The mutational landscape of adenoid cystic carcinoma

PMID: 23685749 · DOI: 10.1038/ng.2643 · Journal: Nature Genetics (2013)

TL;DR

Whole-exome (n=55) or whole-genome (n=5) sequencing of 60 adenoid cystic carcinoma (ACC) tumor/normal pairs from MSKCC revealed a remarkably low somatic mutation rate (0.31 non-silent mutations/Mb) but striking pathway-level convergence. MYB-NFIB translocations were confirmed in 57% of cases by FISH, with additional MYB-pathway lesions in another 8%. Mutations preferentially targeted chromatin-state regulators (35% — including SMARCA2, CREBBP, KDM6A), DNA-damage genes (TP53, UHRF1, ATM, BRCA1), PKA-pathway components (27% — RYR2/3, PTPRG/H/J/K), and FGF/IGF/PI3K signaling (30% — PIK3CA, PTEN, FGFR4, FGF16, IGFBP2), nominating the PI3K pathway as a candidate therapeutic target in a previously untreatable salivary malignancy.

Cohort & data

• 60 ACC tumor/normal pairs: 55 by whole-exome capture (Agilent SureSelect 51MB) and 5 by whole-genome sequencing on Illumina HiSeq 2000.
• Cancer type: ACYC (adenoid cystic carcinoma — salivary gland).
• Dataset: acyc_mskcc_2013; dbGaP accession phs000612.v1.p1.
• 2,221 Gb of mapped sequence; 92.4% of target covered at ≥10x; mean exome coverage 106x, mean genome coverage 37x.
• Methods: whole-exome-seq, whole-genome-seq, targeted re-sequencing validation (SOLiD + MiSeq), FISH for MYB-NFIB on tissue microarrays with 3-color BAC probes, ExomeCNV and Affymetrix SNP 6.0 for copy number, GISTIC2.0, Illumina HumanHT-12 expression arrays (n=23), CREST for structural variants, MuTect + SomaticSniper for SNVs, GATK Indel Detector for indels, CHASM for driver prediction (FDR 0.35).
• Samples obtained with IRB-approved consent; microdissected to >70% tumor purity; aligned to hg19/GRCh37 with BWA.

Key findings

• Mean 22 somatic mutations per sample; 0.31 non-silent mutations/Mb — comparable to neuroblastoma/hematologic malignancies, far below HNSCC (see PMID:21798893) and colon cancer (see PMID:22810696).
• Ti/Tv ratio = 1.1, atypical for most adult solid tumors.
• 710 distinct nonsynonymous mutations validated across 643 genes (1-36 per tumor).
• MYB-NFIB translocations detected by FISH in 57% (34/60) of samples; somatic mutation frequency correlated with solid histology (Wilcoxon p = 4.0 × 10⁻²).
• CHASM nominated drivers in PIK3CA, TP53, PTEN, SMARCA2, KDM6A, and CREBBP; 8 tumors had no CHASM driver call.
• Pathway enrichment: chromatin regulators (35% of tumors, q = 4.5 × 10⁻³), DNA-damage response (q = 5.6 × 10⁻³), PKA signaling (27%, q = 4.2 × 10⁻³), FGF/IGF/PI3K (30%, q = 2.4 × 10⁻²).
• GISTIC2.0 identified recurrent high-level losses at 6q24, 12q13, and 14q. 14q-loss associated with solid histology (Fisher p = 2.0 × 10⁻⁴); 6q24 loss enriched for advanced stage (p = 4.0 × 10⁻²).
• One tumor harbored a tandem duplication within FGFR2; RT-PCR did not detect a fusion transcript.
• Markedly decreased expression of TP53 transcriptional targets in TP53-pathway-altered tumors (binomial p = 1.0 × 10⁻⁴).
• PI3K-altered samples were enriched for solid histology, the most aggressive ACC variant (Fisher p < 1.6 × 10⁻³). p-AKT and p-PRAS40 IHC and GSEA confirmed functional PI3K activation in PIK3CA/PTEN-mutant tumors (p < 1.0 × 10⁻³).
• Notch pathway altered in 13% of samples (NOTCH1 5%; FOXP2 3%; FBXW7 R465H; DTX4; CNTN6).
• Cell-adhesion / axon-guidance genes mutated in 29% (NTNG1, SEMA3G, SEMA5A, FAT3, FAT4 truncations). FAT4 RNAi knockdown in HSG, HSY, HTB-41 and HFF-1 cells significantly increased growth.
• KDM6A H3K27me3 demethylase assay: mutant KDM6A lost demethylase activity and growth-suppressive function vs wildtype; some mutants showed dominant pro-growth phenotype.
• Hotspot IDH1 R132H mutation identified in one tumor.

Genes & alterations

Gene	Alteration	Frequency	Finding
MYB	t(6;9) fusion with NFIB; exon 10 splice/coding mutations; 5 homozygous deletions	57% fusion, 8% other	Disrupts leucine-rich negative regulatory domain; likely constitutive activation
NFIB	t(6;9) fusion partner; truncating CTF/NFI mutations (Y249*, P390fs); 4 homozygous deletions	with MYB fusions	Independent role suggested by point mutations
SMARCA2	Missense in Helicase C domain (T1126I, G1132V, G1164W)	5%	SWI/SNF catalytic subunit; CHASM driver
SMARCE1	HMG-box missense Y73C	2%	Likely DNA-binding LOF
ARID1A	Single mutation	2%	SWI/SNF complex
ATRX	Single mutation	2%	Chromatin remodeler
CREBBP	KAT11 domain missense (R1446C, I1453N, W1472S)	7%	Histone acetyltransferase; CHASM driver
EP300	Splice-site mutation in KAT11 domain (G1429_splice)	—	CREBBP co-activator
KDM6A	Missense	7%	Loss of H3K27me3 demethylase activity confirmed in vitro; CHASM driver
KMT2C (MLL3)	Missense	—	Histone methyltransferase
ARID5B	Mutation	—	Histone modification complex
TP53	Missense P151S + nonsense R213, R342 in DNA-binding/tetramer motifs	5%	Functionally validated by p53-pathway expression loss
UHRF1	2 mutations + 3 homozygous deletions	8%	p53-dependent DNA-damage checkpoint ubiquitin ligase
TXNIP	Frameshift L129fs in arrestin domain	—	Often repressed in cancer
ATM	Missense (2)	—	DNA-damage response
BRCA1	Missense (2)	—	DNA-damage response
PIK3CA	3 missense at COSMIC hotspots	5%	Functional p-AKT/p-PRAS40 activation confirmed
PTEN	R130fs + K144Q, both in catalytic domain	—	Co-occurring in one tumor
FOXO3	Mutation	7%	PI3K-pathway transcription factor
FGFR4	Mutation	—	Receptor tyrosine kinase
FGFR2	Tandem duplication (1 tumor)	—	Similar to hematologic malignancy ITDs; no fusion transcript by RT-PCR
FGF16	Mutation	—	FGF ligand
IGFBP2	Mutation	—	IGF-axis modulator
IL17RD	Mutation	—	PI3K inhibitor via FGFR interaction
NOTCH1	3 missense + 1 nonsense	5%	Notch-pathway activation by GSEA trend
FBXW7	R465H	—	Tumor suppressor; targets MYC and NOTCH1
FOXP2	Mutation	3%	—
RYR3	Recurrent mutation	7%	Intracellular calcium channel; PKA pathway
RYR2	Recurrent mutation	2%	Intracellular calcium channel
PTPRG	Nonsense E736* (phosphatase domain)	—	Tyrosine phosphatase tumor suppressor
PTPRH	Nonsense W602*	—	Tyrosine phosphatase tumor suppressor
PTPRJ	Missense	—	Tyrosine phosphatase tumor suppressor
PTPRK	Frameshift L457fs + 4 homozygous deletions	—	Tyrosine phosphatase tumor suppressor
HSPG2 (perlecan)	Recurrent mutation	7%	Basement-membrane proteoglycan; modulates FGF
IDH1	Hotspot R132H	1 tumor	Catalytic-domain hotspot
FAT4	Truncating mutations	—	Functional growth-suppressor knockdown phenotype
FAT3	Truncating mutations	—	Protocadherin
MGA	Mutation	—	MYB-pathway gene

Clinical implications

• Authors explicitly nominate the FGF/IGF/PI3K axis as a candidate therapeutic vulnerability in ACC: “delineate a previously undescribed subset of ACC patients which we hypothesize may benefit from agents targeting this pathway.”
• Chromatin-modifier mutations in 35% of tumors raise the possibility of epigenetic therapy benefit in a histology with no established systemic agent.
• Prior trials of EGFR/c-KIT-directed agents (e.g., lapatinib, imatinib) have not shown substantive responses, motivating genomically informed approaches.
• TP53-pathway functional deficiency is biomarker-supported by loss of canonical p53 target expression.
• MYB-NFIB fusion presence is established as the dominant recurrent structural lesion in ACC and should be considered in molecular diagnostics for salivary tumors.

Limitations & open questions

• 60-sample cohort is the largest published for ACC at the time but limits power for rare drivers; 8 tumors had no CHASM-designated driver and may harbor non-exonic or non-coding alterations.
• Recurrent translocations beyond MYB-NFIB cannot be excluded; only one FGFR2 tandem duplication was observed and lacked RT-PCR fusion-transcript evidence.
• “Verified ACC cell lines are needed to further substantiate the clinical utility of mutations identified here” — functional validation depended on non-ACC salivary cell lines (HSG, HSY, HTB-41) and surrogate models.
• Therapeutic predictions (PI3K inhibitors, epigenetic agents) remain hypotheses; no clinical correlates are reported in this study.
• Whether NFIB has an oncogenic role independent of its MYB fusion partnership is suggested but not mechanistically resolved.
• Functional consequence of mutations in histone genes (HIST1H2AL R18L, HIST1H1E K75N) is left open.

Citations from this paper used in the wiki

• “Adenoid cystic carcinomas (ACCs) are among the most enigmatic of human malignancies. These aggressive salivary cancers frequently recur and metastasize despite definitive treatment, with no known effective chemotherapy regimen.” (Abstract)
• “We determined the ACC mutational landscape and report the exome or whole genome sequences of 60 ACC tumor/normal pairs. These analyses revealed a low exonic somatic mutation rate (0.31 non-silent events/megabase).” (Abstract)
• “Interestingly, mutations selectively involved chromatin state regulators, such as SMARCA2, CREBBP, and KDM6A, suggesting aberrant epigenetic regulation in ACC oncogenesis.” (Abstract)
• “We observed MYB-NFIB translocations and somatic mutations in MYB-associated genes, solidifying these aberrations as critical events.” (Abstract)
• “We identified recurrent mutations in the FGF/IGF/PI3K pathway that may potentially offer new avenues for therapy (30%).” (Abstract)
• “MYB translocations occurred in 57% of samples (34/60).” (Page 2)
• “Despite low overall mutation frequency, 35% of ACC tumors were mutated in chromatin regulators.” (Page 3)
• “Functional activation of the PI3K pathway was observed in all ACC tumors harboring PIK3CA or PTEN mutations, but not in wild-type tumors.” (Page 5)
• “Significant increases in growth occurred in all cells tested following FAT4 knockdown but not in controls.” (Page 6)

This page was processed by crosslinker on 2026-05-09.