4.2.1. Growth Factor Signaling

Merkel cell carcinoma is a rare but aggressive skin tumor that grows rapidly and, if untreated, may double in size within a week (Becker et al. 2009a; Houben et al. 2009). Normal cells require mitogenic growth signals in order to move from a quiescent state into an active proliferative state (Hanahan and Weinberg 2000). These signals are transmitted into the cell via transmembrane receptors that bind distinctive classes of signaling molecules. In contrast, tumor cells are able to generate their own growth signals and are not as dependent on exogenous growth stimulation. Many oncogenes in cancer cells act by mimicking normal growth factor signals. Further, growth factor receptors are overexpressed or modified in many cancers, resulting in an enhanced response to circulating levels of exogenous growth factor signals or constitutive expression. Relevant growth factor signaling changes reported for Merkel cell carcinoma include a novel single heterozygous base change in exon 10 of the platelet-derived growth factor (PDGF) receptor and heterozygous loss of chromosome 10 or the long arm of chromosome 10 where the tumor suppressor phosphatase and tensin homologue (PTEN) is encoded (Fernández-Figueras et al. 2007; Houben et al. 2009; Swick et al. 2008; Van Gele et al. 2001; Van Gele et al. 1998). However, it was not clear if the base change in PDGF represented a true mutation or a single nucleotide polymorphism (Houben et al. 2009; Swick et al. 2008). No mutations were observed that affected the mitogen activated protein kinase (MAPK) pathway (Houben et al. 2006).

4.2.2. Evading Growth Suppressors

Normal tissues maintain cellular quiescence and homeostasis through multiple antiproliferative signals or growth suppressors (Hanahan and Weinberg 2000; 2011). In addition to inducing and sustaining growth-stimulatory signals, cancer cells must evade powerful antigrowth signals, many of which depend on the actions of tumor suppressor genes. In particular, the retinoblastoma and p53 tumor suppressor pathways are interconnected and operate as central control nodes to regulate cell proliferation, senescence, and apoptosis (Hanahan and Weinberg 2000; 2011; Houben et al. 2009; Yamasaki 2003). Retinoblastoma transduces growth-inhibitory signals originating primarily from outside the cell while p53 receives input from stress sensors, e.g., DNA damage, within the cell (Hanahan and Weinberg 2011; Houben et al. 2009). When in a hypophosphorylated state, retinoblastoma protein blocks cell proliferation by sequestering and altering the function of E2F transcription factors that control the expression of genes essential for progression from G1 to S phase. Phosphorylation of retinoblastoma by cyclin/cyclin-dependent kinase complexes causes dissociation of the retinoblastoma-E2F complex and cell-cycle entry (Sihto et al. 2011). The retinoblastoma and/or p53 pathways are dysregulated in virtually all human tumors (Yamasaki 2003).

Several studies have shown that the retinoblastoma pathway is critical to Merkel cell carcinoma pathogenesis while p53 mutations are rare (Bhatia et al. 2010; Borchert et al. 2014; Cimino et al. 2014; Harms et al. 2013; Higaki-Mori et al. 2012; Houben et al. 2012; Kuwamoto 2011; Lassacher et al. 2008; Sahi et al. 2014; Sihto et al. 2011). Merkel cell carcinoma retinoblastoma expression has a strong positive association with MCV DNA and MCV large T-antigen (LT) expression, suggesting that retinoblastoma inhibition is important for MCV-induced tumorigenesis (Sihto et al. 2011). Although LT expression was not associated with expression of phosphorylated retinoblastoma, LT binds to retinoblastoma, thus reducing retinoblastoma-E2F complex formation and inhibiting its cell-cycle regulation function. The oncogenic role of LT is discussed further in Section 4.3.2.

Cimino et al. (2014) reported that the retinoblastoma pathway was dysregulated in both MCV-negative and MCV-positive Merkel cell carcinoma cases and proposed two separate pathways of Merkel cell carcinoma oncogenesis. In MCV-positive cases, the retinoblastoma protein is functionally inactivated as described above. In MCV-negative cases, Merkel cell carcinoma tumors had truncating, nonsense mutations in the retinoblastoma gene. Mutations in the p53 gene were found primarily in LT- or retinoblastoma-negative tumors suggesting possible involvement of p53 in MCV-negative tumors (Sihto et al. 2011). Additionally, hypermethylation of the p14ARF promoter DNA has been reported in about 40% of Merkel cell carcinoma samples (Lassacher et al. 2008). Silencing of p14ARF could cause inactivation of the p53 pathway. Asioli et al. (2007) also reported that 25 of 47 cases of Merkel cell carcinoma were positive for p63 expression (a member of the p53 family) and that these cases demonstrated a more aggressive clinical course. In contrast, Higaki-Mori et al. (2012) showed no significant correlation of MCV infection and survival with p63 expression. However, the p63 gene is frequently amplified or overexpressed in human cancers (Asioli et al. 2007).

4.2.3. Apoptosis

Apoptosis is controlled by both intrinsic and extrinsic signaling pathways and involves counterbalancing pro- and antiapoptotic members of the Bcl-2 family of regulatory proteins (Adams and Cory 2007; Hanahan and Weinberg 2011). Bcl-2, and related proteins inhibit apoptosis by binding to and suppressing proapoptotic triggering proteins (Bax and Bak) that are embedded in the mitochondrial outer membrane. Overexpression of anti-apoptotic proteins or loss of pro-apoptotic signals results in loss of tissue homeostasis and supports oncogenesis by allowing cancer cells to evade programmed cell death. Bcl-2 overexpression is a common finding in many tumors and has been observed in approximately 67% to 85% of Merkel cell carcinoma tumors examined (Feinmesser et al. 1999; Houben et al. 2009; Kennedy et al. 1996; Sahi et al. 2012). However, Bcl-2 protein expression was not correlated with the MCV status of the tumors (Sahi et al. 2012). The presence of MCV also was associated with deregulated expression of the Bcl-2 gene in several cases of non-small-cell lung cancer (Lasithiotaki et al. 2013). Bcl-2 expression was downregulated in MCV-positive lung tumors compared with MCV-negative tumors (p = 0.05) or healthy tissue (p = 0.047), and the Bax/Bcl-2 ratio was 0.97 for the lung cancer group compared to 8.06 for the controls. A Bax/Bcl-2 ratio of <1 is associated with a lower apoptotic index (Brambilla et al. 1996). In addition, Bcl-2 inhibition was associated with Merkel cell carcinoma tumor shrinkage in an in vivo SCID mouse/human Merkel cell carcinoma xenograft model (Schlagbauer-Wadl et al. 2000).

The antiapoptotic protein survivin was also overexpressed in Merkel cell carcinoma tissue and was found to have a critical role in the survival of MCV-positive Merkel cell carcinoma cells (Arora et al. 2012b; Dresang et al. 2013; Kim and McNiff 2008; Sahi et al. 2012). A sevenfold increase in mRNA encoding the survivin oncoprotein was reported for MCV-positive compared with MCV-negative Merkel cell carcinoma tumors (Arora et al. 2012b). Xie et al. (2014) reported that decreased transcript and protein detection of the survivin gene in MCV-negative Merkel cell carcinoma cells was due to overexpression of microRNA (miRNA) miR-203. miR-203 functions as a tumor suppressor that is downregulated in certain cancers, and its expression was significantly lower in MCV-positive tumors compared with MCV-negative tumors. Nuclear staining for survivin was also associated with an aggressive clinical course and poor prognosis (Kim and McNiff 2008).

4.2.4. Angiogenesis

New blood vessel growth, or angiogenesis, is essential to sustain neoplastic development. To accomplish this growth, an “angiogenic switch” is almost always activated and remains on, causing normally quiescent vasculature to sprout new vessels to supply the growing tumor (Hanahan and Weinberg 2011). Increased expression of vascular endothelial growth factor (VEGF) via activation of hypoxia-inducible factor (HIF-1) is commonly observed in cancer. Although HIF pathway activation was not demonstrated with Merkel cell carcinoma, there was a significant association between metastatic tumor spread and elevated expression of VEGF (Fernández-Figueras et al. 2007). Another study reported that VEGF-A, VEGF-C, and VEGF-receptor 2 were expressed in 91%, 75%, and 88%, respectively, of the 32 Merkel cell carcinomas examined (Brunner et al. 2008).

4.2.5. Immune Evasion

The initial event in MCV-induced Merkel cell carcinoma is most likely a loss of immune surveillance for the virus as evidenced by increased risk in immunosuppressed populations (Amber et al. 2013; Hughes and Gao 2013; Moore and Chang 2010). MCV-related cases of Merkel cell carcinoma display vigorous antibody responses to MCV structural proteins; however, Merkel cell carcinoma tumors do not express detectable amounts of MCV VP1 capsid protein (Pastrana et al. 2009). These data suggest that the strong humoral responses in Merkel cell carcinoma patients are primed by an unusually robust MCV infection rather than Merkel cell carcinoma tumor viral antigen expression and that loss of cellular immune control may allow more extensive viral spread before tumor development (Moore and Chang 2010; Pastrana et al. 2009). Other mechanisms of immune evasions include downregulation of genes associated with the innate immune response (e.g., CCL20, CXCL-9, IL-2, IL-8, TANK) (Moens et al. 2015; Stakaitytė et al. 2014), downregulation of Toll-like receptor 9 (TLR9) (Griffiths et al. 2013), and disruption of inflammatory signaling via inhibition of the NF- κB essential modulator (NEMO) adaptor protein (Shahzad et al. 2013). TLR9 is a key receptor in the host innate immune response that senses viral or bacterial dsDNA. Mouchet et al. (2014) compared transcriptional profiles in MCV-positive Merkel cell carcinoma cells and normal Merkel cells and reported that most of the downregulated genes were related to immune interactions.

A common feature of oncogenic viruses is that they persist in the host as a latent or pseudo-latent infection that generally does not replicate to form infectious virus particles (also known as lytic replication) in tumor cells (Moore and Chang 2010). Latent infection serves as an immune evasion strategy that allows the virus to hide from the immune system by turning off unnecessary viral proteins that might be detected by cell-mediated immunity. When latent viruses switch to lytic infection, virus replication generates pathogen-associated molecular patterns that trigger DNA damage responses and innate immune signaling. These cellular responses result in death of the infected cell and release of infectious virions. Integration of small DNA tumor viruses, such as MCV, into the nascent tumor cell eliminates their ability to replicate as virions (a state of pseudo-latency). Monoclonal integration of viral DNA within individual tumors provides the primary evidence that MCV causes most cases of Merkel cell carcinoma (Moore and Chang 2010; Pastrana et al. 2009). These concepts are discussed further in the following sections.

4.3.1. Presence and Persistence of MCV in Merkel Cell Carcinoma

MCV has been detected in approximately 70% to 97% of Merkel cell carcinoma cases (Amber et al. 2013; Feng et al. 2008; IARC 2013; Rodig et al. 2012; Sihto et al. 2009; Spurgeon and Lambert 2013; Stakaitytė et al. 2014). The data suggest that Merkel cell carcinoma can develop through both a virus-mediated pathway in which MCV promotes tumorigenesis and possibly, in a minority of cases, a nonvirus-mediated pathway (Amber et al. 2013; Martel-Jantin et al. 2012). However, some have suggested that the presence of MCV in Merkel cell carcinoma is more common than reported and that improved detection methods might reveal that all Merkel cell carcinoma specimens contain MCV DNA (Rodig et al. 2012). Martel-Jantin et al. (2012) also reported that MCV detection was much higher in fresh-frozen biopsies than in formalin-fixed paraffin-embedded biopsies. Carter et al. (2009) reported that while 77% (24/31) of Merkel cell carcinoma samples were positive for MCV, 92% (22/24) of these patients were positive for antibodies to MCV. These results raise the possibility that MCV is involved in Merkel cell carcinoma initiation but may not be required to maintain the cancer phenotype and may explain why some Merkel cell carcinomas are negative for MCV DNA. Virus-negative carcinomas might indicate advanced tumor stages with a secondary loss of virus genomes (Niller et al. 2011). It is well documented that viral genomes, either inserted into the host cellular DNA or co-replicating with it in episomal form, can be subsequently lost from neoplastic cells (i.e., a “hit and run” mechanism). The number of copies of the MCV genome found integrated in Merkel cell carcinoma ranges widely from less than one to several thousand copies (Bhatia et al. 2010; DeCaprio and Garcea 2013; Laude et al. 2010; Rodig et al. 2012; Shuda et al. 2009).

4.3.2. Viral Oncogenes and Maintenance of the Malignant Phenotype

The MCV genome consists of about 5,400 base pairs that are divided into early and late coding regions separated by a noncoding regulatory region (see Section 1.1.2) (Chang and Moore 2012; Spurgeon and Lambert 2013; Stakaitytė et al. 2014). The early coding region expresses overlapping nonstructural transcripts from a single T antigen locus that are differentially spliced to form small T (sT) and large T (LT) antigens. The late region encodes MCV structural proteins and a miRNA (MCV-miR-M1-5p) that is encoded antisense to the LT coding region. The miRNA is thought to downregulate expression of the early genes and may have a role in cellular transformation (Lee et al. 2011; Seo et al. 2009). Viral miRNA was detected in half of MCV-positive tumors but was not detected in MCV-negative tumors (Lee et al. 2011).

The sT and LT antigens are critical for viral replication, manipulation of the host cell cycle, and cellular transformation (Amber et al. 2013; Angermeyer et al. 2013; Stakaitytė et al. 2014). About 75% of Merkel cell carcinoma tumor samples are positive for LT, while about 92% are positive for sT (Stakaitytė et al. 2014). In normal infections in permissive cells, MCV completes its replication cycle in the cell nucleus to form virions without inducing tumorigenesis. The T antigens are transcribed immediately upon entry into the nucleus of the host cell and induce the cell to enter S-phase. However, mutagenic events can cause MCV integration into host cell DNA and truncation of LT (see In vitro studies Section below). In these circumstances, expression of sT and mutated LT dysregulate cell proliferation and prevent apoptosis primarily through interactions with retinoblastoma and eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) (Houben et al. 2012; Hughes and Gao 2013; Shuda et al. 2011). Expression of truncated LT also inhibits key responses to UV radiation-induced DNA damage (i.e., DNA repair and cell-cycle defects) and suggests that progressive MCV-mediated genomic instability contributes to Merkel cell carcinoma (Demetriou et al. 2012). This may explain why most cases of Merkel cell carcinoma occur on chronically sun-exposed skin. Mutational analyses have yet to identify other signature mutations in Merkel cell carcinoma (Erstad and Cusack 2014). However, Van Gele et al. (1998) reported that Merkel cell carcinoma cases showed a characteristic pattern of chromosomal gains and losses that were similar to that seen in small-cell lung carcinoma. The roles of LT and sT in cellular transformation and oncogenesis are discussed in the following sections.

In vitro Studies

All LT sequences recovered from primary Merkel cell carcinoma tumors or tumor-derived cell lines harbor signature mutations (Borchert et al. 2014; Schmitt et al. 2012; Shuda et al. 2008; Stakaitytė et al. 2014). These mutations cause premature truncation of the entire C-terminal domain, which leads to the loss of domains associated with viral replication (i.e., origin binding domain and the ATPase/helicase region). Deletions of C-terminal LT sequences appear to be a highly specific surrogate marker for MCV-induced malignancy (Schmitt et al. 2012). Although the sites of mutations are randomly distributed from different tumors, the retinoblastoma-binding motif is preserved (Borchert et al. 2014; Shuda et al. 2008). Integration of MCV genomes with full-length LT capable of initiating host DNA replication would result in unlicensed replication, replication fork collision, DNA breakage, and cytopathic cell death (Shuda et al. 2008; Stakaitytė et al. 2014). Full-length MCV LT also showed a decreased potential to support cellular proliferation, focus formation, and anchorage-independent cell growth via activation of host DNA damage responses and upregulation of p53 downstream target genes (Li et al. 2013).

Infected cells containing a wild-type episomal MCV genome can be transformed into a tumor cell containing multiple copies of an integrated mutant viral genome via two distinct models (DeCaprio and Garcea 2013). LT truncation and amplification of viral genome copy number may occur before or after random integration into the host genome (Figure 4-1). If wild-type MCV is integrated into the host genome, then it must be followed by an LT mutation to disable viral replication. The integrated mutant genome could subsequently undergo copy number amplification. In cases where the LT mutation occurs first, the mutant genome could undergo rolling-circle amplification prior to integration (DeCaprio and Garcea 2013; Stakaitytė et al. 2014). Therefore, at least two mutation events are required prior to tumorigenesis and this may explain why Merkel cell carcinoma is rare. In either case, there is a strong selection pressure within the Merkel cell carcinoma tumors to eliminate viral replication capabilities and retain only replication-deficient copies of MCV.

Figure 4-1

Models of MCV-induced Cell Transformation

Cheng et al. (2013) also demonstrated that truncated LT was more efficient than wild-type LT at inducing cellular proliferation. Knockdown of T antigen expression in MCV-positive Merkel cell carcinoma cell lines induced cell-cycle arrest and apoptosis in vitro and regression of established xenograft tumors in vivo (Houben et al. 2012; Houben et al. 2010). These effects were largely due to the interaction of LT with retinoblastoma. LT also was required for MCV-positive Merkel cell carcinoma cell growth and survival (Angermeyer et al. 2013; Shuda et al. 2011). Borchert et al. (2014) reported that truncated LT antigens exhibit a very high binding affinity for retinoblastoma and that both wild-type and truncated LT antigens could transform baby rat kidney epithelial cells. However, truncated LT antigen did not bind to p53 or reduce p53-dependent transcription. Since the constructs used in this study were likely able to express both LT and sT antigens, sT might have contributed to the transformation events. Liu et al. (2011) also identified human Vam6p (hVam6p) cytoplasmic protein as a novel target for MCV LT. MCV LT translocates hVam6p to the nucleus, sequestering it from its normal function in lysosomal processing. Although this study suggested that hVam6p sequestration was more likely to play a role in MCV replication than in tumorigenesis, the data were insufficient to rule out possible contributions to cell growth and proliferation.

sT expression was sufficient to induce cell transformation, loss of contact inhibition, and anchorage-independent growth in rodent fibroblast, and serum independent growth in human fibroblasts (Angermeyer et al. 2013; Shuda et al. 2011). Silencing of sT expression by sT-specific short hairpin RNAs lead to variable degrees of growth retardation; however, these effects were not sT specific because MCV-negative cell lines were similarly affected. MCV sT-induced cell transformation may be mediated by reducing the turnover of hyperphosphorylated 4E-BP1 (a downstream component of the PI3K-Akt-mTOR signaling pathway) (Shuda et al. 2011; Stakaitytė et al. 2014). The PI3K-Akt-mTOR signaling cascade is an important pathway in cell-cycle regulation that is overactive in many cancers and is often targeted by oncogenic viruses. Hyperphosphorylation prevents 4E-BP1 from sequestering the eukaryotic cap-dependent translation initiation factor 4E (eIF4E), allowing free eIF4E to form the cap assembly on mRNA and initiate translation. eIF4E is part of a multisubunit eIF4F complex (composed of eIF4E, eIF4A, and eIF4G). Overexpression of eIF4F can induce cell transformation in rodent and human cells in vitro (Avdulov et al. 2004; Lazaris-Karatzas et al. 1990; Stakaitytė et al. 2014). An alternative pathway for regulation of cap-dependent translation during mitosis is through cyclin-dependent kinase 1 (CDK1) hyperphosphorylation of 4E-BP1 (Shuda et al. 2015). sT-induced cell transformation was reversed by expression of a constitutively active mutant 4E-BP1 protein that could not be inactivated by MCV sT (Stakaitytė et al. 2014). sT also contributes to LT expression by blocking proteasomal degradation of LT by the cellular SCF^Fbw7 E3 ligase (Kwun et al. 2015; Kwun et al. 2013). sT inhibits E3 ligase through its LT stabilization domain (LSD) and consequently stabilizes other cellular Fbw7 targets such as the cell-cycle regulators c-myc and cyclin E.

These data suggest a synergistic role for both sT and mutated LT antigens during Merkel cell carcinoma tumorigenesis (Borchert et al. 2014; Houben et al. 2012; Shuda et al. 2011). sT may be essential for initial cell transformation and stabilizing LT, while LT is necessary for subsequent survival and proliferation of transformed cells (Stakaitytė et al. 2014). Some of the primary molecular targets and biological effects associated with MCV LT and sT antigens are illustrated in Figure 4-2.

Figure 4-2

Molecular Targets and Biological Effects of MCV LT and sT Antigens

4.3.3. MCV as a Major Risk Factor for Merkel Cell Carcinoma

Although MCV infection is common, Merkel cell carcinoma is rare (Chang and Moore 2012). Early clinical findings identified immunosuppression as a major risk factor for Merkel cell carcinoma and pointed toward an infectious etiology. Epidemiological studies (see Section 3) support an association of Merkel cell carcinoma cases with MCV infection. MCV antibody levels are significantly higher in MCV-positive Merkel cell carcinoma cases compared with healthy controls that are MCV seropositive which suggests that development of Merkel cell carcinoma is preceded by an unusually robust MCV infection (Agelli et al. 2010; Pastrana et al. 2009). MCV also is monoclonally integrated into the host genome in Merkel cell carcinoma primary tumors and metastases providing strong evidence that viral infection precedes clonal expansion of the neoplastic cell (Feng et al. 2008; Laude et al. 2010). Viral genome integration is a typical feature of virus-mediated oncogenesis and refutes the possibility that MCV is merely a coincidental passenger infection in Merkel cell carcinoma (Chang and Moore 2012; Kuwamoto 2011). These data, combined with several studies showing that expression of MCV T antigens are required to sustain tumor growth (see Section 4.2.2), provide strong evidence that MCV is a major risk factor for Merkel cell carcinoma. The key events identified for MCV-induced Merkel cell carcinoma are shown in Figure 4-3.

Figure 4-3

Key Events Leading from MCV Infection to Merkel Cell Carcinoma