Profile of basal cell carcinoma mutations and copy number alterations - focus on gene-associated noncoding variants

Abstract

ABSTRACTBasal cell carcinoma (BCC) of the skin is the most common cancer in humans, characterized by the highest mutation rate among cancers, and is mostly driven by mutations in genes involved in the hedgehog pathway. To date, almost all BCC genetic studies have focused exclusively on protein-coding sequences; therefore, the impact of noncoding variants on the BCC genome is unrecognized. In this study, with the use of whole-exome sequencing of 27 tumor/normal pairs of BCC samples, we performed an analysis of somatic mutations in both protein-coding sequences and gene-associated noncoding regions, including 5’UTRs, 3’UTRs, and exon-adjacent intron sequences. Separately, in each region, we performed hotspot identification, mutation enrichment analysis, and cancer driver identification with OncodriveFML. Additionally, we performed a whole-genome copy number alteration analysis with GISTIC2. Of the >80,000 identified mutations, ∼50% were localized in noncoding regions. The results of the analysis generally corroborated the previous findings regarding genes mutated in coding sequences, including PTCH1, TP53, and MYCN, but more importantly showed that mutations were also clustered in specific noncoding regions, including hotspots. Some of the genes specifically mutated in noncoding regions were identified as highly potent cancer drivers, of which BAD had a mutation hotspot in the 3’UTR, DHODH had a mutation hotspot in the Kozak sequence in the 5’UTR, and CHCHD2 frequently showed mutations in the 5’UTR. All of these genes are functionally implicated in cancer-related processes (e.g., apoptosis, mitochondrial metabolism, and de novo pyrimidine synthesis) or the pathogenesis of UV radiation-induced cancers. We also found that the identified BAD and CHCHD2 mutations frequently occur in melanoma but not in other cancers via The Cancer Genome Atlas analysis. Finally, we identified frequent deletion of chr9q, encompassing PTCH1, and unreported frequent copy number gain of chr9p, encompassing the genes encoding the immune checkpoint ligands PD-L1 and PD-L2. In conclusion, this study is the first systematic analysis of coding and noncoding mutations in BCC and provides a strong basis for further analyses of the variants in BCC and cancer in general.Author summaryThe study is the first systematic analysis of both coding and noncoding mutations in basal cell carcinoma (BCC), the most common and the most highly mutated human cancer. Noncoding mutations accounted for ∼50% (∼40K mutations) of all mutations detected by the standard WES approach. Among the genes frequently mutated in noncoding regions are: BAD with a hotspot in the 3’UTR, DHODH with a hotspot in the Kozak sequence, and CHCHD2 with mutations in 5’UTR, all genes functionally related to cell metabolism and apoptosis. Analysis of copy number alterations showed frequent chr9q deletion, encompassing PTCH1 (the key BCC tumor suppressor), and frequent copy number gain of chr9p, encompassing the genes of the immune checkpoint proteins PD-L1 and PD-L2.