A molecular case–control study of the Merkel cell polyomavirus in colon cancer

INTRODUCTION

The Merkel cell carcinoma is a rare skin cancer that originates from neuroendocrine cells. This pathological condition has been related recently to infection with a novel human polyomavirus, the Merkel cell polyomavirus [Feng et al., 2008; Zur Hausen, 2008; Becker et al., 2009]. Like SV40 [Vilchez and Butel, 2004], Merkel cell polyomavirus appears to play a causal role in tumorigenesis. Tumor induction and expansion is preceded by a monoclonal integration of the virus into target cells. Nonsense mutations or deletions in the early genomic region lead to the expression of truncated large T antigen, loss of viral DNA replication, and cell transformation [Shuda et al., 2008].

The assumed oncogenic power of the Merkel cell polyomavirus has prompted research to explore its role in several human cancers, namely in non-Merkel skin tumors, in neuroblastoma and other malignancies of the central nervous system, in lung cancer, and mesothelioma. A series of high-grade neuroendocrine tumors from different body sites was also investigated, with negative results [Duncavage et al., 2009; Gandhi et al., 2009; Sastre-Garau et al., 2009].

Other polyomaviruses, namely JCV and SV40, have been linked previously to colon cancer [Markovics et al., 2005; Campello et al., 2010; Gjoerup and Chang, 2010]. The hypothesis that polyomaviruses and gastrointestinal malignancies are associated relies on several reports in which sequences of JCV were detected in cancerous tissue, with different rates of prevalence at distinct times and in different geographic areas. In addition, a number of pathogenetic steps in the tumorigenesis associated with JCV have been described [White and Khalili, 2004]. A further rationale for the investigation of Merkel cell polyomavirus in colon cancer is that its epidemiological pattern is similar to that of JCV. It has been demonstrated recently that DNA sequences of both Merkel cell polyomavirus and JCV can be detected in sewage and environmental samples at high concentrations, thus supporting the fecal-oral route of their transmission [Bofill-Mas et al., 2010; Loyo et al., 2010].

Investigations of the possible association between Merkel cell polyomavirus and colon cancer are rare. Studies have been carried out solely on formalin-fixed paraffin-embedded archival tissues, and never compared with samples from control subjects enrolled concurrently [Kassem et al., 2009; Loyo et al., 2010]. In addition, the presence of Merkel cell polyomavirus in pre-cancerous lesions of the colon has never been explored.

To address these issues, a prospective molecular case–control study was undertaken to detect the Merkel cell polyomavirus in patients with colon cancer and in their relatives, who were used as controls.

MATERIALS AND METHODS

RESULTS

Overall, 11 sequences of Merkel cell polyomavirus, encompassing both regions tested, were detected in samples of colon tissue. No blood sample tested positive in either the cases or controls. The distribution of positive results by group and according to pathological features is reported in Table I. The rate of prevalence was 6.3% in the cases and 8.8% in the controls. This difference was not statistically significant in the univariate analysis. In addition, the rates of prevalence for pre-neoplastic lesions, namely adenomas and polyps, were 18.8% and 11.2%, respectively. When adjusted for age, no significant association was found between the rate of prevalence in cases and controls (prevalence of Merkel cell polyomavirus in colon cancer versus prevalence in controls: OR = 4.47; 95%CI = 0.61–28.66; P = 0.147). Furthermore, no significant difference was recorded when cancer was compared with pre-neoplastic disorders (OR = 0.53; 95%CI = 0.11–2.57; P = 0.426), and when pre-neoplastic tissues were compared with normal tissues (OR = 1.71; 95%CI = 0.35–8.34; P = 0.503). Interestingly, the sole neuroendocrine tumor discovered did not harbor Merkel cell polyomavirus.

The amplification products of regions LT3 and VP1 were analyzed by sequencing to confirm their identification and to assess sequence variability. The sequences of the conserved LT3 region were identical to almost all those published, and in particular to sequences of strains isolated in France (not shown). Two representative sequence patterns of the variable region VP1 are reported in Table II, along with the sequences of prototype strains and of other isolates retrieved from GenBank. The strain MCV-ITA61 was identical to the isolate MKT-31 and to other sequences discovered in France and Australia (GenBank accession numbers FM864211, AM992902, FJ009188/9). The strain MCV-ITA71 was identical to MKT23 and to some American and European sequences (GenBank accession numbers FJ392560, FJ649206, AM992904/5). The level of identity of isolates in this survey was higher than 99% with the reference strain MCC350. Overall, the Merkel cell polyomavirus sequence patterns obtained in this study reproduced those of the genotypes that are circulating worldwide, and especially in France.

DISCUSSION

The molecular case–control study reported herein was the first to focus on the role of Merkel cell polyomavirus in incident cases of colon cancer enrolled consecutively. Tissue samples were deep-frozen to ensure the preservation of viral components, because the detection of polyomavirus DNA may be impaired when archival fixed samples are tested [Touzé et al., 2009]. Similarly, the consecutive enrolment of controls was conducted from a cohort of subjects who underwent colonoscopy for cancer screening because they were the first-degree relatives of each patient. The enrolment of controls from relatives of the cases allowed correction for two exposure factors that could have been operating. First, past, or present exposures to environmental determinants could have been, at least in part, common to the cases and controls. Second, genetic determinants of colon cancer could have been distributed homogeneously in the cases and their relatives. Hence, the control of both factors may have contributed to the disclosure of the putative role of the oncogenic polyomavirus.

In this study, Merkel cell polyomavirus DNA was detected in patients with colon cancer, but the rate of prevalence did not differ significantly from that in the well-structured control group. In addition, no Merkel cell polyomavirus sequence was detected in blood samples, thus excluding possible systemic involvement sustained by infection of peripheral blood mononuclear cells (PBMCs) [Shuda et al., 2009].

The rates of prevalence that are reported herein were compared with those found in other surveys, in particular with those obtained using standard PCR to assess or confirm the presence of polyomavirus DNA by amplifying the same regions [Feng et al., 2008; Bialasiewicz et al., 2009; Goh et al., 2009; Wieland et al., 2009]. The findings of this survey were very similar to those of Feng et al. [2008], who reported a rate of prevalence of 8% in selected tissues from the gastro-intestinal tract, with the exclusion of tumors [Feng et al., 2008]. Conversely, in an archival series of cases of colon adenocarcinoma, no Merkel cell polyomavirus was detected [Sastre-Garau et al., 2009]. In a small retrospective case–control study, the rates of prevalence that were found by means of real-time PCR were 16% in specimens of colon cancer and 20% in control tissues. The study also demonstrated that the viral load was high in the polyomavirus-related cancer (i.e., Merkel cell carcinoma), whereas it was very low in tissues from other sites, for example, the urogenital tract. Tissues from the digestive tract had an intermediate viral load [Loyo et al., 2010]. Thus, the sensitivity of a standard PCR appears to be satisfactory when specimens of colon cancer, pre-cancerous lesions and normal mucosa are compared, and in particular when fresh tissues are investigated. Collectively, these findings do not support the hypothesis that there is an etiopathogenetic relationship between Merkel cell polyomavirus and colon cancer. The results point towards a simple involvement of the gastrointestinal tract through the fecal-oral route of transmission of the virus [Feng et al., 2008; Gandhi et al., 2009; Loyo et al., 2010].

This study involved a novel evaluation of pre-cancerous tissues, namely adenomas and polyps, in which the rates of prevalence were similar to those of the other intestinal tissues examined. The meaning of these findings remains speculative. The recovery of Merkel cell polyomavirus sequences from tumoral, pre-cancerous and normal tissues should be considered to represent the transient presence of an infectious agent transmitted by the fecal-oral route [Bofill-Mas et al., 2010]. Nevertheless, the possibility of the genomic integration of Merkel cell polyomavirus in permissive pre-cancerous cells could be considered, and its significance in the oncogenic process could then be established. In any case, the oncogenic mechanism of Merkel cell polyomavirus differs from that of other polyomaviruses, including JCV, whose detection in adenomas and polyps has been considered to be an important step in the “hit-and-run” theory of tumorigenesis [Jung et al., 2008; Shuda et al., 2008].

In conclusion, the findings of this study are consistent with the hypothesis that the ubiquitous Merkel cell polyomavirus [Carter et al., 2009] is present fairly frequently in the gastrointestinal tract, and that its causal role in colon cancer, in addition to the role of other human and simian polyomaviruses, should be reassessed.