Oncofusion

start / graph / fusion-tcf3-pbx1

Fusionstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q11-comparative

TCF3::PBX1 (E2A-PBX1), t(1;19)(q23;p13.3)

also called E2A-PBX1 · E2A::PBX1 · E2A-PRL · t(1;19) ALL · der(19)t(1;19)

A swap between chromosomes 1 and 19 replaces the DNA-reading end of a master B-cell switch with the DNA-reading end of a developmental patterning protein, so the resulting hybrid switches on a completely foreign set of genes at full power. Children with this leukaemia used to do badly and now do well on intensive chemotherapy, but their leukaemia has an unusual habit of coming back in the brain and spinal fluid.

The same thing, for a scientist

t(1;19)(q23;p13.3) fuses the amino-terminal two-thirds of TCF3/E2A (P15923), containing both transactivation domains AD1 and AD2, to the PBX1 homeodomain (P40424), replacing TCF3's own bHLH DNA-binding module with a heterologous homeodomain [E1]. The product is an approximately 85 kDa obligate transcriptional activator: strong, constitutively-on E2A activation domains aimed at PBX1/HOX consensus sites, with simultaneous loss of one TCF3 allele's normal B-lineage function. Mechanistically this is **transcription-factor corruption in the activating direction** — the opposite polarity to ETV6::RUNX1 — and the AD1 domain is the surface that recruits CBP/p300. It occurs in roughly 2-6% of childhood B-cell precursor ALL, most often as the unbalanced der(19)t(1;19), in older children (median age ~7 y) with a pre-B (cytoplasmic mu-positive, CD19+CD10+) immunophenotype [E3]. Historically poor-risk, it is now treated as intermediate-to-favourable with contemporary intensive therapy (5-year EFS 84.2% at St Jude, 83.3% in Taiwan, 79% in the Nordic cohort), but with a distinctive and statistically independent excess of isolated CNS relapse (9.0% vs 1.0% at 5 years) [E2][E3][E4]. There is no fusion-directed agent.

What it looks like
loading 1B72…
PDB 1B72 ↗
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 lost1 deeper

Kept from TCF3/E2A (5-prime partner): the amino-terminal two-thirds of the protein, which carries activation domain 1 (AD1) — the CBP/p300-binding surface, also the LXXLL-like docking module — and activation domain 2 (AD2). These are among the most potent activation domains in the B-lineage transcription factor repertoire, because their day job is driving the entire early-B programme.

Lost from TCF3: the basic helix-loop-helix (bHLH) DNA-binding and dimerisation domain at the C-terminus 1. The fusion therefore cannot bind E-boxes and cannot heterodimerise with TAL1, LYL1 or the other class-II bHLH partners.

Kept from PBX1 (3-prime partner): the homeodomain plus the C-terminal region. PBX1's homeodomain is a cooperative DNA-binder: it forms ternary complexes with HOX proteins on bipartite sites (PDB 1B72 shows PBX1 with HOXB1 on DNA) 5.

Lost from PBX1: the N-terminal PBC-A/PBC-B domains that mediate interaction with MEIS/PREP cofactors and that contain PBX1's own regulatory and nuclear-export controls.

The arithmetic: the fusion is a DNA-binding-domain swap. TCF3 keeps the accelerator and loses the steering; PBX1 supplies new steering and loses its own governor. Because both retained pieces are constitutively active — E2A ADs do not need a ligand or a signal, and the homeodomain does not need PBX1's cofactor domains — the result is a permanently-on activator at the wrong addresses.

Why this is the mirror image of ETV6::RUNX1

Both are transcription-factor fusions in paediatric B-ALL, and comparing them is the cleanest available lesson in polarity.

ETV6::RUNX1TCF3::PBX1
5-prime contributescorepressor recruitment (N-CoR)activation domains (CBP/p300)
3-prime contributesDNA binding (Runt)DNA binding (homeodomain)
net effect at targetsrepresses genes that should be onactivates genes that should be off
cell of originfetal early-B, in uteropre-B, postnatal, older child
age peak2-5 y~7 y, spread across childhood

The DNA-binding module always comes from the 3-prime partner in both, but the sign of the transcriptional output is inverted. Any mechanism-class node that lumps them as 'transcription factor fusions' loses the most interesting fact about them. The graph therefore assigns both to tf-repression-corruption only where repression is the mechanism, and marks TCF3::PBX1 as the activating-polarity case.

Mechanism: a constitutively-on activator plus TCF3 haploinsufficiency

Two things happen at once, and both matter.

1. Gain of function (dominant). The chimera occupies PBX1/HOX-type sites and delivers E2A activation domains there. In reporter and transformation assays the E2A activation domain is required for transforming activity, and the PBX1 homeodomain supplies target specificity. This is a genuine neomorph: neither parental protein activates this gene set.

2. Loss of function (recessive-flavoured). One TCF3 allele is destroyed. TCF3/E2A is the obligate driver of early B-cell commitment (Eb boxes at the IgH enhancer, EBF1 and PAX5 induction). Losing half the dose in a lineage that is exquisitely TCF3-dose-sensitive is a second hit built into the first, and it is the reason this fusion is B-lineage-restricted despite PBX1 being a broadly expressed developmental gene.

No catalytic activity, no ligand pocket, no condensate. Structurally the fusion is an intrinsically disordered activation region welded to a 60-residue helical DNA-binding domain. The only solved surfaces are the PBX1 homeodomain on DNA in a HOX ternary complex (1B72) 5 and the wild-type TCF3/E47 bHLH that the fusion deleted (3U5V, 2QL2, 6MGN) 6. That is an awkward structural situation: the retained functional module (the AD) has no structure, and the structured module (the homeodomain) is shared with wild-type PBX1.

Clinical phenotype, age and the CNS problem

Disease: B-cell precursor ALL with a pre-B immunophenotype — cytoplasmic mu heavy chain positive, CD19+, CD10+, often CD34-negative and CD20-positive. That is a later stage of B-cell development than the pro-B/early-B cell of origin of ETV6::RUNX1 or TCF3::HLF, which is why this fusion sits high on the developmental-order scale in this beat.

Frequency: 5.7% of paediatric B-ALL in the Taiwanese TPOG cohort (64/1129) 4; 1.8% of all paediatric ALL in the Nordic NOPHO cohort (47 cases) 3; 41/735 (5.6%) of B-cell precursor ALL at St Jude 2. The spread between 1.8% and 5.7% is partly denominator (all ALL vs B-ALL) and partly ascertainment.

Cytogenetic form: commonly unbalanced — the der(19)t(1;19), i.e. loss of the reciprocal derivative — in 62% of Nordic cases versus 32% balanced 3. Balanced versus unbalanced did not affect EFS or OS 3.

Age: median 7 years 3. Older than ETV6::RUNX1; this is not a toddler leukaemia.

Prognosis — the rehabilitation story: originally classified as high risk. With contemporary intensive, risk-stratified therapy it is no longer:

  • St Jude, 41 cases: 5-year EFS 84.2% +/- 7.1% versus 84.0% +/- 1.8% for other B-precursor ALL (p = 0.63) 2.
  • Taiwan TPOG, 64 cases treated in the high-risk arm: 5-year EFS 83.3% +/- 4.8%, statistically indistinguishable from TEL-AML1 (85.2%) or high hyperdiploidy (84.0%) 4.
  • Nordic NOPHO, 47 cases: 5- and 10-year EFS 0.79, OS 0.85 and 0.82 — but overall survival was significantly worse than t(12;21) and high hyperdiploidy (p = 0.004), i.e. salvage after relapse is less successful 3.

The CNS signal is the clinically distinctive finding. At St Jude, t(1;19) patients had essentially no isolated haematological relapse (0 vs 8.3% at 5 years, p = 0.06) but a markedly higher cumulative incidence of CNS relapse: 9.0% +/- 5.1% versus 1.0% +/- 0.4% (p < 0.001), and t(1;19) was an independent risk factor for isolated CNS relapse on multivariate analysis 2. The Taiwanese cohort saw the same direction (8.7% vs 5.8% vs 4.1%) without reaching significance 4. The Nordic cohort, by contrast, saw nine marrow relapses and zero CNS relapses 3.

That disagreement is important and is not resolved. The most likely explanation is differences in CNS-directed therapy between protocols (cranial irradiation policy, intrathecal schedule, high-dose methotrexate) rather than differences in the biology — which, if true, means the CNS tropism is real but maskable by adequate CNS prophylaxis.

Druggability: undrugged, with one plausible handle

Status: undrugged. No TCF3::PBX1-directed agent exists.

Why it is hard:

  • The oncogenic module is an intrinsically disordered activation domain. There is nothing to bind.
  • The DNA-binding module is a wild-type PBX1 homeodomain, shared with normal PBX1, which is required for skeletal, pancreatic and haematopoietic development.
  • There is no enzyme anywhere in the protein.

The one plausible handle is the AD1-CBP/p300 interaction, i.e. attacking the coactivator rather than the fusion: p300/CBP bromodomain and HAT inhibitors exist, and the retained AD1 is a documented p300-binding surface (see the p300 TAZ2:E2A-AD1 complex, PDB 2MH0) 6. That is a coactivator-dependency strategy of the same general shape as menin inhibition for KMT2A fusions — go after the recruited host factor, not the fusion — and it is the reason this node is worth an edge to the coactivator-dependency concept even though nothing is in trials.

In practice the clinical lever for this fusion is not a drug but CNS-directed therapy intensity, and the open question is whether MRD-guided CNS prophylaxis can close the relapse gap 23.

Why this fusion does not draw

This entry's 5' domain reference does not resolve, and it is not being renamed. The TCF3 portion retained here spans the N-terminal transactivation region including AD1 and AD2, with the break falling before the bHLH. protein-tcf3-domains carries AD1, the leucine-zipper region and the bHLH, but not AD2 — and UniProt P15923 does not annotate AD1 or AD2 as named regions at all, listing only bHLH (549-602), a leucine-zipper region (389-425) and six disordered stretches. So there is no fetched coordinate for AD2 in any database this project uses, and inventing one to make a viewer lane draw would trade an honest gap for a plausible fabrication. The lane stays unresolved until someone sources AD2's boundaries from the primary literature and adds them to the protein entry.

Facts and figures
gene5TCF3
gene3PBX1
cytobandt(1;19)(q23;p13.3)
frequency note5.7% is 64/1129 B-ALL in the Taiwanese TPOG cohort [E4]; 1.8% is the share of ALL overall in the Nordic NOPHO cohort [E3]; St Jude reported 41/735 (5.6%) of B-cell precursor ALL [E2]. The apparent range is mostly denominator choice (all ALL vs B-ALL) plus ascertainment method.
prognosisRehabilitated from historically high-risk to intermediate/favourable by intensification: 5-year EFS 84.2% +/- 7.1% (St Jude) [E2], 83.3% +/- 4.8% (Taiwan, treated in the high-risk arm) [E4] and 0.79 at 5 and 10 years (Nordic) [E3]. The residual problem is relapse site and salvage: an independent excess of isolated CNS relapse (9.0% vs 1.0% at 5 years) at St Jude [E2], and overall survival significantly worse than t(12;21) or high hyperdiploidy in the Nordic cohort despite equal EFS (p=0.004) [E3].
uniprot5P15923
uniprot3P40424
PDB noteNo structure of the TCF3::PBX1 fusion exists, and there is unlikely ever to be a full one: the retained TCF3 activation domains are intrinsically disordered. 1B72 is the wild-type PBX1 homeodomain with HOXB1 on DNA (the retained 3-prime module). 2MH0 is the p300 TAZ2 domain bound to the E2A activation domain peptide — the most directly druggability-relevant surface in the whole node, because it is the retained module engaging a host coactivator. 3U5V, 2QL2 and 6MGN are the TCF3/E47 bHLH domain, included to document the module the fusion DELETES.
mechanism class noteThe vocabulary has no 'tf-activation-corruption' term, so this is filed under tf-repression-corruption as the nearest transcription-factor-corruption class. Be aware that the POLARITY is inverted relative to ETV6::RUNX1: TCF3::PBX1 is a constitutively-ON activator (E2A activation domains aimed at PBX1/HOX sites), not a repressor. This is flagged for the Queen as a gap in the mechanism-class vocabulary.
druggability classundrugged
stage of hematopoiesisPre-B cell (cytoplasmic mu heavy chain positive, CD19+CD10+, often CD34-negative) - a LATER stage than ETV6::RUNX1 or TCF3::HLF, which is the main reason this fusion ranks late on the developmental-order scale; presents in older children at a median age of 7 years [E3]
age peak~7 y (median), spread across childhood rather than a sharp toddler peak
developmental order rank9
domain coordinate noteDomain-coordinate note (coordinator, 2026-09-19): some domain names listed for the retained or lost segments have NO verified residue range in the reference protein, so the 3D viewer draws those lanes only partially. For TCF3 (P15923) UniProt annotates just four features - the 9aaTAD motif (19-27), the NLS (170-176), the leucine zipper (389-425) and the bHLH domain (549-602). 'Activation domain 2 (AD2)' and the N-terminal transactivation region are real and well described in the literature but have no UniProt coordinates, so they are named here in prose and cannot be rendered. Fixing this needs residue ranges taken from primary structural or mutagenesis work, not a rename.
diseases
[
  "pediatric B-cell precursor ALL (pre-B, cytoplasmic mu positive)",
  "adolescent B-ALL",
  "adult B-ALL (rare)"
]
frequency
{
  "pediatric_b_all_pct": 5.7,
  "pediatric_all_pct": 1.8
}
breakpoints
[
  {
    "id": "bp-tcf3-exon16-pbx1-exon2",
    "gene5_exon_last": 16,
    "gene3_exon_first": 2,
    "frequency_pct": 95,
    "population": "pediatric B-cell precursor ALL with t(1;19)",
    "domains_retained_5": [
      "activation domain 1 (AD1 / ETAD1, CBP-p300-binding)",
      "activation domain 2 (AD2)",
      "TCF3 N-terminal two-thirds"
    ],
    "domains_lost_5": [
      "basic helix-loop-helix (bHLH) DNA-binding and dimerisation domain",
      "TCF3 C-terminus"
    ],
    "domains_retained_3": [
      "PBX1 homeodomain (DNA binding, HOX-cooperative)",
      "PBX1 C-terminal region"
    ],
    "domains_lost_3": [
      "PBC-A domain (MEIS/PREP cofactor interaction)",
      "PBC-B domain",
      "PBX1 N-terminal regulatory region"
    ],
    "phenotype_notes": "A single, highly stereotyped junction: identical E2A-PBX1 mRNA junctions were found across independent t(1;19) lines [E1]. Exon numbers follow the canonical TCF3 and PBX1 transcripts and should be read as a convention label; the functional description (N-terminal two-thirds of E2A joined upstream of the PBX1 homeodomain) is the reliable statement [E1]. Most cases carry the UNBALANCED der(19)t(1;19) with loss of the reciprocal product (62% vs 32% balanced), which has no detectable prognostic effect [E3]."
  }
]
PDB IDS
[
  "1B72",
  "2MH0",
  "3U5V",
  "2QL2",
  "6MGN"
]
mechanism class
[
  "tf-repression-corruption"
]
How this could bite us back
  • The obvious target is the p300/CBP coactivator, not the fusion. But p300/CBP are global coactivators required for normal haematopoiesis and for the function of dozens of essential transcription factors; p300/CBP inhibitors carry broad on-target transcriptional toxicity, and this is exactly the failure mode that makes coactivator-dependency strategies harder than they look.
  • The PBX1 homeodomain is wild-type and shared with normal PBX1, which is required for skeletal, pancreatic and haematopoietic development. Nothing aimed at the homeodomain or its DNA site can be fusion-selective.
  • Intensification is the current answer, and intensification is itself the toxicity. These children already receive high-risk-arm therapy in most protocols [E4], so any further CNS intensification (cranial irradiation, more intrathecal therapy) buys relapse reduction at the cost of neurocognitive late effects in a population with 80%+ survival.
  • The Nordic data show EFS parity but significantly worse overall survival than favourable subtypes [E3], meaning relapse in this subtype is harder to salvage. A strategy that only reduces first relapse without improving salvageability may not move overall survival.
  • Because TCF3 haploinsufficiency is intrinsic to the lesion, any therapy that further impairs B-lineage transcription factor dosage (including some epigenetic agents) risks compounding a defect the leukaemia already exploits.
Still open
  1. Is the CNS tropism real biology or a protocol artefact? St Jude found an independent 9-fold excess of isolated CNS relapse while the Nordic cohort found zero CNS relapses in 47 cases [E2][E3]. A pooled, protocol-adjusted analysis with CNS-directed therapy as a covariate would settle it, and if the tropism is real, the next question is which fusion-driven adhesion or chemokine programme (for instance integrin or CXCR4 axis genes) mediates meningeal homing.
  2. Does the retained AD1-p300/CBP interaction constitute a genuine coactivator dependency, as menin is for KMT2A fusions? If degrading or blocking p300/CBP selectively kills TCF3::PBX1 blasts in isogenic models, this fusion joins the coactivator-dependency class and becomes tractable; if not, it is a genuinely undruggable activator.
  3. How much of the pathology is the gain-of-function neomorph and how much is TCF3 haploinsufficiency? A TCF3 heterozygous-null comparison in the same human B-progenitor background would separate the two, and the answer determines whether restoring TCF3 dosage is a therapeutic idea at all.
  4. Why is this fusion pre-B rather than pro-B when both parental proteins are expressed earlier? Is the pre-B stage where the E2A activation domains are most abundantly licensed, or is it the only stage that tolerates the fusion's programme?
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    t(1;19) produces an approximately 85 kDa chimeric protein consisting of the amino-terminal two-thirds of E2A, with the E2A helix-loop-helix DNA-binding domain replaced by the homeodomain of a chromosome-1 homeoprotein (PBX1), and identical fusion junctions recur across independent t(1;19) cell lines.

    Chromosomal translocation t(1;19) results in synthesis of a homeobox fusion mRNA that codes for a potential chimeric transcription factor · Nourse J, Mellentin JD, Galili N, Wilkinson J, Stanbridge E, Smith SD, Cleary ML · Cell 1990
    PMID 1967982DOI
    What this evidence cannot show

    Cell-line based (three t(1;19) lines) rather than a primary-patient series, and the chimeric protein's transcription-factor function was predicted from sequence at the time rather than demonstrated; the 85 kDa product and the AD1/AD2 retention were confirmed by later functional work.

    Evidence that argues against this (2)
    • Junction invariance is a cell-line observation that does not hold in patients. Izraeli et al. applied RNA-based PCR to 21 children with the E2A/PBX1 rearrangement at diagnosis and found that 3 of them (14%) expressed a VARIANT chimeric transcript alongside the expected one, carrying an identical 27-bp insertion at the E2A/PBX1 junction whose translation replaces Val478 with 10 amino acids; the ratio of the two species differed between patients but stayed constant within a patient across disease stages, and the inserted sequence occurs in no other human transcript. The fusion junction is therefore reproducible at the genomic breakpoint but not at the protein-coding level in a seventh of cases, which is exactly the kind of heterogeneity three immortalised lines cannot show. Izraeli S et al. 1992
    • The domain vocabulary this entry uses for the TCF3 side — AD2, 'N-terminal transactivation region', 'N-terminal two-thirds' — is literature language with no counterpart in UniProt P15923, which annotates neither AD1 nor AD2. Any statement here about which transactivation domains are retained rests on the primary literature's own boundaries, and those differ between papers. UniProt Consortium 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    With contemporary therapy, children with t(1;19)/TCF3-PBX1 had 5-year EFS 84.2% (equal to other B-precursor ALL) and essentially no isolated haematological relapse, but a significantly higher 5-year cumulative incidence of CNS relapse (9.0% vs 1.0%, p<0.001), with t(1;19) an independent risk factor for isolated CNS relapse.

    Increased risk for CNS relapse in pre-B cell leukemia with the t(1;19)/TCF3-PBX1 · Jeha S, Pei D, Raimondi SC, Onciu M, Campana D, Cheng C, Sandlund JT, Ribeiro RC, Rubnitz JE, Howard SC, Downing JR, Evans WE, Relling MV, Pui CH · Leukemia 2009
    PMID 19282835DOI
    What this evidence cannot show

    Only 41 t(1;19) patients, so the 9.0% CNS relapse estimate carries a standard error of 5.1% — the confidence interval is wide. Four successive single-institution protocols with differing CNS-directed therapy are pooled, so the CNS signal is partly a function of which protocols the t(1;19) cases fell into.

    Evidence that argues against this (1)
    • The Nordic NOPHO cohort of 47 t(1;19) cases reported nine bone-marrow relapses and NO CNS relapses at all, directly contradicting the CNS-tropism claim and suggesting the effect is protocol-dependent rather than intrinsic. Andersen MK et al. 2011
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 47 Nordic t(1;19)/der(19)t(1;19) cases (1.8% of all paediatric ALL), median age was 7 years, the translocation was unbalanced in 62% and balanced in 32%, 5- and 10-year EFS was 0.79, and overall survival was significantly worse than for t(12;21) and high hyperdiploidy (p=0.004) despite similar EFS.

    Paediatric B-cell precursor acute lymphoblastic leukaemia with t(1;19)(q23;p13): clinical and cytogenetic characteristics of 47 cases from the Nordic countries treated according to NOPHO protocols · Andersen MK, Autio K, Barbany G, Borgstrom G, Cavelier L, Golovleva I, Heim S, Heinonen K, Hovland R, Johannsson JH, Johansson B, Kjeldsen E, Nordgren A, Palmqvist L, Forestier E · Br J Haematol 2011
    PMID 21902680DOI
    What this evidence cannot show

    Cases were identified cytogenetically across two protocols spanning 1992-2007, so therapy is heterogeneous and partly pre-MRD-era; three cases were neither clearly balanced nor unbalanced. The EFS-versus-OS divergence (similar EFS, worse OS) rests on relatively few relapse events.

    Evidence that argues against this (1)
    • A cohort more than twice the size finds no survival deficit. Asai et al. analysed 112 paediatric TCF3-PBX1 BCP-ALL patients across two Japanese protocols and reported 3-year event-free and overall survival of 85.4 +/- 3.9% and 89.0 +/- 3.5% (JACLS ALL02, n=82) and 5-year EFS and OS of 82.8 +/- 7.0% and 86.3 +/- 6.4% (CCLSG ALL2004, n=30), describing them as comparable to western reports, with age, initial white-cell count and NCI risk having no impact on OS. The Nordic finding of significantly worse OS than t(12;21) and high hyperdiploidy despite similar EFS rests on relatively few events in 47 patients and is not reproduced here. Asai's series also raises a different problem for this subtype: 8 of 9 JACLS relapses occurred during the maintenance phase and neither day-15 nor day-33 marrow status nor MRD identified those patients in advance. Asai D et al. 2014
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 1129 Taiwanese paediatric B-ALL patients, 64 (5.7%) had t(1;19)/TCF3-PBX1; treated mainly in the high-risk arm they achieved 5-year EFS 83.3% +/- 4.8%, indistinguishable from TEL-AML1 (85.2%) or high hyperdiploidy (84.0%), with a non-significant trend toward more CNS relapse (8.7% vs 5.8% vs 4.1%).

    Pediatric acute lymphoblastic leukemia with t(1;19)/TCF3-PBX1 in Taiwan · Yen HJ, Chen SH, Chang TY, Yang CP, Lin DT, Hung IJ, Lin KH, Chen JS, Hsiao CC, Chang TT, Chang TK, Peng CT, Lin MT, Jaing TH, Liu HC, Jou ST, Lu MY, Cheng CN, Sheen JM, Chiou SS, Hung GY, Wu KH, Yeh TC, Wang SC, Chen RL, Chang HH, Yang YL, Chen SH, Cheng SN, Chang YH, Chen BW, Hsieh YL, Huang FL, Ho WL, Wang JL, Chang CY, Chao YH, Lin PC, Chen YC, Liao YM, Lin TH, Shih LY, Liang DC · Pediatr Blood Cancer 2017
    PMID 28436581DOI
    What this evidence cannot show

    Not randomised: t(1;19) cases were deliberately assigned to the more intensive high-risk arm, so the favourable EFS is conditional on that intensification and does not show the fusion is intrinsically low risk. Thirteen of 64 were treated off-protocol by physician preference. The CNS trend was not statistically significant.

    Evidence that argues against this (1)
    • The 'intensification handles it' reading does not extend past childhood. Ribera et al. analysed adolescents and adults with t(1;19)(q23;p13)/TCF3-PBX1 enrolled on paediatric-inspired, MRD-oriented trials: MRD clearance and overall survival were no different from other BCP-ALL patients, but the cumulative incidence of relapse was significantly higher, particularly among patients aged 35 or over carrying additional cytogenetic alterations, and the authors conclude those patients might need additional or intensified therapy such as immunotherapy in first complete remission with or without subsequent transplant. An EFS of 83.3 +/- 4.8% obtained by deliberately assigning children to a high-risk arm is therefore a property of that arm in that age band, not evidence that the fusion is intrinsically low risk, and the same fusion carries excess relapse risk once the host is older. Ribera J et al. 2021
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    The PBX1 homeodomain retained in the fusion binds DNA cooperatively in a ternary complex with a HOX protein, as shown by the PBX1-HOXB1-DNA crystal structure.

    What this evidence cannot show

    Verified by fetching the RCSB entry record. This is the wild-type PBX1 homeodomain with a HOX partner, not the fusion protein; whether TCF3::PBX1 uses HOX-cooperative sites in leukaemic cells is a separate question that the structure does not answer.

    Evidence that argues against this (1)
    • The homeodomain whose DNA chemistry 1B72 describes is DISPENSABLE for transformation by the fusion. Monica et al. built E2a-Pbx delta-HD, a chimera lacking the PBX1 homeodomain entirely: it transformed NIH 3T3 fibroblasts efficiently and induced malignant lymphomas in transgenic mice, whereas mutation of the E2A transactivation motifs AD1 and AD2 significantly impaired transformation and hyperexpression of wild-type or truncated PBX1 alone did not transform at all; either the homeodomain or the PBX1 residues flanking it could support transformation when fused to E2A. Their conclusion is that E2a-Pbx1 may act through protein-protein interactions even without homeodomain-DNA contact. Citing a wild-type PBX1-HOXB1-DNA ternary complex as evidence about the fusion therefore imports a requirement that was directly tested and not supported. Monica K et al. 1994
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Structures exist for the TCF3/E47 bHLH domain that the fusion deletes (PDB 3U5V, 2QL2, 6MGN) and for the p300 TAZ2 domain in complex with the E2A activation domain ETAD1 that the fusion retains (PDB 2MH0), establishing AD1 as a coactivator-binding surface.

    What this evidence cannot show

    Verified by fetching the RCSB entry records. 2MH0 covers a short E2A activation peptide bound to an isolated p300 domain, so it evidences the interaction surface but says nothing about affinity or occupancy in leukaemic chromatin. The bHLH structures are of the deleted module and are included for the domain map, not as evidence about the fusion.

    Evidence that argues against this (1)
    • AD1 is not a dedicated coactivator surface, and a peptide-in-TAZ2 structure cannot show which occupancy dominates. Gow et al. found that the p300/CBP target motif within E2A activation domain 1 is ALSO the ETO-family COREPRESSOR target motif, that E2A-specific amino acid changes in that motif specifically REDUCE corepressor interaction relative to the HEB paralogue, and that replacing E2A-AD1 with HEB-AD1 abolished the ability of E2A-Pbx1 to activate target genes and to transform cells; conversely a separate downstream ETO-interacting sequence carries the silencing of wild-type E2A by AML1-ETO in t(8;21) cells and by ETO-2 in normal haematopoietic cells, and deleting it abrogates that repression. Describing 2MH0 as 'establishing AD1 as a coactivator-binding surface' states one of at least two occupancies of the same short motif, and the reason TCF3::PBX1 activates rather than represses is that E2A's version of that motif binds corepressors WEAKLY - a sequence fact the structure does not contain. Gow CH et al. 2014
Connected to