A Landscape of Driver Mutations in Melanoma
PMID: 22817889 · DOI: 10.1016/j.cell.2012.06.024 · Journal: Cell (2012)
TL;DR
Whole-exome sequencing of 121 melanoma tumor/normal pairs identified six novel significantly mutated melanoma genes (PPP6C, RAC1, SNX31, TACC1, STK19, and ARID2) using a permutation-based statistical framework (InVEx) that leverages intronic mutation rates to control for the high UV-induced passenger mutation load in melanoma. The RAC1 P29S hotspot mutation was shown to be an activating, gain-of-function event directly attributable to UVB mutagenesis, providing definitive genomic evidence linking UV-induced DNA damage to oncogenic driver mutations in melanoma.
Cohort & data
- 121 qualified tumor/normal pairs from 135 melanoma patients (15 primary tumors, 30 metastatic samples, 76 short-term cultures from metastatic tissue).
- Subtypes: 95 cutaneous, 5 acral, 2 mucosal, 1 uveal, 18 unknown primary.
- Dataset: skcm_broad.
- Assay: whole-exome sequencing with solution-phase hybrid capture; 103-fold mean target coverage.
- Copy number: Affymetrix SNP 6.0 arrays analyzed with GISTIC.
- Validation: Sequenom mass spectrometric genotyping in extension sets (n=59, n=63, n=175).
Key findings
- Median somatic mutation rate was 14.4 coding mutations/Mb (higher than any other tumor type at the time); 82.2% were UV-signature C>T transitions.
- The InVEx permutation framework identified 11 genes with significant functional mutation burden (q<=0.2): BRAF, NRAS, PTEN, TP53, CDKN2A, MAP2K1, PPP6C, RAC1, SNX31, TACC1, and STK19.
- RAC1 P29S was the third most frequent hotspot after BRAF and NRAS mutations, validated at 3.9% prevalence (14/355 patients across discovery + extension sets).
- RAC1 P29S is a gain-of-function mutation that shifts RAC1 toward the active GTP-bound state (demonstrated by PAK1 PBD pull-down assays in HEK293FT cells and immortalized melanocytes).
- PPP6C mutations (9% of discovery set) clustered around R264, with 60% in a 12-amino-acid window; pattern suggests gain-of-function alteration of phosphatase activity.
- STK19 D89N hotspot found in 4-5% of melanomas.
- ARID2 harbored significant loss-of-function burden (7%); together with ARID1A, ARID1B, and SMARCA4, 13% of samples had LoF mutations in SWI/SNF complex components.
- NF1 LoF mutations were enriched in BRAF/NRAS wild-type melanomas (25% vs 2%, p=5.8e-3).
- 83% of samples harbored hotspot or COSMIC-recurrent mutations in BRAF (n=73) or NRAS (n=27), mutually exclusive (p=3e-14).
- PTEN mutation/deletion co-occurred with BRAF (44%) but rarely with NRAS (4%, p=4.9e-5).
- 46% of all driver mutations were attributable to UV-induced damage (C>T or G>T); 67% when excluding BRAF/NRAS.
Genes & alterations
| Gene | Alteration | Finding |
|---|---|---|
| BRAF | V600E/K/R hotspot | 60% of samples; mutually exclusive with NRAS |
| NRAS | Q61L/R/K hotspot | 22% of samples; mutually exclusive with BRAF |
| RAC1 | P29S hotspot (C>T) | 3.9% prevalence; gain-of-function, UVB-induced |
| PPP6C | R264C, S270L hotspots | 9%; clustered near catalytic-regulatory interface |
| STK19 | D89N hotspot | 4-5%; predicted kinase, unknown function |
| SNX31 | Distributed missense in FERM-like domain | Sorting nexin; potential Ras effector |
| TACC1 | Distributed near C-terminus/TACC domain | 7%; interacts with Aurora A |
| ARID2 | Nonsense/LoF truncations | 7%; SWI/SNF chromatin remodeling |
| NF1 | LoF mutations | 25% of BRAF/NRAS-WT; RAS pathway negative regulator |
| PTEN | Mutation/focal deletion | Co-occurs with BRAF (44%); rare in NRAS-mutant |
| TP53 | Missense/LoF | 19%; enriched in cases without CDKN2A/p14ARF loss |
| CDKN2A | p16INK4a LoF, p14ARF mutations | >20% (LoF in 12%); RB and p53 pathway deregulation |
| EZH2 | COSMIC-recurrent | Epigenetic regulator; 17% with chromatin modifier mutations |
| IDH1 | COSMIC-recurrent | Epigenetic regulator |
| KIT | V559A, focal amplification | In BRAF/NRAS-WT subset |
| CDK4 | R24 activating, focal amplification | Cell cycle checkpoint |
Clinical implications
- BRAF V600 mutations (50% of melanomas) predict response to RAF inhibitors such as vemurafenib.
- Activating KIT aberrations may predict response to tyrosine kinase inhibitors such as imatinib.
- RAC1 P29S, PPP6C R264C, and STK19 D89N are described as “potentially targetable” recurrent mutations.
- NF1 inactivation in the BRAF/NRAS-WT subtype suggests aberrant MAPK activation that could be therapeutically exploitable (e.g., MEK inhibition).
- The RB pathway is deregulated through somatic mutations in at least 24% of samples (via CDKN2A, RB1, CDK4), informing CDK4/6 inhibitor rationale.
Limitations & open questions
- Extension/validation sets were limited to mass spectrometric genotyping of specific hotspots rather than comprehensive sequencing.
- 76/121 samples were short-term cultures rather than primary tissue, which may introduce culture-derived artifacts.
- Functional validation was performed only for RAC1 P29S; PPP6C, STK19, SNX31, and TACC1 mutations await experimental confirmation of oncogenic activity.
- The statistical framework (InVEx) depends on sufficient intronic coverage from exome capture baits, limiting its applicability to WGS data without modification.
- STK19 has unknown protein function, making therapeutic targeting speculative.
- Acral, mucosal, and uveal subtypes are underrepresented (n=5, 2, 1 respectively).
Citations from this paper used in the wiki
- “Six novel melanoma genes (PPP6C, RAC1, SNX31, TACC1, STK19 and ARID2), three of which - RAC1, PPP6C and STK19 - harbored recurrent and potentially targetable mutations.” (Abstract)
- “RAC1 P29S hot spot mutation in melanoma was validated to be 3.9% (14/355 patients)” (Results, RAC1 section)
- “PAK1 PBD pull-down revealed a significantly higher fraction of RAC1(P29S) in the GTP-loaded active state when compared to wild-type” (Results, RAC1 functional section)
- “13% (16/121) of the discovery samples harbored a LoF mutation in a component of the SWI/SNF complex” (Results, ARID2 section)
- “putative loss-of-function NF1 mutations occurred in 5 of 21 of these tumors (25%) compared to 2 of the remaining 100 samples (2%) (p = 5.8 x 10-3)” (Results, landscape section)
- “46% were caused by C>T (37%) or G>T (9%) mutations characteristic of UVB/UVA-induced mutations” (Results, UV mutagenesis section)
This page was processed by paper-compiler on 2026-05-06.