Dear Editor,

It has been demonstrated, that DNA from the Merkel cell polyoma virus (MCV) is monoclonally integrated in the genome of Merkel cell carcinoma (MCC) cells in the majority of tumors.1 In this respect, Bathia et al. recently reported an observation in THE JOURNAL which suggests that two subgroups of MCC can be distinguished on the basis of the abundance of MCV; moreover, these subgroups differed in their expression of cancer related proteins, i.e. the Retinoblastoma protein (RB).2 The authors report that MCV DNA load was less then one copy per 300 cells in 14 of 23 MCC tumors (60%), and that these tumors were characterized also by loss of RB expression. In the remaining samples, which were characterized by high levels of RB expression, the estimated viral load was always higher than 1 copy per 20 cells. Noteworthy, only in two cases viral load was higher than 1 copy per 2 cells. In consequence, the authors speculate on a possible mechanism in MCC with a minority of infected cells contributing to transformation of uninfected neighbouring cells by paracrine mechanisms.2 However, the results of Bathia and colleagues are in contrast to another recently published paper describing the viral load in MCV+ MCCs to be generally >1 copy per tumor cell as measured by quantitative PCR.3 Moreover, immunohistological analyses of MCV+ MCCs demonstrated that nearly every tumor cell expressed the MCV large T-antigen (LTA) protein. Confirming this latter report, our own immunohistochemical analysis of MCC tissues confirms this pattern of LTA expression, i.e. virtually every tumor cell of MCV+ MCCs is expressing LTA protein (Fig. 1a). Moreover, we analyzed genomic DNA purified from 50 MCC tissue samples for the presence of MCV by Real time PCR. Relative quantification of the samples was calculated by the ΔΔCt method normalized to the repetitive DNA elements LINE1; genomic DNA of a MCC cell line that harbours at least two concatemerized copies of the MCV genome in every cell served as calibrator which allowed to approximate the minimal copy number per cell in the tissues. To this end, MCV DNA was undetectable in only 7 of the 50 samples (14%). In those samples with detectable MCV DNA the median of the estimated minimal copy numbers is approximately 1 copy per cell (Fig. 1b). Indeed, 63% of the MCV positive cases (27 samples) demonstrate minimal copy numbers higher than 0.5 copies/cell. Even some of the values below 0.5 may still be in accordance with 1 MCV copy per tumor cell as some of the tissue samples harbour high proportions of non tumour cells. In only 7 MCV+ samples we estimated a MCV copy number of less than 1 per 20 cells; thus, we can not confirm the findings of Bhatia et al. with respect to the viral load being very low in the majority of MCC cases.

thumbnail image

Figure 1. Presence of MCV in MCC tissues. (a) Immunohistochemical staining of MCC tumor tissues demonstrates expression of the MCV large T-antigen (LTA) in virtually every tumor cell of an MCV+ MCC. The LTA specific antibody CM2B43 was applied to formalin fixed and paraffin embedded Merkel cell carcinoma tissues. (b) Relative quantification of MCV DNA in genomic DNA derived from MCC tissues. The Real time PCR method using Taqman technology has been described in detail previously.7 For an estimation of the minimal number of viral copies per cell, DNA derived from a MCV+ MCC cell line which contains a concatamer of at least two MCV copies was used as calibrator. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

In addition to the differences in virus copy number, we obtained also different results with respect to RB expression; Bhatia and colleagues observed RB negative MCC cases in 60%. In contrast, we could not detect loss of RB expression in a substantial proportion of MCC cases. In fact, we detected RB protein in all samples in virtually all tumor cells when applying immunohistochemical analysis to 50 different paraffin-embedded MCC tumor tissues (Fig. 2a). This series included also 7 cases negatively tested for MCV DNA and 7 cases with relatively low approximated copy numbers indicating that viral abundance in MCC tissues is not correlated with RB expression (Fig. 1b). The presence of RB in the vast majority of MCC tissues was further substantiated by use of a phospho-RB specific antibody which detected the protein in immunohistochemical analysis in a subset of 12 MCC samples (Fig. 2a). Notably, the staining pattern with only some of the tumor cells displaying phosphorylation of RB is consistent with RB phosphorylation being a cell cycle dependent event. Indeed, this has been described both, for non virus-related but also for T-antigen transformed cells with the difference that in the latter the phosphorylation occurs after rather than before entry into the S phase.4 Interestingly, phospho-RB could be detected despite the fact that considerable levels of the CDK inhibitor p16INK4a were expressed in all 50 analyzed MCC tissues (Fig. 2a). Cell cycle dependent RB phosphorylation in the presence of high levels of p16INK4a was also confirmed in three different MCV MCC cell lines. These cell lines increased phospho-RB upon release from cell cycle arrest (Fig. 2b).

thumbnail image

Figure 2. Expression and phosphorylation of the Retinoblastoma protein in MCV-positive and -negative MCC (a) Immunohistochemical staining of MCC tumor tissues. The antibodies applied were α-RB (RB-1441-PO, Dako), α-phospho-RB ser807/811 (Cell Signalling) and α-p16Ink4a (Clone 16P04, Neo Markers). Depicted are stainings of MCV+ and MCV MCC tissues, as indicated. (b) Western blot analysis of total cell lysates derived from the MCV MCC cell line MCC13. The cells were cell cycle arrested by growing them to confluency. Cells were harvested at confluency (0 h) or after splitting at the indicated time points. The antibodies applied were α-phospho-RB ser608 (Cell Signalling) and α-p16Ink4a (Clone 16P04, Neo Markers) and α-beta-tubulin (Sigma Aldrich). [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

These different observations may be explained by the different geographical localisation of population from whom the MCC tumors were derived. Indeed, Garneski and colleagues investigated the presence of MCV DNA in 16 and 21 formalin-fixed MCC tissues from patients in North America and Australia, respectively.5 Using real-time PCR with primers sets designed to amplify regions of the MCV genome corresponding to the small and large T-antigens, MCV DNA was detected in 43% of the total 37 MCC tissues. When samples from North America and Australia were considered separately, the prevalence of MCV DNA in MCC tissues was greater for those samples obtained from North America (69%) as compared to Australia (24%) (p = 0.009).

In conclusion, our analysis demonstrates that RB expression is a general feature of both MCV+ and MCV MCC cells. Inactivation of the RB pathway is proposed to occur in virtually all tumors; this may be achieved in MCV+ MCC cells by expression of the MCV LTA, which is suspected to inactivate RB.6 For MCV MCCs the mechanism is less clear. Loss of p16INK4a, which binds to CDK4 and CDK6 and thereby prevents phosphorylation of RB at the G1/S transition, is a critical RB pathway inactivating event in several tumors. However, our data suggest that in MCV MCC neither this nor loss of RB expression take place. Instead, the observed cell cycle dependent phosphorylation of RB in the presence of high levels of p16INK4a indicate a similar RB inactivating event in both MCV+ and MCV MCC. However, for the MCV cases this event is to date completely unclear, but it is tempting to speculate that it may be caused also by a virus.


We highly appreciate that Patrick Moore made his LTA specific antibody—prior to its publication—available to us. Roland Houben was supported by the Wilhelm-Sander Stiftung (grant 2007.057.1).


  1. Top of page
  • 1
    Feng H, Shuda M, Chang Y, Moore PS. Clonal integration of a polyomavirus in human Merkel cell carcinoma. Science 2008; 319: 10961100.
  • 2
    Bhatia K, Goedert JJ, Modali R, Preiss L, Ayers LW. Merkel cell carcinoma subgroups by merkel cell polyomavirus DNA relative abundance and oncogene expression. Int J Cancer 2009. PMID: 19551862
  • 3
    Shuda M, Arora R, Kwun HJ, Feng H, Sarid R, Fernandez-Figueras MT, Tolstov Y, Gjoerup O, Mansukhani MM, Swerdlow SH, Chaudhary PM, Kirkwood JM, Nalesnik MA, Kant JA, Weiss LM, Moore PS, Chang Y. Human Merkel cell polyomavirus infection I. MCV T antigen expression in Merkel cell carcinoma, lymphoid tissues and lymphoid tumors. Int J Cancer 2009. PMID: 19499546
  • 4
    Ludlow JW, Shon J, Pipas JM, Livingston DM, DeCaprio JA. The retinoblastoma susceptibility gene product undergoes cell cycle-dependent dephosphorylation and binding to and release from SV40 large T. Cell 1990; 60: 387396.
  • 5
    Garneski KM, Warcola AH, Feng Q, Kiviat NB, Leonard JH, Nghiem P. Merkel cell polyomavirus is more frequently present in North American than Australian Merkel cell carcinoma tumors. J Invest Dermatol 2009; 129: 246248.
  • 6
    Shuda M, Feng H, Kwun HJ, Rosen ST, Gjoerup O, Moore PS, Chang Y. T antigen mutations are a human tumor-specific signature for Merkel cell polyomavirus. Proc Natl Acad Sci USA 2008; 105: 1627216277.
  • 7
    Becker JC, Houben R, Ugurel S, Trefzer U, Pfohler C, Schrama D. MC polyomavirus is frequently present in Merkel cell carcinoma of European patients. J Invest Dermatol 2009; 129: 248250.

Roland Houben, David Schrama, Miriam Alb, Claudia Pföhler, Uwe Trefzer, Selma Ugurel, Jürgen C. Becker.