The genomic landscape of cutaneous SCC reveals drivers and a novel azathioprine associated mutational signature - PubMed (original) (raw)
doi: 10.1038/s41467-018-06027-1.
Jun Wang 2, Ai Nagano 3, Ludmil B Alexandrov 4, Karin J Purdie 5, Richard G Taylor 6, Victoria Sherwood 6, Jason Thomson 5, Sarah Hogan 5, Lindsay C Spender 6, Andrew P South 7, Michael Stratton 8, Claude Chelala 3, Catherine A Harwood 5, Charlotte M Proby 6, Irene M Leigh 9
Affiliations
- PMID: 30202019
- PMCID: PMC6131170
- DOI: 10.1038/s41467-018-06027-1
The genomic landscape of cutaneous SCC reveals drivers and a novel azathioprine associated mutational signature
Gareth J Inman et al. Nat Commun. 2018.
Abstract
Cutaneous squamous cell carcinoma (cSCC) has a high tumour mutational burden (50 mutations per megabase DNA pair). Here, we combine whole-exome analyses from 40 primary cSCC tumours, comprising 20 well-differentiated and 20 moderately/poorly differentiated tumours, with accompanying clinical data from a longitudinal study of immunosuppressed and immunocompetent patients and integrate this analysis with independent gene expression studies. We identify commonly mutated genes, copy number changes and altered pathways and processes. Comparisons with tumour differentiation status suggest events which may drive disease progression. Mutational signature analysis reveals the presence of a novel signature (signature 32), whose incidence correlates with chronic exposure to the immunosuppressive drug azathioprine. Characterisation of a panel of 15 cSCC tumour-derived cell lines reveals that they accurately reflect the mutational signatures and genomic alterations of primary tumours and provide a valuable resource for the validation of tumour drivers and therapeutic targets.
Conflict of interest statement
The authors declare no competing interests.
Figures
Fig. 1
Number of somatic mutations and mutation signatures across 40 cSCC samples in the moderate/poor and well-differentiated groups. a Number of nonsynonymous, synonymous and UTR mutations across 40 samples. b Mutation signature compositions across 37 of the 40 samples with sufficient mutation depth, in terms of moderate/poor and well-differentiated groups. IC immunocompetent, IS immunosuppressed, Aza confirmed azathioprine exposure, NC no confirmed azathioprine exposure, SE sun exposed site, U unknown if sun exposed site. c A novel signature, termed signature 32, predominately C > T mutations (75%) in combination with C > A, T > A, and T > C mutations, was identified as one of the dominant signatures. This signature is putatively associated with azathioprine treatment
Fig. 2
Somatic copy number aberrations (SCNA) and LOH events in 40 cSCC samples. a OncoPrint of copy gain, loss and copy-neutral (CN)-LOH segments across moderate/poor and well-differentiated groups. IS immunosuppressed, IC immunocompetent. b Box-and whisker plot indicating the percentage of the genome altered by CNA/LOH across the two groups. Individual values were shown by the blue dots. Moderate/poor differentiated group had significantly higher percentage of altered genome than well-differentiated group (Student’s _t_-test, two sided, p = 0.018)
Fig. 3
Twenty-two SMGs identified in 40 cSCC samples. a Venn diagram of overlap of significantly mutated genes as assessed by MutsigCV, OncodriveFM and OncodriveClust. b Mutation OncoPrint of the 22 SMGs identified by at least two of the three methods, with their aberrational frequencies also indicated with percentage bars. IS immunosuppressed, IC immunocompetent. c Pie chart of the mutational signature contribution to SMG mutations
Fig. 4
Clonality analysis of cSCC tumour samples. Clonality analysis of 35 cSCC exomes using EXPANDS (a) and SciClone (b). Clonality analysis of 22 SMGs identifying clonal and subclonal nonsynonymous mutations using EXPANDS (c) and SciClone (d). The number of nonsynonymous mutations in each gene was also shown
Fig. 5
Expression profiles of driver genes in five independent gene expression data sets. a Expression heatmap of 20 SMGs that were also expressed across normal, AK, in situ and invasive SCC samples from GSE42677. b Log2 fold changes (FC) of AK vs. normal and SCC vs. normal of 20 SMGs from GSE42677, Data set 1 (Lambert), GSE2503, GSE84293 and GSE32628. Normal normal skin, NSE non-sun exposed normal skin, SE Sun exposed skin, AK actinic keratosis. Significant genes in 3 out of 5 data sets were highlighted in green. Within the heatmap, red colour indicates the upregulation in AK or SCC compared to normal control, which blue colour indicates the downregulation in AK or SCC in relation to normal skin
Fig. 6
Expression profiles of genes significantly amplified or deleted in cSCC. a Expression heatmap of significantly amplified and upregulated genes (n = 23), and deleted and downregulated genes (n = 9) in the in situ vs. normal comparison that were also expressed across normal, AK, in situ and invasive SCC samples from GSE42677. b Venn diagrams of overlap across five gene expression data sets of significantly amplified genes and overexpressed (upper panel) and significantly deleted and downregulated genes (lower panel). c Log2 fold changes (FC) of pairwise comparisons of AK vs. normal and SCC vs. normal across the five data sets for 19 amplified/upregulated and 5 deleted/downregulated expressed genes shared across at least three data sets. Within the heatmap, red colour indicates the upregulation in AK or SCC compared to normal control and blue colour indicates the downregulation in AK or SCC in relation to normal skin
Fig. 7
Comparison of significantly mutated genes and pathways between moderate/poor and well-differentiated groups. a Group-specific SMGs from the randomisation test based on MutSigCV and OncodriveFM _p_-values, respectively. Moderate/poor group-specific genes were marked with red circles, while well-differentiated group-specific genes were marked with blue circles. The size of the circle corresponds to its significance compared to that expected by chance. b Mutation OncoPrint of group-specific genes that were expressed across 40 cSCC samples in moderate/poor and well-differentiated groups. Significance level (-log10(_p_-value)) from the randomisation test was also shown for each gene as bar charts. Genes derived from the OncodriveFM statistics were shown in dark grey, while genes derived from MutSigCV statistics were in light grey. Genes at the top panel appeared to be more moderate/poor group specific, while genes at the bottom panel were more well-differentiated group specific. c Significantly mutated KEGG signalling pathways and d biological processes between moderate/poor and well-differentiated groups. Significant pathways and processes in moderate/poor group only were marked with a dotted square. Pathways and biological processes in the bar charts were sorted with terms significant in both groups at the top, and terms significant in moderate/poor group only at the bottom (indicated with red dashed boxes)
Fig. 8
Mutation signatures, somatic CNA/LOH and mutation profiles for identified SMGs in cSCC tumour samples in SCC cell lines. a Mutation signatures across all SCC cell lines. Again, signature 7 and 32 were the most dominant signatures. IC immunocompetent, IS immunosuppressed, Aza confirmed azathioprine exposure, NC no confirmed azathioprine exposure, SE sun exposed site, U unknown if sun exposed site b OncoPrint of copy gain, loss and CN-LOH segments. c Mutation OncoPrint for 22 SMGs detected in cSCC tumour samples across cell lines, ordered by aberration frequency. d Mutation OncoPrint for WD/MD/PD driver genes detected in cSCC tumour samples across cell lines, ordered by frequency. MD/PD-specific genes found in cSCC samples are highlighted in orange
References
Publication types
MeSH terms
Substances
Grants and funding
- 13044/CRUK_/Cancer Research UK/United Kingdom
- 250170/ERC_/European Research Council/International
- 29802/CRUK_/Cancer Research UK/United Kingdom
- A13044/CRUK_/Cancer Research UK/United Kingdom
LinkOut - more resources
Full Text Sources
Other Literature Sources
Medical
Research Materials