The Mutational Landscape of Head and Neck Squamous Cell Carcinoma

Authors

Stransky N

Egloff AM

Tward AD

Kostic AD

Cibulskis K

Sivachenko A

Kryukov GV

Lawrence M

Sougnez C

McKenna A

Meyerson M

Lander ES

Getz G

Golub TR

Garraway LA

Grandis JR

Doi

10.1126/science.1208130

PMID: 21798893 · DOI: 10.1126/science.1208130 · Journal: Science (2011)

TL;DR

Whole-exome sequencing of 74 head and neck squamous cell carcinoma (HNSCC) tumor-normal pairs revealed a tobacco-associated mutational spectrum, confirmed known driver genes (TP53, CDKN2A, PIK3CA, PTEN, HRAS), and identified novel recurrent mutations in NOTCH1 (11%), NOTCH2/3, IRF6, and TP63 that converge on disruption of the squamous differentiation program. HPV-positive tumors had roughly half the mutation rate of HPV-negative tumors.

Cohort & data

  • 92 HNSCC patients initially sequenced; 74 passed quality filters for analysis (18 excluded for stromal admixture).
  • Cancer type: HNSC (oral cavity, oropharynx, hypopharynx, larynx, sinonasal cavity).
  • Dataset: hnsc_broad.
  • Assay: whole-exome-seq (150x mean coverage, 87% loci >20x); two tumors also underwent whole-genome-seq (31x mean coverage).
  • 89% reported tobacco use; 79% alcohol use; 14% HPV-16 positive by PCR/ISH.
  • Copy number profiled by SNP arrays.

Key findings

  • Mean non-synonymous mutation rate: 3.3 mutations/Mb (range 0.43-17.1); HPV-positive tumors had lower rates (2.28 vs 4.83 mutations/Mb; p = 0.004).
  • G-to-T transversion frequency (mean 12%) characteristic of tobacco exposure; HPV-negative laryngeal cancers had highest rates (p = 0.008 for mutation rate, p < 0.0001 for G-to-T fraction vs other sites).
  • MutSig identified 39 significantly mutated genes (FDR q < 0.1).
  • NOTCH1 mutated in 11% of tumors with predominantly loss-of-function mutations (nonsense, splice-site, missense in ligand-binding domain); NOTCH2/NOTCH3 mutated in an additional 11%.
  • A squamous differentiation gene set (NOTCH1, IRF6, TP63) was the top enriched functional category among significantly mutated genes (q < 0.25).
  • At least 30% of cases harbored mutations in genes regulating squamous differentiation.
  • SYNE1 mutated in 20%, SYNE2 in 8% (nuclear polarity regulators upstream of NOTCH1).
  • CASP8 mutated in 8%, DDX3X in 4% (apoptosis-related).
  • PRDM9 (11%) and EZH2 (6%) showed significant mutation rates (histone methyltransferases).
  • HPV detected by PathSeq in 19% of tumors; inverse correlation between HPV and TP53 mutation (p = 0.006).
  • Validation rate: 89.7% overall by mass spectrometric genotyping; 95.7% for mutations with allelic fraction >20%.

Genes & alterations

Gene Alteration Frequency Significance
TP53 Mutations, deletions Most common q < 0.1; nearly universal inactivation (mutation or HPV)
CDKN2A Mutations, deletions Significant q < 0.1; known HNSCC driver
PIK3CA Mutations Significant q < 0.1; known HNSCC driver
PTEN Mutations Significant q < 0.1; known HNSCC driver
HRAS Mutations Significant q < 0.1; known HNSCC driver
NOTCH1 Loss-of-function (nonsense, splice, missense in LBD) 11% q < 0.1; novel in HNSCC
NOTCH2 Non-synonymous mutations 11% (combined with NOTCH3) q < 0.25
NOTCH3 Mutations, focal deletion 11% (combined with NOTCH2) q < 0.25
IRF6 Mutations Significant q < 0.25; squamous differentiation
TP63 Mutations Significant q < 0.25; squamous differentiation
CASP8 Mutations 8% Apoptosis suppression
EZH2 Mutations 6% Histone methyltransferase
CCND1 Amplification Frequent Copy number alteration
EGFR Amplification Rare Copy number alteration
ERBB2 Amplification Rare Copy number alteration

Clinical implications

  • Loss-of-function NOTCH1 mutations suggest that gamma-secretase inhibitors (which suppress NOTCH signaling) could promote rather than inhibit HNSCC, consistent with a clinical trial halted due to skin cancer in the treatment arm.
  • Patients receiving gamma-secretase inhibitors may require monitoring for head/neck squamous malignancies.
  • TP53 inactivation (via mutation or HPV) appears nearly universal, representing a core vulnerability or resistance mechanism.
  • Disruption of the squamous differentiation program may represent a therapeutic target, though synthetic lethal approaches would be needed to exploit dependencies arising from NOTCH inactivation or TP63 alteration.

Limitations & open questions

  • Cohort is predominantly tobacco-exposed (89%) which may limit generalizability to non-tobacco HNSCC.
  • Only 74 samples passed quality filters; statistical power limited for low-frequency events.
  • Functional validation of novel mutations (IRF6, SYNE1/2, PRDM9) was not performed.
  • HPV detection by sequencing showed discordance with PCR in 3 cases (possible contamination or low viral load).
  • The study preceded modern understanding of tumor mutational burden as a biomarker for immunotherapy response.
  • Whether the squamous differentiation disruption is therapeutically targetable remains an open question.

Citations from this paper used in the wiki

  • “The majority exhibited a mutational profile consistent with tobacco exposure; human papilloma virus was detectable by sequencing of DNA from infected tumors.” (Abstract)
  • “At least 30% of cases harbored mutations in genes that regulate squamous differentiation (e.g., NOTCH1, IRF6, and TP63), implicating its dysregulation as a major driver of HNSCC carcinogenesis.” (Abstract)
  • “The mutation rate of HPV-positive tumors was approximately half of that found in HPV-negative HNSCC (mean of 2.28 mutations/Mb compared with 4.83 mutations/Mb; p = 0.004)” (Results)
  • “point mutations affecting [NOTCH1] occurred in 11% of the HNSCC tumors” (Results)
  • “we found non-synonymous point mutations in NOTCH2 or NOTCH3 in 11% of the samples” (Results)

This page was processed by paper-compiler on 2026-05-06.