PIK3R1 in KMT2A-rearranged infant ALL - the only cooperating lesion in this disease that was functionally proved to be activating, and it is an ITD
also called PIK3R1 · GRB1 · p85-alpha · PI3-kinase subunit p85-alpha · PI3K regulatory subunit alpha · P27986
PIK3R1 makes the regulatory half of a growth-signalling enzyme, and it is damaged in about one in ten infant KMT2A leukaemias. It matters out of proportion to that number because it is the only extra mutation in this disease that anyone actually proved makes cells grow - by putting the patient's own mutant gene into cells and showing they no longer needed a growth signal.
The same thing, for a scientist
This entry closes an open forward reference: `comut-pik3r1-kmt2a-r` was cited elsewhere in this graph and had never been written. PIK3R1 (P27986, 724 aa) encodes p85-alpha, the regulatory subunit that holds the PI3K catalytic subunit inhibited until a receptor engages it. In infant MLL-rearranged ALL it is listed among the **non-recurrent** kinase/PI3K/RAS genes in the 22-case whole-genome discovery cohort, alongside FLT3, NF1 and PTPN11 [E1]. Two facts lift it above that description. **The lesion class is unusual.** Two cases carried novel PIK3R1 mutations, **including an internal tandem duplication in the inter-SH2 region** [E1]. An ITD is a length-changing lesion in a non-hotspot region; it is the class that short-read SNV calling handles worst, and the class a hotspot panel cannot see at all. **And it is the only cooperating lesion in this disease with a functional demonstration.** The authors cloned the patients' own alleles - PIK3R1-Dup from INF018 and PIK3R1-R1a*/R1b from INF070 - into BaF3 cells and showed **factor-independent growth**, with FLT3 N676K as the parallel case [E1]. Nothing else in the co-mutation set of this leukaemia has been tested that way. Every other 'activating' label in the pathway rests on the mutation having been shown activating **in some other disease**. **The frequency the paper actually states for the PI3K complex is 11%**, for PIK3CA and PIK3R1 together, described as the first demonstration of recurrent activating mutations targeting the PI3K complex in this disease [E1]. **Where the PI3K arm goes at relapse is the same story as RAS, once.** In case INF073 the PIK3CA mutant allele frequency fell from 39% to 15% between diagnosis and relapse [E1] - neither retained nor lost, which is the middle outcome that a binary 'lost at relapse' framing erases.
This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.
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.
Why an ITD in this gene matters more than its frequency
The inter-SH2 region of p85-alpha is the coiled-coil that binds and inhibits the p110 catalytic subunit. Lesions there - deletions, insertions, duplications - release inhibition without touching the kinase. Three consequences specific to this beat:
1. It is assay-invisible in the wrong hands. An internal tandem duplication is not a single-nucleotide variant. Standard short-read SNV pipelines under-call insertions of this size, and a hotspot panel targeting PIK3CA codons 542, 545 and 1047 has no PIK3R1 content at all. Any cohort that reports 'no PI3K pathway lesions' from a hotspot panel has produced no evidence about PIK3R1.
2. It joins a pattern. This disease's documented lesions include a homozygous NF1 deletion, focal PAX5 losses, a PAX5-KANK1 fusion with no WGS read support, CDKN2A/B deletions and now a PIK3R1 ITD. Five of the recurrent or notable secondary lesions in the quietest genome in paediatric oncology are not point mutations. The mutation-count framing of this disease systematically under-reports it.
3. INF018 carried two of them. The same case that carried the homozygous NF1 deletion carried the PIK3R1 duplication 1. In a disease whose average is 1.3 non-silent mutations in the dominant clone, a single case with two independent pathway lesions is worth naming, and it is consistent with the finding that where two pathway mutations coexist they usually sit in different subclones.
What the BaF3 demonstration does and does not establish
Does. That the specific alleles found in these patients confer cytokine-independent growth on an IL-3-dependent murine pro-B line - i.e. they are gain-of-function, not passengers, in a cell that can report it.
Does not. That they cooperate with a KMT2A fusion. BaF3 factor independence is a sensitivity assay for signalling activation, not a leukaemogenesis assay; it contains no KMT2A fusion, is murine, and is famously permissive - a long list of alleles score positive in BaF3 and do nothing in a patient. The contrast with the KRAS experiment is instructive: KRAS(G12V) is unambiguously activating and still did not initiate leukaemia with MLL-AF4 in human HSPCs. Activating and cooperating are different claims, and this disease is the clearest place in oncology where they come apart.
So the correct reading is: PIK3R1 mutations in infant KMT2A-rearranged ALL are real gain-of-function events of an unusual lesion class, present in a minority of patients, whose contribution to the leukaemia is unestablished.
Accession resolution, and the alias worth flagging
gene_exact:PIK3R1 AND organism_id:9606 AND reviewed:true with gene_primary and length:
| accession | primary | synonyms | protein name | length |
|---|---|---|---|---|
| P27986 | PIK3R1 | GRB1 | Phosphatidylinositol 3-kinase regulatory subunit alpha (PI3-kinase subunit p85-alpha) | 724 |
One row, primary matches. The synonym GRB1 is the flag. It is one character from GRB2 (P62993, growth factor receptor-bound protein 2, 217 aa), a different adaptor in the same signalling pathway that appears constantly in the same papers. A free-text or alias-matching lookup on GRB1 near GRB2 is the kind of near-miss that produces a confident wrong answer, and the length difference - 724 against 217 - is the check that catches it, so both accession and length are asserted here.
Domains fetched from the same record's feature table on the same day: SH3 at 3-79, Rho-GAP at 113-301, SH2 1 at 333-428, SH2 2 at 624-718. The inter-SH2 region where the reported ITD sits is not annotated as a Domain in UniProt; it lies between the two annotated SH2 domains, i.e. between residues 428 and 624. It is recorded in data.inter_sh2_region rather than in data.domains, because a boundary derived from flanking annotations should not be drawn as if it were a curated domain.
| uniprot | P27986 |
|---|---|
| gene symbol | PIK3R1 |
| length AA | 724 |
| alphafold ID | AF-P27986-F1 |
| cytoband | 5q13.1 |
| closes forward reference | comut-pik3r1-kmt2a-r, cited elsewhere in this graph and unwritten until this cycle |
| resolution note | gene_exact:PIK3R1 AND organism_id:9606 AND reviewed:true with fields=accession,gene_primary,gene_synonym,protein_name,length returned one row: P27986, primary gene name PIK3R1, sole synonym GRB1, 724 aa. Fetched from the UniProt REST API on 2026-09-20; domain coordinates from the same record's feature table the same day. |
| druggability | PI3K inhibitors exist and none has been tested in KMT2A-rearranged infant ALL in any study read for this beat. A p85-alpha inter-SH2 lesion releases p110 inhibition, so the pharmacological response would be catalytic-subunit inhibition rather than anything directed at PIK3R1 itself, and the same subclonal-target and paediatric-access problems apply as for MEK inhibition. |
domains — the numbers behind the chart
[
{
"name": "SH3",
"start": 3,
"end": 79,
"pfam": null
},
{
"name": "Rho-GAP",
"start": 113,
"end": 301,
"pfam": null
},
{
"name": "SH2 1 (N-terminal SH2)",
"start": 333,
"end": 428,
"pfam": null
},
{
"name": "SH2 2 (C-terminal SH2)",
"start": 624,
"end": 718,
"pfam": null
}
]inter sh2 region
{
"note": "the coiled-coil that binds and inhibits the p110 catalytic subunit, and the site of the internal tandem duplication reported in this disease; NOT annotated as a Domain feature in UniProt P27986",
"derived_bounds": "between residues 428 and 624, taken from the boundaries of the two annotated SH2 domains",
"why_not_in_domains": "a boundary derived from flanking annotations should not be drawn by a figure as if it were curated"
}cooperating lesion profile
{
"disease": "infant KMT2A-rearranged B-ALL",
"status": "non-recurrent as a gene; PIK3CA plus PIK3R1 together present in 11% of infant MLL-R cases",
"lesion_class": "point mutations plus an internal tandem duplication in the inter-SH2 region - a length-changing lesion in a non-hotspot region",
"functional_status": "the ONLY cooperating lesion in this disease demonstrated activating on the patients' own alleles, by BaF3 factor independence",
"cooperation_status": "unestablished - no experiment has tested a PIK3R1 allele with a KMT2A fusion",
"relapse_behaviour": "the PI3K arm's only data point is PIK3CA in INF073 falling from 39% to 15% MAF, i.e. neither retained nor lost",
"case_INF018": "carried both the homozygous NF1 deletion and the PIK3R1 duplication"
}- PI3K inhibition in an infant acts on a pathway required for normal growth and metabolism, and the adult PI3K-inhibitor experience is dominated by hyperglycaemia, rash and colitis - an unattractive profile to add to an infant leukaemia backbone for a lesion present in a minority of a minority.
- The BaF3 result is the strongest functional evidence in this beat and it is also the assay most likely to overcall: a long list of alleles confer factor independence in BaF3 and do nothing in patients. Treating it as evidence of cooperation would repeat the exact error the KRAS cord-blood experiment exists to warn against.
- The lesion is an ITD in a non-hotspot region. A trial stratifying on 'PI3K pathway mutation' using a hotspot panel would miss it while reporting a PI3K-pathway genotype, which is worse than not testing.
- Relieving p85 inhibition of p110 is one of several routes into the same pathway output, so a patient without a PIK3R1 lesion may have the same pathway activation by another route - and a patient with one may have it in 10% of their blasts.
- Does a PIK3R1 inter-SH2 ITD cooperate with a KMT2A fusion? The allele exists as a cloned construct in the original study and the experiment - fusion plus allele in a human progenitor - has never been run.
- How many published KMT2A-rearranged cohorts used a caller capable of detecting an inter-SH2 ITD? The lesion class is the one short-read SNV pipelines handle worst, and no cohort states its indel-calling sensitivity.
- What is the PIK3R1 mutation frequency separately from PIK3CA? The published figure is the pair at 11%, and every downstream citation of it inherits the pooling.
- Was the PIK3R1 lesion in INF018 in the same cell as the NF1 homozygous deletion? Two independent pathway lesions in one case, in a disease averaging 1.3 mutations, is either a genuinely two-hit patient or two subclones, and the distinction matters.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In infant MLL-rearranged ALL, PIK3R1 was among the non-recurrent kinase/PI3K/RAS mutated genes in the 22-case whole-genome discovery cohort; two cases carried novel PIK3R1 mutations including an internal tandem duplication in the inter-SH2 domain; the patient-derived PIK3R1 alleles (Dup from INF018, R1a* and R1b from INF070) and FLT3 N676K were functionally demonstrated to be activating by conferring factor-independent growth on the IL-3-dependent murine line BaF3; and activating mutations in PIK3CA and PIK3R1 together were present in 11% of cases, described as the first demonstration of recurrent activating mutations targeting the PI3K complex in this disease.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, Wang J, et al. · Nature Genetics 2015What this evidence cannot show
The 11% figure is for PIK3CA and PIK3R1 combined across the infant MLL-R cohort, not for PIK3R1 alone, and this entry does not split it. The functional data sit in supplementary figures 25-29 which were not retrieved; the main-text statement of the result was read. BaF3 factor independence is a signalling-activation assay in a murine pro-B line with no KMT2A fusion, and is permissive. Full text read via efetch db=pmc (PMC4553269) after Europe PMC fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- Activating is not cooperating, and this disease is where the two come apart most clearly: KRAS(G12V) is unambiguously activating and still failed to initiate leukaemia with MLL-AF4 in human cord-blood CD34+ HSPCs, enhancing only extramedullary engraftment and migration. Prieto C et al. 2016
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the same cohort, in matched diagnosis and relapse pairs, the PIK3CA mutant allele frequency in case INF073 decreased from 39% to 15%, alongside pathway mutations maintained or increased in two cases and lost in two others, of five informative pairs.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, Wang J, et al. · Nature Genetics 2015What this evidence cannot show
One case. The observation concerns PIK3CA rather than PIK3R1 and is recorded here because it is the PI3K arm's only relapse data point in this disease; it should not be transferred to PIK3R1. A fall from 39% to 15% MAF is a change in clonal composition, and without purity measurement at both timepoints it could also reflect a difference in blast content between the two samples.
Evidence that argues against this (1)
- A partial decline is neither retention nor loss, and the binary 'frequently lost at relapse' framing used across this graph and the literature erases exactly this middle outcome - of five informative pairs, only two were losses. Andersson AK et al. 2015
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Targeted duplex sequencing of 20 infants and 7 children with KMT2A-rearranged ALL identified PIK3R1 and PIK3CA among the genes recurrently altered in this disease, with 62% of the 87 cancer-implicated mutations below standard NGS resolution, only 33 exceeding 2% cellular prevalence, and 31 of those 33 in PI3K/RAS genes.
Duplex Sequencing Uncovers Recurrent Low-frequency Cancer-associated Mutations in Infant and Childhood KMT2A-rearranged Acute Leukemia. · Pilheden M, Ahlgren L, Hyrenius-Wittsten A, et al. · HemaSphere 2022What this evidence cannot show
PIK3R1 appears in the abstract's list of altered genes without a per-gene count or prevalence, so no PIK3R1 frequency is quoted from it here. A duplex sequencing panel detects sequence variants; whether it would call an internal tandem duplication in the inter-SH2 region depends on panel and caller design and the abstract does not say. 27 patients. Abstract read.
Evidence that argues against this (1)
- Since 31 of the 33 mutations above 2% cellular prevalence were 'typically KRAS/NRAS', the PI3K-arm contribution in that study sits mostly in the sub-2% fraction, which argues that PIK3R1 lesions are among the least clonal in a disease where clonality is already the central problem. Pilheden M et al. 2022
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProtKB lists one reviewed human record with primary gene name PIK3R1: P27986, 724 aa, sole gene synonym GRB1, protein name phosphatidylinositol 3-kinase regulatory subunit alpha / PI3-kinase subunit p85-alpha, with domains SH3 at 3-79, Rho-GAP at 113-301, SH2 1 at 333-428 and SH2 2 at 624-718; the inter-SH2 region is not annotated as a Domain feature.
UniProtKB reviewed human entry P27986 (P85A_HUMAN), retrieved by gene_exact:PIK3R1 with fields accession, gene_primary, gene_synonym, protein_name, length and the domain feature table · UniProt Consortium · UniProt 2026What this evidence cannot show
Fetched 2026-09-20. Coordinates are for the canonical 724-residue p85-alpha isoform; PIK3R1 also produces the shorter p55-alpha and p50-alpha isoforms from alternative promoters, and a paper numbering residues from one of those will disagree with everything here. The inter-SH2 boundary used in prose (between 428 and 624) is derived from the flanking annotated SH2 domains and is deliberately kept out of the domains array for that reason.
Evidence that argues against this (1)
- The synonym GRB1 is one character from GRB2 (P62993, 217 aa), a different adaptor that appears in the same literature; an alias-matching lookup can resolve one to the other and only the length check would catch it. UniProt Consortium 2026
the PI3K arm of the kinase/PI3K/RAS category the pathway entry summarises
PIK3CA plus PIK3R1 in 11% of infant MLL-R cases; an inter-SH2 ITD in two of them
- co-occurs with →NF1 - the only tumour-suppressor loss in this disease, and it is a deletion, so a mutation panel cannot see it
case INF018 carried both the homozygous NF1 deletion and the PIK3R1 duplication
the two alleles functionally proved activating in BaF3 in this disease were PIK3R1 and FLT3 N676K
an ITD in a non-hotspot region: another lesion class the mutation-count framing under-reports
the eleventh candidate, written after the prose-reference checker showed it as an open forward reference