Recently we published a report on HPV positive head and neck cancer that was perhaps too late to be considered for this article.
Analysis of the cancer genome atlas sequencing data reveals novel properties of the human papillomavirus 16 genome in head and neck squamous cell carcinoma.
Nulton TJ, Olex AL, Dozmorov M, Morgan IM, Windle B.
Oncotarget. 2017 Feb 7. doi: 10.18632/oncotarget.15179.
The evidence presented in this manuscript overwhelmingly suggests that HPV16 genomes are not integrated in the large majority of HPV+HNSCC. What does happen on occasion is that the viral DNA joins with human DNA and by a variety of mechanisms discussed in the paper excises the human DNA and forms a circular genome consisting of viral and human DNA. This DNA mostly replicates in an E1-E2 dependent manner. The manuscript presents evidence supporting this statement. Therefore there are three types of HPV+HNSCC when it comes to viral genome status: 1) the viral genome is integrated; 2) the viral genome is episomal; 3) the viral genome is episomal and replicating as a viral-human hybrid. These are roughly split evenly in data from The Cancer Genome Atlas and over two thirds of the HPV+HNSCC therefore replicate episomally. There is very little evidence for so called "mixed" tumors in HPV+HNSCC, i.e. a tumor having both integrated and episomal viral genomes. Category 3 can be mis-categorized if you define integration by the presence of a viral-human hybrid sequence read which explains earlier reports discussing two-thirds to three quarters of the tumors having integrated DNA. Our more recent data (not published) suggests that integrated tumors are in fact becoming rarer and rarer, perhaps due to the cessation of smoking. We also predict that truly integrated tumors have a worse outcome clinically. Therefore it is important that the status of the viral genome is characterized correctly when considering de-escalation therapy for HPV+HNSCC.