SV40 Multiple Tissue Infection and Asbestos Exposure in a Hyperendemic Area for Malignant Mesothelioma

Malignant mesothelioma is one of the most intriguing human cancers from both an epidemiologic and a biological perspective. An epidemic of malignant mesothelioma is taking place in Western countries as well as in Japan and Australia. The mortality rate due to malignant mesothelioma has increased dramatically since the 1970s, with an expected peak rate over the next 20 to 30 years (1). A complex process is thought to sustain pleural carcinogenesis, whereby genomic damage mechanisms work over a long latency period (2). Exposure to asbestos, either occupational or social, represents the major recognized risk factor. Notwithstanding, a proportion of mesothelioma cases cannot be related to asbestos exposure/consumption, suggesting that alternative factors and/or associated cofactors, including genetic predisposition, are probably involved (3).

During the search for carcinogenic agents distinct from asbestos, simian polyomavirus SV40 has emerged as a notable candidate. SV40 is a strong carcinogen in experimental animal models, with pleural cancer arising in the majority of hamsters challenged intrapleurally with SV40. Presumably, the biological effects of SV40 on mesothelioma development are driven by the viral T-antigen (T-ag) protein, which increases the susceptibility of target cells to DNA damage, enhances genetic instability, and expands the transformed clones (4).

Surveys of SV40 presence in human malignant mesothelioma archival tissues have yielded differing results (5, 6). Several factors could explain the broad variability of reported SV40 detection. First, a high level of local asbestos exposure/consumption could mask the role of SV40, if any, as a carcinogenic agent. Alternatively, in such a situation, a cofactor role for the virus might be discerned (7). Second, the circulation of SV40 in distinct areas should be established. Exposure to contaminated early polio vaccines would not have led to universal human involvement by SV40. For example, countries such as Austria, Finland, and Turkey are believed to be unaffected by vaccine contamination, whereas, in other areas (e.g., eastern Europe), the contamination was a long-lasting event (8, 9). These differences would result in geographic variations in exposure to SV40. Third, the accuracy and specificity of molecular tools have been questioned, emphasizing that careful controls of molecular testing are necessary (6, 10).

In Italy, massive occupational exposure to asbestos occurred in restricted areas, and northeastern Italy (provinces of Trieste and Gorizia) may be considered as hyperendemic for mesothelioma. The standardized mortality rates due to mesothelioma among male residents were 11.6 × 10⁻⁵ in the province of Gorizia and 7.4 × 10⁻⁵ in the province of Trieste, compared with a national rate of 2.2 × 10⁻⁵ in the male population (ISTISAN Report, 2002). Preliminary evidence of SV40 infections in residents of this area has been obtained in immunocompromised hosts (11).

We did a pilot, retrospective study to assess the presence of SV40 in a selected cohort of patients affected with malignant mesothelioma. Archival formalin-fixed paraffin-embedded (FFPE) samples, including mesothelioma and other available tissues (i.e., lung, kidney, liver, brain, and breast), were tested for SV40 using a rigorously controlled molecular approach. We discovered evidence of SV40 infection in some malignant mesothelioma patients, supporting a plausible cofactor role for the virus in disease in this hyperendemic area for malignant mesothelioma.

Patients. Nineteen patients diagnosed with pleural mesothelioma were selected whose archival samples of cancer and other tissues were available. Twelve were males (mean age, 72; range, 47–85 years) and seven were women (mean age, 64; range, 40–80 years). Mesothelioma was diagnosed as the primary cancer in 14 patients, whereas, in 5 females, mesothelioma followed an initial breast cancer. All patients' records were searched in the National Mesothelioma Registry (ReNaM, Italy). According to ReNaM guidelines and/or other criteria, including compensation for recognized occupational mesothelioma, asbestos exposure was established for 17 of 19 patients.

Samples and histology. Tissue blocks from 19 FFPE samples of mesothelioma were examined. All mesothelioma samples contained a large majority of malignant cells and minimal nonmalignant cells or necrotic debris. The mesothelioma samples were composed of 6 epithelioid, 12 mixed, and 1 sarcomatoid histotypes. The histologic examination and classification of tumors were done according to WHO criteria.

A total of 48 other FFPE tissues from the malignant mesothelioma patients were available for virological examination. Both normal and cancer samples were included: 8 kidney and 3 brain tissue samples free of neoplastic cells; 15 lung samples, 6 of which were infiltrated with tumor cells; 18 liver samples, 6 of which were infiltrated; and 4 breast cancer samples. Thirty cervical samples from women with human papillomavirus–related cancer were included as controls.

DNA extraction. Sections (nos. 4–6), each 10 microns thick, from FFPE samples were treated with xylene at 50°C for 1 h, followed by washing with 100% ethanol for 10 min. DNA extraction was done using the High Pure Template Preparation kit (Roche Applied Science), following the manufacturer's protocol. DNA was eluted in 100 μL TE buffer and stored at −20°C until analyzed. Samples were processed in a dedicated plasmid-free room.

PCR and sequence analysis. Qualitative PCR using primers PYVfor-PYVrev, which recognize a conserved sequence in the T-ag NH₂-terminal coding region of polyomaviruses JCV, BKV, and SV40, was done as described previously (12). JCV-Mad-1, BKV-Dunlop, and SV40-VA45-54 recombinant plasmids served as positive controls.

SV40 quantification was done by real-time PCR for the detection of SV40 T-ag sequence using the ABI Prism 7700 Detection System following a recently described protocol (13). Positive control DNA was added outside the core facility after sealing the tubes containing the master mix and negative controls. Standard dilutions, fixed from 10 to 10⁷ copies of the T-ag target gene, were run in duplicate for each experiment.

A quantitative real-time PCR assay was done for the cellular RNase P gene. The detection of the cellular gene served both to verify the suitability of a DNA sample for analysis and as a reference gene for determining the human cell equivalents present in each DNA sample. The primer and probe components of the Taqman RNase P control reagents kit (Applied Biosystems) were used. To avoid PCR contaminations, pre-PCR and post-PCR were set up in a dedicated laboratory, with dedicated tools and supplies.

Samples positive by qualitative PCR with primer set PYVfor-PYVrev were submitted to sequencing reaction. The products were sequenced using a BigDye Terminator Cycle Sequencing kit, version 3.1 (Applied Biosystems) and an automated sequencer (ABI Prism 310).

Statistical methods. Continuous variables were summarized using arithmetic mean for age and geometric mean for viral load. Differences between continuous data were tested by the Mann-Whitney U test. Dichotomous variables were tabulated in 2 × 2 contingency tables, and the differences between proportions were tested by the likelihood ratio χ² statistic or Fisher's exact test, when appropriate.

The majority of malignant mesothelioma patients examined in this study (17 of 19) had a documented history of exposure to asbestos (Table 1). Three of 19 (15.8%) of the mesothelioma specimens were SV40 positive. Virus was detected only in malignant mesothelioma of asbestos-exposed patients [3 of 17 (17.6%)]. Four of 32 (12.5%) normal tissue samples were found to contain SV40 sequences, whereas none of 12 tissue samples containing infiltration of cancer cells were virus positive. Breast cancer specimens from 4 malignant mesothelioma patients, as well as cervical scrapings from 30 cervical cancer patients, also were negative for SV40 DNA. Considering any site of viral detection, 5 of 19 (26.3%) patients were SV40 positive.

The distribution of SV40 DNA in normal and cancerous tissues is detailed (Table 2). It is noteworthy that for two of three patients with SV40-positive malignant mesothelioma, other sites were also found to harbor SV40 DNA (i.e., liver and kidney normal tissues). In addition, two other patients, both with virus-negative malignant mesothelioma samples, showed the presence of SV40 in liver tissues. Among the normal tissues examined, SV40 footprints were detected in 3 of 18 (16.7%) liver samples and in 1 of 8 (12.5%) kidney samples. Tissues containing cancer cells as a fraction of total specimen tested negative for SV40 genomic signals.

Direct PCR sequencing was possible for five of seven positive samples. The sequences matched that of SV40 for nucleotides (nt) mapping between 4572 and 4425 of the T-ag NH₂-terminal domain; the amplicons did not contain the 9 nt insert present in BKV and JCV DNA sequences in this region.

SV40 viral loads in positive tissue samples were quantified by real-time PCR (Table 3). The single-copy cellular RNase P gene was used to normalize viral genome copies to the number of cell equivalents in each sample tested. This allows a calculation of the proportion of infected cells in a given tissue. Virus-positive mesotheliomas had a higher mean proportion of SV40-infected cells than did positive normal tissues (12.6% versus 5%; P = 0.045), based on an assumption of one viral genome copy per cell.

This exploratory study was designed to investigate the presence of SV40 in a selected series of malignant mesothelioma patients, the majority of whom had a well-established history of occupational exposure to asbestos. SV40 was detected in 3 of 19 (15.8%) mesothelioma samples, all 3 positive samples being from patients classified as asbestos exposed. However, only two nonexposed patients were included in this series. Thus, in the context of the debate about the presence and role of SV40 in malignant mesothelioma, this survey is consistent with those discovering the presence of the virus in tumor tissue, including specimens from the United States, Japan, Sweden, and Italy (14–16). In Italy, a recent case-control study that used a conventional PCR assay detected SV40 in 42.1% of malignant mesothelioma samples, as well as in 33.3% of bladder urotheliomas (7). In our study, a real-time quantitative PCR assay for SV40 and for the reference cellular gene RNase P was used (13), providing viral loads and fresh insights into the virus-host relationship. This study was unique because, in addition to mesothelioma tissues, other tissue samples from the malignant mesothelioma patients, including liver, kidney, lung, breast, and brain, were analyzed. A significant proportion (12.5%) of these normal tissues was found to be SV40 positive. To the best of our knowledge, this is the first time that simultaneous SV40 infections of different organs in the same malignant mesothelioma patient have been observed.

The meaning of our findings is uncertain as the survey was not designed to explore etiopathogenetic issues about SV40 and malignant mesothelioma. However, several considerations may be advanced. Very few reports of SV40 tissue loads are available, and the methodologies used are not always comparable. Based on the technology used in this study, mesothelial tissue seemed to have a higher viral content than normal tissues, consistent with in vitro experimental data (16). A caveat to the calculation of percentage infected cells in different tissues, which assumes one genome copy per cell, is that productively infected cells would likely contain multiple copies of viral DNA, effectively reducing the calculated number of infected cells.

An interesting finding of this survey was the lack of SV40 footprints in tissues infiltrated with mesothelioma cells, as well as in four breast cancers that preceded the onset of malignant mesothelioma. This issue deserves further evaluation. If confirmed, the finding may be consistent with a role for the virus as an asbestos cofactor in the transformation process (7), where retention of viral markers is not essential in advanced cancers. Recent experimental data support the cocarcinogen theory, showing that the presence of SV40 permitted a lower asbestos threshold limit and shortened the latency period of tumors in hamsters (17).

Finally, the detection of SV40 in normal tissues of malignant mesothelioma patients deserves comment. Whereas virus was found in pleural transformed cells, the normal lung tissues were consistently free from infection, indicating stable virus localization in serosal cells without the involvement of contiguous tissues, as shown previously in tissue microdissection studies (18). The involvement of kidney may reflect the nephrotropic properties of SV40, as renal infections in humans and in monkeys are commonly viewed as a natural reservoir for polyomaviruses (4). The detection of SV40 in normal liver tissue from malignant mesothelioma patients has not been described previously. As a viral genomic signal has been detected in peripheral blood mononuclear cells of immune-compromised patients and malignant mesothelioma patients (11, 19), SV40 infections may undergo reactivation and become detectable in liver, possibly due to viral clearance from the blood by Kupffer cells (20).

In conclusion, findings reported here indicate that malignant mesothelioma patients from this area in Italy warrant further investigation of SV40 involvement. A prospective evaluation of fresh samples collected from incident cases, well matched with suitable controls, should reveal new insights into the role of the simian polyomavirus in a highly endemic area for malignant mesothelioma.

Grant support: Istituto di Ricovero e Cura a Carattere Scientifico Burlo Garofolo, Friuli Venezia Giulia Region and NIH grant CA104818.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.