Merkel Cell Carcinoma (MCC) is a rare, aggressive neoplasm, which commonly involves the skin, but can subsequently metastasize to lymph nodes and other organs. These tumors are believed to arise from neuroendocrine-derived mechanoreceptor Merkel cells, located in the basal layer of the epidermis, that form synapse-like contacts with enlarged nerve terminals.
Although rare, the incidence of MCC has tripled over the past 15 years to approximately 1500 new MCC cases in the U.S. each year. MCC is primarily seen in the elderly Caucasian population 65 years of age and above, and shows a slight predominance in males. In addition to UV exposure, immunosuppression is a risk factor for MCC and the cancer occurs more commonly than expected among AIDS and post-transplant patients. This strong association between immunosuppression and MCC is highly suggestive for an infectious origin for this disease. While surgery for localized lesions can be curative, MCC is one of the most lethal skin cancers if it disseminate. MCC has also been reported as a primary or secondary malignancy for multiple myeloma, chronic lymphocytic leukemia (CLL), melanoma, non-Hodgkin lymphoma and other cancers.
Merkel Cell Polyomavirus (MCV)
The Chang/Moore lab is interested in the search for oncogenic viruses as potential etiologic agents in human malignancies. In 2008 we identified a novel polyomavirus in Merkel cell carcinoma (MCC) tissues (Feng, Shuda et al. 2008). A direct sequencing approach called digital transcriptome subtraction (DTS) was used to identify foreign transcripts with homology to polyomavirus T antigens in the Merkel cell tumor genome. MCV genome was detected in 8 out of 10 MCC tumors used in this initial study. Importantly, the viral genome was found to be clonally-integrated into the tumor genomes indicating that it was present prior to the cancer cell clonal expansion. Two full-length viral genomes, MCC350 and MCC339 were deposited in Genbank (EU375803 and EU375804, respectively) and DNA clones are available through the AIDS Reasearch Reagent Repository. Since these genomes were cloned from tumors and are nonpermissive for viral replication, it is likely that mutations are present in these genomes that are not present in wild-type viruses. MCV shares significant homology to viruses belonging to the MuPyV subgroup and is very closely related to the African Green Monkey lymphotrophic polyomavirus (LPyV) and more distantly to other known human polyomaviruses and SV40.
Tumor-specific MCV LT antigen
The MCV T antigen locus encodes for four differentially spliced mRNA transcripts corresponding to polyomavirus large T antigen (LT), and small T antigen (ST, encoded by two transcripts) as well as an additional isoform 57kT, which may represent an analogue to the SV40 17kT transcript (Shuda, Feng et al. 2008). The MCV LT contains conserved features of other polyomavirus LT proteins:(1) the conserved region 1 (CR1), which is functionally similar to the cell transforming E1A region of adenoviruses (2) the DnaJ domain, which binds heat shock chaperone 70 (Hsc70) (3) the retinoblastoma tumor suppressor (Rb) protein family binding motif LXCXE, which induces cell cycle progression and (4) the OBD and helicase/ATPase domains are required for viral DNA replication. We find that tumor-derived LT transcripts (LTT) from integrated viruses have premature stop codon mutations or deletions that eliminate the helicase/ATPase domain(Shuda, Feng et al. 2008). Some LTT also delete the OBD as seen in MCC 350. However, this is not the case in all tumor-derived LTs, which indicates that the OBD does not benefit nor negatively influence tumor progression. MCV positive control tissues from various body sites do not show these LT truncation mutations. It is reasonable to believe that the wild type virus is maintained episomally at these sites.
Does MCV Cause Merkel Cell Carcinoma?
There is substantial evidence that suggests a causative role for MCV in the pathogenesis of most MCC tumors. There are a subset of MCC (20-30%), which do not harbor MCV and apparently have a distinct pathoetiology.
1. Monoclonal Integration of MCV in MCC
In the initial report Feng and colleagues found 7 out of 10 MCC tumors from different patients to be positive for MCV by southern blotting (a contamination free method used to detect specific DNA sequences in DNA samples via hybridization). This high rate of MCV positivity has been confirmed many other laboratories around the world (Kassem, Schopflin et al. 2008; Becker, Houben et al. 2009; Bhatia, Goedert et al. 2009; Busam, Jungbluth et al. 2009; Garneski, Warcola et al. 2009). Around 75% of MCC cases are positive for MCV DNA. Southern blotting and 3??T-RACE also shows that MCV is clonally integrated into the MCC genome within tumors. This suggests that MCV infection preceded initial tumor development and monoclonal expansion of cells. Viral integration is analogous to that seen in high-risk human papillomavirus integration in cervical cancer cells. MCV genome integration occurs at different sites in the genome in different individual cases. In one case studied by Feng et al. a metastatic tumor had the same integration pattern as the primary tumor showing that the metastasis arose from a single cancer cell derived from the primary tumor. Although MCV integrates at different sites, it is currently unknown whether or not virus integration interrupts common cellular pathways that prevent tumor cell growth.
2. MCV is Specific for MCC
Numerous surveys on other tumor types have been performed including other neuroendocrine tumors, cutaneous melanomas, basal cell carcinomas, sqamous cell carcinomas and hematolymphoid malignancies. Most of these studies have found that MCV only in MCC. In cases where MCV is found in non-MCC tumors, it generally is at low abundance suggesting incidental infection (Shuda, Arora et al. 2009).
3. Copy Number of MCV in MCC (Biological Gradient)
Quantitative PCR studies (as well as Southern blotting) from our lab show that the virus is present at > 1 copy per cell in MCC tumors positive for the virus. Most non-tumor tissues are negative for virus genome. Some non-MCC tumor tissues have been found to harbor virus but at 2-3 logs lower than that found in MCC suggesting incidental infection (Shuda, Arora et al. 2009). These data show that MCV is specifically associated with MCC and that the virus is in a high copy number in tumors. Virus genome can also be occasionally detected by PCR from tissues without tumors (gut, respiratory tract) suggesting that it natively infects these tissues. It is not known whether MCV infection is associated with non-cancerous diseases.
4. Signature Truncation Mutations in Tumor Derived Viruses
Tumor-derived MCV possess premature stop codon mutations resulting in truncation of the large T protein. These peculiar mutations are specific for MCV found in cancers and absent from viruses derived from non-tumor tissues (wild-type viruses). The truncation mutations eliminate the T antigen helicase activity but retain the LXCXE??”retinoblastoma protein-binding motif as well as other N-terminal motifs found in the T antigens (Shuda, Feng et al. 2008). Hence, these viruses retain ability to regulate the host cell cycle and inhibit the retinoblastoma protein gene family but can no longer initiate replication, and the virus would be lost if it were not integrated into the host tumor cell genome. For this reason MCV cannot be a passenger virus that secondarily infects the tumor. These two independent mutation events ??” virus integration and T antigen truncation seem to play a mechanistic role in the development of MCC and may also explain why MCC is relatively rare. It is likely that skin exposure to UV or ionizing radiation increases risk for MCC by enhancing T antigen and virus integration mutations.
5. MCV T antigen expression is specific to Merkel cell carcinoma (CM2B4 staining)
A monoclonal antibody raised against T antigen protein(CM2B4, developed an tested in our laboratory) shows uniform nuclear staining in MCC tumor cells. Surrounding non-neoplastic cells do not express MCV T antigen. A large number of other tumors tested via CM2B4 immunohistochemistry based screening do not show MCV large T antigen protein expression (Shuda, Arora et al. 2009).
MCV, however, is a common human infection
Serological studies from our laboratory (Tolstov, Pastrana et al. 2009) and from other groups suggest that MCV is a common infection of humans that causes MCC on rare occasions. Together with the Buck Laboratory (NIH), we developed a serologic assay made from virus-like particles (VLP), generated by expression of MCV VP1 and VP2 proteins. A VLP enzyme-linked immunosorbent assay does not cross-react with human BKV or murine polyomaviruses, suggesting that MCV VLP is a suitable antigen for a blood test. MCC patients have significantly higher titers (more IgG but not IgM antibodies) against MCV than various control population. The subset of MCV-negative MCC patients have significantly lower average MCV antibody titers than those of MCV-positive MCC patients. Age-specific MCV prevalence increases from ~50% among children aged 15 years or younger to ~80% among persons older than 50 years. These results suggest that MCV is a common but previously unrecognized human infection.
MCV Origin Replication Studies
MCV replicationis another area of interest for our laboratory since loss of viral replication capacity is intimately tied into MCV tumorigenesis. Early studies of other animal polyomaviruses showed that tumors and cancer cell transformation occur in cell lines or animals that are non-permissive or semi-permissive for virus replication. Cells that are fully permissive for polyomavirus replication generally die in the process of generating fully-infectious virions. Like other polyomaviruses, the helicase domain of large T (LT) assembles on the origin and opens the origin sequence toallow initiation of cellular polymerase-mediated replication.
To investigate MCV replication, we have cloned the MCV origin into a plasmid which can be transfected together with full-length, wild-type MCV T antigen into eukaryotic cells, and replication of the plasmid is measured by Southern blotting. The results of this replication assay reveal that the minimal MCV core origin is 71 base pairs in length. This sequence contains a poly T tract as seen in SV40, as well as the typical polyomavirus GAGGC repeats that serve as binding sites for LT. Mutation of the 8 pentanucleotide sequences within the core origin showed that there are 4 pentanucleotide sequences that are indispensable for replication. However, unlike in SV40 these are not arranged in a head to head fashion, implying that LT binding and helicase hexamer formation may proceed by a different process than seen in SV40. Consistent with this, MCV and SV40 T antigens do not activate replication of each other??Ts viral origins (Kwun, Guastafierro et al. 2009).
Effects of Tumor-derived LT mutations on origin binding
As described previously, tumor-derived MCV LT is unable to initiate MCV origin replication due to truncating mutations in the LT protein that disrupt the helicase domain. In addition, we have found single nucleotide mutations in the origin of a tumor-derived strain that also prevent replication by full-length LT. We compared origin-binding properties for wild-type LT protein (TAg206.wt) on the wild-type origin (Ori339) and the mutated Ori350 (Ori350), as well as mutant LT proteins from tumor derived viral strains (MCV339 and MCV350). MCV339 LT (TAg339) protein retains the OBD but has a truncated helicase domain, whereas the MCV350 LT (TAg350) mutation eliminates both the OBD and helicase domains. Quantitative chromatin immuno-precipitation (ChIP) assays demonstrate that wild-type LT (TAg206) efficiently binds to the wild-type origin Ori339 but binding to Ori350, possessing the single base pair mutation in a critical pentanucleotide sequence, is reduced ~50%. Origin-binding by MCV339 and MCV350 LT proteins is comparable to vector alone controls for both viral origins suggesting that structural or enzymatic features of the MCV LT helicase domain are required for efficient recognition of the DNA element (Kwun, Guastafierro et al. 2009).
Other factors that influence viral replication
We also find that MCV LT cDNA expression alone is not sufficient for efficient replication, but replication efficiency is restored to that of the wild type genomic LT by co-expression of LT cDNA with ST cDNA. Co-expression of LT with 57kT does not increase replication over LT alone, and neither ST nor 57kT individually have replication capacity. On LT (but not ST), the HPDK motif within the LT DnaJ domain is required for replication suggesting that recruitment of heat-shock proteins are involved in MCV origin replication. For ST, mutations in its PP2A-recognition motif eliminate the enhancing activity of ST.
Model for Tumor Cell Evolution of MCV
Since all tumor-derived LT sequences analyzed so far show premature stop codons or sequence deletions that eliminate the helicase domain of LT, this suggests that there is strong selective pressure to delete the helicase domain with viral integration. It is reasonable to believe that through the process of natural selection only those cells with integrated viral genomes that delete the helicase domain become tumor cells. In addition, we observe a point mutation in the OBD of a MCV genome isolated from MCC, which eliminates replication completely. This suggests that viral replication has to be eliminated upon viral integration for tumorigenesis to occur.
One explanation for this is that MCV LT initiates promiscuous origin replication that is unlicensed by the cell and thus generates multiple copies viral DNA fragments within a given cell cycle. Integration of such an adventitious origin into the host genome is likely to generate replication fork collisions and DNA fragmentation. A rare cell that has accidental integration of MCV genome can only be rescued if corresponding mutations that eliminate the LT helicase activity occur. If these mutations leave intact the oncogenic N-terminal domains of T antigen, then the cell is at risk for uncontrolled replication. It is likely that as the tumor cells expand, additional cellular mutations that enhance cell proliferation and prevent apoptosis are acquired resulting in the mature Merkelc cell cancer cell.
MCV teaching slides are available here.
Becker, J. C., R. Houben, et al. (2009). “MC polyomavirus is frequently present in Merkel cell carcinoma of European patients.”J Invest Dermatol 129(1): 248-50.
Bhatia, K., J. J. Goedert, et al. (2009). “Merkel cell carcinoma subgroups by merkel cell polyomavirus DNA relative abundance and oncogene expression.” Int J Cancer.
Busam, K. J., A. A. Jungbluth, et al. (2009). “Merkel cell polyomavirus expression in merkel cell carcinomas and its absence in combined tumors and pulmonary neuroendocrine carcinomas.”Am J Surg Pathol 33(9): 1378-85.
Feng, H., M. Shuda, et al. (2008). “Clonal integration of a polyomavirus in human Merkel cell carcinoma.” Science319(5866): 1096-100.
Garneski, K. M., A. H. Warcola, et al. (2009). “Merkel cell polyomavirus is more frequently present in North American than Australian Merkel cell carcinoma tumors.”J Invest Dermatol 129(1): 246-8.
Kassem, A., A. Schopflin, et al. (2008). “Frequent detection of Merkel cell polyomavirus in human Merkel cell carcinomas and identification of a unique deletion in the VP1 gene.”Cancer Res 68(13): 5009-13.
Kwun, H. J., A. Guastafierro, et al. (2009). “The minimum replication origin of merkel cell polyomavirus has a unique large T-antigen loading architecture and requires small T-antigen expression for optimal replication.”J Virol83(23): 12118-28.
Shuda, M., R. Arora, et al. (2009). “Human Merkel cell polyomavirus infection I. MCV T antigen expression in Merkel cell carcinoma, lymphoid tissues and lymphoid tumors.” Int J Cancer 125(6): 1243-9.
Shuda, M., H. Feng, et al. (2008). “T antigen mutations are a human tumor-specific signature for Merkel cell polyomavirus.” Proc Natl Acad Sci U S A 105(42): 16272-7.
Tolstov, Y. L., D. V. Pastrana, et al. (2009). “Human Merkel cell polyomavirus infection II. MCV is a common human infection that can be detected by conformational capsid epitope immunoassays.” Int J Cancer 125(6): 1250-6.