Analysis of The Cancer Genome Atlas sequencing data reveals novel properties of the human papillomavirus 16 genome in head and neck squamous cell carcinoma

Abstract

Human papillomavirus (HPV) DNA is detected in up to 80% of oropharyngeal carcinomas (OPC) and this HPV positive disease has reached epidemic proportions. To increase our understanding of the disease, we investigated the status of the HPV16 genome in HPV-positive head and neck cancers (HNC). Raw RNA-Seq and Whole Genome Sequence data from The Cancer Genome Atlas HNC samples were analyzed to gain a full understanding of the HPV genome status for these tumors. Several remarkable and novel observations were made following this analysis. Firstly, there are three main HPV genome states in these tumors that are split relatively evenly: An episomal only state, an integrated state, and a state in which the viral genome exists as a hybrid episome with human DNA. Secondly, none of the tumors expressed high levels of E6; E6*I is the dominant variant expressed in all tumors. The most striking conclusion from this study is that around three quarters of HPV16 positive HNC contain episomal versions of the viral genome that are likely replicating in an E1-E2 dependent manner. The clinical and therapeutic implications of these observations are discussed.

Keywords: HPV16; TCGA; genomic structure; head and neck cancer; integration.

Figure 1. HPV16 Genomic DNA Profiles

Plots are shown for three HNC samples with HPV16 DNA showing the HPV DNA present in each sample in copies per cell. The results displayed are based on the content for a 20-base interval across the 7904 bp HPV16 genome. The three samples are: A. TCGA-CV-5442, B. TCGA-CR-5249, and C. TCGA-CR-6482.

Figure 2. HPV16 RNA Expression Profiles

Plots are shown for three HNC samples with HPV16 DNA showing the HPV RNA sequence expression in each sample in units of Reads Per Kilobase per Million (RPKM). The results displayed are based on the content for a 20-base interval across the 7904 bp genome. The three samples are: A. TCGA-CV-5442, B. TCGA-CR-5249, and C. TCGA-CR-6482.

Figure 3. Human Genomic DNA Profiles of Amplified Regions

Plots are shown of copies per cell for three human DNA regions with junctions to HPV DNA and amplification in three different samples. The results displayed are based on the DNA content for a 1000-base interval across the genomic region shown for TCGA-CR-6482 and TCGA-BA-4077, and a 500-base interval for TCGA-CV-5971. A. Plot shows a region of Chr2 amplified in TCGA-CR-6482. The red arrow indicates the site of HPV recombination with human DNA and junctions between HPV and human DNA. The blue arrows indicate the sites for the ends of the human DNA amplicon and the sites that have recombined with each other to form an excision junction. B. Plot shows a region of Chr14 amplified in TCGA-BA-4077. The green arrows indicate the sites of the ends of the human DNA amplicon and the sites that have recombined with HPV DNA to form HPV-human DNA junctions. C. Plot shows a region of Chr20 amplified in TCGA-CV-5971. The green arrows indicate the sites for the ends of the human DNA amplicon and the sites that have recombined with HPV DNA to form HPV-human DNA junctions.

Figure 4. A Proposed Structure for the HPV Dimer Human DNA Hybrid Episome

A circular structure with HPV16 and human DNA is shown. Two copies of HPV DNA are linked as a dimer in direct tandem, shown in blue. A partial deletion in one copy of the HPV genome is shown as a gap. Human DNA, shown in red, is joined to the HPV DNA at the sites of the partial deletion. Regions for the HPV genes are shown as color-coded bars.

Figure 5. Proposed Mechanistic Models of HPV Recombining with the Human Genome and Excising

A. A diagram with a series of steps showing the integration of the HPV genome into the human genome, unscheduled replication from the HPV origin to form a replication bubble, excision of the HPV genome along with human DNA to form a circle, and homologous recombination repair of the human genome. The green and blue colored regions on the human DNA represent sites of excision and human-to-human DNA end joining. B. A diagram with a series of steps showing breakage of human and HPV genomic DNA, joining of HPV and human DNA, excision of HPV and human DNA to form a circle, and homologous recombination repair of the human genome. The green and blue colored regions on the human DNA represent sites of breakage and joining of human DNA to HPV DNA. Single lines represent double-stranded DNA. Black lines represent human genomic DNA. Red lines represent HPV genomic DNA. Small black arrows mark the sites of DNA breakage and excision. Wavy lines indicate DNA crossovers.

Figure 6. E7 Gene Expression Versus HPV Copy Number

A plot showing no relationship between HPV copy number and E7 gene expression. Copy number per cell and E7 gene expression level in each HNC sample were plotted. The R² shows no significant correlation.