Abstract 3087: Identification of EBV infection and its life cycle in pan-cancer samples

Abstract

Abstract Viral infection is associated with 10-15 percent of human cancers and EBV (human herpesvirus 4 (HHV-4)) contributes to cancer development in diverse cancer types. The life cycle of EBV is divided into latent and lytic cycles. Lytic replication cycle begins when early transcription factors are activated, and initiation complex is formed. Subsequently, lytic antigens are expressed once infected cells enter the lytic cycle. But the contribution of each cycle in cancer development is not well known yet. Hence we explored EBV integration and gene expression pattern by analyzing TCGA whole transcriptome sequencing (WTS) data.We investigated the integration of EBV by using FASTQ-formatted TCGA tumor WTS data. We designed a three-step EBV detection workflow. 1) Overall viral integration into human genome sequence was detected by VirusSeq. 2) “Spliced Transcripts Alignment to a Reference” algorithm was used for mapping against a hybrid reference genome of human and EBV. 3) To quantify the expression of EBV-related genes, we applied “RNA-Seq by Expectation Maximization” algorithm to obtain Transcripts Per Million values. Furthermore, we validated this detection workflow in EBV-positive cell lines from the public data repository and analyzed mutation profiles by using Mutation Annotation Format files.We investigated 851 WTS samples over 23 cancer types and scanned viral presentations by mapping against four types of EBV strains (Human HHV-4 complete wild-type, GD1, AG876, and artificial join). Consequently, we detected 88 samples mapped against the hybrid genome, and identified 46 of them were EBV-infected. By performing gene expression analysis, we recognized that 39/46 samples were expressing EBV lytic genes. Three types of cancers (Colorectal adenocarcinoma (COAD), Rectum adenocarcinoma (READ), and Stomach adenocarcinoma (STAD)) showed EBV lytic gene expression. Along with STAD, both BZLF1 and BALF2 were highly expressed in COAD; BZLF1 encodes an early transcription factor, and BALF2 is one of the viral genes constructs the initiation complex. Using our method, we also identified the expression of EBV lytic and latent genes in EBV-positive cell lines, hence validating our results. In addition, we found out an association between mutational events and EBV infection in COAD by profiling mutational signatures: increased C to T transitions and TP53 disruptions in EBV-infected samples. Overall, we could identify biologically relevant EBV-infected cancer samples considering EBV life cycle. We hereby drew two major conclusions based our findings. First, we established an analysis workflow to find out biologically meaningful EBV infection in cancer by quantifying gene expression of composing the EBV life cycle, beyond a simple viral integration. Second, we discovered EBV expression pattern in small number of colon cancer patients. Our integrative data will provide clues for comprehensively understanding of EBV-associated cancers. Citation Format: Hyojin Song, Hogune Im, Sung-Soo Yoon, Youngil Koh. Identification of EBV infection and its life cycle in pan-cancer samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3087.