Viral oncogenesis in tumours of the central nervous system: reality or random association? A retrospective study on archived material

Abstract Central nervous system (CNS) tumours have devastating effects and are recurrent, with dismal prognosis (gliomas) or life‐threatening by the compression effect (meningiomas). This disease's aetiology remains debatable. Over the last decade, the hypothesis that human viruses may be implicated in these tumours has been proposed. In this study, our aim is to examine the presence of 11 viruses in the most frequent CNS primary tumours. Using polymerase chain reaction (PCR), we assessed the viral presence in archived, paraffin‐embedded tumour tissues from 114 patients with glioma and meningioma and in the brain tissue from 40 controls lacking tumour pathology. We focused on candidate neuro‐oncogenic types (herpesviridae and polyomaviruses) and on human papillomavirus (HPV). HPV presence, for which involvement in these tumours was hardly investigated, was found to be associated with both tumour categories compared with controls (glioma, p = 0.032; meningioma, p = 0.032), whereas the presence of the neuro‐oncogenic viruses was found in a negligible number of both categories, suggesting a lack of association with the tumour presence. Moreover, our study reveals a positive correlation between HPV presence and glioma malignancy, and a negative correlation with meningioma grading. Our results suggest that the presence of HPV seems to be significantly associated with primary tumours of the CNS and its meninges.

The aggressiveness of glioblastomas is both histological and clinical, determined by their localization within a rigid space and adjacent to vital areas; for example, patients with glioblastoma, the most aggressive glial primary tumour, have a 5-year survival rate of 5.1% and a median survival of around 10 months. 3 Recent approaches for glioblastoma therapy address some pathogenic features and are based on immunological and genetic-mediated techniques, adding to classical surgery, radio and chemotherapies. [4][5][6] For meningiomas, standard therapy comprises surgery with or without adjuvant radiation, depending on the tumour grade and the degree of resection. 7 For any other tumour type, knowing the precise aetiology and pathogenesis would be an essential step to treat, cure or even prevent their onset. Little is known about the aetiology of this brain proliferative condition, and currently, the only established risk factors are ionizing radiation and some rare genetic-related syndromes. 8 Moreover, there is growing evidence and consensus regarding the association of certain viruses with brain tumours. 9 Although some viruses have recently been proposed as therapeutic methods for aggressive glial tumours, [10][11][12][13] the viral role as an etiological agent is not sufficiently understood.
Overall, approximately 12% of human cancers are supposed to have a viral aetiology. 14 Human viral oncogenesis is probably characterized by three main features: (a) oncoviruses are necessary but not sufficient for developing cancer; (b) virus-dependent cancers occur many years after acute infection and after persistent chronic infections; and (c) the immune system can act both towards immunosuppression and chronic inflammation. 15,16 Viruses activate various intracellular metabolic pathways, with a progressive accumulation of cellular markers that constitute the malignant phenotype. This is achieved by the accumulation of somatic oncogenic alterations (oncogenic strokes) caused by spontaneous mutations or mutations induced by exposure to environmental carcinogens regarding the host's genetic background and the selective pressure imposed by the tissue microenvironment. 17 The viral oncogenic activity is most recently considered to be oncomodulatory. 18 Several types of viral infections have been suggested to be potentially associated with CNS tumours, especially glioblastomas. Herpesviridae is a large family of DNA viruses that comprises Herpes Simplex 1 and 2 (HHV1 and HHV2), Varicella zoster virus (HHV-3), Epstein-Barr virus (EBV or HHV-4), Cytomegalovirus (CMV or HHV-5), Roseolovirus (HHV-6 and HHV7) and Kaposi's sarcoma-associated herpesvirus (KSHV or HHV-8). These neurotropic viruses in the CNS control the activation of glial cells that release pro-inflammatory mediators and increase oxidative stress, which are associated with neurodegenerative disease pathogenesis. Indeed, the presence of many of these viruses has been related to Alzheimer's disease, 19 Parkinson's disease 20 and epilepsy. 21 Regarding cancer, although EBV and KSHV are officially recognized as carcinogens, other viruses from this family have also been associated with different types of cancer, but their role in carcinogenesis remains unclear. 22 Emerging evidence has demonstrated the presence of human CMV proteins and nucleic acids in brain tumours in both adults and children, 23 detecting the virus presence, particularly in patients with high-grade gliomas. However, data are contrasting and the need for research into the involvement of this virus in gliomas has recently been argued 24,25 since the CMV presence seems limited only to glioblastoma and not present in other types of glial tumours. 26,27 Moreover, the real presence of CMV in gliomas is questioned, 28,29 and single-cell sequencing does not clearly highlight the presence of this virus 30 Also, the results regarding the role of the Epstein-Barr virus (HHV-4) in glioma genesis are controversial. 31 Although its presence was detected at higher frequency than other herpesviruses (HHV-5, HHV-6 and HHV-8) in pilocytic astrocytoma, it was not considered responsible for tumorigenesis in those cases. 32 Also, HHV-6 is potentially involved in glial oncogenesis. The ability of HHV-6 to integrate into a chromosomal region that is highly relevant to carcinogenesis, the ability of the HHV-6 ORF-1 protein to bind to p53, and its detection of early and late antigens in adult primary and recurrent CNS tumours (more frequently in glial tumours) 33 make this virus a potential key factor in glioma cases. 34 However, further studies are considered mandatory. 35 HHV-8 or Kaposi's sarcoma-associated herpesvirus is well known for its oncogenic role in a wide category of haematological malignancies. However, since it also has a strong neurotropism, 36 it can modulate the proliferation of glioma stem cells. 37 Moreover, HHV-2, EBV, and HHV-6 have been detected in highgrade glioma tissues, and HHV-1 and HHV-2 immunoreactivity have been found in some cases, although these were not detected in their tumour tissue. 38 Human polyomaviruses represent a diverse group of human pathogens comprising John Cunningham virus (JCV), BK virus and simian virus 40 (SV40), which generally cause asymptomatic infection in healthy individuals. JC and BK viruses are mostly associated with cutaneous, mucosal tumours and bladder carcinoma. 39 These viruses have been involved in neural oncogenesis, mainly in glioblastoma. 38 In particular, glioblastoma with a small-cell neuronallike component was mainly related to JCV. 40 However, the role of this virus in the carcinogenesis of brain tumours remains partially understood. 41,42 Regarding SV40, most studies on its presence have been performed in the sera of glioma patients without assessing its presence within the brain tissue. 43 Moreover, differences in the detection methods (viral DNA, mRNA and expression of viral oncoproteins) are considered essential in affirming or denying SV40 involvement as an oncogenic agent. 44 Finally, a potential involvement of HPV in glioblastoma has been suggested 27 since this virus has been found to have prognostic implications in glioblastomas. 45

| ME THODS
In this retrospective study, we aimed to detect the presence of various potential oncogenic viruses in the most frequent tumours of the brain and their meninges (glioblastoma and meningioma) compared to patients lacking tumour pathology.

| Sample collection
In this study, we enrolled 154 individuals. Fifty-six patients were diagnosed with meningioma and 58 with glioma (44 glioblastoma, nine grade two diffuse astrocytoma and five anaplastic astrocytoma). Forty patients represented the control group, deceased by other causes and lacking any brain tumour. All study specimens were collected from the Neuropathology Department of the National Institute of Neurology and Neurovascular Diseases. This study was conducted following the guidelines of the Declaration of Helsinki and was approved by the Ethics Committee of the National Institute of Pathology 'Victor Babes' (Registration number 53 of 6 December 2017). All the patients or their relatives signed the written informed consent. Only patients with primary tumours of the brain and meninges were enrolled. We excluded metastases (with biology specific to the tissue of origin), glandular (pineal, pituitary) or accessory tissue (eye) tumours.
The first section was stained with haematoxylin-eosin, and a second histopathological diagnosis was established by an expert pathologist to confirm the initial classification of the cases. All tumours were classified and graded according to the WHO classification of CNS tumours. 46 Samples with small tissue presenting large necrotic areas or artefacts (ie intraoperative electrical mark) detected by microscopic examination were excluded.
The initial series comprised all primary tumour categories of the brain, including oligodendroglioma, ependymoma and choroid plexus tumours, with a total of 200 cases. However, since most of these other entities were represented by a limited number of cases, we restricted our research to the most common and aggressive brain tumour, grade IV glioblastoma.8963.
The sociodemographic and clinical data of the selected cohorts are reported in Table 1.

| DNA isolation, virus detection and HPV genotyping
A number of 5-10 slides (50 μm) were cut from each sample according to the volume. Before DNA isolation, to remove the paraffin, the The presence of HPV has been detected by PCR using consensus primers targeting the L1 region, GP5+/GP6+ primer set, 47 as described by Yoshida et al. 48 This generated a 140-bp-long fragment from the HPV L1 structural gene. The amplified DNA was s8998u9ubjected to electrophoresis on 2% agarose gel and then visualized by ultraviolet illumination using ethidium bromide. For all the samples, the DNA integrity was verified by PCR amplification of β-Globin gene as previously reported. 48 PCR reactions that were negative after the first round were subjected to a further 30 rounds of PCR amplification using the same primers (auto-nested PCR).

| Statistical analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS version 17.0). Continuous variable (Age) was tested with the t test. For categorical variables, the frequency, with percentages, was reported and assessed for the two groups by chi-square test if the number of patients in a subgroup was ≥5, and when the number was less than 5 in any cell, the Fisher exact test was applied.

| RE SULTS
A total of 154 individuals were involved in the study, including 79 females (51.3%) and 75 males (48.7%). The mean age of the study population was 58.89 years (SD ±14.01). The participants were grouped into three groups according to the diagnosis of meningioma (n = 56), glioma (n = 58, including 44 glioblastoma and 14 astrocytoma) and control (n = 40). The groups were homogenous for sex but not for age since the control group was older than the experimental group  Table S1).
The presence of HPV in the 32 positive cases was validated using an IVD kit. Specifically, after the first round of PCR with this kit and after the reverse hybridization, in all the analysed samples, we visualized the band of amplification for the endogenous control, demonstrating that the reaction succeeded and the integrity of DNA. After the auto-nested PCR, we obtained, as expected the band of the endogenous control, and the band for the conserved HPV sequence (universal HPV band), demonstrating the presence of HPV. However, none of the 40 genotypes tested were found in our groups.
An analysis examining the HPV frequency in age categories (age ≤60 and age >60) was conducted in the entire selected cohort and in the three groups. No statistical difference was found between the two age categories in any analysed group ( Table 2).
The distribution of HPV-positive and HPV-negative cases in more stringent age categories is reported in (Additional file 1: Table S2).
The presence of HPV in sex categories was also investigated, finding that, in the entire cohort (N = 154), the number of HPVpositive cases was higher in the female group (26.6%) than in the male group (14.7%) but without statistical significance (χ 2 = 3.318, p = 0.069). This trend of difference between males and females was also observed in the control, glioma and meningioma groups, again without statistical significance. (Additional file 1: Table S3).    Table 3.

| DISCUSS ION
In this study, we explored the presence of infectious agents belong- No statistical association was observed between the other selected viruses and the brain tumours analysed in our study.
Our results on HPV in glioma and glioblastoma parallel those reported by Vidone and collaborators 45 [50][51][52] Of note, some of these studies were performed without control cases, leading to a partial conclusion about the oncogenic role of these viruses in glioma and glioblastoma. 53 The lack of positivity for HHV-8 (even in the control tissue) disagrees with other studies, which affirm its positivity in 63.3% of healthy postmortem brain tissue. 36 In this cohort study, HHV-6 was more frequently detected in low-grade glioma (7.1%), almost as in control cases (10%), compared to glioblastoma (2.3%) or meningioma (1.78%). This agrees with other studies that found a higher presence of this HHV type in lowergrade gliomas. 54 Herpes viruses, mainly HHV-6, followed by HHV-1, HHV-3,   57 However, these data derived from seric determinations and the presence of these polyomaviruses in the brain tissue were not determined in parallel. Hence, the viruses could be latent only in circulating mononuclear blood cells or in the cells of the proximal renal tubule, reactivating in immunocompromised individuals.

| CON CLUS IONS
Overall, our findings, although based on a small number of cases, show a significantly increased frequency of HPV in glioblastoma and meningioma compared to normal brain tissue and suggest the non-involvement of herpes and polyomaviruses in the investigated brain tumours. However, the detection of a virus within the tumour tissue is not sufficient to affirm its role as a causative agent. Further studies on larger cohorts are needed to clarify the role of these infectious agents in brain tumours and to develop new preventive or therapeutic strategies. Since HPV has been found to have prognostic implications in glioblastomas, and the HPV vaccine in order to prevent cervical cancer cases has established success, a brain tumours prevention through vaccine may be considered.

ACK N OWLED G EM ENTS
Research and publication of the present study were funded by

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article and its additional information files.