Oncofusion

start / graph / gene-igf2bp1

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6-interactome-rna-layer

IGF2BP1 — the IMP paralogue that belongs to a DIFFERENT fusion, and the cleanest control the family offers

also called IMP1 · IMP-1 · CRD-BP · ZBP1 · VICKZ1 · insulin-like growth factor 2 mRNA-binding protein 1

IGF2BP1 is the third sibling in the IGF2BP family, and in childhood leukaemia it belongs to a different genetic subtype — ETV6-RUNX1 — not to MLL/AF4. Because of that it works as a built-in control when asking whether the family matters to the MLL/AF4 fusion.

The same thing, for a scientist

IGF2BP1 is the third IMP-family m6A reader. It is a real leukaemia gene - it maintains leukaemia stem cell properties in AML by regulating HOXB4, MYB and ALDH1A1 [E2], synergises with ETV6-RUNX1 in B-ALL [E3], and an oncofetal SNHG29-IGF2BP1 RNA axis is a lineage-restricted dependency in acute megakaryoblastic leukaemia [E4] - but it is not this disease's paralogue. In B-ALL entities, over-expressed IGF2BP1 characterises ETV6/RUNX1-positive disease while MLL/AF4-positive disease is marked by IGF2BP2 and IGF2BP3 [E1]. This graph's dependency scan reproduces the split: IGF2BP1 is **+0.095** in the four B-lymphoid KMT2A-rearranged lines while its two paralogues are negative [E5]. The entry exists because a family question needs a family member that should not respond, and this is it.

Go deeper
The value of a paralogue that goes the other way

Most of what this beat can say about the IGF2BP family is soft: expression associations, modest Chronos differences, a mouse deletion in a chimeric construct. Soft evidence gets much harder when a member of the same family, measured the same way in the same experiment, moves in the opposite direction.

Two independent datasets do that here. The 2011 B-ALL expression survey assigns IGF2BP1 to ETV6/RUNX1 and IGF2BP2 plus IGF2BP3 to MLL/AF4 1. This project's DepMap scan, which knows nothing about that paper, gives IGF2BP1 a positive B-lymphoid index-arm difference (+0.095) against negative differences for the other two (-0.123 and -0.135) 5.

The agreement is not strong evidence - four lymphoid lines, a 2011 qPCR cohort - but it is the kind of agreement that is hard to manufacture, because the two measurements have no shared machinery and no shared authors.

The corollary is the useful part: a pan-IMP-family explanation of IGF2BP3's role in MLL-Af4 leukaemia is disfavoured. If the family were acting as an interchangeable unit, the member that is not co-induced should still be able to substitute, and the subtype specificity of the expression pattern argues it does not.

What IGF2BP1 does where it does belong

In AML, IGF2BP1 maintains leukaemia stem cell properties by regulating HOXB4, MYB and ALDH1A1 2 - a self-renewal programme readout, and worth noting because it is the same kind of phenotype reported for IGF2BP3 in MLL-Af4 and for MSI2 in MLL-AF9: an effect on leukaemia-initiating function rather than on bulk proliferation. That is the phenotype class a two-week CRISPR dropout cannot see, and IGF2BP1's own DepMap row (-0.057 mean, 2.2% of lines depleted) is the worked example.

In B-ALL it synergises with ETV6-RUNX1 to drive oncogenic signalling 3, and in acute megakaryoblastic leukaemia a SNHG29-IGF2BP1 axis is described as a lineage-restricted dependency 4. Three fusion contexts, three different partners, and none of them KMT2A.

Facts and figures
uniprotQ9NZI8
length AA577
gene symbolIGF2BP1
uniprot resolutiongene_exact:IGF2BP1 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as IGF2BP1; length 577 aa checked
expression in b allmarks ETV6/RUNX1-positive disease; low or negative in healthy marrow and blood
pubmed counts
(MLL OR KMT2A) AND (CRD-B…0.000(MLL OR KMT2A) AND (CRD-BP OR ZBP1 OR IMP1): 0(MLL OR KMT2A) AND IGF2BP15.000(MLL OR KMT2A) AND IGF2BP1: 5IGF2BP1 AND leukemia25.0IGF2BP1 AND leukemia: 25
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": -0.0573,
  "pct_lines_depleted": 2.2,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": 0.0952,
  "kmt2a_r_myeloid_diff": -0.0272,
  "null_percentile": 68.78
}
pubmed counts — the numbers behind the chart
{
  "(MLL OR KMT2A) AND IGF2BP1": 5,
  "(MLL OR KMT2A) AND (CRD-BP OR ZBP1 OR IMP1)": 0,
  "IGF2BP1 AND leukemia": 25
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.0818,
  "chronos_mean_other_leukaemia": -0.0991,
  "difference": 0.0173,
  "pan_essentiality_all_lines": -0.0573,
  "selectivity_rank": 11385,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
How this could bite us back
  • Using IGF2BP1 as a negative control assumes it is not expressed in t(4;11) blasts. That rests on a 2011 transcript survey and has not been checked at protein level in this disease.
  • IGF2BP1 turns up as a dependency in AML, ETV6-RUNX1 B-ALL and AMKL. A paralogue that is a dependency everywhere it is examined is a poor control for anything.
  • If the family compensates, a selective IGF2BP3 inhibitor would select for IGF2BP1 or IGF2BP2 up-regulation, and no published work monitors that.
Still open
  1. Does depleting IGF2BP1 in a KMT2A::AFF1 line do anything at all? The comparison has never been run and it is the control the family question needs.
  2. Is IGF2BP1 induced when IGF2BP3 is lost in MLL-Af4 cells, which would convert 'different subtype' into 'reserve capacity'?
Every source behind this entry (6)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Over-expressed IGF2BP1 characterises ETV6/RUNX1-positive B-ALL, while MLL/AF4-positive B-ALL is characterised by over-expression of IGF2BP2 and IGF2BP3; over-expression of IGF2BP1 specifically is an inherent characteristic of ETV6/RUNX1-positive blasts on follow-up dynamics.

    PMID 21414819DOI
    What this evidence cannot show

    RT-qPCR, single centre, 2011, transcript only. Abstract read; full text not read.

    Evidence that argues against this (1)
    • Subtype-specific expression does not establish subtype-specific function. IGF2BP1 could be perfectly capable of substituting for IGF2BP3 in an MLL/AF4 cell and simply not be transcribed there, which would make the paralogue-control argument weaker than it looks. Oncofusion knowledge graph (internal analysis) 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    IGF2BP1 maintains leukaemia stem cell properties in acute myeloid leukaemia by regulating HOXB4, MYB and ALDH1A1.

    PMID 31768017DOI
    What this evidence cannot show

    Abstract-level; AML rather than KMT2A-rearranged B-lymphoid disease. A stem-cell-property readout, which is the phenotype class most sensitive to assay choice.

    Evidence that argues against this (1)
    • DepMap gives IGF2BP1 a Chronos mean of -0.057 with only 2.2% of 1,208 lines depleted, which would read as near-dispensable. The contradiction is the beat's recurring one: a self-renewal requirement is invisible to a two-week proliferation dropout, and citing either number without the other misrepresents the gene. Oncofusion knowledge graph (internal analysis) 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    IGF2BP1 synergises with ETV6-RUNX1 to drive oncogenic signalling in B-cell acute lymphoblastic leukaemia.

    PMID 37670323DOI
    What this evidence cannot show

    Abstract-level. Establishes the ETV6-RUNX1 association functionally rather than only by expression, which is what makes it useful as a contrast to the MLL/AF4 case.

    Evidence that argues against this (1)
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    An oncofetal SNHG29-IGF2BP1 RNA axis is a lineage-restricted dependency in acute megakaryoblastic leukaemia.

    PMID 42380666DOI
    What this evidence cannot show

    Abstract-level; a fourth disease context. Cited to establish the pattern that IGF2BP1's dependencies are lineage-restricted and attach to specific fusion or lineage contexts.

    Evidence that argues against this (1)
    • That the same paralogue is a dependency in AML, in ETV6-RUNX1 B-ALL and in AMKL undercuts the 'belongs to a different fusion' framing: IGF2BP1 appears in whichever context is examined, and its absence from the MLL/AF4 literature may reflect nobody having looked rather than a real boundary. Stoskus M et al. 2011
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives IGF2BP1 a mean gene effect of -0.057 across 1,208 lines with 2.2% of lines depleted, and a POSITIVE difference of +0.095 in the four B-lymphoid KMT2A-rearranged index lines (-0.027 myeloid) against 43 other leukaemia lines, at the 68.8th percentile of a 458-gene random null - while IGF2BP2 and IGF2BP3 are negative in the same arm.

    What this evidence cannot show

    Four lymphoid lines. The positive sign is within the null and is not itself a finding; the reported observation is that the three paralogues separate in the same direction as an independent expression dataset.

    Evidence that argues against this (1)
    • Three paralogues and four lines is the kind of comparison that produces a clean-looking split by chance. The agreement with the expression data is the only reason it is recorded, and neither dataset alone would support it. Stoskus M et al. 2011
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q9NZI8 (IF2B1_HUMAN) is human IGF2BP1, 577 amino acids; resolved by gene_exact query with the primary gene name read back as IGF2BP1 and the length checked.

    What this evidence cannot show

    One reviewed human hit with matching primary gene name. Heavily aliased: CRD-BP, ZBP1 and IMP1 are all in use, and ZBP1 collides with the Z-DNA binding protein 1 innate-immunity gene, so that alias must be title-checked in any search.

    Evidence that argues against this (1)
    • Searched this session, `(MLL OR KMT2A) AND (CRD-BP OR ZBP1 OR IMP1)` returns 0 records - so the alias expansion, which recovered tenfold more literature for the septins, recovers nothing here. The absence of IGF2BP1 from the KMT2A literature is real and not a symbol artefact. NCBI PubMed (queries run 2026-09-20) 2026
Connected to