Patient-individual cancer cell lines and tissue analysis delivers no evidence for a role of DNA virus infection on colorectal cancer oncogenesis

Background Several DNA viruses are highly suspicious to have oncogenic effects in humans. This study investigates the presence of potentially oncogenic viruses such as SV40, JCV, BKV and EBV in patient-derived colorectal carcinoma (CRC) cells typifying all molecular subtypes of CRC. Methods Sample material (gDNA and cDNA) of a total of 49 patient-individual CRC cell lines and corresponding primary material from 11 patients, including normal, tumor-derived and metastasis-derived tissue were analyzed for sequences of SV40, JVC, BKV and EBV using endpoint PCR. In addition, the susceptibility of CRC cells to JCV and BKV was examined using a long-term cultivation approach of patient-individual cells in the presence of viruses. Results No virus-specic sequences could be detected in all specimens. Likewise, no morphological changes were observed and no evidence for viral infection or integration could be provided after long term CRC cell cultivation in presence of viral particles. Conclusions In summary, the presented data suggest that there is no direct correlation between tumorigenesis and viral load and consequently no evidence for a functional role of the DNA viruses included into this analysis in CRC development.


Background
The ndings gathered in recent years on the relationship between tumorigenesis and several microorganisms, especially viruses, allow a new approach in the prevention and therapy of tumors [1]. For cervical carcinoma associated with human papilloma viruses (especially types 16 and 18, approximately 70% of all cases), there are already worldwide preventive measures and vaccines against the HPV highrisk types in place [2]. If there was a causal relationship between the development of colorectal cancer (CRC) and a viral infection, it would be possible to create new therapeutic and preventive approaches. In 2012 alone, nearly 1.4 million people worldwide developed CRC and 694,000 died as a result. This makes CRC the third leading cause of cancer and the fourth most common cause of cancer death worldwide [3].
In Germany, CRC is the third most common form of cancer among men and the second most among women and lifetime risk to develop a CRC is 5 to 7% [4].
The rst oncogenic viruses were discovered in chicken in 1911. Peyton Rous of the Rockefeller Institute was able to isolate a retrovirus (subsequently named after him Rous sarcoma virus) from tumors of diseased chicken and he could demonstrate their malignant potential by causing the same sarcomas after injecting viral particles into healthy chicken. Rous sarcoma virus belongs to the same family of Retroviridae as avian leukosis virus, which is able to cause leukemia in chicken. Later, it could be shown that murine polyomavirus and simian virus 40 (SV40) could cause tumors in experimental animals. SV40 also has a carcinogenic effect on human cells [1]. It is well known that at least 6 different viruses could cause cancer in humans [1]. For the year 2002 it was estimated, that approximately 10-15% of all cancers worldwide were caused by viral infections [5].
Polyomaviruses are icosahedral, nonenveloped viruses with a size of 45 nm. The capsid is composed of 72 pentameric capsomers and surrounds a circular, double-stranded viral DNA genome of approximately 5.5kbp. The virus particles are taken up by endocytosis into the target cell and transported there to the nucleus [6]. It has been repeatedly con rmed that SV40 is oncogenic both in vitro and in vivo [7]. Also for the human polyomaviruses JC virus (JCV) and BK virus (BKV), direct carcinogenic effects are still under debate [8]. Regarding JCV, several case control studies have been conducted to establish the association between JCV seropositivity and speci c human tumor types, including CRC [9,10,11,12,13,14,15]. BK virus DNA has been detected in several human cancers, including cancer of the head and neck, lung, urogenital tract, liver and pancreas [16,17].
Epstein-Barr-Virus (EBV) is one of the human pathogenic viruses of the family Herpesviridae. The virus particles are enveloped and have a genome of double-stranded DNA. Seroprevalence in adulthood is over 98%. Though, a direct connection between EBV infection and CRC has not yet been established [18].
The aim of our study was to investigate the presence of potentially oncogenic viruses such as SV40, JCV, BKV and EBV in patient-speci c CRC cells, selected to represent all molecular subtypes of CRC.
Furthermore, the susceptibility of low-passage, patient-derived CRC cells to the human polyomaviruses JCV and BKV was investigated.

Methods
Tissue sampling, cell line establishment and cell culture Primary CRC resection specimens, specimens of metastases and normal tissue were received fresh from surgery, with informed written patient consent. Tissues were immediately vitally frozen in freezing medium (fetal calf serum (FCS) containing 10% DMSO) at -80°C for subsequent examinations. All cell lines were established from patient derived tumor material as previously described by [19] (used cell lines are listed in Table 1). Cell lines were cultured in T75 culture asks using Dulbecco's Modi ed Eagle Medium (DMEM)/Ham´s F12 medium supplement with 10% FCS and 2mM L-glutamine as described before [20]. All procedures involving patient material were approved by the Ethics Committee of the gDNA and RNA isolation gDNA was isolated from cell pellets using the "Wizard Genomic DNA Puri cation Kit" from Promega according to the manufacturer's instructions. The measurement of the isolated gDNA was carried out on the NanoDrop 1000 from Thermo Scienti c. The RNA was isolated from cell pellets using the "GeneMATRIX Universal RNA Puri cation Kit" from EURx according to the manufacturer's instructions.
The measurement of the isolated RNA was then carried out on NanoDrop 1000. cDNA Synthesis cDNA was synthesized from RNA using the "Reverase Kit (M-MuLV, RNase H minus)" from Bioron according to the manufacturer's instructions. The nal cDNA concentration was 12.5 ng/μl after synthesis.
Endpoint PCR RT-PCR was performed for all primary materials and established cell lines to be analyzed. There were gDNA, cDNA and RNA from the different samples.

RT-PCR for SV40 and JC/BK
Primers used for SV40 bind to the speci c SV40 nucleotide sequence for the large T antigen (PCR product size: 172bp). Used primers for JC/BK virus bind to the common speci c JC/BK nucleotide sequence for the large T antigen (PCR product size JC/BK: 178bp and 181bp). As a positive control for SV40 gDNA, served the SV40 immortalized broblast cell line LCT 7. For JC/BK a positive control loop test material with virus particles was used, which was provided by the Institute of Medical Microbiology, Virology and Hygiene.

RT-PCR for EBV
Primers used for EBV bind to speci c EBV nucleotide sequence (PCR product size: 153bp). As a positive control for EBV gDNA served the EBV immortalized B cell line HROC212.
RT-PCR for β-actin ß-actin was used as an unregulated household gene to verify that samples are gDNA free after RNAisolation (PCR product size: cDNA: 210bp, gDNA: 700bp). gDNA of cell line HROC111Met1 T0 M2 served as positive control for ß-actin. The annealing temperature was 58° C. All samples were applied to a 2% gel after ampli cation.

Agarose gel electrophoresis
To illustrate fragment sizes, a 100 bp DNA ladder (Carl Roth) was used. For electrophoresis, a voltage of 80V (15min) and 100V (45min, Elektrophoresis Power Supply, Biometra) was set. Detection took place on the BIO-Print ST4 from VILBER.

Viral infection of cell lines
To test the susceptibility of CRC cells to the JC and BK virus, HROC257 cells were infected with these viruses. The ratio of virus particles and target cells was calculated on the basis of the MOI (= multiplicity of infection) according to the following formula:

Results
Primary aim of this study was the detection of the potentially oncogenic viruses SV40, JCV, BKV and EBV in patient-individual low passage (<P50) CRC cell lines. Therefore, gDNA and cDNA from established and characterized cell lines and corresponding patient tumor (tu) and normal (n) tissues, as well as metastases (met) were examined by RT-PCR.

Analysis of SV40 in cell lines and primary tissue samples
For the detection of SV40, altogether 49 individual patient-derived colorectal tumor cell lines were examined. No speci c band pattern could be shown for the SV40 virus in all cell lines (Figure 1).
The positive control shows a clear speci c band at 172bp. In order to clarify whether SV40 sequences are detectable at RNA level, additionally cDNA from the same cell lines was examined. Again, no viral sequences could be ampli ed with the positive control showing a band at 172bp (not shown). However, examination of primary tumor material revealed signals in HROC277tu and HROC296tu ( Figure 2).
PCR product sizes were about 170bp and 300bp, respectively. To clarify the speci city of the detected bands, a reampli cation of the PCR products was carried out with the same primers. However, the initially detected band of HROC296 (300bp) could not be reampli ed, while the reampli cation product of HROC277 exhibited an unspeci c smear and must therefore also be regarded as negative (Figure 3). To verify these ndings, a highly sensitive qPCR (TaqMan®) from HROC348 Met and HROC212 cell lines was performed. However, no JC or BK viral load could be detected in any of the samples and the initial results could thus not be con rmed. In analogy to the SV40 analysis, cDNA was also assayed for JCV and BKV. All cDNA samples were negative (data not shown).

Analysis of JC and BK virus in cell lines and primary tissue samples
The examination of primary tumor material revealed a signal at approximately 380bp for the samples HROC107 n/tu, HROC126 n, HROC183 n/tu and HROC212 n ( Figure 5). As product size was greater than the expected 178bp or 181bp for JC/BK virus, a reampli cation PCR was performed with the PCR product of the primary materials. Here, the result of the rst RT-PCR could be con rmed. All samples of the reampli cation showed a product size of approximately 380bp. To check the speci city of this ampli ed nucleotide sequence of the primary patient materials, the reampli ed amplicons were sequenced by Sanger sequencing. This analysis identi ed a human DNA sequence from clone RP11-203L2 (chromosomes 9q21.11-21.2) with a 99% identi cation probability. Thus, in none of the investigated primary materials, JCV or BKV were detectable.

Analysis of EBV in cell lines and primary tissue samples
RT-PCR revealed several ampli cation products at about 700bp in size in the examined cell lines and primary tissue samples with no difference between tumor and normal tissue (data not shown). The product size is about 700bp and is thus by far larger than the expected product size of 150bp. A reampli cation failed to reproduce the initial 700bp bands in all cases. Similar to the observation for SV40, either an unspeci c smear or no signal at all was obtained. In conclusion, it can be assumed that the 700bp amplicons are non-speci c.

Susceptibility of HROC257 to JC and BK virus
To test the susceptibility of CRC cells to infection with JC and BK virus, the cell line HROC257 was infected with JC and BK viruses with an MOI of 0.39 as explained before. Growth and morphology of the cells were documented photographically ( Figure 6). Particular attention was paid to the following parameters: changes in growth behavior / inhibition of proliferation, cell lysis or potential transformatory effects, re ected in the morphology of the cells. During cultivation of the HROC257 cells in the presence of the JCV and BKV, no morphological changes could be observed in comparison to untreated control cells. As shown in Figure 6, the growth behavior of infected and uninfected cells is similar. In order to further analyze susceptibility to JCV and BKV, these HROC257 cells were subjected to the above described RT-PCR, 40 days after viral infection. Both, cell culture supernatant and isolated cellular gDNA were used as templates. In positive controls, the analysis of cell culture supernatant showed bands at 178bp (JCV) and 181bp (BKV). But no products were detectable in supernatant of neither JCV nor BKV treated cells (data not shown). This indicates that no viral particles were present in the supernatant. In addition, isolated cellular gDNA of virus-treated HROC257 cells were tested for JCV and BKV after 40 days of incubation. But again no viral amplicon could be detected (data not shown). These results indicate that HROC257 were not susceptible to JC or BK virus infection.

Discussion
Recent ndings shed light into the relationship between tumor development and infections by several viruses. This culminates in the development of preventive vaccinations against cervical carcinoma associated with HPV high-risk types [2]. A causal relationship between the development of CRC and viral infections has repeatedly been suggested and this could, if proved, open new avenues for similar therapeutic and preventive approaches.
Our study aimed at deciphering a potential involvement of several oncogenic DNA viruses in CRC development. The starting hypothesis was that if viruses would have an active role in the development of CRC, their footsteps in form of virus-encoded nucleotide sequences should be present in patients' CRC cells. In order to avoid any bias by non-tumor cells present in the tumor microenvironment, we focused on patient-individual, low passage CRC cell lines; but corresponding patient tissue samples were analyzed in addition. Using highly sensitive PCR approaches, we investigated the natural occurrence of the polyomaviruses SV40, JCV and BKV as well as the herpesvirus EBV -the most frequently discussed viruses with a potential role in CRC development [9,15,21,13,10,22,23,24,25]. First, both patientindividual CRC cell lines (n = 49) and corresponding tissue samples from patients (normal and tumor tissue, as well as liver, lung and peritoneal metastases) were analyzed for the presence of viral sequences by end-point PCR. But in none of the tested samples, viral sequence ampli cation was successful.
Regarding a potential SV40 infection, none of the 49 low-passage patient-derived cell lines was positively tested for SV40. This nding excludes SV40 infection at least as a frequent event in CRC cells. It is well known that SV40 might act as a cofactor in the carcinogenesis of several carcinomas; but its involvement in CRC oncogenesis has been examined poorly up to now [26]. Only one case control study by Campello et al. analyzed the involvement of SV40 in CRC pathogenesis [9]. However, only six out of 94 analyzed patients' tumors were tested positive for SV40 and the authors could not delineate a relationship between viral infection and CRC. An explanation for the differences to the results of the present study could be the extracellular presence of viral particles -either between the cells, more likely in in ltrating lymphocytes or macrophages, and possibly even in the lumen of the gut.
An infection of the analyzed cell lines with the polyomaviruses JCV and BKV could be ruled out likewise. Hitherto existing data on prevalence of JCV infections in CRC are very heterogeneous. Several small case control studies were reported. A study from Portugal could show a prevalence of 90% JCV infections in samples of CRC lesions [15] and in another study from Greek about 60% of the samples tested positive for JCV [21]. A Chinese study of 137 samples described a rate of infection in CRC of 40.9% [13]. But in contrast, a large American study failed to prove JCV infections in 233 patient samples [27] and a study from Iran described a positive detection rate of only 1.46% (2 of 70 patients) [28]. Our own results are pretty much in line with the latter ndings and also indicate that JCV is unlikely involved in CRC development.
Worldwide ninety percent of adults have been exposed to BKV. BKV DNA has been detected in several human cancers, including carcinoma of the lung, urogenital tract, liver, head and neck, in rhabdomyosarcoma and in brain tumors [17,29]. However, reports on detection of BKV DNA in CRC tissues are very scarce. A study by Giuliani and colleagues could detect BKV DNA in 9% out of 66 samples [22] and Casini et al. found BKV DNA in several CRC tissues [10]. Again, in our study population in none of the examined cell lines or tissue samples BKV DNA could be detected.
Finally, all cell lines and primary materials were examined for the presence of EBV. Besides non-speci c PCR products, presence of EBV DNA could not be veri ed. All analyzed samples must therefore be considered as EBV negative. Militello et al. described a very low positive detection frequency for EBV DNA in cancerous and cancer-adjacent mucous samples [30] and a study by Cho et al. could not detect EBV in 274 CRC specimen [31]. In contrast, several studies reported a high EBV DNA positivity of CRC tissues [23,24,25]. Interestingly, a study by Fiorina et al. demonstrated EBV in 23 of 44 CRC samples but they could prove that this was restricted to latency in the lymphoid in ltrate of the tumors [32]. This latter observation delivers the best explanation for the inconsistencies between the before mentioned results. EBV most effectively infects B cells and rarely other lymphoid cell populations. In cultured epithelialderived cells with no contaminating lymphocytes left over from the initial culturing process, a crosscontamination of the PCR results can safely be excluded. Of note, such a large panel of patient-individual cell lines with corresponding tissue samples has never been tested before. The primary CRC tissues included into our analysis were obtained from the invasive margins which are especially rich in (vital) tumor cells -but in ltrating T cells and also B cells are also readily detectable [33]. Thus, we cannot exclude sporadic positive results in EBV testing due to EBV-positive lymphocytes present in the tumor microenvironment, if a larger sample set of our primary CRC tissue collection would be analyzed.
Based on the data obtained in this work for SV40, JCV, BKV and EBV, it can be concluded that none of the tested viruses are likely to have an obvious general role in CRC development. Possible reasons for the variation in the detection rate of these viruses in our and several of the aforementioned studies, are beside the lymphocytic contamination of the specimens [34], differences in the detection limit of the applied detection systems but possibly also the genetic background of the different patient populations.
In order to verify the general negative ndings on a more functional level, we tested exemplary the susceptibility of the CRC cell line HROC257 T0 M1 to JCV and BKV infection. However, even after an extended cultivation period subsequent to viral inoculation, no evidence for a successful viral infection could be provided. This led us to conclude that CRC cells are not susceptible to JC or BK virus infection and thus further supports the negative screening results.

Conclusions
In conclusion, the presented data suggest that there is no direct correlation between tumorigenesis and viral load of SV40, JVC, BKV and EBV and consequently no evidence for a functional role of the DNA viruses included into this analysis in CRC development.

Consent for publication
Not applicable Availability of data and material The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests   Reampli cation PCR for detection of SV40. Material: PCR product of the primary material HROC277tu and HROC296tu. Product size of the speci c amplicon is 172bp.  Detection of JC and BK by endpoint PCR. Shown are the results of the analysis of primary material.