Intra-Patient Genomic Variations of Human Papillomavirus Type 31 in Cervical Cancer and Precancer

Abstract

Human papillomavirus type 31 (HPV31) is detected less frequently in cervical cancer than two major causative types, HPV16 and HPV18. Here, we report a comprehensive analysis of HPV31 genome sequences in cervical lesions collected from Japanese women. Of 52 HPV31-positive cervical specimens analyzed by deep sequencing, 43 samples yielded complete genome sequences of around 7900 base pairs and 9 samples yielded partially deleted genome sequences. Phylogenetic analysis showed that HPV31 variant distribution was lineage A in 19 samples (36.5%), lineage B in 28 samples (53.8%), and lineage C in 5 samples (9.6%), indicating that lineage B variants are dominant among HPV31 infections in Japan. Deletions in the viral genome were found in the region from the E1 to L1 genes, but all the deleted genomes retained the E6/E7 genes. Among intra-patient nucleotide variations relative to a consensus genome sequence in each sample, C-to-T substitutions were most frequently detected, followed by T-to-C and C-to-A substitutions. High-frequency, intra-patient mutations (>10%) in cervical cancer samples were found in the E1, E2, and E7 genes, and all of them were nonsynonymous substitutions. The enrichment of high-frequency nonsynonymous substitutions strongly suggests that these intra-patient mutations are positively selected during the development of cervical cancer/precancer.