Impact of viral presence in tumor on gene expression in non-small cell lung cancer

Youngchul Kim, Christine M. Pierce, Lary A. Robinson

In our recent study, most non-small-lung cancer (NSCLC) tumor specimens harbored viral DNA but it was absent in non-neoplastic lung. However, their targets and roles in the tumor cells remain poorly understood. We analyzed gene expression microarrays to identify genes and pathways differentially altered between virus-infected and uninfected NSCLC tumors. Gene expression microarrays of 30 primary and 9 metastatic NSCLC patients were preprocessed through a series of quality control analyses. Linear Models for Microarray Analysis and Gene Set Enrichment Analysis were used to assess differential expression. Various genes and gene sets had significantly altered expressions between virus-infected and uninfected NSCLC tumors. Notably, 22 genes on the viral carcinogenesis pathway were significantly overexpressed in virus-infected primary tumors, along with three oncogenic gene sets. A total of 12 genes, as well as seven oncogenic and 133 immunologic gene sets, were differentially altered in squamous cell carcinomas, depending on the virus. In adenocarcinoma, 14 differentially expressed genes (DEGs) were identified, but no oncogenic and immunogenic gene sets were significantly altered. In bronchioloalveolar carcinoma, several genes were highly overexpressed in virus-infected specimens, but not statistically significant. Only five of 69 DEGs (7.2%) from metastatic tumor analysis overlapped with 1527 DEGs from the primary tumor analysis, indicating differences in host cellular targets and the viral impact between primary and metastatic NSCLC. The differentially expressed genes and gene sets were distinctive among infected viral types, histological subtypes, and metastatic disease status of NSCLC. These results support the hypothesis that tumor viruses play a role in NSCLC by regulating host genes in tumor cells during NSCLC differentiation and progression.