XPO1 mutation in acute leukaemia, counted with denominators and a positive control: 1 of 739 B-ALL, 0 of 584 paediatric AML, 0 of 1,103 adult AML
also called XPO1 E571K in acute leukaemia cohorts · cBioPortal XPO1 leukaemia scan
We asked a public cancer-genomics database how often the XPO1 hot spot is mutated in childhood and adult acute leukaemia, and checked first that the same query finds it where it is known to be common. It is essentially absent from acute leukaemia and abundant in CLL, exactly as the literature implies.
The same thing, for a scientist
**Method.** cBioPortal public API, `/molecular-profiles/{study}_mutations/mutations/fetch` with `entrezGeneIds:[7514]` (XPO1) against each study's `_all` sample list, run 2026-09-20. Denominators are each study's `_sequenced` sample list, read from `/studies/{id}/sample-lists?projection=DETAILED`. Every returned `proteinChange` was read, not just counted. | study | sequenced n | XPO1 mutations | protein changes | |---|---:|---:|---| | **cll_broad_2022** (CLL, Broad, Nat Genet 2022) | 1,075 | **52** | A1041V, A636S, A805T, D624G, **E571G, E571K, E571Q, E571V**, G579V | | **cllsll_icgc_2011** (CLL, ICGC, Nat Genet 2011) | 105 | **2** | E571K, E571V | | **bll_target_gdc** (B-lymphoblastic leukaemia/lymphoma, TARGET GDC 2025) | 739 | **3** | **E571K**, S384_P385insND, T1030_S1031delinsIPFVR | | aml_target_gdc (paediatric AML, TARGET GDC 2025) | 584 | **0** | - | | all_phase2_target_2018_pub (paediatric ALL phase II, TARGET 2018) | 150 | **0** | - | | all_stjude_2016 (ALL, St Jude) | 73 (all samples) | **0** | - | | alal_target_gdc (acute leukaemias of ambiguous lineage) | 251 (all samples) | **0** | - | | aml_ohsu_2022 (adult AML, Cancer Cell 2022) | 903 | **0** | - | | laml_tcga_pub (adult AML, TCGA NEJM 2013) | 200 | **0** | - | | aml_target_2018_pub (paediatric AML, TARGET 2018) | 1,025 (all samples) | **0** | - | **The positive control is the point (rule 67).** The identical query shape returns **52** XPO1 mutations in one CLL cohort, dominated by codon 571 in four different substitutions. The instrument returns ones. The zeros in acute leukaemia are therefore measurements, not query failures. **The one B-ALL hit, stated exactly.** `TARGET-10-PAPAGK-04A`, an XPO1 **E571K** missense at chr2:61,492,337 C>T, tumour alt/ref counts **42/167** (variant fraction about 20%), sample type suffix 04 (recurrent, bone marrow). Clinical attributes exposed by the resource: B-lymphoblastic leukaemia/lymphoma, `Lymphoid leukemia, NOS`, TARGET-ALL-P2, female, MUTATION_COUNT 25. **The record exposes no fusion or cytogenetic annotation, so whether this patient was KMT2A-rearranged cannot be determined from this resource.** Its `AGE` field reads 18 while `DAYS_TO_BIRTH` reads -853 (about 2.3 years); the two are inconsistent and neither is used here. **Conclusion.** XPO1 E571 mutation is a mature-lymphoid-neoplasm event. In acute leukaemia it is a single occurrence in 739 sequenced B-ALL samples and absent from 1,687 sequenced AML samples across three cohorts. **Any XPO1 rationale in KMT2A-rearranged leukaemia has to be about wild-type protein.**
What this analysis does not bound
- Rule 78 (count what the join dropped). Nothing was dropped by symbol resolution here: the query keys on the Entrez id 7514 directly, not on a gene symbol, so there is no multi-mapping exposure. The denominators are the studies' own sequenced lists; where a study had no
_sequencedlist the all-sample count is used and is labelled as such in the table. - Panel coverage. Several of these cohorts are whole-exome or targeted rather than whole-genome. A targeted panel that does not include XPO1 would return zero for a reason that has nothing to do with biology. cBioPortal's
_sequencedlist is the study's own statement of which samples were profiled on the mutation platform, which is the best available control for that and is not a guarantee that XPO1 was on every panel. - KMT2A status. No cohort queried here exposes fusion calls through this endpoint, so the analysis cannot say whether any XPO1-mutant sample was KMT2A-rearranged. It can say that with one mutation in 739 B-ALL samples, the maximum possible overlap is one patient.
- Infants. None of these cohorts is infant-specific. The question 'is XPO1 mutated in infant KMT2A-rearranged ALL' is not answered by this analysis and no cohort found in this beat answers it.
| script | manual; cBioPortal public REST API via scratchpad helper q61_cbio.py, 2026-09-20 |
|---|---|
| status | run |
| date | 2026-09-20 |
| gene identifier used | Entrez 7514 (XPO1) - not a symbol, so no multi-mapping exposure (rule 78) |
| positive control | cll_broad_2022: 52 XPO1 mutations in 1,075 sequenced samples, codon 571 dominant, through the identical query shape |
| result summary | XPO1 E571 mutation is a mature-lymphoid event. One E571K in 739 sequenced B-ALL samples; zero in 1,687 sequenced AML samples across three cohorts. Any XPO1 rationale in KMT2A-rearranged leukaemia concerns wild-type protein. |
inputs
[
"cBioPortal /molecular-profiles/{study}_mutations/mutations/fetch",
"entrezGeneIds [7514]",
"/studies/{id}/sample-lists?projection=DETAILED"
]counts — the numbers behind the chart
{
"cll_broad_2022": {
"sequenced": 1075,
"xpo1_mutations": 52
},
"cllsll_icgc_2011": {
"sequenced": 105,
"xpo1_mutations": 2
},
"bll_target_gdc": {
"sequenced": 739,
"xpo1_mutations": 3,
"e571k": 1
},
"aml_target_gdc": {
"sequenced": 584,
"xpo1_mutations": 0
},
"aml_ohsu_2022": {
"sequenced": 903,
"xpo1_mutations": 0
},
"laml_tcga_pub": {
"sequenced": 200,
"xpo1_mutations": 0
},
"all_phase2_target_2018_pub": {
"sequenced": 150,
"xpo1_mutations": 0
},
"aml_target_2018_pub": {
"all_samples": 1025,
"xpo1_mutations": 0
},
"all_stjude_2016": {
"all_samples": 73,
"xpo1_mutations": 0
},
"alal_target_gdc": {
"all_samples": 251,
"xpo1_mutations": 0
}
}single b all case
{
"sample": "TARGET-10-PAPAGK-04A",
"protein_change": "E571K",
"genomic": "chr2:61492337 C>T",
"tumor_alt_ref": "42 / 167",
"kmt2a_status": "not exposed by this resource - unknown"
}not bounded
[ "capture-panel coverage of XPO1 per study", "KMT2A status of the one mutant B-ALL case", "infant-specific cohorts (none queried)" ]
- A counted zero across AML cohorts is a strong argument that XPO1 mutation is not a KMT2A-r biomarker, and it is NOT an argument that XPO1 is irrelevant here - the whole mechanism in this beat runs on wild-type protein.
- If panel coverage is uneven, the zeros are softer than they look, and the analysis has no coverage control.
- Was XPO1 on the capture panel of every cohort returning zero? The cleanest resolution is a per-study coverage check at chr2:61.49 Mb, which this analysis did not do.
- Is the single TARGET B-ALL E571K case KMT2A-rearranged? One lookup in the TARGET clinical supplement would settle it and the resource used here does not carry the field.
- Is there any infant leukaemia cohort with XPO1 sequenced? None of the cohorts queried is infant-specific, and infant KMT2A-rearranged ALL has one of the lowest mutation burdens of any cancer, so the prior for any XPO1 mutation there is very low.
- The variant fraction in the one B-ALL case is about 20% - subclonal. Would it be seen at diagnosis, or is it a relapse-acquired event? The sample suffix 04 indicates recurrent disease.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
Querying cBioPortal for XPO1 (Entrez 7514) mutations on 2026-09-20 returned 52 mutations in 1,075 sequenced CLL samples (Broad 2022) dominated by codon 571 substitutions E571K, E571G, E571Q and E571V, 2 in 105 sequenced CLL samples (ICGC 2011), 3 in 739 sequenced B-lymphoblastic leukaemia/lymphoma samples (TARGET GDC 2025) of which one was E571K, and 0 in each of paediatric AML (TARGET GDC, 584 sequenced), adult AML (OHSU 2022, 903 sequenced), adult AML (TCGA, 200 sequenced) and paediatric ALL phase II (TARGET 2018, 150 sequenced).
cBioPortal for Cancer Genomics - public API, XPO1 (Entrez 7514) mutation query across leukaemia studies · cBioPortal for Cancer Genomics · cBioPortal public API 2026What this evidence cannot show
Run 2026-09-20 against the live public API; counts are of mutation RECORDS, and a sample with two XPO1 variants contributes two. Denominators are each study's own `_sequenced` sample list where one exists. The analysis cannot control for whether XPO1 was on every study's capture panel, and a targeted panel omitting XPO1 would produce a zero for a technical reason. The CLL arm is the positive control and it returns a large number through the identical query shape, which is the rule-67 requirement. No fusion or cytogenetic annotation is exposed through this endpoint, so KMT2A status of any mutant sample is unknown.
Evidence that argues against this (1)
- The zeros in AML are only as good as each study's capture panel, which this analysis did not verify. More fundamentally, XPO1 mutation frequency says nothing about whether WILD-TYPE XPO1 matters in these diseases: the DEK::NUP214 work reports a strong, drug-responsive XPO1 dependency in primary AML samples that carry no XPO1 mutation at all. Mutation frequency is the wrong instrument for the question this beat is really asking. Kaya F et al. 2025
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The recurrent XPO1 hot spot is predominantly E571K with E571G, was first identified in chronic lymphocytic leukaemia by whole-genome sequencing, and is characteristic of mature B-cell neoplasms.
Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. · Puente XS, Pinyol M, Quesada V, et al. · Nature 2011What this evidence cannot show
Cited here as the prior that makes the CLL arm a valid positive control, not as a measurement in acute leukaemia. Abstract-level for this beat; the specific residue identity comes from PMID 33451349, cited in `mechanism-xpo1-e571k-hotspot-change-of-function`.
Evidence that argues against this (1)
- METHODOLOGICAL DEFEATER: 'characteristic of mature B-cell neoplasms' is partly a statement about which assay was applied to which disease. Liang et al. profiled 49 evaluable archival formalin-fixed Hodgkin lymphoma specimens with a targeted panel and recovered XPO1 (E571); they then tested 13 patients' plasma cell-free DNA by droplet digital PCR and confirmed the mutation in 31%. Two assays on the same disease give two different pictures of how characteristic the hot spot is, and the acute-leukaemia denominators this entry counts against come from panel and exome data in samples with high blast fractions rather than from a 0.05%-sensitivity assay. The cohort arithmetic in this entry is a floor on prevalence, not an estimate of it. Liang WS et al. 2019
counted cohort frequencies with a CLL positive control
1 of 739 sequenced B-ALL, 0 of 1,687 sequenced AML