Investigation of adenovirus occurrence in pediatric tumor entities

Abstract

Adenoviruses (AdVs) contain genes coding for proteins with transforming potential, and certain AdV serotypes have been shown to induce tumors in rodents. However, data on the possible oncogenicity of AdVs in humans are scarce. We have therefore employed a real-time quantitative PCR (RQ-PCR) assay permitting highly sensitive detection of all 51 currently known human AdV serotypes to screen more than 500 tumor specimens derived from 17 different childhood cancer entities including leukemias, lymphomas, and solid tumors. Most tumor entities analyzed showed no evidence for the presence of AdV sequences, but AdV DNA was detected by RQ-PCR in different brain tumors including 25/30 glioblastomas, 22/30 oligodendrogliomas, and 20/30 ependymomas. Nonmalignant counterparts of AdV-positive brain tumors, including specimens of ependymal cells, plexus choroideus, and periventricular white matter, were screened for control purposes and revealed the presence of AdV DNA in most specimens tested. Identification of the AdV types present in positive malignant and nonmalignant brain tissue specimens revealed predominantly representatives of species B and D and, less commonly, C. To exclude contamination as a possible cause of false-positive results, specimens with AdV sequences detectable by PCR were subsequently analyzed by in situ hybridization, which confirmed the PCR findings in all instances. The central nervous system appears to represent a common site of AdV infection with virus persistence, thus providing the first evidence for the possible contribution of AdVs to the multistep process of tumor pathogenesis in brain tissue.

FIG. 1.

Real-time PCR analysis of AdV-positive brain tumors. (a) In glioblastoma and oligodendroglioma, AdV species B was most commonly detected, whereas in ependymoma species D was predominant. The predominant occurrence of AdV species D in ependymoma was statistically significant (P < 0.001), while the preponderance of AdV species B in the other brain tumor entities did not achieve statistical significance. n. a., not available for AdV species identification due to inadequate DNA quality; CNS, central nervous system. (b) Amplification curves of the hexon gene in AdV-positive brain tumors (glioblastomas) generally showed threshold cycles in a range between 35 and 38, as depicted in the displayed examples of six brain tumor specimens positive for AdV species B. The single-copy control gene, the beta-2-microglobulin gene, revealed threshold cycles at or around 27, indicating a difference in copy numbers from AdV target genes of up to 3 logs. Other viral genes, such as E1A, E1B, and E4orf1, were detected in all AdV-positive brain tumor samples (Table 2), at the same level as the hexon gene.

FIG. 2.

Detection of AdV DNA in brain tumor tissue by ISH. (a to c) AdV-positive cells in ependymoma (a), glioblastoma (b), and oligodendroglioma (c) are visualized by a color reaction (panels a, b, and c, 1,000-fold magnification). Cells were counterstained with nuclear fast red. (d) Loosely scattered AdV-positive cells in an ependymoma specimen. (e) Positive control: biotinylated oligoprobes against ALU repeats. (f) Negative control: biotinylated pSP vector DNA (panels d, e, and f, 500-fold magnification).