High-throughput sequencing analysis of a “hit and run” cell and animal model of KSHV tumorigenesis

Abstract

Kaposi's sarcoma (KS), is an AIDS-associated neoplasm caused by the KS herpesvirus (KSHV/ HHV-8). KSHV-induced sarcomagenesis is the consequence of oncogenic viral gene expression as well as host genetic and epigenetic alterations. Although KSHV is found in all KS-lesions, the percentage of KSHV-infected (LANA+) spindle-cells of the lesion is variable, suggesting the existence of KS-spindle cells that have lost KSHV and proliferate autonomously or via paracrine mechanisms. A mouse model of KSHVBac36-driven tumorigenesis allowed us to induce KSHV-episome loss before and after tumor development. Although infected cells that lose the KSHV-episome prior to tumor formation lose their tumorigenicity, explanted tumor cells that lost the KSHV-episome remained tumorigenic. This pointed to the existence of virally-induced irreversible oncogenic alterations occurring during KSHV tumorigenesis supporting the possibility of hit and run viral-sarcomagenesis. RNA-sequencing and CpG-methylation analysis were performed on KSHV-positive and KSHV-negative tumors that developed following KSHV-episome loss from explanted tumor cells. When KSHV-positive cells form KSHV-driven tumors, along with viral-gene upregulation there is a tendency for hypo-methylation in genes from oncogenic and differentiation pathways. In contrast, KSHV-negative tumors formed after KSHV-episome loss, show a tendency towards gene hyper-methylation when compared to KSHV-positive tumors. Regarding occurrence of host-mutations, we found the same set of innate-immunity related mutations undetected in KSHV-infected cells but present in all KSHV-positive tumors occurring en exactly the same position, indicating that pre-existing host mutations that provide an in vivo growth advantage are clonally-selected and contribute to KSHV-tumorigenesis. In addition, KSHV-negative tumors display de novo mutations related to cell proliferation that, together with the PDGFRAD842V and other proposed mechanism, could be responsible for

A non-coding RNA network involved in KSHV tumorigenesis

“…Unsupervised analysis of 14 samples based on miRNAs expression profiles shows how they cluster together in an unsupervised way according to their predefined features ( Figure 4A). Interestingly, the samples cluster in the same pattern as when the analysis was made for lncRNAs ( Figure 1B) and also for all host genes in our previous study [Naipauer et al, 2020]. The distance among groups is reflexed in the number of DE miRNAs ( Figure 4B; S4 Table).…”

“…To identify changes in host lncRNAs expression profile we analyzed the number of differentially expressed (DE; FC>1.5, pvalue <0.05) lncRNAs in key biological comparisons that were detected by RNA-sequencing analysis of: two KSHV (+) cells, two KSHV (-) cells, six KSHV (+) tumors, two KSHV (-) tumor cells and three KSHV (-) tumors (S1 Table). This mouse model allows for unique experimental comparisons in the same cell and KS-like mouse tumor types: 1) KSHV (-) cells versus KSHV (+) cells can be used to study KSHV mediated effects in vitro, 2) KSHV (-) tumors versus KSHV (+) tumors can be used to dissect the role of ncRNAs in tumorigenesis by comparing tumors driven by KSHV versus tumors driven by host mutations, 3) KSHV (+) cells grown in vitro and in tumors can be used to study in vitro versus in vivo variations induced by micro-environmental cues, and 4) KSHV (-) tumor cells versus KSHV (-) tumors can be used to study in vitro versus in vivo variations in the absence of KSHV [Naipauer et al, 2020]. We first analyzed lncRNAs expression in these comparisons and found that the highest number of DE lncRNAs was observed in KSHV (+) tumors in both comparisons versus KSHV (-) tumors and versus KSHV (+) cells ( Figure 1D; S1 Table).…”

“…We performed heat map representations of all or top-50 DE microRNAs -according to each comparison-for all the 4 biological relevant comparisons mentioned previously ( Figure 4C). To identify the experimentally supported targets from our previous published work [Naipauer et al, 2020] for the DE miRNAs identified in this study we employed DIANA TARBASE v8 [Karagkouni et al, 2018]. Next, we selected those targets whose expression antagonizes with that of its miRNA in the corresponding comparison ( Figure 5A; S5 Table).…”

“…Using deep RNA sequencing of KSHV Bac36 transfected mouse endothelial cells (mECK36) and tumors [Mutlu et al, 2007], we have previously analyzed the host and viral transcriptome to characterize mechanisms of KSHV dependent and independent sarcomagenesis as well as the contribution of host mutations [Naipauer et al, 2020]. We now decided to study in this model, in a genome-wide fashion, the ncRNAs landscape to better understand the relationship between mRNAs, lncRNAs and miRNAs in shaping KSHV tumorigenesis mechanisms.…”

Designer nucleases to treat malignant cancers driven by viral oncogenes

“…There are recent postulations for a ‘hit-and-run’ hypothesis suggesting that the accumulation of other genomic mutations could supersede the tumor’s dependence on the viral oncogenes resulting in a virus-independent tumor (Reviewed in [ 130 ]). This virus-independence has been observed with the spontaneous loss of EBV episomes in a Burkitt’s lymphoma (BL) cell line [ 131 ], or KSHV transformed cell lines that are episome negative but still tumorigenic in mice [ 132 ], and even in instances of HPV-related cancers, often considered a clear example of oncogenic addiction, known HPV18 transformed cell lines can lack HPV DNA [ 133 ]. Future studies should focus on clarifying the context and extent to which these cancers are dependent on viral elements, as this insight would allow for the cognizant application of nucleases in the treatment of virus-associated cancers.…”