial-eligible and much less accurate above 80. Hoglund M et al. 2013
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In paediatric Ph+ ALL, chemotherapy plus continuous imatinib gave 5-year disease-free survival of 70% +/- 12%, not significantly different from sibling-donor (65% +/- 11%) or unrelated-donor (59% +/- 15%) transplant, showing no transplant advantage.

    Long-term follow-up of imatinib in pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia: Children's Oncology Group study AALL0031 · Schultz KR, Carroll A, Heerema NA, Bowman WP, Aledo A, Slayton WB, Sather H, Devidas M, Zheng HW, Davies SM, Gaynon PS, Trigg M, Rutledge R, Jorstad D, Winick N, Borowitz MJ, Hunger SP, Carroll WL, Camitta B · Leukemia 2014
    PMID 24441288DOI
    What this evidence cannot show

    Non-randomised comparison of small cohorts (n = 28 chemotherapy-plus-imatinib, 21 sibling BMT, 13 unrelated BMT) with wide confidence intervals; transplant assignment was by donor availability, not randomisation, so selection bias runs in both directions. MRD analysis was underpowered.

    Evidence that argues against this (1)
    • A later prospective intergroup cohort running the same strategy did not reproduce the number. EsPhALL2010 gave imatinib from day 15 of induction and continuously through chemotherapy, transplanted only 59 of 155 patients (38%) in first complete remission, and reached 5-year event-free survival of 57.0% (95% CI 48.5-64.6) and 5-year overall survival 71.8% (63.5-78.5) - materially below the 70% 5-year disease-free survival quoted here. Forty patients (26%) relapsed and 41 (26%) died, 25 of those deaths (61%) in continuous complete remission, i.e. from toxicity rather than leukaemia. The trialists' own conclusion was that early continuous imatinib with intensive chemotherapy 'might increase toxicity', not that transplant is dispensable. Biondi A et al. 2018
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The BCR N-terminal oligomerisation domain retained in every BCR::ABL1 fusion forms a coiled-coil that drives fusion oligomerisation, and its structure is solved (PDB 1K1F), yet it has never yielded a clinical drug.

    What this evidence cannot show

    Verified by fetching the RCSB entry record. The structure covers only the isolated ~72-residue oligomerisation domain. The statement that no clinical agent targets it is an assertion about the absence of approvals, which is harder to evidence than a positive claim; peptide and small-molecule oligomerisation disruptors have been reported preclinically.

    Evidence that argues against this (1)
    • 'Never yielded a clinical drug' is right about the clinic and misleading about tractability. Dixon et al. redesigned the coiled coil itself: charge-repulsion mutations lowered its self-oligomerisation while added salt bridges raised its affinity for the Bcr-Abl coiled coil, and introducing the optimised mutant into K562 cells decreased Bcr-Abl phosphorylation, reduced proliferation and increased caspase-3/7 activity and DNA fragmentation, outperforming the wild-type coiled coil in every assay. The interface is disruptable by a designed protein; what has not been achieved is a small molecule against it. For the lab's menin-interface argument this locates the failure in modality rather than in the interface, which is a materially different claim from 'oligomerisation interfaces are undruggable'. Dixon AS et al. 2011
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Crystal structures of the ABL1 kinase domain bound to imatinib exist and show the type II, DFG-out binding mode reaching past the gatekeeper Thr315: 2HYY is the HUMAN ABL1 kinase domain (UniProt P00519) and 1IEP is the MURINE c-Abl kinase domain (UniProt P00520).

    What this evidence cannot show

    Verified by fetching the RCSB entry records on 2026-09-19. Both cover the isolated kinase domain, not the fusion protein; 2HYY is a four-copy structure whose interpretation of the P-loop conformation has been discussed in the literature. 1IEP is murine c-Abl, not human ABL1 - see counter_evidence and the correction section dated 2026-09-19.

    Evidence that argues against this (1)
    • 1IEP is not a human structure. Fetched from RCSB on 2026-09-19, its single polymer entity is 'PROTO-ONCOGENE TYROSINE-PROTEIN KINASE ABL', source organism Mus musculus, UniProt P00520, 293 residues, X-ray at 2.1 A. Only 2HYY is human ABL1 (UniProt P00519, 273 residues, 2.4 A). Mouse and human kinase domains are effectively identical across the imatinib site, so the type-II DFG-out conclusion survives intact; what does not survive is the species attribution, which was applied to both entries at once - exactly the failure mode that accession-assembled structure lists reliably produce. The claim has been corrected to attribute the human structure to 2HYY and to name 1IEP as murine. RCSB Protein Data Bank 2002
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review

    Before tyrosine kinase inhibitors, only 20-30% of children with Ph+ ALL were cured with chemotherapy alone and about 60% with closely matched allogeneic transplant, making it one of the poorest-risk paediatric ALL subgroups.

    PMID 21091149DOI
    What this evidence cannot show

    A narrative review, so the historical cure rates are summarised across heterogeneous eras and regimens rather than drawn from one cohort; the 88% 3-year EFS it quotes from early AALL0031 was later revised downward by the mature 5-year analysis in [E5], which is a useful reminder that early trial reports of this kind overstate.

    Evidence that argues against this (1)
    • Those are era figures, and they had already moved before tyrosine kinase inhibitors arrived. In 610 children with Philadelphia-positive ALL treated between 1995 and 2005 WITHOUT any TKI, 7-year event-free survival was 32.0% +/- 2.0% and overall survival 44.9% +/- 2.2%, significantly better than the 25.0% +/- 3.0% and 36.0% +/- 3.0% of the 1986-1996 cohort (p=.007 and p=.017), on improved chemotherapy and transplantation alone. The same analysis found matched-unrelated-donor transplant equivalent to matched-sibling transplant, overturning the earlier conclusion that only sibling donors helped. The pre-TKI baseline against which imatinib's benefit is read is therefore not a fixed 20-30%; part of the apparent gain belongs to chemotherapy, donor availability and supportive care. Arico M et al. 2010
  9. E9established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Ponatinib overcomes the T315I gatekeeper mutation (70% major cytogenetic response in T315I chronic-phase CML versus 51% in dasatinib/nilotinib-resistant patients, with no single resistance mutation detected) but caused serious arterial thrombotic events in 9% of patients.

    A phase 2 trial of ponatinib in Philadelphia chromosome-positive leukemias · Cortes JE, Kim DW, Pinilla-Ibarz J, le Coutre P, Paquette R, Chuah C, Nicolini FE, Apperley JF, Khoury HJ, Talpaz M, DiPersio J, DeAngelo DJ, Abruzzese E, Rea D, Baccarani M, Muller MC, Gambacorti-Passerini C, Wong S, Lustgarten S, Rivera VM, Clackson T, Turner CD, Haluska FG, Guilhot F, Deininger MW, Hochhaus A, Hughes T, Goldman JM, Shah NP, Kantarjian H · N Engl J Med 2013
    PMID 24180494DOI
    What this evidence cannot show

    Single-arm phase 2 in heavily pretreated adults at the 45 mg starting dose, median follow-up only 15 months; the vascular event rate rose substantially with longer follow-up after this report and led to dose reduction and label changes, so 9% understates the eventual risk. Only 32 Ph+ ALL patients, no children.

    Evidence that argues against this (1)
    • The 9% came from an early data cut. In the final 5-year analysis of the same PACE trial (270 chronic-phase CML patients, 56.8 months median follow-up), the CUMULATIVE incidence of arterial occlusive events reached 31% - more than three times the figure quoted - while the exposure-adjusted incidence of new events fell from 15.8 to 4.9 per 100 patient-years between years 1 and 5. Protocol-wide dose reductions were introduced in October 2013 specifically to manage this. Efficacy also reads differently at 5 years: 60% major cytogenetic response overall and 40% major molecular response, with 73% estimated 5-year overall survival. A single-timepoint toxicity percentage for a drug taken indefinitely is not a comparable quantity to a response rate measured once. Cortes JE et al. 2018
  10. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Clinical resistance to imatinib in CML arises at the drug-binding surface of the oncoprotein itself: in 6 of 9 patients analysed, resistance was associated with a single amino acid substitution of the threonine in the ABL1 kinase domain known to form a critical hydrogen bond with the drug, and that threonine-to-isoleucine substitution alone was sufficient to confer STI-571 resistance in a reconstitution experiment; in the other 3 patients resistance came from progressive BCR-ABL gene amplification, and in all cases BCR-ABL signal transduction was reactivated.

    Clinical resistance to STI-571 cancer therapy caused by BCR-ABL gene mutation or amplification. · Gorre ME, Mohammed M, Ellwood K, Hsu N, Paquette R, Rao PN, Sawyers CL · Science 2001
    PMID 11423618DOI
    What this evidence cannot show

    Nine patients, mostly advanced-phase, in the first year of imatinib use; the mutational spectrum found later is far wider than one residue, and the 6-of-9 proportion is not a population estimate. Both routes described here keep the oncoprotein as the driver, which is why this paper reads as confirmation that the target is right - it cannot speak to resistance that bypasses BCR-ABL altogether.

    Evidence that argues against this (1)
    • Not all imatinib failure is target-level, and the part that is not was visible before the mutation data. Primitive quiescent Lin-negative CD34-positive Philadelphia-positive cells from chronic-phase CML patients remained viable through three days of culture with STI571 with or without growth factors, including in the most sensitive cases where the drug killed almost all dividing cells, and the surviving undivided population was confirmed to belong to the leukaemic clone. A drug-binding-surface mutation is therefore one route out of a larger set, and the persistent-stem-cell route requires no genetic change at the interface at all - which is the route that has no analogue in the mutation-centric story this entry tells. Graham SM et al. 2002
  11. E11established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The same escape geometry has now been observed for a chromatin drug in KMT2A-rearranged and NPM1-mutant leukaemia: somatic MEN1 mutations at the revumenib-menin interface were identified in patients with acquired resistance to menin inhibition, the same interface mutations recurred in xenograft models and in an unbiased base-editor screen, and the mutants attenuate drug-target binding while leaving the interaction with the natural ligand MLL1 intact and preventing inhibitor-induced eviction of menin and MLL1 from chromatin.

    MEN1 mutations mediate clinical resistance to menin inhibition. · Perner F, Stein EM, Wenge DV, Singh S, Kim J, Apazidis A, Rahnamoun H, Anand D, Marinaccio C, Hatton C, Wen Y, Stone RM, Schaller D, Mowla S, Xiao W, Gamlen HA, Stonestrom AJ, Persaud S, Ener E, Cutler JA, Doench JG, McGeehan GM, Volkamer A, Chodera JD, Nowak RP, Fischer ES, Levine RL, Armstrong SA, Cai SF · Nature 2023
    PMID 36922589DOI
    What this evidence cannot show

    Phase 1 clinical material, so the number of resistant patients sequenced is small and the mutations are selected under one compound; whether the same residues govern escape from the structurally distinct menin inhibitors (ziftomenib, bleximenib, enzomenib, the covalent agents) is not established by this work. Note the asymmetry that makes the analogy imperfect and is developed in the analogy section: the mutated gene here is wild-type MEN1, a host gene the leukaemia did not create, not the fusion.

    Evidence that argues against this (1)
    • Menin-inhibitor resistance does not require the drug, which breaks the strict parallel with imatinib. In serial samples from two patients with multiply relapsed KMT2A::AFF1 ALL, heavily pretreated samples were much less sensitive to menin inhibition than samples taken earlier in the same patients' disease course, although NONE of the patients had ever received a menin inhibitor; transcriptomic analysis showed sustained on-target efficacy against the canonical menin-inhibitor targets in the resistant cells, and the emergent RAS-pathway and TP53 mutations were not sufficient to induce resistance in vitro. Imatinib resistance is selected BY imatinib; menin-inhibitor resistance can be present before the class is ever given, which is an argument about therapy sequencing that the BCR::ABL1 analogy does not generate. Mahdavi L et al. 2026
Connected to