Le mélanome est un cancer de la peau très agressif. Microphthalmia-associated transcription factor (MITF) est un facteur de transcription clé contrôlant le développement de la lignée mélanocytaire, ainsi que la prolifération et l’invasion des cellules de mélanome. Pour mieux comprendre les fonctions de MITF, nous avons identifié ses cofacteurs impliques dans la régulation de la transcription. Nous avons montré que le complexe de remodelage de la chromatine NURF interagit avec MITF. Ma thèse a consisté à élucider le rôle de NURF dans le mélanome et les mélanocytes. La perte de BPTF, la principale sous-unité de ce complexe, induit un arrêt de la prolifération et une entrée en senescence des cellules de mélanome. Nous avons montré que BPTF et MITF coopèrent pour réguler l’expression de gènes impliqués dans la prolifération and invasion suggérant que BPTF et un cofacteur de MITF. De façon inattendue, l’inactivation de BPTF spécifiquement dans les mélanocytes entraine la perte progressive et totale de la pigmentation du pelage en raison de l’incapacité des cellules souches mélanocytaire à produire une descendance fonctionnelle. C’est la première fois que l’interaction fonctionnelle entre NURF et MITF est démontrée in vitro, complétée par des observations phénotypique uniques in vivo, contribuant à la compréhension de la biologie des mélanocytes et du mélanome. / Melanoma is a highly aggressive form of skin cancer. Microphthalmia-associated transcription factor (MITF) isa key regulator of development of the melanocyte lineage and proliferation and invasion of melanoma cells.To further elucidate the functions of MITF, we identified factors co-regulating transcription with MITF. We identified the NURF chromatin-remodeling complex as MITF interactor. My thesis aims to elucidate the role of NURF in melanoma and melanocytes. Loss of BPTF, the principal subunit of the complex, led to arrest of proliferation and entry into senescence of melanoma lines. We showed BPTF and MITF co-regulate genes involved in proliferation and invasion suggesting that BPTF acts as cofactor for MITF. Remarkably, the mouse model of melanocyte-specific BPTF ablation led to progressive and complete loss of coat pigmentation due to the inability of the melanocyte stem cells to produce functional progeny. This is the first report of NURF-MITF functional interaction in vitro, complemented with a unique in vivo phenotype, both adding to a general understanding of melanocyte and melanoma biology.

Malignant melanoma (MM) is the deadliest skin cancer with an ever-increasing incidence. New treatments have improved the prognosis for patients with advanced MM. Still, most patients do not respond, and the side effects can be severe, underlining the need for better therapies. The overall aim of this thesis was to evaluate new means to improve the treatment for patients with advanced MM. Immunostimulatory gene therapy (AdCD40L) was evaluated in a clinical study and BRAF-inhibitory treatment in rare cases of BRAF-mutated MM. Due to its immunogenicity, MM is an attractive target for immunostimulatory gene therapy. AdCD40L is an adenovirus carrying the human gene for CD40 ligand, which in different ways can stimulate the immune system to combat cancer. We conducted a Phase I/IIa study with AdCD40L in patients with metastatic MM having received established treatments. In cohort 1 (n=6), four weekly, intratumoural AdCD40L injections were given. In cohort 2 (n=9), low dose cyclophosphamide was added to increase the immune response. Since irradiation may act synergistically with immunotherapy, patients in cohort 3 (n=9) also received a single fraction of radiotherapy (8 Gy). This fraction was given towards the lesion selected for injections. The primary objectives were to assess the feasibility and safety of AdCD40L-treatment and secondarily its anti-tumour effects. Patients were thoroughly assessed for toxicity. The anti-tumour response was evaluated by imaging techniques (FDG-PET/CT, DW-MRI scans), tumour biopsies and blood tests. Plasma protein markers were measured with a multiplex platform. Another objective was to evaluate the potential of DW-MRI and FDG-PET/CT for prediction of AdCD40L treatment response, in terms of overall survival (OS). AdCD40L was well tolerated with mild transient reactions. Local and distant responses in PET/CT scans along with a significantly better 6-month survival in the cohorts that received cyclophosphamide conditioning were observed. Effector lymphocyte responses were elicited. All patients had an increased T effector/T regulatory-cell ratio and death receptors were significantly up-regulated post therapy. Inflammatory cytokines and other plasma proteins were altered in favourable ways by the AdCD40L treatment. The analyses support that the functional DWI parameters may be better early predictors of OS than the established metabolic and morphologic criteria of FDG-PET/CT and CT/MRI, respectively. In conclusion, the stimulation of the CD40 pathway to initiate anti-tumour immunity is a promising treatment alternative for MM patients. However, further studies with developed treatment schemes are warranted. In the first report ever on treatment of a pregnant patient with a BRAF-inhibitor, the therapy was initiated in the second trimester. The treatment with vemurafenib enabled prolonged gestation, hence reducing the risk of immaturity-related complications. Further, we report the first case worldwide of a patient with metastatic conjunctival melanoma who benefitted from treatment with vemurafenib. Additional studies are needed to assess the efficacy of BRAF -inhibitors in the different subtypes of ocular melanoma.

Abstract Skin is the largest organ in our body protecting us from ultraviolet radiation and xenobiotics. UV-radiation is a common cause of squamocellular carcinoma and melanoma of the skin that cause morbidity and mortality world wide. Reactive oxygen species are constantly formed by, for example, cellular respiration and UV-radiation, and they can readily react with virtually any macromolecule within cell structures causing damage to DNA, proteins and lipids. Oxidative stress (OS) is a homeostatic process that is dysregulated in cancer cells to their benefit. Nuclear factor erythroid-2-related factor 2 (Nrf2) is the main regulator of antioxidant response and it has been shown to be upregulated in various cancers enabling their survival and growth. By using immunohistochemistry we studied the change and prognostic significance of OS markers in melanoma from paraffin embedded patient samples. Nrf2 expression is increased in melanoma, associating with deeper invasion and a worse melanoma-specific outcome. In addition, epithelial-to-mesenchymal transition markers Slug, Twist and Zeb1 showed altered expression levels in relation to invasion and metastasis associating also with Nrf2. With the help of target inhibition molecules Vemurafenib and MEK-inhibitor CI-1040, In vitro study showed that BRAF- and NRAS-mutations might activate Nrf2. Furthermore, Nrf2-regulated antioxidant enzyme peroxiredoxin I showed decreased expression in malignant melanomas and metastases compared to benign naevi. Intriguing findings were made from the surrounding structures of melanomas e.g. loss of expression of an oxidative lesion marker 8-hydroxy-2’-deoxyguanosine in adjacent endothelial cells associated with worse melanoma-specific survival. Changes in the expression of adhesion molecules claudins 1-5 and 7 were studied in the progression of cutaneous squamous cell carcinomas and preneoplastic lesions. Change in claudin composition can alter epidermal permeability and cell polarity. Efficiency of oncological treatment modalities is frequently based on oxidative stress damage. Nrf2-inhibition could offer the means to increase the sensitivity of cancerous tissue to oxidative insults and hinder proliferative and survival signalling. Later research should focus on the relation of Nrf2 with other signalling and observations made from the tumour microenvironment. / Tiivistelmä Iho on elimistön suurin elin, ja se suojaa meitä auringon ultravioletti (UV)-säteilyltä ja muilta ulkoisilta tekijöiltä. UV-säteily on yhteinen etiologinen tekijä ihon levyepiteelikarsinoomalle ja melanoomalle, jotka aiheuttavat maailmanlaatuisesti paljon sairastavuutta ja kuolleisuutta. Reaktiivisia happiradikaaleja muodostuu esimerkiksi soluhengityksestä ja UV-säteilystä, ja ne voivat reagoida minkä tahansa makromolekyylin kanssa aiheuttaen vaurioita solun perimäainekseen, proteiineihin ja lipidirakenteisiin. Oksidatiivisen stressin (OS) säätely on tärkeä homeostaattinen prosessi, joka vinoutuu syöpäsolujen hyödyksi. Nuclear factor erythroid-2-related factor 2 (Nrf2) on antioksidanttivasteen pääsäätelytekijä, ja sen ilmentyminen on lisääntynyt useissa syövissä lisäten syöpäsolun selviytymistä ja kasvua. Tutkimme potilasaineiston ja immunohistokemian avulla OS:n merkkiaineiden muutoksia melanoomassa ja niiden merkitsevyyttä taudin ennusteelle. Nrf2:n ilmentyminen on lisääntynyt melanoomassa liittyen syvempään invaasioon ja huonompaan tautispesifiseen ennusteeseen. Lisäksi epiteliaali-mesenkymaalitransition merkkiaineiden, Slug, Twist ja Zeb1 ekspression muutoksia havaittiin syvyyskasvun ja metastasoinnin yhteydessä assosioituen myös Nrf2 ilmentymiseen. In vitro- tutkimus osoitti spesifisten inhibiittoreiden avulla, että BRAF- ja NRAS-mutaatiot saattavat aktivoida Nrf2 melanoomassa. Myös Nrf2:n säätelemän entsyymin peroksiredoksiini I:n ilmentyminen on vähentynyt melanoomassa ja metastaaseissa verrattuna hyvänlaatuisiin pigmenttiluomiin. Merkittäviä muutoksia havaittiin myös melanoomaa ympäröivistä rakenteista, esimerkiksi OS:n vauriomarkkerin 8-hydroksi-2’-deoksiguanosiinin vähentynyt ilmentyminen endoteelisoluissa liittyi huonompaan tautispesifiseen ennusteeseen. Lisäksi tutkimme soluväliliitosproteiinien klaudiinien 1–5 sekä 7 ilmentymistä levyepiteelikarsinoomissa ja niiden esiasteissa. Klaudiinien muutokset voivat vaikuttaa ihon permeabiliteettiin ja solujen polarisaatioon. Onkologisten hoitomuotojen teho perustuu usein happiradikaalien aiheuttamiin vaurioihin. Nrf2-inhibitio voisi tarjota keinon lisätä syöpäkudoksen herkkyyttä näille vaurioille sekä estää syöpäsolun selviytymissignalointia. Tulevat tutkimukset tulisivat keskittyä Nrf2 signaloinnin ja muun solusignaloinnin välisiin suhteisiin sekä havaintoihin kasvaimen mikroympäristön muutoksista.

L'infiltration des mélanomes par les macrophages (TAM) est souvent corrélée à un mauvais pronostic. Cependant, les mécanismes qui régulent le recrutement et la fonction de ces cellules restent encore mal compris. Des études récentes ont montré un rôle majeur de la sphingosine kinase 1 (SK1) tumorale, l'enzyme qui produit la sphingosine-1-phosphate (S1P), dans le remodelage du stroma associé à la tumeur. Le but de ce projet a été d'étudier le rôle de la SK1 tumorale sur le microenvironnement inflammatoire, et en particulier les macrophages, lors du développement des tumeurs mélaniques. In vitro, nous avons montré que l'inhibition de SK1 dans les cellules de mélanome : 1) bloque la migration des macrophages. A l'inverse, la surexpression de cette kinase dans les cellules tumorales stimule la migration des cellules inflammatoires. Cet effet est dépendant de la S1P et de sa fixation sur les récepteurs S1PR présents à la surface des macrophages ; 2) réduit la sécrétion de TGF-ß et 3) stimule la différenciation des macrophages vers un phénotype M1 antitumoral. Ce phénomène n'est pas dépendant de la S1P, ni des S1PRs, mais de la sécrétion de TGF-ß par les cellules tumorales. En effet, la différenciation macrophagique peut-être réversée par l'addition de TGF-ß recombinant dans le milieu de sécrétion des cellules tumorales. In vivo, nos résultats montrent que la greffe orthotopique, i.e. intradermique, de cellules de mélanome murin invalidées pour la SK1 à des souris syngéniques C57BL/6 est associée à une réduction de la croissance tumorale, comparée à des souris ayant reçu des cellules de mélanome contrôles. De plus, l'invalidation de la SK1 tumorale conduit à une augmentation significative de l'expression de cytokines anti-tumorales ainsi qu'à une polarisation Th1 au sein de la tumeur. Ce phénomène s'accompagne d'une réduction du pourcentage de macrophages M2 CD206+MHCIIlow, et à l'inverse, d'une augmentation du pourcentage de macrophages M1 CD206-MHCIIhigh ainsi que de lymphocytes T CD4+ et CD8+ infiltrés dans la tumeur. Ces résultats suggèrent un rôle clé de la SK1 tumorale dans le recrutement et la polarisation des macrophages dans les mélanomes. Ainsi, l'axe SK1/ TGF-ß pourrait constituer une cible thérapeutique prometteuse dans le contrôle de la croissance de cette tumeur. / Melanoma infiltration by macrophages (TAM) is often correlated with poor prognosis. However, the mechanisms that regulate the recruitment and function of these cells remain poorly understood. Recent studies have shown a major role of tumor sphingosine kinase 1 (SK1), the enzyme that produces sphingosine-1-phosphate (S1P), in tumor stroma remodeling. The aim of this project was to investigate the role of tumor SK1 on the inflammatory microenvironment, particularly macrophages, during the development of melanoma. In vitro, we showed that the inhibition of SK1 in melanoma cells: 1) blocks macrophage migration. Conversely, overexpression of this kinase in tumor cells stimulates the migration of inflammatory cells. This effect is dependent on S1P binding to its receptors (S1PR) on the macrophage surface; 2) reduces the secretion of TGF-ß and 3) stimulates macrophage differentiation towards an antitumor M1 phenotype. The latter phenomenon does not depend on S1P nor S1PRs, but on the secretion of TGF-ß by tumor cells. Indeed, macrophage differentiation can be reversed by adding recombinant TGF-ß in the tumor cell-conditioned medium. In vivo, our results showed that orthotopic injection, i.e. intradermal, of murine melanoma cells invalidated for SK1 in C57BL / 6 syngenic mice was associated with a reduction in tumor growth compared to mice having received control melanoma cells. Furthermore, the invalidation of tumor SK1 leads to a significant increase in the expression of anti-tumor cytokines and a Th1 polarization within the tumor. This phenomenon is accompanied by a reduction in the percentage of CD206+MHCIIlow M2 macrophages, and conversely, an increase in the percentage of M1 macrophages CD206-MHCIIhigh as well as CD4+ and CD8+ cells infiltrated into the tumor. These results suggest a key role of tumor SK1 in the recruitment of macrophages and their polarization in melanoma. Thus, the axis SK1 / TGF-ß could be a promising therapeutic target in controlling the growth of this tumor.

Les facteurs de transcription inducteurs de la transition épithélio-mésenchymateuse (EMT) sont fréquemment réactivés de manière aberrante dans de nombreux cancers. Dans les carcinomes, ils favorisent dans les étapes précoces le développement tumoral en inhibant les systèmes de sauvegarde tels que la sénescence ou l'apoptose et à terme dans les étapes finales, promeuvent la formation de métastases. Dans le cadre du mélanome nous avons mis en évidence des fonctions antagonistes de ces facteurs : il était connu que ZEB2 et SNAIL2 sont impliqués dans la délamination de la crête neurale et la détermination mélanocytaire mais de manière inattendue ZEB2 et SNAIL2 sont aussi exprimés dans les mélanocytes adultes normaux et leur expression est diminuée au cours de la progression maligne, au profit des facteurs ZEB1 et TWIST1. Ce changement de profil d'expression est un facteur de mauvais pronostic pour les patients. Mes travaux ont permis de montrer que cet effet antagoniste passe par la régulation de MITF, le facteur clef du développement mélanocytaire. ZEB2 et SNAIL2 exercent une fonction oncosuppressive en activant l'expression de MITF et en promouvant la différenciation mélanocytaire. Au contraire, ZEB1 et TWIST1 jouent un rôle oncogénique en inhibant MITF et favorisent l'acquisition de propriétés de cellules souches. J'ai poursuivi l'étude de la fonction oncogénique de Twist1 in vivo dans un modèle murin de mélanome induit par l'oncogène BRAFV600. J'ai ainsi mis en évidence que l'expression conjointe de Twist1 et de BRAFV600 entraîne la formation de mélanomes agressifs, dédifférenciés et invasifs. Enfin, j'ai analysé et caractérisé l'implication de ces facteurs dans le processus de résistance aux thérapies ciblées anti-BRAF dans les mélanomes BRAFV600E. J'ai démontré que ZEB1 peut contribuer à la résistance aux inhibiteurs de BRAF. En effet, l'expression de ZEB1 est augmentée dans des cellules de mélanomes résistantes (innée ou acquise), par rapport aux cellules sensibles. Tandis que l'expression de ZEB1 favorise l'émergence de cellules résistantes, cibler ZEB1 augmente la sensibilité des cellules à l'inhibiteur de BRAF et resensibilise les cellules résistantes, mettant en évidence l'intérêt de cette combinaison d'un point de vue thérapeutique / Embryonic transcription factors inducers of Epithelial to Mesenchymal Transition (EMT-TFs) are frequently and aberrantly reactivated in many cancers. In carcinomas they promote, in early stage, tumor development by inhibiting the failsafe programs of the cell like senescence and apoptosis, and in final stages, they promote metastases formation. In melanoma we highlighted antagonist’s functions of these factors: it was known that ZEB2 and SNAIL2 were involved in delamination of neural crest cells and melanocytes determination but unexpectedly, ZEB2 and SNAIL2 are also expressed in normal adult melanocytes and their expression is decreased during malignant progression, for the benefit of ZEB1 and TWIST1 factors. This change in expression profiling is a factor of poor prognosis for patients. The results I provide show that this antagonistic effect might go through the regulation of MITF, the key factor in the development of melanocytic. ZEB2 and SNAIL2 exert an oncosuppressive function by activating the expression of MITF and promoting melanocytic differentiation. On the contrary, ZEB1 TWIST1 plays an oncogenic role in inhibiting MITF and promotes the acquisition of stem cells properties. In order to go deeper in the study of TWIST1 oncogenic functions, I used an in vivo Melanoma murine model induced by the BRAFV600 oncogene. I thus put in evidence that the combined expression of TWIST1 and BRAFV600 leads to the formation of aggressive, dedifferentiated and invasive melanoma. Finally, I analyzed and characterized the involvement of these factors in the process of resistance to targeted therapies against BRAFV600E melanomas. I demonstrated that ZEB1 may contribute to BRAF inhibitors resistance. Indeed, ZEB1 expression is increased in resistant (innate or acquired) melanomas cells, compared to sensitive cells. While ZEB1 expression promotes the emergence of resistant cells, targeting ZEB1 increases the sensitivity to BRAF inhibitor and sensitizes resistants cells, highlighting the interest of this combination from a therapeutic point of view

Melanoma is one of the most aggressive types of cancer. It originates from the transformation of melanocytes present in the epidermal/dermal junction of the human skin. It is commonly accepted that melanomagenesis is influenced by the interaction of environmental factors, genetic factors, as well as tumor-host interactions. DNA photoproducts induced by UV radiation are, in normal cells, repaired by the nucleotide excision repair (NER) pathway. The prominent role of NER in cancer resistance is well exemplified by patients with Xeroderma Pigmentosum (XP). This disease results from mutations in the components of the NER pathway, such as XPA and XPC proteins. In humans, NER pathway disruption leads to the development of skin cancers, including melanoma. Similar to humans afflicted with XP, Xpa and Xpc deficient mice show high sensibility to UV light, leading to skin cancer development, except melanoma. The Endothelin 3 (Edn3) signaling pathway is essential for proliferation, survival and migration of melanocyte precursor cells. Excessive production of Edn3 leads to the accumulation of large numbers of melanocytes in the mouse skin, where they are not normally found. In humans, Edn3 signaling pathway has also been implicated in melanoma progression and its metastatic potential. The goal of this study was the development of the first UV-induced melanoma mouse model dependent on the over-expression of Edn3 in the skin. The UV-induced melanoma mouse model reported here is distinguishable from all previous published models by two features: melanocytes are not transformed a priori and melanomagenesis arises only upon neonatal UV exposure. In this model, melanomagenesis depends on the presence of Edn3 in the skin. Disruption of the NER pathway due to the lack of Xpa or Xpc proteins was not essential for melanomagenesis; however, it enhanced melanoma penetrance and decreased melanoma latency after one single neonatal erythemal UV dose. Exposure to a second dose of UV at six weeks of age did not change time of appearance or penetrance of melanomas in this mouse model. Thus, a combination of neonatal UV exposure with excessive Edn3 in the tumor microenvironment is sufficient for melanomagenesis in mice; furthermore, NER deficiency exacerbates this process.

The cancers of the skin, melanoma and the keratinocyte cancers, basal cell andsquamous cell carcinomas (BCC and SCC), are among the most common cancersin white populations. While ultraviolet radiation (UVR) is their principal cause,links with non-UVR-related factors have also been noted. Ultimately, theinteraction of these elements results in malignancy however, understanding oftheir specific contributions remains incomplete. This thesis reports findings fromsix studies aiming to investigate gaps in current knowledge of the role of UVR andnon-UVR-related risk factors on skin cancer. The papers are groupedaccording to the aspects of skin cancer epidemiology and aetiology they address. The first two papers address the descriptive epidemiology of melanoma inEngland, a country with low ambient solar UVR. They arise from ecologicalstudies using national melanoma registration data and document rising trends inmelanoma incidence by anatomic site (Paper 1), and by region of residence andsocio-economic deprivation (Paper 2). Their findings were consistent with thesuggestion that increases in recreational UVR exposure are driving rises inmelanoma rates. These results emphasise both the need to closely monitor UVRexposure and melanoma trends and the importance of public health campaigns. The second group of three papers considers the assessment of associations ofnutritional factors with keratinocyte cancer. Two studies use data from aprospective cohort to evaluate the relationship between dietary intake (Paper 3)and blood concentrations (Paper 4) of omega-3 and omega-6 polyunsaturatedfatty acids (PUFA) in relation to BCC and SCC risk. Associations with both PUFAtypes were observed. In addition, Paper 5, a three-way correlational assessment,demonstrated that questionnaire and blood circulating levels of omega-3 PUFAwere highly correlated with measures of skin bioavailability. Collectively, thesestudies give evidence for associations of these nutrients with skin cancer and forthe utility of both intake and biomarker measures for assessing the relationships. The final paper explores the relationship between a widely cited non-UVR riskfactor, namely scars and cancers of the skin. It reports a systematic review of allpublished observational studies quantifying this association. While innumerablecase reports were found, quantitative analyses were rare. The review identifieda major gap in the literature where knowledge of scar malignancies is notevidence-based, but rather founded mainly on cumulative anecdotal reporting. Taken together, this body of published work highlights the largely unrecognisedcomplexity of the aetiology of cancers of the skin. Future research must bebroad in scope in order to advance understanding of the interaction betweenUVR and other risk factors and to provide a base for health messages aimed atreducing the burden of these malignancies.

Die vorliegende Arbeit befasst sich dem Glykosaminoglykan (GAG) Hyaluronan (HA) und dessen Einfluss auf die Entwicklung des malignen Melanoms (MM). Das in der extrazellulären Matrix (ECM) reichlich vorkommende HA wird schon seit dem späten 20. Jahrhundert genauerer im Zusammenhang mit der Tumorentwicklung untersucht. Gegensätzliche Eigenschaften von HA, die zum einen Tumore fördern und zum anderen ihnen entgegenwirken, wurden seither veröffentlicht. Allgemeiner Konsens ist, dass das HA-Molekulargewicht über die verschiedenen Effekte von HA entscheidet. Momentan ist jedoch nicht ausreichend belegt, wie HA auf das Tumorwachstum einwirkt und welche HA-Größen dies speziell begünstigen. Diese Untersuchung basiert auf einem in vivo Knockout von der Hyaluronan Synthase 2 (Has2) in der Maus, der die Produktion von hoch-molekulargewichtigen HA (HMW-HA) weitestgehend einschränkte. Das Wachstum vom experimentellem MM wurde in Abwesenheit der meisten HMW-HA untersucht. Diese Arbeit versuchte den Beweis zu erbringen, dass das lokale Wachstum und die Metastasenbildung der MM-Zellen abhängig von der vorhandenen HMW-HA in der Nähe des Tumors ist. Der Has2-knockout in der Mauslinie C57BL6 wurde verifiziert und nach einem veränderten Phänotyp der Mäuse untersucht. Obwohl nahezu dreiviertel der absoluten HA Menge durch den Knockout fehlte, zeigten die Mäuse keinen offensichtlichen Veränderungen. Erst eine Messung der Haut-Permeabilität deutete auf eine womöglich verstärkte Ausbildung der Stratum corneum hin. Eine direkte Auswirkung vom fehlendem HA auf das Tumorwachstum und der Metastasierungsrate konnte nicht ausreichend belegt werden. Das verwendete Mausmodell sowie die Wahl des experimentellen Tumors werden in dieser Arbeit kritisch diskutiert. Parallel durchgeführte in vitro Versuche mit teilweise artifiziellen 3D Matrizen, die unterschiedliche Mengen von HA enthielten, zeigten einen stärkeren Einfluss von niedermolekulare HA (LMW-HA) auf die Proliferation und Invasion von MM Zellen. Diese Beobachtungen stimmen mit dem derzeitigen Konsens überein, dass LMW-HA aktivierende Signaltransduktion auslöst und HMW-HA eher homöostatisch wirkt.:TABLE OF CONTENT Zusammenfassung Summary Table of content List of figures List of tables Abbreviations 1 Introduction 1.1 Structures of the skin 1.2 The malignant melanoma 1.3 The tumor microenvironment (TME) 1.4 Hyaluronan 1.4.1 HA Structure 1.4.2 HA anabolism 1.4.3 HA catabolism 1.4.4 HA binding proteins 1.4.5 HA cell surface receptors 1.5 Hyaluronan in malignant melanoma 1.6 Aim of the thesis 2 Methods 2.1 Murine malignant melanoma cell lines 2.2 Inducible Has2-knockout mice 2.2.1 The Cre/loxP System 2.2.2 Genetical modification for the Has2-knockout 2.2.3 4-Hydroxytamoxifen induction of knockout 2.2.4 Experimental tumor model in mouse 2.2.5 Intravenous tumor injection 2.3 Primary Has2-knockout fibroblasts for in vitro experiments 2.3.1 Isolation of primary fibroblast from mouse skin tissue 2.3.2 Induction of Has2-ko fibroblasts 2.4 Molecular analysis 2.4.1 PCR analysis of genome DNA 2.4.2 Quantification of gene expression 2.4.3 Microarray analysis 2.4.4 Quantification of Has2 protein levels 2.4.5 Quantification of HA 2.4.6 HA size analysis by agarose gel electrophoresis 2.5 Physical analysis of the skin 2.5.1 Skin elasticity measurement 2.5.2 Skin permeability measurement 2.5.3 Skin hydration and TEWL measurement 2.6 Histological analysis 2.6.1 Cryo-fixed samples 2.6.2 Immunofluorescence staining 2.6.3 FFPE samples 2.6.4 HA staining 2.6.5 Histochemical images 2.7 Physiological analysis of MM cell proliferation with BrdU-Assay 2.8 Statistical analysis 3 Materials 3.1 Mouse lines 3.2 Cell lines 3.3 Buffer and solution recipes 3.4 Chemicals 3.5 Molecular-biological reagents and enzymes 3.6 Primers 3.7 Antibodies 3.8 Kits 3.9 Devices and tools 3.10 Disposables 3.11 Software 4 Results 4.1 The Has2-knockout mouse model 4.1.1 Has2-knockout characterization 4.1.2 Incomplete Has2 deletion 4.1.3 Effects of the HA knockdown 4.1.4 Has2-knockout efficiency in other organs 4.1.5 Effects of Has2-knockout on gene expression 4.2 In vitro Has2-knockout and effect on MM cells 4.2.1 In vitro Has2- and HA-knockdown 4.2.2 Has2-ko fibroblast conditioned media decreased MM proliferation 4.2.3 MM conditioned media influenced fibroblast’s HA secretion 4.2.4 The transition towards in vivo tumor experiments 4.3 In vivo Tumor experiments 4.3.1 HA threshold, correlation, and exclusions 4.3.2 Effects of HA knockdown on primary tumor volume and weight 4.3.3 Tumor histology and HA localization 4.3.4 HA fragments in tumors, healthy-, and tumor-related-skin samples 4.3.5 Metastasis formation 5 Discussion 5.1 HA knockdown 5.2 HA knockdown phenotype 5.2.1 Skin stiffness 5.2.2 Skin water homeostasis 5.3 Paracrine interactions between MM and fibroblasts 5.4 HA thresholding 5.5 Tumor readouts 5.6 In vitro ECM models 5.7 Metastasis 5.8 Alternative tumor models with stronger stromal interaction 5.9 The presented results considering current HA-Tumor research 6 Conclusion 7 Literature Danksagung Lebenslauf Akademischer Werdegang Stipendium und Auszeichnung Publikation und Posterpräsentation Publikationen Vorträge und Posterpräsentationen Eigenständigkeitserklärung / The present work addresses the glycosaminoglycan (GAG) hyaluronan (HA) and its influence on the development of malignant melanoma (MM). HA, which is abundant in the extracellular matrix (ECM), has been studied closely in relation to tumor development since the late 20th century. Opposing properties of HA, on the one hand promoting tumors and the other hand counteracting them, have since been published. The general consensus is that HA molecular weight determines the various effects of HA. However, there is insufficient evidence on how HA affects tumor growth and which HA sizes specifically promote it. This study is based on an in vivo knockout of hyaluronan synthase 2 (Has2) in mice, which largely restricted the production of high molecular weight HA (HMW-HA). Growth from experimental MM was examined in the absence of most HMW-HA. This work sought to provide evidence that local growth and metastasis of MM cells is dependent on the presence of HMW-HA in the vicinity of the tumor. Has2 knockout in the mouse line C57BL6 was verified and examined for an altered phenotype of the mice. Although nearly three-quarters of the absolute HA amount was absent due to the knockout, the mice showed no obvious change. Only a measurement of skin permeability indicated a possible increased barrier function of the stratum corneum. A direct effect of the lack of HA on tumor growth and metastasis rate could not be sufficiently demonstrated. The applied mouse model and the choice of experimental tumor are critically discussed in this work. Parallel in vitro experiments with partially artificial 3D matrices containing different amounts of HA showed a stronger influence of low molecular weight HA (LMW-HA) on proliferation and invasion of MM cells. These observations are consistent with the emerging consensus that LMW-HA triggers activating signal transduction and HMW-HA is more homeostatic.:TABLE OF CONTENT Zusammenfassung Summary Table of content List of figures List of tables Abbreviations 1 Introduction 1.1 Structures of the skin 1.2 The malignant melanoma 1.3 The tumor microenvironment (TME) 1.4 Hyaluronan 1.4.1 HA Structure 1.4.2 HA anabolism 1.4.3 HA catabolism 1.4.4 HA binding proteins 1.4.5 HA cell surface receptors 1.5 Hyaluronan in malignant melanoma 1.6 Aim of the thesis 2 Methods 2.1 Murine malignant melanoma cell lines 2.2 Inducible Has2-knockout mice 2.2.1 The Cre/loxP System 2.2.2 Genetical modification for the Has2-knockout 2.2.3 4-Hydroxytamoxifen induction of knockout 2.2.4 Experimental tumor model in mouse 2.2.5 Intravenous tumor injection 2.3 Primary Has2-knockout fibroblasts for in vitro experiments 2.3.1 Isolation of primary fibroblast from mouse skin tissue 2.3.2 Induction of Has2-ko fibroblasts 2.4 Molecular analysis 2.4.1 PCR analysis of genome DNA 2.4.2 Quantification of gene expression 2.4.3 Microarray analysis 2.4.4 Quantification of Has2 protein levels 2.4.5 Quantification of HA 2.4.6 HA size analysis by agarose gel electrophoresis 2.5 Physical analysis of the skin 2.5.1 Skin elasticity measurement 2.5.2 Skin permeability measurement 2.5.3 Skin hydration and TEWL measurement 2.6 Histological analysis 2.6.1 Cryo-fixed samples 2.6.2 Immunofluorescence staining 2.6.3 FFPE samples 2.6.4 HA staining 2.6.5 Histochemical images 2.7 Physiological analysis of MM cell proliferation with BrdU-Assay 2.8 Statistical analysis 3 Materials 3.1 Mouse lines 3.2 Cell lines 3.3 Buffer and solution recipes 3.4 Chemicals 3.5 Molecular-biological reagents and enzymes 3.6 Primers 3.7 Antibodies 3.8 Kits 3.9 Devices and tools 3.10 Disposables 3.11 Software 4 Results 4.1 The Has2-knockout mouse model 4.1.1 Has2-knockout characterization 4.1.2 Incomplete Has2 deletion 4.1.3 Effects of the HA knockdown 4.1.4 Has2-knockout efficiency in other organs 4.1.5 Effects of Has2-knockout on gene expression 4.2 In vitro Has2-knockout and effect on MM cells 4.2.1 In vitro Has2- and HA-knockdown 4.2.2 Has2-ko fibroblast conditioned media decreased MM proliferation 4.2.3 MM conditioned media influenced fibroblast’s HA secretion 4.2.4 The transition towards in vivo tumor experiments 4.3 In vivo Tumor experiments 4.3.1 HA threshold, correlation, and exclusions 4.3.2 Effects of HA knockdown on primary tumor volume and weight 4.3.3 Tumor histology and HA localization 4.3.4 HA fragments in tumors, healthy-, and tumor-related-skin samples 4.3.5 Metastasis formation 5 Discussion 5.1 HA knockdown 5.2 HA knockdown phenotype 5.2.1 Skin stiffness 5.2.2 Skin water homeostasis 5.3 Paracrine interactions between MM and fibroblasts 5.4 HA thresholding 5.5 Tumor readouts 5.6 In vitro ECM models 5.7 Metastasis 5.8 Alternative tumor models with stronger stromal interaction 5.9 The presented results considering current HA-Tumor research 6 Conclusion 7 Literature Danksagung Lebenslauf Akademischer Werdegang Stipendium und Auszeichnung Publikation und Posterpräsentation Publikationen Vorträge und Posterpräsentationen Eigenständigkeitserklärung

Les thérapies ciblées ont amélioré la survie des patients atteints de mélanome cutané métastatique, toutefois certains patients échappent au traitement. La réponse au traitement met en jeu des paramètres dépendants du patient, de la biologie de la tumeur, et de la survenue de mécanismes de résistance largement étudiés.Le but de ce travail est d’étudier deux marqueurs mesurables en pratique clinique lors du suivi des patients, pour leur potentiel prédictif de réponse aux inhibiteurs de BRAF en monothérapie, l’un biologique et l’autre pharmacologique : la charge allélique mutée (CAM), définie par le rapport muté/total, des oncogènes BRAF et NRAS, et la concentration plasmatique de vemurafenib (CPV).Nous rapportons pour la première fois la prévalence et les mécanismes du déséquilibre allélique des oncogènes BRAF et NRAS, défini par une CAM élevée (>60%) ou faible (≤30%), par opposition à une CAM « hétérozygote » équilibrée (entre 30 et 60%). Des arguments in vitro et in vivo soutiennent l’hypothèse qu’une CAM BRAFV600E >60% puisse être associée à une meilleure réponse aux inhibiteurs de BRAF. La CAM de NRAS ne semble pas être un marqueur pronostique au moment du diagnostic de mélanome, toutefois sa valeur prédictive sur la réponse aux inhibiteurs de MEK sera intéressante à étudier.Nous montrons pour la première fois l’impact du dosage de la CPV lors du suivi des patients traités en monothérapie. Une CPV basse serait prédictive d’une progression ; ainsi, la CPV semble être un marqueur prédictif original de réponse, pour lequel la valeur seuil ainsi que les mécanismes pharmacogénétiques impliqués restent à déterminer. / Targeted therapies have improved survival in patients with metastatic cutaneous melanoma, but some patients escape treatment. Response to treatment involves parameters dependent of the patient, the tumor biology, and the occurrence of resistance mechanisms extensively investigated. The aim of this work is to study two measurable markers in clinical practice in patient monitoring, for their predictive potential of response to BRAF inhibitors monotherapy; the first one is biological and the other one is pharmacological: mutant allele burden (MAB), defined by the ratio mutant/total, of BRAF and NRAS oncogenes, and plasma vemurafenib concentration (PVC).We report for the first time the prevalence and the mechanisms of BRAF and NRAS oncogenic allelic imbalance, defined by a high (>60%) or a low (≤30%) MAB, as opposed to a balanced “heterozygous” MAB (30 to 60%). In vitro experiments and in vivo data support the hypothesis that a BRAFV600E MAB >60% may be associated with a better response to BRAF inhibitors. NRAS MAB does not seem to be a baseline prognostic marker, but its predictive value of response to MEK inhibitors will be interesting to investigate.We demonstrated for the first time the impact of CPV assay during the monitoring of patients treated with monotherapy. A low CPV would be predictive of progression; thus, CPV appears to be an original predictive marker of response, for which the threshold value and the involved pharmacogenetic mechanisms remain to be determined.

Les fibroblastes associés au cancer (CAF) jouent un rôle central dans un processus complexe d'interaction entre les tumeurs et le stroma et favorisent la croissance tumorale. Des preuves émergentes suggèrent que ces fibroblastes sont impliqués dans l'altération de la réponse immune anti-tumorale. Cependant, les mécanismes immuno-modulateurs sous-jacents dépendants de ces fibroblastes ne sont encore que très partiellement définis. Au cours de ma thèse, j’ai mis en évidence que les fibroblastes associés au mélanome diminuent la susceptibilité des cellules tumorales de mélanome à la lyse induite par les cellules Natural killer (NK) par un mécanisme dépendant de la sécrétion de métalloprotéinases matricielles (MMPs) actives. Cette sécrétion de MMPs réduit l'expression de deux ligands du récepteur activateur NKG2D, MICA/B, à la surface des cellules tumorales et diminue par conséquent l'activité cytotoxique des cellules NK dépendante de NKG2D contre les cellules tumorales de mélanome. D’autre part, grâce à une approche génomique globale, mon travail a montré que l’hypoxie au sein du stroma tumoral pourrait augmenter les capacités immuno-modulatrices des CAFs en modifiant l’expression d’un ensemble de gènes qui codent pour des protéines immunosuppressives. L’ensemble de ces résultats démontrent donc que les CAFs sont des déterminants essentiels modifiant la susceptibilité des cellules tumorales aux cellules tueuses mais qu’il existerait également un dialogue entre le microenvironnement hypoxique et les CAFs leur permettant d’augmenter leur potentiel immunosuppresseur. / Cancer-associated fibroblasts (CAF) play a central role in a complex process of interaction between tumors and stroma and promote tumor growth. Emerging evidence suggest that these fibroblasts are involved in the alteration of the anti-tumor immune response. However, the underlying immunomodulatory mechanisms dependent on these fibroblasts are still only partially defined. During my thesis, I demonstrated that melanoma-associated fibroblasts decrease the susceptibility of melanoma tumor cells to Natural killer (NK) cell lysis through a mechanism dependent on the secretion of active matrix metalloproteinases (MMPs). This secretion of MMPs reduces the expression of the two NKG2D ligands, MICA/B at the surface of the tumor cells and consequently decreases the NKG2D-dependent cytotoxic activity of NK cells against melanoma tumor cells. On the other hand, using a global genomic approach, my results suggested that hypoxia within the tumor stroma could increase the immunomodulatory capacities of CAFs by modifying the expression of a set of genes that encode for immunosuppressive proteins. Together, our results show that CAFs are essential determinants modifying the susceptibility of tumor cells to killer cells but that there is also a crosstalk between the hypoxic microenvironment and the CAFs allowing them to increase their immunosuppressive potential.