Les cellules tumorales circulantes (CTCs) sont des cellules tumorales provenant de la tumeur primaire et/ou des métastases que l’on retrouve dans la circulation sanguine. Les plus agressives d’entre elles peuvent envahir les organes distants pour former des métastases. Leur faible nombre parmi la multitude de cellules sanguines rend difficile leurs détections et leurs études. C'est pourquoi, le challenge actuel est de pouvoir mettre en culture ces cellules.Dans le cadre de l’éude clinique nationale COLOSPOT, notre laboratoire a pu recueillir des échantillons de patients atteints d’un cancer colorectal métastatique. Grâce à ces prélèvements sanguins, 9 lignées dérivées de CTCs ont pu être établies : CTC-MCC-41, CTC-MCC-41.4, CTC-MCC-41.5A-G.Dans ce projet de thèse, les 9 lignées cellulaires ont été caractérisées au niveau du génome, du transcriptome, du protéome, du sécrétome et fonctionnel, et comparées à des lignées cellulaires tumorales primaires et métastatiques connues, comme effectuée précédemment sur la lignée CTC-MCC-41 (Cayrefourcq et al. 2015)Cette analyse très complète a montré malgré des profils génétiques très différents, toutes les lignées CTCs ont les caractéristiques d’un phénotype intermédiaire épithélial/mésenchymal, des propriétés de cellules souches, la mutation BRAFV600E et la capacité d’éviter divers processus de lutte contre les cellules tumorales comme la résistance à l’anoïkis et l’échappement au système immunitaire. Les études fonctionnelles ont montré que les CTC-MCC pouvaient induire rapidement la formation de tubes avec des cellules endothéliales in vitro, signe d'un potentiel angiogénique.La seconde partie de ce travail de thèse a été d’étudier la transition épithélio-mésenchymateuse (EMT) in vitro. Ce phénomène est une étape clé du processus métastatique des CTCs et implique diverses transformations des cellules à divers niveaux : morphologique, protéique et transcriptomique. Trois méthodes différentes ont été testées pour induire l’EMT au sein des lignées CTCs impliquant deux différents modes d’induction et deux modes de culture. Ces changements ont pu être observés dans les lignées témoins, validant les expérimentations effectuées. Cependant, l’EMT n’a pas été clairement observée sur les lignées CTCs.En conclusion, ces analyses suggèrent que les CTC coliques cultivés à partir de biopsies liquides séquentielles, effectuée durant le traitement d’un même patient, ont des caractéristiques communes. Mais la sélection des clones, avec un phénotype distinct, résistant au traitement, a été observée. D'autres études avec ces lignées CTC-MCC sont en cours, évaluant leur capacité à induire des tumeurs résistantes à des médicaments spécifiques ou à analyser la contribution épigénétique. Ces données peuvent fournir des indications pour la découverte de nouveaux biomarqueurs permettant d'identifier les sous-populations de CTC les plus agressives et pour la mise au point de nouveaux médicaments pour inhiber les CTCs initiatrices de métastases dans le cancer du côlon. / Circulating tumor cells (CTCs) are tumor cells that have been shed from the primary tumor and/or metastases into the bloodstream. The most aggressive ones can invade distant organs to form metastases. Their low number among the multitude of blood cells makes difficult their detection and study. This is why the current challenge in this field of expertise is to be able to culture them ex vivo.In the national COLOSPOT clinical study, our team was able to collect samples of patients with metastatic colorectal cancer. From blood samples of only one patient, 9 cancer cell lines derived from CTCs could be established: CTC-MCC-41, CTC-MCC-41.4, CTC-MCC-41.5A-G.In this project, the 9 CTC-MCC lines have been characterized at the genome, transcriptome, proteome, secretome and functional levels, and compared with primary and metastatic commercial colon cancer cell lines, as previously done on the CTC-MCC-41 line (Cayrefourcq et al., Cancer Res. 2015)These analyses have shown that despite their very different genetic profiles, all CTCs have the characteristics of an epithelial/mesenchymal intermediate phenotype, stemcell like characteristics, with BRAFV600E mutation, and the ability to avoid biological processes such as the resistance to anoïkis and the escape to the immune system. Moreover, functional studies have shown that all CTC-MCC lines can rapidly induce tubes formation with endothelial cells in vitro, a sign of an angiogenic potential.The second part of this thesis work was to study the epithelial-to-mesenchymal transition (EMT) in vitro. This phenomenon is a key step in the metastatic process and involves several cell transformations at various levels: morphological, proteomic and transcriptomic. Three different methods have been tested to induce EMT within these CTC-MCC lines involving two different induction and culture modes. These changes could be observed in the control lines, validating the experiments carried out. However, EMT has not been clearly observed yet on the CTC-MCC lines.In conclusion, this longitudinal study suggest that colorectal CTCs cultured from sequential liquid biopsies, performed during treatment of the same patient, have common characteristics. However, our results strongly suggest that no clonal selection, with a distinct phenotype, resistant to treatment, has occurred. Further studies with these CTC-MCC lineages are in process, evaluating their ability to induce in vivo drug-resistant tumors or to analyze the epigenetic contribution. These data may provide guidance for the discovery of new biomarkers to identify the most aggressive CTC subpopulations and for the development of novel drugs to inhibit metastases-competent CTCs in colon cancer.

Doctor of Philosophy / Biochemistry and Molecular Biophysics Interdepartmental Program / Anna Zolkiewska / Cancer cells exploit key signaling pathways in order to survive, proliferate, and metastasize. Understanding the intricacies of the aberrant signaling in cancer may provide new insight into how to therapeutically target tumor cells. The goal of my research was to explore the role of two modulators of transmembrane signaling, the secretory pathway and cell surface proteolysis, in the aggressiveness of breast cancer cells. To study the role of the secretory pathway, I focused on the family of endoplasmic reticulum (ER) chaperones. I found that several ER chaperones were upregulated in breast cancer cells grown under anchorage-independent conditions as mammospheres versus those grown under adherent conditions. Furthermore, certain members of the protein disulfide isomerase (PDI) family were consistently upregulated in two different cell lines at both the mRNA and protein levels. Knocking down these PDIs decreased the ability of the cells to form mammospheres. I demonstrated that the requirement for PDI chaperones in mammosphere growth is likely due to an increased flux of extracellular matrix (ECM) components through the ER. Next, I examined the role of cell surface proteolysis in modulating the aggressiveness of breast cancer cells. Cell-surface metalloproteases release soluble growth factors from cells and activate the corresponding growth factor receptors. I determined that specific metalloproteases (ADAM9 or ADAM12), modulate the activation of Epidermal Growth Factor Receptor (EGFR). I demonstrated that EGFR activation enhances the CD44⁺/CD24⁻ cell surface marker profile, which is a measure of cancer cell aggressiveness. I found that the MEK/ERK pathway, which is a downstream effector of EGFR activation, modulates the CD44⁺/CD24⁻ phenotype. When DUSP4, a negative regulator of the MEK/ERK pathway, is lost, activation of EGFR by metalloproteases no longer plays a significant role in cancer cell aggressiveness. This indicates that the ligand dependent activation of the EGFR/MEK/ERK pathway is a critical step in DUSP4-positive aggressive breast cancer. Finally, I examined the importance of metalloproteases in the regulation of Programmed-death ligand 1 (PD-L1), a transmembrane protein expressed by some cancer cells that plays a major role in suppressing the immune system. I demonstrated that cell-surface metalloproteases have the ability to cleave PD-L1 and release its receptor-binding domain to the extracellular environment. Collectively, these data indicate that (a) ER chaperones support anchorage-independent cell growth, (b) metalloproteases are important in regulation of an aggressive phenotype through the EGFR/MEK/ERK pathway, and (c) metalloproteases cleave PD-L1, a key component of immunosuppression in cancer.

abstract: Following diagnosis of a glioblastoma (GBM) brain tumor, surgical resection, chemotherapy and radiation together yield a median patient survival of only 15 months. Importantly, standard treatments fail to address the dynamic regulation of the brain tumor microenvironment that actively supports tumor progression and treatment resistance. Moreover, specialized niches within the tumor microenvironment maintain a population of highly malignant glioblastoma stem-like cells (GSCs). GSCs are resistant to traditional chemotherapy and radiation therapy and are likely responsible for near universal rates of tumor recurrence and associated morbidity. Thus, disrupting microenvironmental support for GSCs could be critical to more effective GBM therapies. Three-dimensional (3D) culture models of the tumor microenvironment are powerful tools for identifying key biochemical and biophysical inputs that may support or inhibit malignant behaviors. Here, we developed synthetic poly(N-isopropylacrylamide-co-Jeffamine M-1000® acrylamide) or PNJ copolymers as a model 3D system for culturing GBM cell lines and low-passage patient-derived GSCs in vitro. These temperature responsive scaffolds reversibly transition from soluble to insoluble in aqueous solution by heating from room temperature to body temperature, thereby enabling easy encapsulation and release of cells in a 3D scaffold. We also designed this system with the capacity for presenting the cell-adhesion peptide sequence RGD for adherent culture conditions. Using this system, we identified conditions that promoted GBM proliferation, invasion, GSC phenotypes, and radiation resistance. In particular, using two separate patient-derived GSC models, we observed that PNJ scaffolds regulated self-renewal, provided protection from radiation induced cell death, and may promote stem cell plasticity in response to radiation. Furthermore, PNJ scaffolds produced de novo activation of the transcription factor HIF2α, which is critical to GSC tumorigenicity and stem plasticity. All together, these studies establish the robust utility of PNJ biomaterials as in vitro models for studying microenvironmental regulation of GSC behaviors and treatment resistance. / Dissertation/Thesis / Doctoral Dissertation Biomedical Engineering 2017

Submitted by JÚLIO HEBER SILVA (julioheber@yahoo.com.br) on 2016-12-16T18:14:58Z No. of bitstreams: 2 Tese - Constanza Thaise Xavier Silva - 2016.pdf: 3305701 bytes, checksum: 6bb860eedef867f3085559fcca9d7190 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2016-12-26T12:28:18Z (GMT) No. of bitstreams: 2 Tese - Constanza Thaise Xavier Silva - 2016.pdf: 3305701 bytes, checksum: 6bb860eedef867f3085559fcca9d7190 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-12-26T12:28:18Z (GMT). No. of bitstreams: 2 Tese - Constanza Thaise Xavier Silva - 2016.pdf: 3305701 bytes, checksum: 6bb860eedef867f3085559fcca9d7190 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-12-13 / Cutaneous melanoma is a type of skin cancer that originates in melanocytes, predominantly affecting young and middle-aged adults. Several evidence suggests that cancer stem cells exist for melanoma. The POU5F1 / OCT4, NANOG genes have been studied as cancer stem markers. OCT4 and NANOG are involved in the maintenance of pluripotency and self-renewal of undifferentiated embryonic stem cells. This study aims to evaluate the epidemiological, clinical-pathological and expression of the OCT4 and NANOG proteins in cutaneous melanoma samples. We selected 102 cases for the epidemiological, clinical and pathological study, diagnosed between the years 2004 to 2008. For survival study, patients with a followup of up to 60 months were selected, with a recorded death. Of the 102 cases evaluated, 62.7% were female and 37.3% male; With mean age of 57.2 years and 63.1 years respectively (p= 0.0026). The most prevalent age at diagnosis of melanoma was between 51 and 70 years (44.1% p= 0.023). In this study, there was a predominance of lesions located in the trunk (32.3%). Histopathological examination of the type of tumor growth showed a predominance of the superficial extensive type in 52.9% of the cases. According to the Breslow index, lesions with ≤1.0 mm predominated in 39.2% of the individuals, followed by lesions> 4.0 mm in 23.5% of the cases. According to the Clark level 29.4% of the cases were classified in level IV; Followed by 25.5% cases with level V; Clark II in 23.5%; Clark III in 20.6%; And Clark I in only 1 case (1.0%). There were metastases in 47% of the cases and the main localization sites were: lymph nodes, brain, skin and lung. Regarding the clinical evolution of the patients, there were 26 cases of deaths due to melanoma (25.5%). The survival curve calculated at the 60-month follow-up was 73.0%. Univariate analysis revealed significant associations between nuclear overexpression of OCT4 and the following variables: Breslow index with thickness > 2.1 mm (p = 0.021, OR: 2.64, 95% CI: 1.15-6.05 ); Levels of Clark, IV and V (p = 0.001, OR: 4.11, 95% CI: 1.79-9.46); ulceration present (p≤0.0001; OR: 459.0; 95% CI: 51.67-4077.27); Presence of metastases (p <0.0001, OR: 40.25, 95% CI: 12.90- 125.62), and death from cutaneous melanoma (p <0.0001). The significant associations between cytoplasmic hyperexpression of OCT4 and the following variables were: present ulceration (p = 0.015, OR: 2.73, 95% CI: 1.21-6.16); Presence of metastases (p = 0.004, OR: 3.34, 95% CI: 1.47 - 7.62) and death from cutaneous melanoma (p = 0.039, OR 2.67, 95% CI 1.05 - 6,77). And the significant associations between the cytoplasmic hyperexpression of NANOG and the following variables were: present ulceration (p≤ 0.0001); Presence of metastases (p ≤ 0.0001) and death due to cutaneous melanoma (p = 0.030). Our study demonstrated a strong association of the OCT4 / O melanoma cutâneo é um tipo de câncer de pele que tem origem nos melanócitos, afetando predominantemente adultos jovens e de meia-idade. Diversas evidências sugerem a existência células-tronco tumorais para o melanoma. Os genes POU5F1/OCT4, NANOG vem sendo estudados como marcadores células-tronco tumorais. O OCT4 e o NANOG estão envolvidos na manutenção da pluripotência e autorrenovação das células-tronco embrionárias indiferenciadas. Este estudo tem por objetivo avaliar os fatores epidemiológicos, clínico-patológicos e expressão da proteína OCT4 e NANOG em amostras de melanoma cutâneo. Foram selecionados 102 casos para o estudo epidemiológico, clínico e patológico, diagnosticados entre os anos de 2004 a 2008. Para estudo de sobrevida foram selecionadas pacientes com seguimento de até 60 meses, com óbito registrado. Dos 102 casos avaliados foram observados 62,7% do gênero feminino e 37,3% masculino; com média de idade de 57,2 anos e 63,1 respectivamente (p= 0,0026). A idade de maior prevalência ao diagnóstico do melanoma foi entre 51 a 70 anos (44,1% p= 0,023). Nesse estudo, houve predomínio das lesões localizadas no tronco (32,3%). Ao exame histopatológico quanto ao tipo de crescimento do tumor, houve predomínio do tipo extensivo superficial em 52,9% dos casos. Segundo o índice de Breslow predominaram as lesões com ≤1,0mm em 39,2% dos indivíduos, seguidas pelas lesões >4,0mm em 23,5% dos casos. De acordo com o nível de Clark 29,4% dos casos foram classificados no nível IV; seguidas por 25,5% casos com nível V; Clark II em 23,5%; Clark III em 20,6%; e Clark I em apenas 1 caso (1,0%). Houve metástases em 47% dos casos e os principais sítios de localização foram: linfonodos, cérebro, pele e pulmão. Em relação à evolução clínica dos pacientes ocorreram 26 casos de óbitos por melanoma (25,5%). A curva de sobrevida calculada no seguimento de 60 meses foi de 73,0%. A análise univariada revelou associações significativas entre a hiperexpressão nuclear de OCT4 e as seguintes variáveis: índice de Breslow com espessura de > 2,1 mm (p= 0,021; OR: 2,64; IC 95%: 1,15 - 6,05); níveis de Clark, IV e V (p= 0,001; OR: 4,11; IC 95%: 1,79 - 9,46); ulceração presente (p≤ 0,0001; OR: 459,0; IC 95%: 51,67 - 4077,27); presença de metástases (p≤ 0,0001; OR: 40,25; IC 95%: 12,90 - 125,62) e óbito por melanoma cutâneo (p≤ 0,0001). As associações significativas entre a hiperexpressão citoplasmática de OCT4 e as seguintes variáveis foram: ulceração presente (p= 0,015; OR: 2,73; IC 95%: 1,21 - 6,16); presença de metástases (p= 0,004; OR: 3,34; IC 95%: 1,47 - 7,62) e óbito por melanoma cutâneo (p= 0,039; OR: 2,67; IC 95%: 1,05 - 6,77). E as associações significativas entre a hiperexpressão citoplasmática de NANOG e as seguintes variáveis foram: ulceração presente (p≤ 0,0001); presença de metástases (p≤ 0,0001) e óbito por melanoma cutâneo (p= 0,030). Nosso estudo demostraram uma forte associação dos genes OCT4 e NANOG como os piores fatores prognósticos do melanoma cutâneo.

O carcinoma hepatocelular (HCC) é uma neoplasia primária com mau prognóstico e alta taxa de recorrência. Estudos recentes demostram que o HCC pode ser classificado em três subtipos segundo o perfil molecular. Destes subtipos, o HCC pouco diferenciado apresenta pior prognostico. Neste sentido, torna-se de particular interesse o estudo de compostos com efeitos diferenciadores e citotóxicos nas células destas neoplasias pouco diferenciadas. O butirato, um ácido graxo de cadeia curta produzido pela fermentação microbiana da fibra alimentar no intestino, tem demonstrado atividade anti-neoplásica e capacidade moduladora da diferenciação celular em diversos tipos celulares, incluindo linhagens de HCC humano e células progenitoras hepáticas. Assim, objetivou-se neste estudo, caracterizar o efeito do butirato de sódio (NaBu) em duas linhagens de células neoplásicas de rato: uma pouco diferenciada (GP7TB) e a outra, uma linhagem derivada de um HCC diferenciado (JM-1). A linhagem GP7TB mostrou maior resistência ao NaBu (ED50= 7,7 mM) do que as células JM-1 (ED50= 5,2 mM). A redução na viabilidade celular após 72 h de tratamento com NaBu esteve relacionada com a diminuição na proliferação celular e no caso das células GP7TB, de um aumento na apoptose. O tratamento com NaBu induziu alterações morfológicas nas duas linhagens celulares, porém apenas nas células do tipo GP7TB, essas alterações sugerem um processo de diferenciação/transdiferenciação celular. O aumento na expressão de genes envolvidos no controle da pluripotência de células tronco, assim como de alguns marcadores de células tronco, sugere que o NaBu induziu uma reprogramação profunda das células GP7TB. Por outro lado, a redução na expressão de genes relacionados com migração e plasticidade celular assim como de proliferação celular apontam que estas células diminuíram seu potencial invasivo e a capacidade de autorenovação. Embora sejam necessárias análises adicionais para confirmar o efeito observado nos perfis de expressão gênica, os resultados deste estudo sugerem que o NaBu apresenta efeito antineoplásico por meio da redução da proliferação, aumento da apoptose e modulação da expressão de genes associados com a transição epitéliomesenquimal em células com características tronco tumorais. / Hepatocellular carcinoma (HCC) is a primary neoplasia with poor prognosis and high recurrence rate. Recent evidence suggests that HCC can be classified in three different subtypes based on their molecular profile. Among these subtypes, the poorlydifferentiated HCC has the worst prognosis. Therefore, the study of compounds with pro-differentiating and cytotoxic effects on poorly-differentiated neoplastic cells represents a matter of primary concern. Butyrate which is a short-chain fatty acid produced by microbial fermentation in the intestine, has demonstrated anti-neoplastic activity and pro-differentiating potential in several cell types, including, human HCC cell lines and liver progenitor cells. In this study, we aimed to characterize the effect of sodium butyrate (NaBu) on two neoplastic cell lines derived from rats: a poorlydifferentiated cell line (GP7TB) and, a cell line derived from a well-differentiated HCC. GP7TB showed increased resistance to NaBu treatment (ED50= 7.7 mM) compared to JM-1 (ED50= 5.2 mM). The reduction in cell viability observed after 72 h of treatment was explained by a reduction in cell proliferation and, in the case of GP7TB, by increased levels of apoptosis. The NaBu treatment induced morphological alterations in both cell lines. However, only in the case of GP7TB cells, the alterations suggested a differentiation/transdifferentiation process. The up-regulation of genes involved in pluripotency and genes expressing stem cell markers indicated that NaBu triggered a deep reprogramming of GP7TB cells. Besides, a down-regulation in the expression of genes related with cell migration and plasticity suggested that these cells reduced their invasive potential and their self-renewal capacity. Additional analyses are necessary to confirm the observed effect on gene expression profiles. However, the results of this study suggest that NaBu exert anti-neoplastic effects through apoptosis, reduction of cell proliferation and downregulation of genes associated with epithelial-mesenchymal transition of cancer stem-like cells.

L'hétérogénéité tumorale observée dans le cancer du sein est à l'origine des échecs cliniques malgré un important arsenal thérapeutique. Cette hétérogénéité est expliquée par la présence, au sein de la tumeur d'une population minoritaire de cellules, les cellules souches cancéreuses (CSC). Ces CSC sont responsables des résistances aux traitements conventionnelles entrainant des rechutes et des métastases. Un meilleur décryptage des programmes régulant les propriétés cardinales de ces cellules, autorenouvellement et différenciation, est une étape cruciale pour le développement de nouvelles thérapies. L'identité des CSC est régulée par des mécanismes épigénétiques. Les travaux de cette thèse se sont intéressés à l'étude de deux mécanismes épigénétique: la méthylation de l'ADN et les microARNs. Nous avons d'abord identifié une signature de 68 régions hypométhylées dans les CSC. Cette signature présentait un enrichissement de la voie TGF-ß et avait un impact pronostic sur la survie des patientes. Nous nous sommes ensuite intéressés à la régulation des CSC par les miARNs. Nous avons identifié miR-600, un microARN bimodal, régulant la balance autorenouvellement-différenciation. miR-600 régule la voie Wnt, via SCD1. L'identification de l'axe miR-600/SCD1/Wnt ouvre une nouvelle perspective thérapeutique pour cibler les CSC. Nos travaux ont décryptés de nouveaux programmes épigénétiques régulantl'identité le destin des CSC mammaires. Ces travaux pourront être le terreau du développement de nouvelles approches thérapeutiques pour traiter le cancer du sien. / Tumor heterogeneity observed in breast cancer is the main cause of clinical failures despite a significant therapeutic arsenal. This heterogeneity is explained by the presence of a minority population of cells, the cancer stem cells (CSC). CSC are resistant to conventional therapies causing relapse and metastasis. Deciphering programs regulating the cardinal properties of these cells, self-renewal and differentiation, is a crucial step for the development of new therapies. The identity of CSC is regulated by epigenetic mechanisms. The work of this thesis focused on the study of two epigenetic mechanisms: DNA methylation and microRNAs. We first identified a signature of 68 regions hypomethylated in CSC. This signature showed an enrichment of the TGF-ß pathway and had a prognostic impact on patient survival. We then were interested in the regulation of CSC by miRNAs. We identified miR-600, a bimodal microRNAs, regulating the self-renewal-differentiation balance. MiR-600 regulates Wnt pathway via SCD1. The identification of the miR-600/SCD1/Wnt axis opens a new therapeutic perspective to target CSCs. Our work deciphered epigenetic programs, regulating breast CSC-fate and open new perspective to improve breast cancer care.

Les claudines sont des protéines des jonctions serrés (JS) dont l'expression est perturbée en cas de cancer. Dans ce travail, nous montrons d'abord que la surexpression de la claudine-2 agit comme intermédiaire de l'effet pro-tumorigène de la symplékine en coopération avec le facteur de transcription ZONAB dans le cancer colorectal (CCR) (Buchert, Papin et al 2010). Nous avons alors étudié le rôle de la claudine-2 sur les cellules souches cancéreuses (CSC), considérées comme responsables du maintien à longue durée de la tumorigénèse, en analysant les conséquences de sa sous-expression expérimentale dans les CSC coliques putatives exprimant le marqueur LGR5. Sur la lignée de CCR humain DLD1, l'inhibition de claudine-2 par shRNA ou siRNA diminue le nombre de cellules LGR5(+) et augmente l'expression du marqueur de différenciation CK20. La sous-expression de la claudine-2 dans les cellules LGR5(+) diminue leur capacité à survivre en dilution clonale en suspension dans un milieu sans sérum, réduit le nombre et la croissance de tumeurs qu'elles induisent après xénogreffe, et augmente l'expression des clusters de microRNA miR-182/183/203, miR-141/200a et miR-200b/c/429, inhibiteurs du phénotype CSC. Ces résultats suggèrent que la claudine-2 participe au maintien du phénotype CSC des cellules LGR5(+). Nous montrons de plus que le génotype des cellules LGR5(-) isolées est plastique, et que cette population contient aussi une sous-population de type CSC. Celle-ci pourrait exprimer une activité ALDH, et l'inhibition de claudine-2 induit une augmentation paradoxale de la formation de sphéroïdes par les cellules ALDH(+) isolées. La claudine-2 peut donc moduler différemment la capacité d'initiation tumorale des cellules LGR5(+) et ALDH(+). L'inhibition de tumorigénicité induite par la sous-expression de la claudine-2 prédomine dans la population tumorale globale, au sein de laquelle les cellules LGR5(+)/(-) et ALDH(+)/(-) coexistent. / Claudins are tight junction proteins frequently displaying abnormal expression patterns in cancer. In this work, we first showed that claudin-2 overexpression acts as an essential intermediate in the tumor-promoting role of the Symplekin/ZONAB complex in colorectal cancer (CRC). We then set out to analyze the function of Claudin-2 in cancer stem cells (CSC), which are strongly suspected to drive long-term tumorigenicity. To this effect, we experimentally down-regulated claudin-2 expression in LGR5-expressing cells, putative CRC cancer stem cells. We showed that claudin-2 inhibition by shRNA increased the expression of the epithelial differentiation marker CK20 and decreased the number of LGR5(+) cells, without inducing their apoptosis. Claudin-2-depleted LGR5(+) cells were no longer able to survive as single cells in suspension in serum free medium, and formed less and smaller tumors in a xenograft model. Moreover, the inhibition of claudin-2 increased the expression of the miRNA clusters miR-182/183/203, miR-141/200a and miR-200b/c/429, known as potent CSC phenotype inhibitors. Hence, our results suggest that claudin-2 promotes the CSC phenotype of LGR5(+) cells. In addition, we show that the genotype of isolated LGR5(-) cells is unstable and that this population also contains a CSC-like sub-population, potentially displaying a high ALDH activity. Interestingly, claudin-2 down-regulation in isolated ALDH(+) cells induces a paradoxical increase in their sphere-forming ability. Claudin-2 can thus differentially modulate the tumor-initiating capability of LGR5(+) and ALDH(+) cells. The inhibitory effect on tumorigenicitiy induced by claudin-2 depletion is clearly predominant in the global, heterogenous tumor cell population, where LGR5(+)/(-) and ALDH(+)/(-) cells coexist.

TRAIL ligand can trigger apoptosis of permissive human cells via engagement of its two pro- apoptotic receptors TRAIL-R1 (DR4) and TRAIL-R2 (DR5). Its ability to induce apoptosis independently on p53 status and to selectively kill cancer cells in vitro and in vivo made this ligand an attractive target in cancer research. However, acquired resistance of primary cancer cells, unsatisfactory outcome of clinical trials and recent studies arguing that TRAIL might under specific conditions promote cancer progression, opened new plethora of questions, which need to be addressed. Though both receptors DR4 and DR5 are ubiquitously expressed, different types of tumours show preference for either of the receptors. The relative participation of DR4 and DR5 in TRAIL- induced signalling is still largely unknown. To analyse TRAIL receptor-specific signalling, I prepared Strep-tagged, trimerised variants of recombinant human TRAIL ligands with high affinity for either DR4 or DR5 receptor. Using these receptor-specific ligands, I examined a contribution of individual pro-apoptotic receptors to TRAIL-induced signalling pathways. I found that in TRAIL resistant colorectal HT-29 cells but not in pancreatic PANC-1 cancer cells, DISC formation and initial caspase-8 processing proceeded comparably in both DR4- and...

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第19551号 / 医博第4058号 / 新制||医||1012(附属図書館) / 32587 / 京都大学大学院医学研究科医学専攻 / (主査)教授 川口 義弥, 教授 坂井 義治, 教授 羽賀 博典 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

CD133, a pentaspan glycoprotein, has long been known to represent aggressive, stem-like populations across various human malignancies. While its expression correlates with numerous clinical outcomes including disease progression, metastasis, recurrence, and poor overall survival in numerous cancers, little is currently known about its function. In the brain cancer glioblastoma (GBM), CD133-expressing cells have previously been shown to initiate tumours, evade therapy and interestingly, self-renew, a key property of cancer stem cells. With an implied signalling role in driving self-renewal, we aim to elucidate the role of CD133 in glioblastoma. To understand the role of CD133, we aim to study its protein-protein interactions using the proximity-dependent labelling technique known as miniTurboID. By tagging proteins of interest with a promiscuous biotin ligase at both protein termini, potential interactors can be biotinylated and identified by subsequent mass spectrometry. While miniTurboID has traditionally been performed by synthetic transgenes expressing the tagged proteins of interest in commercial cell lines, overexpression may not recapitulate its native function. Thus, using CRISPR technology, we aim to insert the miniTurboID ligase at both the N- and C-terminus of CD133 in patient-derived human GBM lines. Although little is currently known about CD133 function, development of targeted therapies has presented a promising strategy in pre-clinical studies. In the Singh Lab, we previously developed a chimeric antigen receptor T-cell, or CAR-T, comprised of a T-cell expressing a synthetic receptor capable of recognizing a tumor-associated antigen and activating cytolytic-killing directed towards the target cell. Currently, CAR-T therapies are autologous, or patient-derived, in nature which may host a myriad of concerns including patient-specific qualitative and quantitative T-cell dysfunction, inconsistent generation of CAR products, and availability to rapidly progressing patients. To circumvent this concern, “off-the-shelf”, donor-derived or allogeneic CAR-T products may be generated for use in GBM patients. However, in addition to CAR integration, allogeneic products must be additionally modified to eradicate expression of the endogenous TCR, as this would induce a phenomenon known as graft versus host disease, in which healthy tissues are targeted. Thus, in this thesis, we show gene editing potential in human GBMs to perform an endogenous genomic knock-in of miniTurboID. With the identification of interacting proteins, defining the subsequent functionality of CD133 may elucidate oncogenic cellular programs, and highlight common nodes of interaction within divergent cell signaling pathways. To develop an allogeneic CAR-T product, we designed a two-step approach in which the CAR sequence was integrated into the TCR gene for simultaneous knock-out. We later show early pre-clinical efficacy in comparison to traditional autologous CAR-T in our patient-derived models of human GBM. Thus, by using CD133 as a centralizing concept in this thesis, we ultimately hope to develop our biological understanding of CD133, while testing the therapeutic development of a donor-derived CAR-T therapy. / Thesis / Master of Science (MSc) / Glioblastoma (GBM) is one of the most common malignant brain tumors in adults. Despite an aggressive therapy regimen, almost all patients relapse 7-9 months post-diagnosis. Therapy failure and poor patient outcome may be attributed to a small population of cells known as glioblastoma stem cells, or GSCs, that are able to escape therapy and seed disease recurrence. GSCs are most notably identified by the cell surface protein CD133, which has previously been shown to associate with pro-tumor properties including treatment resistance, tumor growth, maintenance, progression and metastasis. While expression of CD133 in cancer has been heavily characterized, little is currently known about its function. One such avenue to understand its mechanism of action in cancer, and more particularly GBM, is to define its interactions with other proteins. Protein-protein interactions play a pivotal part as the backbone of signalling pathways that drive tumor development and growth. Therefore, defining and mapping the CD133 interaction network may help us understand how this protein governs regulation of GSCs, and ultimately, GBM progression. While the biology of CD133 has yet to be elucidated, targeting CD133 on GSCs has presented a promising therapeutic strategy for patients with GBM. Previously in the Singh Lab, we developed an engineered T-cell therapy, known as a CAR-T, that can recognize CD133 to induce tumor cell death. While this showed success in our animal models of human GBM, other considerations must be addressed on its path to clinical development. As of current, CAR-T therapies are generated from T-cells taken from cancer patients. This hosts a myriad of concerns including the quality of patient T-cells, the time and cost to manufacture, and its availability for patients with rapidly progressing disease. To circumvent this issue, donor-derived CAR-T cells can be genetically engineered for safe usage in GBM patients as a readily available, “off-the-shelf” therapy. To define the function of CD133, we have attempted to use a technique known as BioID, which tags the protein of interest with a smaller biotin ligase. This biotin ligase can subsequently tag proteins that come within the vicinity of CD133, that may later be identified by sequencing as potential interactors. As current use of BioID may not reliably mimic the interaction of CD133, we sought to genetically engineer human GBM lines with the BioID protein to more closely resemble tumor-relevant behaviours of CD133. To develop a donor-derived CAR-T therapy, we similarly used genetic engineering of T-cells to ensure specific targeting of tumor cells with CD133, while sparing healthy tissues. By using CD133 as a centralizing concept in this thesis, we ultimately hope to develop our biological understanding of CD133, while testing the therapeutic development of a donor-derived CAR-T therapy.