Comparative molecular profiling of HPV-induced squamous cell carcinomas

Abstract

Approximately 5% of all cancer incidences result from human papillomavirus (HPV) infection. HPV infection most commonly leads to cancers of the anogenital region or oropharynx. It is unknown whether different HPV-mediated cancers collectively share a molecular signature and it is important to determine if there are targetable alterations common to different types of HPV-positive tumors. We analyzed 743 p53 wild-type samples of anal, cervical, oropharyngeal, and vulvar squamous cell carcinomas which underwent multiplatform testing at a commercial molecular profiling service. Expression of 24 proteins was measured by immunohistochemistry (IHC), mutation of 48 genes was determined by next-generation and Sanger sequencing, and copy number alteration for six genes was determined by in situ hybridization. The four cohorts had remarkably similar molecular profiles. No gene had a statistically significant difference in mutation frequency or copy number change between the four different types of squamous cell carcinomas. The only significant differences between cohorts were frequency of ERCC1 and SPARC loss as determined by IHC. In all four cancer types, oncogene mutation and PD-L1 expression was relatively infrequent. The most commonly mutated gene was PIK3CA, with mutations most often affecting the helical domain of the protein and accompanied by concurrent lack of PTEN expression. Loss of MGMT and RRM1 was common among the four cohorts and may be predictive of response to cytotoxic therapies not currently being used to treat these cancer types. The similar molecular profiles of the four cohorts indicate that treatment strategies may be similarly efficacious across HPV-positive cancers.

The manuscript describes our multiplatform testing and analysis of 743 samples of anal, cervical, oropharyngeal, and vulvar squamous cell carcinomas to identify molecular signatures common to HPV-induced cancers. We specifically focused on biomarkers which may be predictive of treatment efficacy and aid in determining treatment regimens. Examination of 79 biomarkers revealed striking similarities between the different cancers types, including similar rates of oncogene mutation, gene amplification, PD-L1 expression, and MGMT and RRM1 silencing which indicates that treatment protocols may be similarly effective in these different cancers.

http://ift.tt/2r3yLcO