In recent years, many studies of herpesviruses and brain tumors have examined the association between CMV and high-grade astrocytomas, but the relationship is still controversial [31]. Also, no previous study has systematically examined CMV, HHV-6A, HHV-6B, and EBV viral DNA in low- and high-grade astrocytomas. The human herpesviruses have the ability to become latent in specific host cell types, primarily in the CNS and lymphoid tissue [32]. A review of recent studies that used immunohistochemistry (IHC), in situ hybridization (ISH), and PCR to detect CMV in GBM tumor specimens showed that results vary from undetected to 100 % positivity [18]. IHC and PCR each focus on different molecular targets, so detection selectivity discrepancies might be contributing to discordant results. Both techniques have their respective advantages and disadvantages. For instance, the interpretation of IHC is complicated due to background pigments, such as hematoidin, hemosiderin, formalin precipitates, and unspecific staining [5]. While PCR only focuses on a selected genomic sequence, off target amplification may result in false positives, and negative PCR results, i.e. the non-amplification of a specific amplicon, cannot be interpreted as the absence of the viral genome.
The absence of CMV in the GBM samples in this present study does not support a CMV genomic DNA and glioblastoma association, in contrast to other studies that have reported a high prevalence of CMV genomic DNA in glioblastoma by PCR [33–35]. Most studies have utilized either a protein or a genomic sequence targeting technique, but not both, to detect CMV. A few studies have proposed that certain onco-viruses might trigger viral DNA recombinogenic activities, which could promote oncogenesis with the secondary loss of the viral genome in tumor cells [36, 37]. Viruses that initiate such “hit and run oncogenesis” in the host may be absent at the time of screening.
Other than screening for known viral sequences or using viral-specific IHC assays, a less biased approach would be to sequence brain tumors and search for viral or non-human nucleic acid sequences, for example using a next-generation sequencing platform [38]. With this approach, a report in 2014 demonstrated EBV and HHV-6 viral sequences in 5/21 (24 %) and 1/21 (5 %) of GBM cases, respectively [16]. These viral positivity levels are consistent with the results of this present study, which detected EBV and HHV-6B in 4/45 (8.9 %) and 6/45 (13.3 %) of GBM cases, respectively. However, since the 2014 study did not investigate a control cohort, this limited the interpretation of the roles of EBV and HHV-6 in their GBM cohort.
Although to our knowledge no studies have specifically investigated HHV-6B in glioblastoma, some studies have suggested the involvement of HHV-6 in patients with gliomas. A 2012 study reported 17 of 40 (42.5 %) glioma samples with detectable HHV-6, compared to one of 13 (7.7 %) controls, using nested-PCR [8]. Furthermore, among the glioma cohort, 7/14 GBM (50 %) were HHV-6 positive. In a separate 2009 study, Crawford and colleagues reported a three-fold increase in HHV-6 positivity by IHC in glial tumors compared to non-glial tumors in an adult cohort, and 14/30 (46.7 %) of CNS tumors were HHV-6 positive by nested-PCR compared to 0/25 control brains [39]. While these data suggest the involvement of HHV-6 in at least a subset of brain tumors, the specific potential oncogenic mechanisms of HHV-6 are not fully understood.
We found EBV to be present at a significantly higher frequency in GBM compared to controls. EBV is an oncovirus that is responsible for the development of Hodgkin’s lymphoma, nasopharyngeal carcinoma and Burkitt’s lymphoma. However, few studies have examined the relationship between EBV and astrocytomas, specifically. In a study by Neves et al. in 2012, EBV was the most frequently detected human herpesviruses in pilocytic astrocytoma (WHO grade I) by PCR, but at fairly low levels (100/100 ng cellular DNA) [40]. Neves and colleagues detected only one sample with CMV, but did not detect any HHV-6. Further epidemiological and molecular studies are needed to investigate the role of EBV in astrocytomas.
Among the glioblastoma cohort in this present study, there were two cases with co-infection of EBV and HHV-6B. Currently, no study has reported EBV and HHV-6 co-infection in glioblastoma, though this has been observed in suspected viral CNS disease samples [41, 42]. Several in-vitro studies indicate that HHV-6 may activate EBV expression, which might promote EBV-associated diseases later in life [43, 44]. A 2015 study reported two encephalitis cases with HHV-6 and EBV co-infection detected by PCR [45]. However, the coinfection of these two herpesviruses in our case cohort cannot be seen as causal agents of the astrocytoma. The detection of HHV-6B and EBV could be a mere consequence of molecular events within the tumor that may be associated with reactivation of latent virus [46].
In our cohort, the sensitivity of multiple human herpesvirus detection did not vary with the biopreservation method. All preservation methods were comparable in terms of virus detection, though fresh-frozen remains the most ideal for viral detection. Some studies have raised concerns about the possibility of DNA degradation upon long-term storage in FFPE. The vast majority of studies used FFPE-preserved tissues for the detection of viral protein or DNA. However, in a study that examined the quality of FFPE extracted DNA, no significant differences were found between DNA extracted from FFPE tissue blocks that had been stored for less than one year and FFPE tissue blocks that had been stored for 1–11 years [47].
Lastly, the ddPCR multiplex assay is a highly precise viral diagnostic tool that enables the absolute quantification of multiple viral target regions [48, 49]. Although multiplex qPCR has a shorter analysis time, high sensitivity, and a broad dynamic range, the multiplex ddPCR assay has lower variability, outstanding accuracy, and provides better quality of quantification in the multiplex assay [49]. A study reported by Hayden in 2013 compared quantitative real-time PCR (qRT-PCR) and ddPCR for the detection of CMV in clinical samples, and reported that ddPCR showed less variability when detecting samples at higher concentrations [50]. Overall, the study indicated that ddPCR is an accurate assay for detecting CMV viral loads in clinical samples. In our study, we compared a multiplex ddPCR assay with a singleplex ddPCR assay, and there was no significant difference in quantification or qualification ability. However, any type of PCR system is prone to the misclassification of viral detection due to low sensitivity or specificity of the primer/probe set. As a result, optimization of the primers and probe becomes critical. Regardless of the PCR type, a multiplex assay can be adopted as a diagnostic tool that quantifies multiple viral targets, while conserving samples in a clinical setting.