MERKEL CELL CARCINOMA, TOWARDS THE IDENTIFICATION OF NOVEL MOLECULAR TARGETS AND THE ESTABLISHMENT OF A THREE-DIMENSIONAL CELL CULTURE MODEL

Merkel cell carcinoma (MCC) is a rare cutaneous neuroendocrine carcinoma. Nevertheless, its incidence has increased in the last few years and this trend is predicted to persist. Furthermore, with a general mortality rate of 33-46% and a high rate of recurrences, it is one of the most aggressive types of skin cancer. Currently, MCC has two recognized etiologies. Merkel cell polyomavirus (MCPyV) DNA is found clonally integrated in the genome of the tumor cells in around 80% of MCCs from the Northern hemisphere. MCPyV-positive (MCPyV+) MCCs constitutively express two viral oncoproteins, the small (sT) and the large (LT) tumor antigens (TAs), which contribute to the uncontrolled proliferation of the transformed cells. The MCPyV-negative (MCPyV-) subtype, which is predominant in Australia and New Zealand, is thought to arise upon accumulation of UV-light induced somatic mutations in proto-oncogenes and tumor suppressors. This is evidenced by a higher mutational burden than that of the MCPyV+ counterpart. In any case, the two subtypes of MCC share oncogenic mechanisms, such as inactivation of the retinoblastoma protein, which may explain the phenotypic similarities. MCC is primarily a disease of the elderly, though its incidence is also higher among immunocompromised individuals. Until recently, the classic treatment modalities consisted of surgery and radiotherapy for local lesions or locally metastatic disease, and chemotherapy for metastasized MCC. Nevertheless, responses to chemotherapy were mostly short-lived and at expenses of an elevated general toxicity. Recently, the advent of immune checkpoint inhibitors, such as avelumab and pembrolizumab, has considerably improved the management of patients with metastasized disease. These immune checkpoint inhibitors induced durable responses with acceptable toxicity, highlighting the importance of tumor infiltrating lymphocytes and the role of the tumor microenvironment in the clinical outcome of patients with MCC. Hence, immunotherapies have become the standard of care for patients with metastatic MCC and their use at earlier stages or as adjuvant is currently under investigation in clinical trials. However, around half of the patients do not respond to the treatment or develop resistance after an initial response. Furthermore, a large proportion of patients, who are immunosuppressed or under immune suppressive treatment, are not eligible for immunotherapy. Similarly, elderly patients often suffer from other comorbidities or the tumors commonly appear at areas, such as the face, that are delicate for intervention with surgery or radiotherapy. Consequently, there is an unmet need for the development of alternative therapeutic approaches to treat patients with MCC that are not amenable to the current treatment possibilities. MCPyV+ MCCs constitutively express both the viral sT and a truncated form of the LT. The truncated LT is unable to initiate viral replication due to lack of the carboxyl (C)-terminal sequence containing the DNA origin binding domain and the DNA helicase activity. These viral proteins have been shown to have important roles in tumorigenesis and in the maintenance of MCPyV+ MCC cells proliferation. Here, current evidence points towards a more predominant role of MCPyV sT in tumor initiation, whereas the LT has been suggested to be involved in maintenance of the oncogenic phenotype. Indeed, experiments with short hairpin RNAs (shRNAs) targeting the viral TAs showed that MCPyV+ MCC cells required the expression of these oncoproteins to maintain cell growth. Therefore, we hypothesized that the most advanced gene-editing tool currently available, CRISPR/Cas9, could be used to target the viral TAs and, consequently, hamper the growth of virus-positive MCCs. Thus, we applied this approach in two MCPyV+ MCC cell lines, MS-1 and WAGA, by designing single-guide RNAs (sgRNAs) targeting only the LT or both, sT and LT. Our results showed that this CRISPR/Cas9-based strategy caused a significant decrease of LT protein levels, impaired cell proliferation and reduced cell cycle progression. Importantly, cells transfected with a non-targeting sgRNA and HEK293T cells, which are MCPyV-, remained unaffected. Hence, our results further validated previous reports regarding the role of MCPyV TAs in the maintenance of the growth of MCC cells. Furthermore, these findings provided strong evidence for the development of a CRISPR/Cas9-based therapeutic approach against MCPyV+ MCC. Targeted therapy presents as an attractive approach against the two subtypes of MCC. Several reports have shown that the PI3K/mTOR pathway is frequently activated in MCC, irrespective of the viral state. Consequently, MCC cell lines and xenografts were shown to be sensitive to MLN0128, a dual mTOR1/2 inhibitor. This drug is currently under clinical investigation for its use in the treatment of MCC and other solid tumors. Although the MAPK/ERK pathway has not been reported to be aberrantly activated in MCC, there is a crosstalk with the PI3K/mTOR pathway and compensatory activation has been largely proved in other malignancies. Therefore, these observations prompted us to test the effects of combining MLN0128 with trametinib, a MEK1/2 inhibitor that is currently used for the treatment of metastatic melanoma with the V600E or V600K changes in the BRAF protein. A panel of three virus-positive and three virus-negative MCC cell lines were treated with combinations of these two compounds. As determined by the Chou-Talalay method, the combined targeting was synergistic in the assayed MCC cells. Thus, a reduced dose of each inhibitor was necessary to reach the same efficacy in inhibiting cell growth than when used as single agents. Although the precise mechanism involved in this synergistic activity has not been completely elucidated, the combined targeting of the PI3K/mTOR and MAPK/ERK pathways appears to be a promising approach to treat MCC and to overcome the limitations of single drug regimens related to general toxicity and drug resistance. One of the main hurdles for the development of new treatments against MCC is the difficulty to perform prospective clinical trials, due to the rarity and rapid evolution of the disease. Furthermore, there is lack of a model able to mimic the process of MCC tumorigenesis in vitro. Although cell lines and patient derived xenografts of MCC have been established, they present certain limitations, owing the physiological differences between mice and humans. Moreover, transgenic animals expressing the viral oncoproteins have not been able to fully resemble virus-positive MCCs thus far. Here, we developed organotypic epithelial raft cultures (OERCs) of MCC by using primary human keratinocytes (PHKs) and both MCPyV+ and MCPyV- MCC cell lines grown on top of an artificial dermal equivalent. The differentiation of the rafts and the growth of MCC cell lines was confirmed through histology and immunohistochemistry. Normal PHKs grown on top of an artificial dermal equivalent at the air-liquid interface stratified and differentiated, reproducing a fully differentiated epithelium. Important differences were noted in the behaviour of the different MCC cell lines tested, such as growth pattern, invasiveness, and requirement of keratinocytes to proliferate. Virus-positive MCC cell lines generally grew exclusively on top of the differentiated epithelium, forming cell clusters. Among the virus-negative MCC cell lines, only MCC14/2 cells showed robust proliferation, even in the absence of keratinocytes and when embedded into the artificial dermis. Gene expression analysis of OERCs of co-cultures of MCC cell lines and PHKs revealed that some genes, such as those encoding distinct types of collagens, were differently expressed in rafts from MCPyV+ cells and rafts from MCC14/2 cells. Furthermore, several molecules were found highly expressed in all the analysed OERCs (e.g., matrix metalloproteinases that are involved in the degradation of the extracellular matrix, thus contributing to cell invasion and migration) and they might serve as potential targets for novel MCC treatments or as diagnostic biomarkers. Despite the need to standardize and adapt this model to high throughput analysis, OERCs represent a suitable tool for evaluating the efficacy and selectivity of new drug candidates against MCC. Furthermore, they can potentially be employed to study the molecular events driving MCC. In summary, the work presented in the current thesis provides new insights into the understanding of MCC and the bases for the investigation of new therapeutic approaches to treat patients suffering from this malignancy, either alone or in combination with the current treatment modalities. Moreover, advances in the development of OERCs will provide clinicians and researchers with new means to test novel drug candidates against MCC and to shed light into the molecular mechanisms involved in tumorigenesis.status: accepte