Epstein–Barr virus (EBV) is a gammaherpesvirus that persistently infects over 90% of adults, usually without consequence (Queen et al., 2013). However, EBV is linked to a diverse array of lymphoid and epithelial malignancies, including Burkitt lymphoma (BL) (Queen et al., 2013). The virus is detected in 30% of sporadic BL (sBL), 50% of immunodeficiency associated BL (ID-BL) and virtually all endemic BL (eBL) (Leoncini et al., 2008). The possible contribution of EBV to BL pathogenesis is largely unknown and it is unclear how directly infection and disease are linked (Rochford et al., 2005). In fact, viral genes seem mostly to have triggering or accessory roles in disease, and are likely to be essential for cancer-cell survival only in the early phase of the neoplastic transformation in non-immunocompromised carriers (Vereide and Sugden, 2010). Moreover, the role of EBV is further confounded by the less than total association of the virus with histologically similar tumors. This may be explained by the hit-and-run hypothesis for viral-induced lymphomagenesis which proposes that after eliciting a heritable change in the gene-expression pattern of the host cell, the genome of tumor viruses may be completely lost (Ambinder, 2000; Minarovits et al., 2011). Following, cancers accumulate vast numbers of host mutations which become the main drivers of oncogenesis, promoting autonomous growth (Minarovits et al., 2011). Thus, it seems inevitable that a cancer, with time, will evolve to be independent from viral gene functions, allowing viral genome loss. This results in an inverse correlation between the number of viral genes expressed in these tumor cells and their associated cellular mutations (Rochford et al., 2005) as it has been recently demonstrated in cell lines and tumor samples (Abate et al., 2015). The main problem with the hit-and-run hypothesis has been lack of evidence in primary tumors; in addition, the studies present in the literature on the role of EBV in tumorigenesis have analyzed mainly EBV-positive cancers. Focusing on virus-positive cancers provide little information about genome loss, and the difficulty of analyzing spontaneous cancers, where the molecular changes driving transformation are almost always unknown, makes firm functional conclusions hard to draw (Stevenson et al., 2010).

To assess the presence of the virus in a specific sample, different approaches can be used, most of which are characterized by a high sensitivity. To date, the most employed methods for diagnostic purposes are immunohistochemistry and in situ hybridization (ISH) for EBV-encoded RNAs (EBER). However, they have a low specificity, and the accuracy of such assays has been recently called into question by molecular studies that showed the presence of the virus in samples previously diagnosed as EBER negative (Gallagher et al., 2003; Qi et al., 2013). More recently, microRNA (miRNA) expression has been shown to be a quite sensitive and specific tool to characterize normal and neoplastic cells, even for pathogens detection (Ma et al., 2016). By miRNA profiling we recently observed the expression of EBV-encoded miRNAs in one EBER-negative BL case (Abate et al., 2015; Piccaluga et al., 2016). Based on that, the aim of the present study was to identify the most suitable method to detect EBV infection in pathology samples by comparing different conventional (immunohistochemistry and EBER-ISH) and non-conventional (EBV-microRNAs detection and EBV viral load measurement) methods. We looked for EBV infection in 10 typical BL cases. Immunohistochemistry and EBER-ISH failed to identify the virus in six samples, whereas microRNAs expression profiling, quantitative reverse transcription PCR and viral load measurement identified a previous EBV exposure in all the specimens, also in those diagnosed as EBV negative by conventional tools. Our findings shed new light on the pathogenesis of EBV-related tumors, highlighting the role of the virus also in “EBV-negative” cases and proposing EBV-miRNAs searching as the most sensitive approach to identify also EBV vestiges. Because of the limitation of using EBV-miRNAs detection routinely, there is a lack of data to determine the true burden of EBV-associated cancers worldwide, and the data available is likely to under represent the epidemiology of EBV infection. Assessing that EBV might be responsible of a larger number of cancers than previously known may open the way for the opportunity of a large scale prevention tactic.

Epstein–Barr virus is thought to play a causative role in the pathogenesis of EBV-positive BL. The etiology of cases lacking its genome in neoplastic cells is poorly understood, but is has been suggested that EBV may use a hit-and-run mechanism in them. EBV supplies a normal B-cell information to survive and grow that are potentially oncogenic (hit). A healthy immune system removes these proto-tumor cells because they express recognized viral antigens (Rochford et al., 2005). EBV can avoid elimination of infected cells by inducing epigenetic alterations and silencing of targeted viral genes in the host cell (run) (Queen et al., 2013; Birdwell et al., 2014). These alterations are stable and heritable and would be maintained also after the loss of the virus (Queen et al., 2013; Birdwell et al., 2014). Such ability of EBV to manipulate the host machinery to silence its own gene expression and to reshuffle the cellular epigenome resulting in long lasting cellular consequences, can act as a mechanistic framework for the hit-and- run oncogenesis (Queen et al., 2013; Birdwell et al., 2014). Accordingly, selection favors a proto-tumor cell that evolves to reduce its dependence on the virus from proliferation and survival by gaining complementary cellular mutations (Ambrosio et al., 2015b). In more evolved tumors, the only viral genes expressed are those poorly immunogenic and non-immunogenic (i.e., EBER and viral-encoded miRNAs). Unfortunately the standard methods to detect EBV often fail to identify also these non-immunogenic molecules, but if the virus is truly responsible for all BL, all cases should theoretically show some evidence of the infection (Queen et al., 2013). Therefore, the challenge is to devise investigative strategies to prove or exclude hit-and-run oncogenesis (Ambinder, 2000).

In our study we investigated this possibility by applying conventional and non-conventional tools on a series of BL cases. In fact, the main problems with the hit-and-run hypothesis are the absence of proof in natural setting that led to track the fate of an infected cell, and the lack of evidence in primary tumors (Stevenson et al., 2010). The only data we have concerns in vitro studies. It has been shown that in some cases of episome loss from BL cell lines in tissue culture, fragments of the EBV genome are incorporated into cellular DNA (Srinivas et al., 1998). Furthermore, in a series of sBL, similar fragments of the EBV genome were found in tumors that by standard criteria (EBER-ISH or EBNA-1 immunohistochemistry) would be classed as virus-negative (Srinivas et al., 1998).

The present study adds new evidence to the existing literature demonstrating that the routine methods employed to identify EBV (i.e., immunohistochemistry and EBER-ISH) are disappointing as we identified EBV genome in cases diagnosed as EBNA-1 and EBER negative (Gallagher et al., 2003; Stevenson et al., 2010). We also provide evidence to propose miRNAs detection as the most specific and sensitive tool to identify even EBV vestiges (e.g., EBV exosomes) and diagnose a previous infection in “EBV-negative” patients. According to the hit-and-run hypothesis, our findings support the possibility that EBV might have contributed to lymphomagenesis in our samples by initiating the oncogenic process. Then EBV genome was lost at each cell cycle and no longer detectable by conventional methods, being the tumor under the selective pressure exerted by the immune system. In fact, every population of proliferating EBV-positive cells loses 8% of the viral genomes each cell cycle; after eight cycles, only 50% of the viral DNA will remain; after 50 cycles, only 1% will persist in the population (Vereide and Sugden, 2009, 2010). The few remaining EBV genome in neoplastic cells would then be responsible for the production of the detected miRNAs. In this regard, selective release and transfer of RNA via exosomes might play a relevant role as recently demonstrated in other settings (Baglio et al., 2016). Though our findings do not represent a definitive proof of the presence of EBV inside the neoplastic cells, they highlight for the first time the possibility that EBV might contribute to the development of more cancers than simply those remaining viral genome-positive. In fact, the virus may impact on host cell homeostasis in various ways by interfering with cellular miRNAs expression and by encoding its own genes and miRNAs (Lenze et al., 2011). EBV-miRNAs may compete with miRNAs machinery and target cellular genes, thus dysregulating key pathways.

Whether confirmed on a larger cohort of cases and different tumor types, the current study may further support the rationale for strengthening the effort toward EBV vaccines that could potentially prevent the development of EBV-associated neoplasms independently of the presence or absence of viral genomes in the neoplastic cells, thus affecting the worldwide epidemiology of lymphomas. This is an idea worth considering, and seems to be realistic, because it has been demonstrated that vaccinia against EBV superficial antigens could protect from or delay EBV infection in infants (Stevenson et al., 2010).

LM analyzed data and performed the experiment, MA analyzed data and draft the work, MP carried out microRNAs profiling, GLB, LDP, and MN performed histology, extraction and quality assessment of the tissue samples. SG, MG, NO, and GDF contributed samples and tools, DG reviewed the manuscript, PP designed the study and analyzed data, LL designed the study, analyzed data and reviewed critically the manuscript. All the authors approved the final version to be published.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.