MCPyV is thought to play a role in MCC tumorigenesis . Although a causal link between MCPyV and other types of malignancy has not been established to date, recent studies have presented evidence of MCPyV detection in several cancers. Our previous findings showed that MCPyV was present in 4/30 cutaneous SCCs (13%) among Japanese patients . A German group showed that 7/28 cutaneous SCCs (25%) were positive for MCPyV . In other studies from North America, 26/177 cutaneous SCCs (15%) and 2/15 SCCs (13%) were positive for MCPyV [7, 34]. Thus, the prevalence of MCPyV in cutaneous SCCs has been confirmed among distinct geographic populations. The present study demonstrated that the prevalence of MCPyV in cervical SCCs is close to that seen in cutaneous SCCs.
For detecting MCPyV, we used two primer sets targeting the LT and VP1 regions, which gave different detection rates. Given the decreased amplification efficiency of larger amplicons by PCR of FFPE tissues, the LTsh primers should have detected MCPyV in more cases than the VP1 primers. Otherwise, PCR amplification might be hampered by mutations or deletions that exist in the primer regions, as suggested by recent studies [9, 40].
The occurrence of false positive PCR results is unlikely. Our PCR runs were always performed using the appropriate controls and the negative controls were consistently negative in all experiments. To confirm that the PCR products contained MCPyV-specific DNA sequences but not artifacts, and to exclude the possibility of cross-contamination, we sequenced all the PCR products. Obvious variations in the DNA sequences were found in the MCPyV VP1 gene. The sequencing results revealed the existence of three variants of the VP1 in our cases. The amino acid substitutions were present at three distinct positions, among which the replacement of glutamic acid with glutamine was found previously between two North American isolates, MCC350 and w162 . Thus, amino acid substitutions are likely to occur frequently in MCPyV. The same was also reported among French MCPyV isolates . On the other hand, amino acid substitutions at other locations would contribute to the antigenic diversity of the Japanese MCPyV. Any potential role of these substitutions remains to be elucidated.
We conducted immunohistochemistry of the MCPyV DNA-positive cervical SCCs and ACs to study the localization of MCPyV. CM2B4, a mouse monoclonal antibody to the MCPyV LT antigen, is available commercially and has been used widely for immunohistochemistry. Recently, a Japanese group generated a rabbit polyclonal antibody targeting a broader LT antigenic region than CM2B4 . In addition to the CM2B4 monoclonal antibody, we employed this polyclonal antibody for immunohistochemistry in some cases. Both antibodies resulted in homogeneous or speckled nuclear staining of the tumor cells, indicating that MCPyV exists in cervical cancer cells. Nonspecific staining of the tissues is unlikely, because no signals were detected in the MCPyV PCR-negative samples and because our immunohistochemical method with the CM2B4 antibody was controlled by testing an isotype-matched control antibody. However, in some cases, weak immunoreactivity against these antibodies was also observed in a few surrounding normal cells. Therefore, we then performed PCR using DNAs extracted from normal tissues of the same patients with MCPyV PCR-positive cervical cancers, but neither the LTsh nor VP1 primers detected MCPyV DNA (data not shown). These findings suggest that the MCPyV genome was also present in nonneoplastic tissues of the uterine cervix at levels not detectable by PCR. The findings of semiquantitative immunohistochemistry did not correspond with the viral copy numbers detected by quantitative real-time PCR. A possible interpretation on the results would be that DNA in archived formalin-fixed tissues might be fragmented in the primer-targeting gene areas, and thereby the viral DNA copy numbers might be underestimated .
Currently, more than 100 HPV types have been identified and classified into high-risk and low-risk types according to the probability of developing a cervical cancer . The high-risk HPV types are regarded as major causes of cervical cancer. Compared with Southeast Asia, Northern Africa, Europe and North America, HPV types 16 and 18 are less common and HPV types 31, 33, 52 and 58 are more common in Japanese patients with cervical cancers . In the present study, the high-risk HPV types were detected predictably in all Japanese MCPyV-positive samples. These were HPV types 16, 31 and 58 in MCPyV-positive SCCs and HPV types 16 and 18 in MCPyV-positive ACs. The HPV typing pattern suggested no direct association of MCPyV with HPV types.
So far, research on the etiology of cervical cancers has focused on HPV. The present study provides the first evidence that MCPyV is present in a subset of HPV-associated cervical cancers. Because the MCPyV LT antigen is considered to be an oncoprotein responsible for the MCPyV-dependent oncogenic pathway [40–43], expression of the MCPyV LT antigen in HPV-associated cervical cancer cells suggests that MCPyV could be a cofactor of HPV for tumor initiation and/or progression. In the present study, an MCPyV-positive status had a tendency to be found at earlier clinical stages, although this needs further study. Recently, Houben et al.  presented new evidence that MCPyV might only be needed for tumor initiation, but additional mutations during tumor progression render LT antigen expression dispensable for MCC carcinogenesis. Although we do not know yet whether MCPyV is such a transient “hit and run” infectious pathogen or just a passenger during the development of cervical cancer, our findings should stimulate further investigations to clarify these important issues.