Reduced local mutation density in regulatory DNA of cancer genomes is linked to DNA repair

Abstract

Carcinogenesis and neoplastic progression are mediated by the accumulation of somatic mutations. Here we report that the local density of somatic mutations in cancer genomes is highly reduced specifically in accessible regulatory DNA defined by DNase I hypersensitive sites. This reduction is independent of any known factors influencing somatic mutation density and is observed in diverse cancer types, suggesting a general mechanism. By analyzing individual cancer genomes, we show that the reduced local mutation density within regulatory DNA is linked to intact global genome repair machinery, with nearly complete abrogation of the hypomutation phenomenon in individual cancers that possess mutations in components of the nucleotide excision repair system. Together, our results connect chromatin structure, gene regulation and cancer-associated somatic mutation.

Figure 1

Relative density of somatic mutations is reduced in DHSs of all analyzed cancer genomes (lung, melanoma, colon, multiple myeloma). Mutation density per (uniquely mappable) bp is shown for 1) DHS maxima defined as plus or minus 75 bp around the peak of DNase I hypersensitivity (marked as DHS peaks), 2) DHSs, 3) 1000 bp flanking regions and 4) overall genome. Mutation density in DHSs is substantially lower in comparison with immediate flanking regions and genome average. The effect is stronger for DHS maxima compared to overall DHSs.

Figure 2

Density of somatic C:G→T:A transition mutations in melanoma samples strongly depends chromatin accessibility in a monotonic and continuous fashion. Density of C:G→T:A transitions per C:G base-pair in 400bp genomic intervals is shown as function of chromatin accessibility in melanocytes measured by the density DNase I cleavages. The dependence is presented separately for introns and intergenic regions, and is equally present in both. Mutation densities are parametrically fitted to a spline function using a Generalized Additive Model Poisson regression model.

Figure 3

Normalization of DHS hypomutation in melanoma genomes with mutated nucleotide excision repair pathway genes. Relative mutation density in DHSs of melanoma genomes is shown for samples with an intact NER system (blue) and samples with non-synonymous mutations in NER pathway genes (red). Non-synonymous changes in NER pathway genes significantly track relative mutation reduction in DHSs (P < 0.0237, Wilcoxon-Mann-Whitney test).

Figure 4

Reduction of mutation density in DHSs and in transcribed regions. Shown for individual melanoma samples (scatter plot) are non-synonymous mutations in genes involved in NER (marked by shape and color corresponding to each gene). Roles of these genes within NER pathway are shown by the diagram on the right. XPG, XPF and LIG1 are core repair components; CETN2 and DDB2 are specific to GGR and are involved in lesion recognition. CSB is specific to TCR and is involved in recruiting NER to the stalled Pol II RNA polymerase. Samples with low level (or no) reduction of somatic mutations in DHSs and carrying non-synonymous changes in genes of core NER components also show low level (or no) reduction of mutation frequency in transcribed regions, suggesting that core NER genes are responsible for both effects.