A high-throughput next-generation sequencing-based method for detecting the mutational fingerprint of carcinogens

Abstract

Many carcinogens leave a unique mutational fingerprint in the human genome. These mutational fingerprints manifest as specific types of mutations often clustering at certain genomic loci in tumor genomes from carcinogen-exposed individuals. To develop a high-throughput method for detecting the mutational fingerprint of carcinogens, we have devised a cost-, time- and labor-effective strategy, in which the widely used transgenic Big Blue mouse mutation detection assay is made compatible with the Roche/454 Genome Sequencer FLX Titanium next-generation sequencing technology. As proof of principle, we have used this novel method to establish the mutational fingerprints of three prominent carcinogens with varying mutagenic potencies, including sunlight ultraviolet radiation, 4-aminobiphenyl and secondhand smoke that are known to be strong, moderate and weak mutagens, respectively. For verification purposes, we have compared the mutational fingerprints of these carcinogens obtained by our newly developed method with those obtained by parallel analyses using the conventional low-throughput approach, that is, standard mutation detection assay followed by direct DNA sequencing using a capillary DNA sequencer. We demonstrate that this high-throughput next-generation sequencing-based method is highly specific and sensitive to detect the mutational fingerprints of the tested carcinogens. The method is reproducible, and its accuracy is comparable with that of the currently available low-throughput method. In conclusion, this novel method has the potential to move the field of carcinogenesis forward by allowing high-throughput analysis of mutations induced by endogenous and/or exogenous genotoxic agents.

Figure 1.

cII mutant frequency and mutation spectrum in UVB-irradiated cells versus control. Mutation analysis of the cII gene in mouse embryonic fibroblasts irradiated with UVB or control was performed using the cII mutagenesis assay, as described in ‘Materials and Methods’. (A) Absolute mutant frequency of each specific type of mutation in the cII gene of UVB-irradiated cells or control as determined by both the new method (NGS) and the conventional method (ABI). Average results (bars) from multiple analyses plus 95% confidence interval (error bars) are shown. (B) Percentage increase in frequency of each specific type of mutation in the cII gene of UVB-irradiated cells or control as determined by both the new method (NGS) and the conventional method (ABI). Results are expressed as ‘induced mutation (%)’, which is calculated as [(mutant frequency of each type of mutation in UVB group − mutant frequency of the respective type of mutation in control group × 100)/(overall induced mutant frequency in UVB group − overall spontaneous mutant frequency in control group)]. Distribution of mutations in the cII gene of UVB-irradiated cells (C) and control (D) as determined by the new method (NGS) and/or the conventional method (ABI).

Figure 2.

cII mutant frequency and mutation spectrum in 4-ABP-treated mice versus control. Mutation analysis of the cII gene in bladder DNA of mice treated with 4-ABP or control was performed using the cII mutagenesis assay, as described in ‘Materials and Methods’. (A) Absolute mutant frequency of each specific type of mutation in the cII gene of 4-ABP-treated mice or control as determined by both the new method (NGS) and the conventional method (ABI). (B) Percentage increase in frequency of each specific type of mutation in the cII gene of 4-ABP-treated mice or control as determined by both the new method (NGS) and the conventional method (ABI). Results are expressed as ‘induced mutation (%)’, which is calculated as [(mutant frequency of each type of mutation in 4-ABP group − mutant frequency of the respective type of mutation in control group × 100)/(overall induced mutant frequency in 4-ABP group − overall spontaneous mutant frequency in control group)]. Distribution of mutations in the cII gene of 4-ABP-treated mice (C) and control (D) as determined by the new method (NGS) and/or the conventional method (ABI).

Figure 3.

cII mutant frequency and mutation spectrum in SHS-treated mice versus control. Mutation analysis of the cII gene in lung DNA of mice exposed to SHS or control was performed using the cII mutagenesis assay, as described in ‘Materials and Methods’. (A) Absolute mutant frequency of each specific type of mutation in the cII gene of SHS-exposed mice or control as determined by both the new method (NGS) and the conventional method (ABI). (B) Percentage increase in frequency of each specific type of mutation in the cII gene of SHS-exposed mice or control as determined by both the new method (NGS) and the conventional method (ABI). Results are expressed as ‘induced mutation (%)’, which is calculated as [(mutant frequency of each type of mutation in SHS group − mutant frequency of the respective type of mutation in control group × 100)/(overall induced mutant frequency in SHS group − overall spontaneous mutant frequency in control group)]. Distribution of mutations in the cII gene of SHS-exposed mice (C) and control (D) as determined by the new method (NGS) and/or the conventional method (ABI).

Figure 4.

Spontaneous cII mutation spectrum in control (sham-treated) mice. Mutation analysis of the cII gene in lung DNA of control mice (clean air exposed) was performed using the cII mutagenesis assay, as described in ‘Materials and Methods’. (A) Absolute mutant frequency of each specific type of mutation in the cII gene of control mice as determined by both the new method (NGS) and the conventional method (ABI). (B) Percentage of each specific type of mutation in the cII gene of control mice as determined by both the new method (NGS) and the conventional method (ABI). (C) Distribution of mutations in the cII gene of control mice as determined by both the new method (NGS) and the conventional method (ABI).

Figure 5.

Reproducibility of the cII mutation spectra established in duplicate/multiplicate samples of carcinogen-treated mice/cells versus control. The hierarchical clustering analysis (A) and the PCA (B) were performed using the Partek Genomics Suite v6.11.1116 (http://www.partek.com).

Figure 6.

Read coverage analysis for cII mutations in carcinogen-treated mice/cells versus control. For each sample, total number of reads was used as benchmark and subsequently, randomly selected 5×, 10×, 20×, 35×, 50× and 100× coverage analyses were performed as described in the text.