Abstract
Intra-host tumor virus variants may influence the pathogenesis and treatment responses of some virally-associated cancers. However, the intra-host variability of Kaposi sarcoma-associated herpesvirus (KSHV), the etiologic agent of Kaposi sarcoma (KS), has to date been explored with sequencing technologies that possibly introduce more errors than that which occurs in the viral population, and these studies have only studied variable regions. Here, full-length KSHV genomes in tumors and/or oral swabs from 9 Ugandan adults with HIV-associated KS were characterized. Furthermore, we used deep, short-read sequencing using duplex unique molecular identifiers (dUMI)–random double-stranded oligonucleotides that barcode individual DNA molecules before library amplification. This allowed suppression of PCR and sequencing errors to ~10−9/base as well as afforded accurate determination of KSHV genome numbers sequenced in each sample. KSHV genomes were assembled de novo, and rearrangements observed were confirmed by PCR and Sanger sequencing. 131-kb KSHV genome sequences, excluding major repeat regions, were successfully obtained from 23 clinical specimens, averaging 2.3x104 reads/base. Strikingly, KSHV genomes were virtually identical within individuals at the point mutational level. The intra-host heterogeneity that was observed was confined to tumor-associated KSHV mutations and genome rearrangements, all impacting protein-coding sequences. Although it is unclear whether these changes were important to tumorigenesis or occurred as a result of genomic instability in tumors, similar changes were observed across individuals. These included inactivation of the K8.1 gene in tumors of 3 individuals and retention of a region around the first major internal repeat (IR1) in all instances of genomic deletions and rearrangements. Notably, the same breakpoint junctions were found in distinct tumors within single individuals, suggesting metastatic spread of rearranged KSHV genomes. These findings define KSHV intra-host heterogeneity in vivo with greater precision than has been possible in the past and suggest the possibility that aberrant KSHV genomes may contribute to aspects of KS tumorigenesis. Furthermore, study of KSHV with use of dUMI provides a proof of concept for utilizing this technique for detailed study of other virus populations in vivo.
Kaposi sarcoma (KS) is a leading cancer in sub-Saharan Africa and in persons with HIV co-infection. Kaposi sarcoma-associated herpesvirus (KSHV, also referred to as human herpesvirus-8, or HHV-8) is the etiologic agent of KS, but the factors that contribute to the development of KS, which occurs in only a small subset of infected individuals, remain largely unknown. While strain differences or mutations in other tumor viruses are known to affect the risk and progression of their associated cancers, whether genetic variation in KSHV is important to the natural history of KS is unclear. Most studies of KSHV diversity have only characterized ~4% of its 165-kb genome, and the observed variation in some studies is likely to have been impacted by PCR or cloning artifacts. To precisely define genomic diversity of KSHV in vivo, we evaluated full-length viral genomes (except the internal repeat regions) using a technique that greatly lowers sequencing error rates and thus measures genomic diversity much more accurately than previous studies. In addition, we extended our analyses to look for potential tumor-specific changes in the KSHV genomes by examining viruses in both tumor and non-tumor tissues. To these ends, we performed highly sensitive, single-molecule sequencing of whole KSHV genomes in paired KS tumors and oral swabs from 9 individuals with KS. We found that KSHV genomes were virtually identical within the 9 individuals, with no evidence of quasispecies formation or multi-strain infection. However, KSHV genome aberrations and gene-inactivating mutations were found to be common in KS tumors, often impacting the same genes and genomic regions across individuals. Whether theses mutations influence KS tumorigenesis or result from genomic instability commonly found in tumors warrants further study. Lastly, aberrant KSHV genomes were found to be shared by distinct tumors within individuals, suggesting the capacity of KS tumor cells to metastasize and seed new lesions.
Collapse