Characterization of high-risk Epstein-Barr virus variants in nasopharyngeal carcinoma

Abstract

Nasopharyngeal carcinoma (NPC) is a rare malignancy in the world, but it is endemic in the region of South China including Hong Kong. It is closely linked to the infection of Epstein-Barr virus (EBV). While the majority of the carriers of EBV do not develop NPC, EBV occasionally infects nasopharyngeal epithelium and promotes NPC development. The differences between the strains involved in NPC and those carried by the healthy population are not clearly understood. Studies on EBV genetics have suggested that EBV derived from NPC patients in endemic regions are genetically distinct from those derived from patients of other diseases or geographic regions where NPC is less prevalent, pointing to the contribution of EBV genetics to NPC. However, the viral genetic variations that lead to a higher risk of NPC are poorly characterized.

This thesis begins with a case-control study comparing the EBV genomes in the biopsies of Chinese NPC patients and the saliva of Chinese healthy donors in Hong Kong. Clustering of the EBV genomes based on their genetic variants revealed that the EBV in Hong Kong Chinese population consists of several major subpopulations that have different levels of risk of NPC. Genome-wide association study (GWAS) identified the associations between NPC and EBV variants near EBER2, which is one of the EBV-encoded small RNAs. The lead variant in this locus was a 4-bp deletion located downstream of EBER2. Variants within EBER2 were predicted to alter the secondary structure of the RNA product of EBER2. Moreover, the risk alleles of NPC-associated variants were found to exclusively exhibit high frequencies in Chinese NPC. Following up the case-control study, a replication study and a meta-analysis reinforced the association signal at the EBER2 locus. Phylogenetic analysis showed that the NPC-associated subpopulations reported in different studies collectively constitute high-risk EBV lineages. Further identification of variants strongly linked to EBER2 and BALF2, which were reported in our and other studies respectively, suggested possible associations in BVRF2 and BARF1. Moreover, the high-risk lineages were characterized by variants in EBER2, BALF2, BVRF2, BARF1 and possibly yet reported causal variants for NPC. Lastly, the value of EBV variants in the saliva in the risk assessment of NPC was explored. Risk prediction models that incorporate any of the EBV variants, in addition to age and sex, have significantly improved abilities to discriminate NPC and healthy saliva.

To conclude, EBV strains carrying the risk alleles in the EBER2 locus are strongly associated with NPC. The high-risk EBV lineages are characterized by variations in EBER2, BALF2, BVRF2, BARF1 and potentially other yet discovered causal variants. An investigation into their functions may help to elucidate the pathogenic roles of EBV. The genetic variations in salivary EBV are predictive of NPC, suggesting their potential utility in non-invasive risk assessment of NPC.