Hypermutation in human cancer genomes: footprints and mechanisms

Key Points Increased spontaneous or environmentally enhanced mutagenesis often correlates with increased mutation load and cancer risk. Mutation loads of individual cancer genomes can differ by several orders of magnitude and the top mutation loads are defined as having a hypermutation phenotype. Recent accumulation of cancer genomics information caused a breakthrough in understanding the origin and mechanisms of hypermutation. Mutation rates and distribution across cancer genomes are influenced by features of genome structure and function, such as replication timing, transcription and chromatin structure. Mutation rates also increase in the vicinity of rearrangement breakpoints Mutation rates can vary depending on local DNA sequence and the kind of genetic change. These parameters are defined as mutation signatures. Statistical analysis of somatic mutation signatures in cancer genomes has deciphered new sources of hypermutation in cancer and has confirmed the role of classic carcinogenic mutagens in cancer hypermutation. Mutation signatures can be identified by large-scale statistical analysis (termed non-negative matrix factorization (NMF)) of complex genome-wide mutation catalogues, as well as through selecting a fraction of mutations enriched with a single mutagenic mechanism. The latter can be achieved by concentrating on groups of closely spaced mutations: that is, mutation clusters. Combining NMF methods with analyses that concentrate on clustered mutations helped to identify a new kind of strong and ubiquitous carcinogenic mutagen that acts endogenously — a subclass of apolipoprotein B mRNA editing enzyme catalytic polypeptide-like (APOBEC) cytidine deaminases. The use of these complementary techniques greatly enhanced the statistical power to analyse APOBEC-mediated mutagenesis in cancer. Merging statistical pattern analysis with mechanistic information is feasible for other sources of mutations for which vast mechanistic knowledge has been accumulated over past decades. This can lead to the identification of new environmental, occupational and endogenous sources of hypermutation, as well as to an understanding of their specific affects in different cancer types, and even in individual cancer samples. Recent analyses of cancer genomes have revealed the occurrence of mutation patterns, which indicate their source. This Review discusses what we have learned, and what is yet to learn, from these data and how our current understanding of cancer mutati