NRAS as a co-mutation in KMT2A-r leukemia — the one co-lesion that is specifically enriched in t(4;11)
also called NRAS in MLL-rearranged ALL · N-RAS co-lesion · NRAS G12/G13 infant ALL
NRAS is the second most commonly broken gene found alongside a KMT2A fusion, and unlike the others it is clearly more common in the specific t(4;11) fusion that causes most infant leukemia — 32% of those cases versus 6% of the rest. Like its cousin KRAS, it appears late, in only some of the cells, and often vanishes when the disease returns.
The same thing, for a scientist
NRAS is the second most frequently co-mutated gene in KMT2A-rearranged leukemia and the only co-lesion with a demonstrated, statistically significant enrichment in the t(4;11)/KMT2A::AFF1 subgroup: 32% of t(4;11)-positive infants versus 6% of the rest (P = 0.01) in the 124-patient Interfant-99/06 cohort. It was recurrent at n = 2 of 22 in the PCGP whole-genome discovery cohort and at n = 3 of 20 in non-infant KMT2A-r leukemia. The observed variant spectrum is narrower than KRAS's — restricted to codons 12 and 13 (G12C, G12D, G12S, G13D, G13R) with no Q61 or atypical alleles in the public curation of the PCGP study. Like KRAS it is subclonal, frequently lost at relapse, and has not been shown to be required for initiation or maintenance.
The t(4;11) enrichment is the distinguishing fact
Across the whole co-mutation beat, almost nothing is specific to a particular fusion partner. NRAS is the exception.
In 124 infant B-ALL cases uniformly diagnosed and treated on Interfant-99/06, NRAS mutations were significantly more frequent in t(4;11)-positive patients: 32% versus 6% (P = 0.01) 1. Together with KRAS, NRAS accounted for all recurrent mutations in that cohort — 38% of the 42-case discovery arm and 34% of the 82-case validation arm carried an activating mutation in one of the two genes 1.
That is a five-fold enrichment tied to a specific fusion, and it invites a mechanistic question the literature has not answered: is there something about the KMT2A::AFF1 program specifically that selects for NRAS rather than KRAS activation? Candidate explanations include lineage (t(4;11) is the most uniformly pro-B of the KMT2A fusions), cell of origin, or simply that NRAS is the RAS isoform most readily activated in early B-lineage progenitors. None has been tested in this context.
It is worth stating plainly what this enrichment does not show. It does not show that NRAS cooperates with KMT2A::AFF1 — enrichment is a selection signal, and the functional cooperation experiments done for KRAS came out negative (see comut-kras-kmt2a-r). No equivalent experiment adding mutant NRAS to KMT2A::AFF1 in human HSPCs has been published, which is a conspicuous gap given that NRAS is the isoform the enrichment points to.
Frequency and variant spectrum1 deeper
| cohort | NRAS finding | source |
|---|---|---|
| PCGP infant KMT2A-r ALL, whole genome (n = 22) | recurrent, n = 2 (second to KRAS at n = 4) | 2 |
| PCGP non-infant KMT2A-r leukemia, 7-19 y (n = 20) | recurrent, n = 3 (tied with KRAS and FLT3) | 2 |
| Interfant-99/06, t(4;11)-positive | 32% | 1 |
| Interfant-99/06, non-t(4;11) | 6% (P = 0.01) | 1 |
| Interfant-99/06 discovery / validation, KRAS+NRAS combined | 38% / 34% | 1 |
| Interfant (Driessen), NRAS+KRAS+BRAF by Sanger (n = 109) | 14% overall, 24% in t(4;11) | 3 |
Variant spectrum. The public curation of the PCGP study (cBioPortal all_stjude_2015, 93 sequenced samples, retrieved 2026-09-18) returns 8 NRAS mutations across 8 samples, with distinct protein changes G12C, G12D, G12S, G13D and G13R 4.
This spectrum is notably tighter than KRAS's in the same dataset. Every NRAS allele observed is at codon 12 or 13; there is no Q61 allele and none of the atypical alleles (A146T, K117N) that appeared in KRAS. Q61 is the dominant NRAS hotspot in melanoma and in adult AML, so its absence here — in a small sample, with the usual caution — is a real difference in mutational context worth noting rather than assuming. One consequence is practical: a hotspot panel covering codons 12/13/61 would capture the full observed NRAS spectrum in this disease, whereas the same panel would miss two of the eight KRAS alleles.
Denominator warning
The 8/93 cBioPortal count is not a frequency for KMT2A-r leukemia. The 93 samples mix infant KMT2A-r, infant non-KMT2A-r, non-infant KMT2A-r and relapse specimens 4. Use it for the allele spectrum only. The clean frequency statements are 32% vs 6% in t(4;11) vs non-t(4;11) infants 1, and n = 2 of 22 whole-genome infant KMT2A-r cases 2.
Clonal behaviour and outcome — the same story as the pathway as a whole
NRAS shares the pathway's defining liabilities, documented in comut-ras-pathway-kmt2a-r.
- Subclonal. 65% (20/31) of activating kinase/PI3K/RAS mutations in the PCGP infant cohort sat below 30% mutant allele frequency at 92% average tumour purity 2. The Interfant cohort independently describes the KRAS/NRAS mutations as subclonal 1.
- Lost at relapse. Two thirds of relapsing Interfant patients lost the diagnostic RAS mutation 1; in PCGP matched pairs the pathway mutation was lost in 2 of 5 informative cases and decreased in a third 2.
- Expressed when present. 100% of the activating pathway mutant alleles were expressed at RNA level irrespective of allele frequency, against 48% of non-silent SNVs generally 2 — so the subclones are functionally active, just not dominant.
- Prognosis is contested. RAS mutation was an independent adverse predictor in the 109-patient Driessen cohort (HR 3.194, 95% CI 1.211-8.429, P = 0.019) 3, but only a non-significant trend in PCGP 2. Neither analysis separates NRAS from KRAS, so there is no NRAS-specific prognostic estimate in the literature. That is a genuine gap, and it is answerable from existing Interfant material.
The Driessen analysis adds a mechanistic clue: RAS-mutant blasts were more resistant to glucocorticoids in vitro (P < 0.05) and presented with higher white-cell counts (P = 0.013) 3, and MEK inhibition reverses glucocorticoid resistance in KMT2A-r blasts of both RAS genotypes 5.
| gene symbol | NRAS |
|---|---|
| hgnc ID | HGNC:7989 |
| entrez gene ID | 4893 |
| uniprot | P01111 |
| cytoband | 1p13.2 |
| pathway | RAS-MAPK |
| role in kmt2a r | second most frequent co-mutated gene; the only co-lesion with demonstrated t(4;11)-specific enrichment |
| genome build note | No coordinates asserted. Protein changes in HGVS p. notation against the canonical NRAS transcript as curated by cBioPortal; source analysis aligned to GRCh37-lite (NCBI build 37). |
frequency
{
"interfant_t4_11_positive_pct": 32,
"interfant_non_t4_11_pct": 6,
"p_value": 0.01,
"pcgp_infant_wgs_discovery": {
"n_mutated": 2,
"n_cohort": 22
},
"pcgp_non_infant_kmt2a_r": {
"n_mutated": 3,
"n_cohort": 20
},
"interfant_kras_plus_nras_discovery_pct": 38,
"interfant_kras_plus_nras_validation_pct": 34
}variant spectrum cbioportal all stjude 2015
{
"n_mutations": 8,
"n_samples_with_mutation": 8,
"n_samples_sequenced": 93,
"denominator_warning": "mixed infant/non-infant, KMT2A-r and non-KMT2A-r, diagnosis and relapse; not a KMT2A-r frequency",
"observed_alleles": [
"G12C",
"G12D",
"G12S",
"G13D",
"G13R"
],
"q61_observed": false,
"atypical_alleles_observed": false,
"retrieved": "2026-09-18"
}clonal behaviour
{
"subclonal": true,
"fraction_pathway_mutations_below_30pct_maf": "20/31 (65%)",
"fraction_relapses_losing_diagnostic_ras": "two thirds",
"mutant_allele_expressed": true
}prognosis
{
"nras_specific_estimate": "none published; all prognostic analyses pool NRAS with KRAS (and BRAF)",
"pooled_ras_hazard_ratio": 3.194,
"pooled_ras_ci95": [
1.211,
8.429
],
"pooled_ras_p": 0.019,
"pcgp_replication": "non-significant trend only"
}functional cooperation
{
"nras_plus_kmt2a_aff1_in_human_hspc": "not published — a conspicuous gap given that the enrichment signal points to NRAS rather than KRAS",
"kras_g12v_plus_kmt2a_aff1": "does not initiate leukemia; enhances extramedullary engraftment and CNS infiltration (see comut-kras-kmt2a-r)"
}- Resistance: NRAS-mutant clones would be expected to survive upstream-directed therapy (e.g. FLT3 inhibition) since NRAS acts downstream of the receptor, making NRAS status a plausible resistance marker for any RTK-directed agent in this disease.
- Futility: NRAS mutation is subclonal and frequently lost at relapse, so an NRAS-directed agent risks attacking a clone irrelevant to the eventual relapse.
- No direct drug: there is no clinical-stage NRAS-selective inhibitor, so intervention necessarily falls back to MEK or ERK inhibition, which is genotype-agnostic and therefore cannot be justified on NRAS status alone.
- Compensation: MEK inhibition relieves feedback suppression of RTKs including FLT3, which is over-expressed in nearly all KMT2A-r blasts, providing a ready-made route to pathway reactivation.
- Essential-function toxicity: NRAS signalling is required for normal haematopoietic and lymphoid development; sustained MAPK suppression in an infant under one year has no established safety profile.
- Why is NRAS specifically, rather than KRAS, enriched in t(4;11)? Lineage, cell of origin, and isoform-specific signalling in early B progenitors are all candidates and none has been tested. This is the sharpest isoform-specific question in the beat.
- Does mutant NRAS cooperate with KMT2A::AFF1 in human HSPCs? The negative KRAS experiment is often generalised to RAS as a class, but the enrichment data point at NRAS and the equivalent NRAS experiment has never been done.
- Is there an NRAS-specific prognostic effect? Every published analysis pools NRAS with KRAS and BRAF. Splitting the existing Interfant cohorts by isoform is a zero-cost reanalysis that has not been reported.
- Is the absence of NRAS Q61 alleles in KMT2A-r infant ALL real? Q61 dominates NRAS mutation in melanoma and adult AML. If the codon-12/13 restriction holds in a larger cohort, it implies a different mutational mechanism or a different selective requirement in the fetal B progenitor.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 124 infant B-ALL cases from Interfant-99/06, NRAS mutations were significantly more frequent in t(4;11)-positive patients than in the rest (32% vs 6%, P = 0.01); recurrent mutations were found exclusively in KRAS and NRAS (38% of the 42-case discovery cohort, 34% of the 82-case validation cohort), were subclonal, and two thirds of relapsing patients lost the diagnostic RAS mutation.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis · Agraz-Doblas A, Bueno C, Bashford-Rogers R, Roy A, Schneider P, Bardini M, et al. · Haematologica 104(6):1176-1188 2019What this evidence cannot show
The 32% vs 6% comparison is a single-cohort finding with P = 0.01 and no stated subgroup denominators in the abstract, so the absolute case counts behind the percentages are not verifiable from the abstract alone. Discovery and validation arms used different assays. No independent replication of the t(4;11)-specific NRAS enrichment has been located.
Evidence that argues against this (1)
- The Driessen cohort found RAS mutations enriched in t(4;11) as a class (~24% vs ~14% overall) but screened NRAS, KRAS and BRAF together and did not report an isolated NRAS enrichment, so the isoform specificity rests on one cohort. Driessen EMC et al. 2013
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
NRAS was recurrently mutated in n = 2 of the 22 whole-genome PCGP infant KMT2A-r ALL cases and n = 3 of the 20 non-infant KMT2A-r leukemias; across the cohort 65% (20/31) of kinase/PI3K/RAS mutations were below 30% mutant allele frequency, all mutant alleles were expressed, and the pathway mutation was lost at relapse in 2 of 5 informative matched pairs.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias · Andersson AK, Ma J, Wang J, Chen X, Gedman AL, Dang J, et al. · Nature Genetics 47(4):330-7 2015What this evidence cannot show
n = 2 of 22 is two events; the ordering of NRAS below KRAS rests on a two-case difference and is not a stable ranking. Clonality inferred from mutant allele frequency. Relapse dynamics from 5 informative pairs.
Evidence that argues against this (1)
- The entry reports the finding that undermines the co-mutation frame and does not draw it: 65% of these lesions sit below 30% mutant allele frequency and the pathway mutation was LOST at relapse in 2 of 5 paired cases. A subclonal lesion that disappears under treatment pressure is a passenger of the expansion rather than a driver of it. Leukemia 2021
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 109 infant ALL patients screened by Sanger sequencing of NRAS, KRAS and BRAF, mutations were present in ~14% overall and ~24% of t(4;11)-positive patients (P = 0.04), were an independent predictor of poor outcome (HR 3.194, 95% CI 1.211-8.429, P = 0.019), and associated with higher white-cell count (P = 0.013) and in vitro glucocorticoid resistance (P < 0.05).
Frequencies and prognostic impact of RAS mutations in MLL-rearranged acute lymphoblastic leukemia in infants · Driessen EMC, van Roon EHJ, Spijkers-Hagelstein JAP, Schneider P, de Lorenzo P, Valsecchi MG, Pieters R, Stam RW · Haematologica 98(6):937-44 2013What this evidence cannot show
Results are pooled across NRAS, KRAS and BRAF and cannot be attributed to NRAS specifically — there is no NRAS-isolated prognostic estimate in this disease. Sanger sequencing under-detects subclones below ~20% allele frequency.
Evidence that argues against this (1)
- The PCGP cohort found only a non-significant trend toward poorer EFS/OS for pathway mutations, so the independent adverse prognostic effect is not replicated. Andersson AK et al. 2015
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
The public curation of the PCGP study contains 8 NRAS mutations across 8 of 93 sequenced samples, with distinct protein changes G12C, G12D, G12S, G13D and G13R, and no Q61 or atypical alleles.
cBioPortal for Cancer Genomics, study all_stjude_2015 (Acute Lymphoblastic Leukemia, St Jude, Nat Genet 2015) — NRAS mutation records retrieved via the public REST API on 2026-09-18 · cBioPortal for Cancer Genomics · Database 2026What this evidence cannot show
8/93 is NOT a KMT2A-r frequency — the 93 samples mix infant KMT2A-r, infant non-KMT2A-r, non-infant KMT2A-r and relapse specimens. The absence of Q61 is from 8 observed alleles and is compatible with chance; it should be treated as a hypothesis about mutational context, not an established difference. Variant lists are deduplicated protein changes, not per-sample genotypes.
Evidence that argues against this (1)
- Eight of 93 is the same PCGP cohort re-counted through a portal, not a second cohort, and the absence of Q61 or atypical alleles is as likely to reflect the calling pipeline's hotspot panel as the biology of the tumours. Ensembl, EMBL-EBI 2026
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MEK inhibitors trametinib, selumetinib and MEK162 impair primary RAS-mutant KMT2A-r infant ALL cells ex vivo and synergistically enhance prednisolone sensitivity in both RAS-mutant and RAS-wild-type cells.
MEK inhibition is a promising therapeutic strategy for MLL-rearranged infant acute lymphoblastic leukemia patients carrying RAS mutations · Kerstjens M, Driessen EMC, Willekes M, Pinhanços SS, Schneider P, Pieters R, Stam RW · Oncotarget 8(9):14835-14846 2017What this evidence cannot show
Ex vivo primary-cell assays only; no in vivo or clinical data; sample numbers per arm not stated in the abstract; results are not broken down by RAS isoform, so no NRAS-specific sensitivity claim can be made. Published in Oncotarget, which has since been delisted from some indexes.
Evidence that argues against this (1)
- Ex vivo sensitivity of primary blasts is the weakest rung of preclinical evidence for a combination: the cells are outside their niche, prednisolone synergy is measured over days against a disease scored at years, and the same report finds RAS-WILD-TYPE samples also responding — which removes the genotype selectivity that would make RAS the reason to give a MEK inhibitor. Leukemia 2021
gene-level dossier: accession, the HRAS1 alias trap, the 0.06-0.34% resistant clone
- co-occurs with →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)