Tumor Virus-Induced Host Network Perturbations Identify Cancer Genes

DNA tumor virus proteins can promote tumorigenesis by physically targeting host cell proteins such as RB and p53. Based on these classical examples, Rozenblatt-Rosen and colleagues hypothesized that interactions between DNA tumor virus proteins and host gene products generally promote tumorigenesis in similar ways as somatic mutations and that a systematic analysis of virus–host interactions might identify additional cancer genes. A yeast two-hybrid screen of viral and human open reading frames (ORF) and tandem affinity purification of epitope-tagged viral proteins expressed in human cells followed by mass spectrometry identified a significantly overlapping set of interactions among human papillomavirus (HPV), Epstein-Barr virus, adenovirus, and polyomavirus proteins and human proteins. The transcriptomes of viral ORF-transduced cell lines were also profiled to identify host genes and pathways that were perturbed by viral proteins. These complementary approaches established a virus–host perturbation network that revealed that specific host genes or pathways were commonly affected by viral proteins from multiple types of DNA tumor viruses, suggesting that these perturbations may be oncogenic events. Many viral proteins also targeted specific host genes or biologic processes, and network analysis and clustering highlighted similarities among viral proteins from different tumor virus types. This observation that homologous proteins within the same tumor virus family, such as HPV E6 proteins, had distinct host interactions provides insight into the biologic mechanisms underlying the different cancer risk associated with each virus type as well as the etiology of viral disease. Importantly, the host targets identified were significantly enriched for known causal cancer genes, suggesting that the virus–host perturbation network could be effectively used in combination with tumor sequencing or other cancer gene discovery methods to prioritize candidate driver mutations.

Rozenblatt-Rosen O, Deo RC, Padi M, Adelmant G, Calderwood MA, Rolland T, et al. Interpreting cancer genomes using systematic host network perturbations by tumour virus proteins. Nature 2012 Jul 18 [Epub ahead of print].