Molecular Profiling of Merkel Cell Polyomavirus-Associated Merkel Cell Carcinoma and Cutaneous Melanoma

Merkel cell carcinoma (MCC) is a rare, high-grade, aggressive cutaneous neuroendocrine malignancy most commonly associated with sun-exposed areas of older individuals. A relatively newly identified human virus, the Merkel cell polyomavirus (MCPyV) has been implicated in the pathogenesis of MCC. Our study aimed to examine nine MCC cases and randomly selected 60 melanoma cases to identify MCPyV status and to elucidate genetic differences between virus-positive and -negative cases. Altogether, seven MCPyV-positive MCC samples and four melanoma samples were analyzed. In MCPyV-positive MCC RB1, TP53, FBXW7, CTNNB1, and HNF1A pathogenic variants were identified, while in virus-negative cases only benign variants were found. In MCPyV-positive melanoma cases, besides BRAF mutations the following genes were also affected: PIK3CA, STK11, CDKN2A, SMAD4, and APC. In contrast to studies found in the literature, a higher tumor burden was detected in virus-associated MCC compared to MCPyV-negative cases. No association was identified between virus infection and tumor burden in melanoma samples. We concluded that analyzing the key morphologic and immunohistological features of MCC is critical to avoid confusion with other cutaneous malignancies. Molecular genetic investigations such as next-generation sequencing (NGS) enable molecular stratification, which may have future clinical impact.


Introduction
Merkel cell carcinoma (MCC) is a rare, high-grade, aggressive primary cutaneous neuroendocrine malignancy causing rapidly enlarging lesions on sun-exposed areas of older Caucasian individuals [1,2]. The incidence of MCC has increased over the past 30 years due to the aging population, the additional reporting, and improvements in diagnostic techniques [3]. The incidence ranges from 0.1 to 1.6 cases per 100,000 people per year [4,5]. The rising incidence with its often rapidly aggressive course underscores a critical need to analyze the histopathologic and the immunohistochemical (IHC) features of the disease. MCC was regarded as a dermal malignancy that often involves the subcutis. The tumor cells of MCC show a characteristic neuroendocrine cytomorphology with scant cytoplasm and uniform round to oval nuclei., show in the majority of cases, positivity for cytokeratin 20 (CK20), synaptophysin, chromogranin A (CHGA), and neurofilament (NF), but often negative for thyroid transcription factor 1 (TTF-1) [6].
A relatively newly identified human virus, the Merkel cell polyomavirus (MCPyV), has been implicated in the pathogenesis of MCCs [7]. The virus appears to be causative in most cases and it was applied as a diagnostic and prognostic marker [6]. MCPyV is present in around 60-80% of MCC and only in 11% of control tissues within their majority lower copy numbers [8]. Some studies published molecular aberrations that can play a role in the virus-associated MCC malignancy such as KIT, PIK3CA, and genes in the Hedgehog signal transduction pathway [9][10][11]. In other forms of MCC in which MCPyV cannot be detected, supposing the ultraviolet light exposure represents key drivers in the carcinogenesis [6]. MCPyV-negative MCCs are described by a higher mutation burden compared to MCPyV-positive MCC, and these mutations can be considered UV-signature variations. The most relevant alterations in virus-negative cases affect NOTCH, RB1, and TP53 genes [12,13]. In geographic regions with less UV exposure, MCC is more common with MCPyV positivity, whereas in areas with relatively high UV exposure, MCPyVnegative MCCs predominate [14].
Due to limited data available for molecular study in MCCs, particularly with nextgeneration sequencing (NGS) methods, the aim of our study was (1) to detect the MCPyV status of nine MCC patients, diagnosed from three separate institutes, (2) to compare the molecular differences between virus-positive and -negative subgroups, (3) to identify MCPyV frequency in other randomly selected cutaneous melanoma, and (4) to elucidate genetic differences between virus-associated and -negative cases and MCPyV-associated melanoma samples as well. For this purpose, histology examination, diagnostic IHC including antibody against large T antigen of MCPyV, MCPyV-specific PCR, and NGS with solid tumor gene panel analysis were performed using MCC and MCPyV-associated melanoma samples. Additionally, reverse-hybridization StripAssay was carried out on melanoma samples detecting BRAF mutation.

Patients' Samples
Altogether, nine formaldehyde-fixed paraffin-embedded tissue (FFPE) MCCs and 60 control melanoma samples were tested. All protocols were approved by the authors' respective Institutional Review Board for human subjects (IRB reference number: 60355/2016/EKU).

DNA Isolation
Genomic DNA was extracted from FFPE tissues using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). The isolations were carried out according to the manufacturer's standard protocol and the DNA was eluted in 50 µL elution buffer. The DNA concentration was measured in the Qubit dsDNA HS Assay Kit using a Qubit 4.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA).

Merkel Cell Polyomavirus Molecular Detection
For virus T antigen and/or viral capsid DNA sequences in the nine MCC and 60 histopathological similar cutaneous malignant melanoma samples, LT1 (large T1), LT3 (large T3), and VP1 (viral capsid protein 1) primer pairs were used. The DNA amplification was carried out according to the earlier study [15]. To demonstrate that the quality and quantity of the DNA samples were acceptable, the human β-globin gene was also amplified.

BRAF StripAssay
Reverse hybridization was carried out using BRAF 600/601 StripAssay according to the manufacturer's protocol (ViennaLab Diagnostics, Vienna, Austria). The assay covers nine clinically relevant mutations in the BRAF gene and is certified for human in vitro diagnostics (IVD). For interpretation, hybridization strips were aligned using the standardized layout supplied with the reagents, and positive bands were identified.

Next-Generation Sequencing
The amount of amplifiable DNA (ng) was calculated according to the Archer PreSeq Raw sequence data were analyzed with Archer analysis software (version 6.2.; Archer DX, Boulder, CO, USA) for the presence of single-nucleotide variants (SNVs) as well as insertions and deletions (indels). For the alignment, the human reference genome GRCh37 (equivalent UCSC version hg19) was built. Molecular barcode (MBC) adapters were used to count unique molecules and characterize sequencer noise, revealing mutations below standard NGS-based detection thresholds. The sequence quality for each sample was assessed and the cutoff was set to 5% variant allele frequency (VAF). VAF is the percentage of sequence reads observed matching a specific DNA variant divided by the overall coverage at that locus. Large insertion/deletion (>50 bp) and complex structural changes could not be captured by the method. The results were described using the latest version of the Human Genome Variation Society nomenclature for either the nucleotide or protein level. Individual gene variants were cross-checked in the COSMIC (Catalogue of Somatic Mutations in Cancer) and ClinVar databases for clinical relevance. We used gnomAD v.2.1.1 population database to compare the significance of each gene alteration, which is included in our Archer NGS analysis system.

Clinical Presentations
The clinical features of the MCC patients are shown in Table 1. The average age of the nine MCC patients was 77.8 (range: 63-89). The gender distribution was six male and three female. The localization of the tumor was more frequent in the sun-exposed regions than in other areas (e.g., gluteus). The average tumor size was 2.4 cm (range: 1-6 cm) and the mean tumor depth was 1.7 cm (range: 0.4-6.7 cm). The oncological treatments included excision, chemotherapy, and/or radiotherapy.

Histological Features Including Immunohistochemistry
The tumor cells of MCC show uniform, small, blue, round cells with high nuclear/ cytoplasmic ratio with granular chromatin and numerous nucleoli ( Figure 1A). Histopathological characteristics of the patient samples are presented in Table 2. Cell proliferation ratio was determined above 50%. CK20 (usually dot-like) ( Figure 1B), synaptophysin, and CHGA positivity were found, while no TTF-1 expression was detected in all cases. Lymphovascular invasion was present in two cases. The p53 IHC was positive in six samples ( Figure 1C). Vimentin and S100 protein positivity was detected in all melanoma samples.

Merkel Cell Polyomavirus Detection
MCPyV was detected with two methods. Five MCC samples were positive for the MCPyV antibody ( Figure 1D). PCR amplification of the viral large T protein LT3 (amplification product size 308 bp) was positive in six cases. Additionally, in one case (case 3), only the viral capsid protein VP1 (351 bp) was positive, as well. The LT1 (439 bp) PCR was negative in all cases. Altogether, seven MCPyV-positive MCC samples were detected using PCR (7/9, 77.8% vs. 5/9, 55.6% MCPyV IHC). IHC staining had limited sensitivity in case 8 because the viral LT genome sequence was detected by PCR. In case 3 the virus LT antigene IHC did not detect positivity, aiming at viral capsid protein being present when using PCR amplification.
Two types of agarose electrophoresis lane intensity were detected and high virus copy number samples were characterized with a robust band (samples 1, 2, 6, and 9), while samples 4 and 8 showed weak stripe and, consequently, low virus copy number. The MCPyV LT3 and VP1 PCR results are presented in Figure 2.

Merkel Cell Polyomavirus Amplification in Melanoma Samples
PCR amplification of LT1, LT3, and VP1 virus genes was carried out on 60 melanoma DNA samples. LT3 was present in four melanoma samples with a low copy number (6.7%), opposite to LT1 and VP1, where no PCR product was amplified.

NGS-Based Mutation Profiling
The 67 genes' solid tumor panel analysis identified a series of alterations in the affected genes. The NGS analysis results are summarized in Table 3.

NGS-Based Mutation Profiling
The 67 genes' solid tumor panel analysis identified a series of alterations in the affected genes. The NGS analysis results are summarized in Table 3.
Variant tumor burden (VTB) was defined with the number of gene variants above 2% VAF (Figure 3). The largest VTB was in cases 4 and 6, while the smallest was in cases 7 and 8. Besides the MCPyV-associated case 7, in the other virus-negative case, case 5, low VTB was determined. In virus-associated melanoma cases, the average tumor burden was detected.

Discussion
In our study, we compared two methods to detect MCPyV in MCC and melanoma samples. The MCPyV monoclonal antibody, targeting the large T-antigen (clone CM2B4), effectiveness ranged from 39% to 90% [6]. The sensitivity of IHC is also related to the preanalytic varying parameters such as tissue fixation as well as viral copy numbers in the tumor cells [16]. Studies comparing PCR to IHC detection of MCPyV usually demonstrate proper concordance. Based on our results, together with other studies, we found PCR to be more sensitive than IHC [17][18][19][20]. Furthermore, we used only PCR to identify MCPyVassociated melanoma samples.
MCPyV-positive tumors have favored overall survival compared to MCPyV-negative MCCs, which also showed aggressive molecular mechanisms [8]. After viral integration, MCPyV stimulates host cell gene mutations and consequently dysregulated cell proliferation. Advances in NGS techniques have enabled the identification of these mutational landscapes that clarify the significant differences between virus-positive and virusnegative cases. As in melanoma and other skin cancers that are associated with UV radiation, these mutations (namely C-to-T pyrimidine dimers) suggest that, at least in MCVnegative tumors, the cell is unable to effectively repair UV-induced damage and subsequently accumulates further mutations.
Studies of molecular profiling on MCCs are scant and even controversial. Increased expression of the KIT receptor tyrosine kinase, in primary MCCs, demonstrated a shorter survival compared to patients with reduced levels of KIT in the malignant cells, although activating mutations in KIT have not been identified in MCCs [9,21]. Other genetic aberrations were also described, such as activating PIK3CA mutations [10], expression of the Hedgehog signaling cascade [11], aberrations of growth factors, and cell proliferation regulation [22,23]. The large T antigen regulates the life cycle of the virus and the cell cycle of the host cells. The latter is caused by interaction with the tumor suppressor gene RB1 and TP53 genes. The small T antigen is capable of stimulating cell proliferation through activation of several signal transduction pathways [24]. Additionally, in MCC tumor cells, small T antigen binds SCF(Fbw7) protein, thereby stabilizing large T antigen, which is a substrate for this E3 ubiquitin ligase [25]. In our study, we identified genetic variants that affected the abovementioned RB1, TP53, and FBXW7 genes only in virus-positive MCC

Discussion
In our study, we compared two methods to detect MCPyV in MCC and melanoma samples. The MCPyV monoclonal antibody, targeting the large T-antigen (clone CM2B4), effectiveness ranged from 39% to 90% [6]. The sensitivity of IHC is also related to the preanalytic varying parameters such as tissue fixation as well as viral copy numbers in the tumor cells [16]. Studies comparing PCR to IHC detection of MCPyV usually demonstrate proper concordance. Based on our results, together with other studies, we found PCR to be more sensitive than IHC [17][18][19][20]. Furthermore, we used only PCR to identify MCPyVassociated melanoma samples.
MCPyV-positive tumors have favored overall survival compared to MCPyV-negative MCCs, which also showed aggressive molecular mechanisms [8]. After viral integration, MCPyV stimulates host cell gene mutations and consequently dysregulated cell proliferation. Advances in NGS techniques have enabled the identification of these mutational landscapes that clarify the significant differences between virus-positive and virus-negative cases. As in melanoma and other skin cancers that are associated with UV radiation, these mutations (namely C-to-T pyrimidine dimers) suggest that, at least in MCV-negative tumors, the cell is unable to effectively repair UV-induced damage and subsequently accumulates further mutations.
Studies of molecular profiling on MCCs are scant and even controversial. Increased expression of the KIT receptor tyrosine kinase, in primary MCCs, demonstrated a shorter survival compared to patients with reduced levels of KIT in the malignant cells, although activating mutations in KIT have not been identified in MCCs [9,21]. Other genetic aberrations were also described, such as activating PIK3CA mutations [10], expression of the Hedgehog signaling cascade [11], aberrations of growth factors, and cell proliferation regulation [22,23]. The large T antigen regulates the life cycle of the virus and the cell cycle of the host cells. The latter is caused by interaction with the tumor suppressor gene RB1 and TP53 genes. The small T antigen is capable of stimulating cell proliferation through activation of several signal transduction pathways [24]. Additionally, in MCC tumor cells, small T antigen binds SCF(Fbw7) protein, thereby stabilizing large T antigen, which is a substrate for this E3 ubiquitin ligase [25]. In our study, we identified genetic variants that affected the abovementioned RB1, TP53, and FBXW7 genes only in virus-positive MCC samples. Additional pathogenic variants were identified in CTNNB1 and HNF1A genes. In one study, the HNF1A gene mutation involvement of the virus-positive MCC pathogenesis was identified as well and, similar to our investigation, FGFR1, ABL1, and JAK3 variants were also demonstrated [26]. Aberrations involving EGFR, FOXL2, and STK11 genes in MCPyV-positive MCCs had not been identified earlier.
In the literature, significant differences in the mutational burden that exists between MCPyV-positive and -negative cases were characterized, an order of magnitude higher for somatic single nucleotide variants per exome in virus-negative tumors compared to viruspositive cases [13]. Another study compared the molecular genetics of MCPyV-positive and -negative tumors and identified that MCPyV-positive MCCs harbor relatively few mutations and do not display a definitive UV-signature, verifying the oncogenic role of T-antigens as dominant drivers for these tumors [27]. One large genomics study in MCC characterized the molecular landscape of immune checkpoint inhibitor response. MCPyV sequences were detected only in low tumor burden cases. The response rate was 50% in high VTB and 41% in low VTB and MCPyV-positive tumors [28]. On the contrary, due to the limited number of MCPyV-negative cases tested, VTB was higher in MCPyV-positive samples in our study.
We found four BRAF-mutant MCPyV-associated melanoma cases in our study subjects, which was discordant with earlier findings, where no virus-positive melanoma cases were found and, thus, the mutation status was not analyzed [29]. In small-cell lung cancer, cancer sharing differentiation markers with MCC and MCPyV was detected in 39% [30]. Additionally, MCPyV was present in Kaposi sarcoma [31], non-melanoma skin cancers [32,33], and other cancers [34].
Several studies dealt with genetic aberrations in melanoma using NGS [35][36][37], and, according to them, the most affected genes are BRAF, NRAS, and KIT. In our study, in the three virus-associated cases, besides BRAF mutation, PIK3CA, CDKN2A, and APC pathogenic variants were identified. Considering the genetic similarity between virusassociated MCC and MCPyV-positive melanoma samples, no significant differentiation was observed in the tumor burden, but the affected genes were different, owing to the randomly confused signal transduction pathways by the virus.
In conclusion, MCC is an aggressive cutaneous tumor that, although rare, is increasing in incidence. We used different methodologies to analyze the samples that could provide important information for potential therapeutic options and diagnostic approaches in the future. In MCPyV-positive MCC, RB1, TP53, FBXW7, CTNNB1, and HNF1A pathogenic variants were identified, while in virus-negative cases only benign variants were found. In contrast to studies found in the literature, a higher tumor burden was detected in virus-associated MCC compared to MCPyV-negative cases. No association was identified between virus infection and tumor burden in melanoma samples. We concluded that analyzing the key morphologic and immunohistological features of MCC is critical to avoid confusion with other cutaneous malignancies. Molecular genetic investigations such as NGS enable molecular stratification, which may have future clinical impact.