Mutations in the HPV16 genome induced by APOBEC3 are associated with viral clearance

Abstract

HPV16 causes half of cervical cancers worldwide; for unknown reasons, most infections resolve within two years. Here, we analyze the viral genomes of 5,328 HPV16-positive case-control samples to investigate mutational signatures and the role of human APOBEC3-induced mutations in viral clearance and cervical carcinogenesis. We identify four de novo mutational signatures, one of which matches the COSMIC APOBEC-associated signature 2. The viral genomes of the precancer/cancer cases are less likely to contain within-host somatic HPV16 APOBEC3-induced mutations (Fisher's exact test, P = 6.2 x 10^-14), and have a 30% lower nonsynonymous APOBEC3 mutation burden compared to controls. We replicate the low prevalence of HPV16 APOBEC3-induced mutations in 1,749 additional cases. APOBEC3 mutations also historically contribute to the evolution of HPV16 lineages. We demonstrate that cervical infections with a greater burden of somatic HPV16 APOBEC3-induced mutations are more likely to be benign or subsequently clear, suggesting they may reduce persistence, and thus progression, within the host.

Fig. 1. Four de novo mutational signatures identified using all variants across the HPV16 genome in women from the PaP cohort.

The x-axis indicates the 5′ and 3′ nucleotides for each of the top panel substitutions for the three base-pair motifs. The y-axis shows the single-base substitution (SBS) composition of each mutational signature by the 96 trinucleotide sequence motifs. For each identified signature, shown as A–D, the similarity was determined to the known COSMIC SBS signatures (https://cancer.sanger.ac.uk/cosmic/signatures/SBS/). The identified signature letter (A–D) and in parentheses the most similar COSMIC SBS signature number along with the cosine similarity are shown along the right y-axis. Cosine similarity ranges from 0 to 1, with a cosine of 1 indicating a perfect match.

Fig. 2. Frequency of the 96 trinucleotide mutation types for variants across the HPV16 genome in women from the PaP cohort.

Illustrated for (a) high variant allele fraction (VAF) and (b) low VAF variants. The x-axis indicates the 5′ and 3′ nucleotides for each of the top panel substitutions for the three base-pair motifs.

Fig. 3. The number of APOBEC3-induced mutations across the HPV16 genome by gene region in women from the PaP cohort.

The plots show only the cases and controls with one or more APOBEC3-induced mutations at a (a) high variant allele fraction (VAF) and (b) low VAF. Each vertical line represents a sample with at least one APOBEC3-induced mutation, colored by the number of mutations observed, as 1–3 (see legend), per viral gene region. Samples with no APOBEC3-induced mutations are not illustrated; the size of the case and control panels correspond to the number of individuals with at least one APOBEC3-induced mutation. The samples are organized along the x-axis by status (case vs. control). Cases are cervical intraepithelial neoplasia grade 3 and cancer cases (CIN3+). The right y-axis represents viral gene regions with the overall frequency of APOBEC3 mutations summarized, taking into account the sample sizes of the cases/controls and potential APOBEC3-mutable sites, for CIN3+ cases in dark gray and controls in light gray. The top panel histogram summarizes the total APOBEC3-induced mutations for the cases and controls across the HPV16 genomes. URR upstream regulatory region, E6 early gene 6, E7 early gene 7, E1 early gene 1, E2 early gene 2, E4  early gene 4, E5  early gene 5, L2  late gene 2, L1  late gene 1.

Fig. 4. Summary of the effects of mutations induced by the activity of APOBEC3.

Viral APOBEC3-induced mutations are illustrated in the circle plot by viral gene region in the inner ring for low variant allele fraction (VAF) somatic mutations in red, and the outer ring for high VAF constitutive variants in blue for all individuals in the PaP cohort (mutations in both cases and controls are illustrated). Modified from the Cancer Genome Atlas Research Network.