DNA secondary structures and epigenetic determinants of cancer genome evolution

950 VOLUME 18 NUMBER 8 AUGUST 2011 NAture StructurAL & moLecuLAr bIoLogY Loss of genomic integrity is a common hallmark of cancer genomes1. Recent technological advances have led to several large-scale cancer genome profiling studies2–5 that have identified genome-wide patterns of alterations in many cancer samples. Notably, DNA breakpoints in cancer genomes, and also in the genomes of apparently healthy subjects, are distributed nonrandomly2,5–7, suggesting that some regions within the human genome—so-called breakpoint hotspots—are exquisitely prone to rearrangement of genetic material. Some of these regions are common across many cancer types, whereas others are specific to particular types, indicating that genomic instability may manifest itself differentially in neoplasms of diverse origin. Many exogenous factors (such as nicotine exposure in lung cancer) and endogenous factors (such as repeat elements) as well as molecular mechanisms can cause double strand breaks and erroneous DNA repair, leading to genomic alterations1,8–10. Under certain circumstances, DNA can adopt non-B conformations, and recently two such secondary structures (H-DNA and Z-DNA) were shown to contribute to DNA damage11–13. Guanine-rich sequences (G3N1–7 G3N1–7G3N1–7G3), which are frequent in the human genome, can adopt four-stranded structures called G-quadruplexes (G4) both in vivo and in vitro14–16. G4 structures obstruct the movement of DNA polymerase17, thereby increasing the risk of DNA breakage or nonallelic homologous recombination. Indeed, G4 structures have been implicated in germline deletion18,19 and recombination20 events. However, the role of G4 structures in genomic instability in cancer has so far not been systematically investigated. In addition to genetic factors, various epigenetic factors are also associated with genomic instability both during the somatic evolution of cancer21,22 and in germline evolution during speciation23. Moreover, epigenetic patterns differ between cell types and thus possess the potential to generate tissue-specific patterns of alterations. Selective epigenetic states such as CpG methylation interact with G4 (refs. 24–26) and other non–B-DNA structures27,28, potentially interfering with their formation and stability. The D4Z4 region, for instance, which is hypomethylated in some cancer types and hypermethylated in others, contains a subregion that is resistant to hypermethylation and harbors G4s motifs29. Furthermore, the CpG dinucleotide frequently resides within G4s, whose CpG methylation is usually low—especially at gene promoters, exons and untranslated regions30. These findings raise the possibility that the mutagenic potential of DNA secondary structures may be modulated by epigenetic states. Here we set out to systematically investigate the role of DNA secondary structures in genomic instability in cancer. We integrated published data on genomic alterations from over 2,700 cancer samples, as well as potentially G-quadruplex–forming sequences (PG4s) and DNA methylation. We propose that hypomethylation and G4 structures together could have a causal role in genomic instability in cancer, thus representing one of the mechanistic bases for tissue-specific mutational landscapes of cancers.