Integration of HBV DNA into the human genome is considered an important molecular event in HCC development20, 21, 24. Recently, a couple of studies have shown evidence of HBV DNA integration in extrahepatic cancers, like ICC9, 14, 28 and NHL8. However, little is known about the integrative status of HBV in GC. In this study, we performed a genome-wide analysis of HBV integration in 10 HBV-positive GC tumors and para-tumors using HIVID and long-read sequencing technologies. To our knowledge, it was revealed for the first time that HBV DNA can integrate into the genome of GC cells. Moreover, combining this with the corresponding in vitro data, we were able to pinpoint several HBV integrated genes, especially SPRY3 and CHD6, as potential driver genes in GC development.
Increasingly more researchers agree that HBV infection is not restricted to the liver. And more evidence suggests that a link exists between chronic HBV infection and GC6, 10, 29. Our previous study concluded that individuals with HBV seropositivity had a significantly increased risk of GC showing an OR of 1.29 (95% CI, 1.22, 1.37), and HBV-positive GC patients had poorer DFS and OS than patients with an HBV-negative status7, suggesting that HBV infection may perform as an adverse biomarker for predicting survival in GC patients. However, the direct evidence of HBV infection in GC tissue is scarce. Only Cui et al. once reported a higher expression levels of HBx protein in GC tissues by IHC analysis, and HBx-positive gastric epithelial cells demonstrated a higher nuclear-to-cytoplasmic ratio than HBx-negative cells13. In this study, HBV DNA was detectable in 70.0% (7/10) GC tissues and 80.0% (8/10) para-tumor tissues. To our knowledge, this is the first time that detected the HBV DNA in GC cells.
HBV integration was the key factor driving the formation of HCC, and research on integrated carcinogenic mechanisms has always been a hot topic30, 31. HBV DNA integration into the host genome occurs at early steps of clonal tumor expansion and induces both genomic instability and direct insertional mutagenesis of diverse cancer-related genes20. However, whether HBV integration events play a role in HBV-related GC has not been studied yet. Our study revealed for the first time that HBV DNA can integrate into the genome of GC cells. By HIVID sequencing, we identified 434 HBV DNA integrations sites, 175 from GC tissues, and 259 from para-tumor tissues. HBV DNA can integrate into para-tumor gastric tissues may suggest that HBV DNA integration is an early event occurring soon after the gastric infection, and has already existed in normal precancerous gastric tissues. Besides, the characterization of integrated HBV DNA segments in the host genome will enable a better understanding of the pathogenic mechanism and the role of HBV in GC. We analyzed the profile of HBV sequence adjacent to a viral-host junction and found that the integrated HBV sequences are significantly enriched in X gene, particularly at the 3’-end of the X gene. HBx, which has pleiotropic properties that support virus gene expression and replication, and contributes importantly to the development of HCC32. Previous studies showed that the integrated DNA of HBV in HCC usually contains the viral basal core promoter (BCP)/enhancer II (Enh II), and the sequence encoding C-terminal truncated HBx which is a pleiotropic transactivator32, 33. Recently, Zhang et al found that the C-terminal truncated HBx could mediate metabolic reprogramming from mitochondrial respiration to aerobic glycolysis by NFACT2-TXNIP signaling pathway, indicating that C-terminal truncated HBx plays a critical role in hepatocarcinogenesis34. The profile of the HBV sequence adjacent to viral-host junction displays a similar pattern in GC and HCC, speculating that the sustained expression of HBx may render the growth advantage to host cells and facilitate the clonal expansion during the tumor initiation and progression stages in GC development.
In addition to HIVID, we also used long-read sequencing to detect HBV integration sites and SVs in the GC samples. To the best of our knowledge, no articles on long-read sequencing for identifying HBV integration events in GC samples have been published. Long-read sequencing analysis enabled us to sequence long chimeric reads that contain at least one fragment of HBV DNA wrapped on both sides of the human genome. Since the sequencing was not based on HBV probe capture, the small number of HBV-human chimeric reads limited our analysis. Only two integration sites were sequenced, which might result in sampling bias. We found that HBV was not integrated with a complete genome, in case 2T, a length of ~ 1000 bp HBV DNA was inserted in the intron of gene LINC02107. In the case 2N, the genome was integrated by HBV DNA with a length of ~ 900 bp. We further revealed the spatial relationship between HBV integration and various SVs, and found an increased proportion of SVs located close to HBV integration sites was observed in DEL and INS, suggesting that HBV integration induces genomic instability and SVs that have substantial functional consequences in GC carcinogenesis. These results warrant further functional investigation in the future.
HBV insertion can result in host genome instability and cis-activation of the adjacent genes. Our results revealed that HBV seemed to have its preferential targets in GC. For instance, SPRY3 was repeatedly inserted by HBV DNA for three times in three distinct samples. Other protein-coding genes, like CHD6, CPNE4, DPP10, KLHL4, RBFOX1, and SLC6A15, were inserted by HBV DNA for more than once in different samples. Genes commonly targeted by HBV in HCC usually play a role in liver carcinogenesis21, 31. Thus, in this study, we first analyzed the mRNA expression and prognosis of the HBV interrupted gene SPRY3 by using the data from TCGA database, and found that SPRY3 had a high expression in GC, which was associated with shorter survival time. We also performed functional analysis of SPRY3 using loss-of-function assays and validated its oncogenic role in GC development. Sprouty (Spry) proteins were involved in RTK-driven signaling pathways, however, SPRY3 is rarely reported in the literature. It has been reported that SPRY3 potentiates a tumorigenic potential of glioblastoma cells by in-vitro functional experiments35. Sutterlüty et al revealed that SPRY3 accelerates osteosarcoma cell proliferation and migration, suggesting that SPRY3 as a candidate for a tumor promoter in osteosarcoma36. Besides, we also focused on the function of HBV integrated gene CHD6 in GC development. CHD6, which was integrated by HBV DNA for twice in the HBV capture sequencing, showed a gene amplification in the long-read sequencing. Moreover, functional analysis of CHD6 using in vitro validation of its oncogenic role in GC development. CHD6, a member of chromodomain helicase DNA binding protein (CHD) family, plays critical role in regulating gene transcription. Although studies of CHD6 in GC have not been previously reported, several papers have shown that CHD6 plays as a potential driver gene in colorectal cancer37 and prostate cancer38. The above results suggested the importance of HBV integration genes in GC, warranting further mechanistic investigation in the future.
There are several limitations to our study. First, we used the paired non-tumor gastric tissues from GC patients as controls, rather than normal gastric cells. From the current research, it is uncertain whether HBV DNA integrations confer the host cell a growth advantage over its neighbors. Second, we only detected 10 HBV-associated GC and para-tumor tissues with HIVID in this study. Therefore, the real pattern of HBV DNA integration in GC awaits further larger-scale investigations. Third, to efficiently identify genes that may be affected by HBV integration, we focused on those genes that are within or closest to HBV breakpoints. The trade-off is that some true target genes outside this range may be excluded. The last but not least, combined with the corresponding in-vitro data, we highlighted SPRY3 and CHD6 as potential driver genes in GC. However, further investigations are needed to fully understand the direct mechanism induced by HBV-integrated sequence.
Taken together, the present study reveals the preference of integration occurring within the whole host and virus genome, and the targeted genes recurrently affected by HBV DNA integration in GC tissues by both HIVID and long-read sequencing technology. Moreover, combined with the bioinformatics analysis and functional experiments, we highlight several HBV recurrent integration genes, especially SPRY3 and CHD6 as potential driver genes for GC. Finally, we revealed the spatial relationship between HBV integration and various SVs. All of these findings provide a new direction to research into the mechanism of HBV-associated GC initiation.