Le rétinoblastome, un cancer pédiatrique de la rétine en développement, est la tumeur intraoculaire la plus fréquente chez l’enfant et représente environ 4 % de tous les cancers infantiles. Bien qu'il s'agisse d'une maladie rare, l'hôpital Curie (centre de référence pour le rétinoblastome en France) accueille environ 50 à 60 nouveaux patients chaque année. Notre groupe a précédemment caractérisé deux sous-types de rétinoblastomes. Les tumeurs de type « cone-like » ou sous-type 1 sont plutôt différenciées et homogènes, présentent une surexpression des gènes liés aux cellules cônes (photorécepteurs) de la rétine, sont diagnostiquées cliniquement plus tôt et regroupent la majorité des formes héréditaires et bilatérales. Les tumeurs « mixed-type » ou sous-type 2, présentent une hétérogénéité intra-tumorale et une surexpression des gènes liés aux cellules des cônes et des cellules ganglionnaires de la rétine, sont enrichies en patients unilatéraux qui sont diagnostiqués cliniquement à des âges plus avancés. Nous avons caractérisé le paysage moléculaire et génomique de 102 rétinoblastomes provenant de trois institutions : l'Institut Curie (France), l'Hôpital Garrahan (Argentine) et l'Hôpital Sant Joan de Déu (Espagne). Le développement d'une signature de méthylation par pyroséquençage pour la classification des échantillons nous a permis d'élargir nos échantillons classés, d'une première série de 72 à notre dernière série de 102 tumeurs. L'analyse du paysage mutationnel de notre série a révélé que les tumeurs du sous-type 2 avaient plus de mutations somatiques par échantillon que les tumeurs du sous-type 1. De plus les gènes BCOR et ARID1A étaient les deux seuls gènes mutés de manière récurrente, et identifiés uniquement dans le sous-type 2. En divisant notre cohorte de tumeurs en sous-type 1 et 2, la distribution des mutations le long de RB1 était significativement différente. Par ailleurs, nous avons identifié une région de la protéine RB1 (dans le Domaine A) enrichie en mutations provenant des tumeurs du sous-type 2, avec très peu de mutations du sous-type 1. En plus, nous avons caractérisé deux événements récurrents de fusion chromosomique perturbant le gène DACH1. Les tumeurs de sous-type 2 sont caractérisées par une surexpression de TFF1, non exprimée dans la rétine normale. L'analyse par immunohistochimie de TFF1 dans des tumeurs localement invasives provenant de l'hôpital Garrahan a révélé la présence de cellules TFF1+ envahissant la région rétrolaminaire du nerf optique. Nous avons exploré un possible rôle oncogène de TFF1 dans le rétinoblastome lié à la survie cellulaire, à la migration cellulaire et à l'invasion cellulaire, qui n'a finalement pas été mis en évidence in vitro. Le sous-type moléculaire 2 regroupe les tumeurs MYCN amplifiées et les tumeurs avec une activation de la voie de signalisation MYC et des gènes cibles de MYC. L'utilisation de JQ1 et OTX015 (inhibiteurs des protéines BET) a fortement réduit la viabilité in vitro de lignées cellulaires de rétinoblastomes représentatives du sous-type 2, avec une régulation négative significative du gène et de la protéine MYC/MYCN. Nos résultats préliminaires suggèrent une nouvelle piste thérapeutique par l'inhibition des protéines BET dans le rétinoblastome. Les modèles précliniques largement utilisés dans la recherche sur le rétinoblastome n'ont pas été caractérisés ou classés au niveau moléculaire. Nous avons utilisé la même approche que pour la classification des tumeurs primaires et avons constaté que la plupart des modèles cellulaires et PDX étudiés étaient classés dans le sous-type moléculaire 2 et partageaient des caractéristiques moléculaires, génomiques et protéiques trouvés dans les tumeurs primaires de ce sous-type moléculaire. En conclusion, nous avons pu caractériser de façon plus approfondie le sous-type 2 des rétinoblastomes, qui semble présenter un phénotype plus agressif et qui est le sous-type représenté dans les modèles précliniques analysés. / Retinoblastoma (RB) is a rare pediatric cancer of the developing retina that represents the most common intraocular tumor in children, and accounts for about 4% of all childhood cancers. Although being a rare disease, the Curie Hospital (the referral center for retinoblastoma in France) treats about 50-60 new patients each year. Our group has previously characterized two retinoblastoma subtypes. The cone-like or subtype 1 tumors rather differentiated and homogenous, presenting an overexpression of genes related to cone photoreceptor retinal cells, clinically diagnosed earlier and grouping the majority of hereditary and bilateral forms. The mixed-type or subtype 2 tumors, displaying an intra-tumoral heterogeneity and showing overexpression of genes related to cone and retinal ganglion cells, are enriched in unilateral patients clinically diagnosed at older ages. The general goal of my thesis was to extend the molecular characterization of these subtype 2 retinoblastomas. We characterized the molecular and genomic landscape of retinoblastoma in a series of 102 primary tumors, integrating samples from three institutions: the Curie Institute (France), the Garrahan Hospital (Argentina) and Sant Joan de Déu Hospital (Spain). The development of a pyrosequencing-based tool for sample classification allowed us to enlarge our classed samples, from an initial series of 72, to our final series of 102 tumors. Analysis of the mutational landscape in our series revealed that tumors from the subtype 2 had significantly more somatic mutations per sample than tumors from the subtype 1. Besides RB1 gene, BCOR and ARID1A where the only two recurrently mutated genes, and identified only in the subtype 2. Distribution of mutations alongside the RB1 gene has so far been analyzed in terms of a single group of retinoblastomas. When splitting our cohort in subtype 1 and subtype 2 tumors, the distribution of mutations was significantly different. Besides, we identified a region of the RB1 protein (in Domain A) enriched in mutations from tumors of the subtype 2, and devoid of mutations of the subtype 1. Besides somatic mutations, we characterized two recurrent chromosomal fusion events disrupting DACH1. Subtype 2 tumors are characterized by an overexpression of TFF1, not expressed in the normal retina. Immunohistochemical analysis of TFF1 in locally invasive tumors coming from the Garrahan Hospital revealed the presence of TFF1+ cells invading the retrolaminar region of the optic nerve. We then explored a possible oncogenic role of TFF1 in retinoblastoma related to cell survival, cell migration and cell invasion, which was not fully uncovered. Molecular subtype 2 regroups the MYCN amplified tumors and tumors with MYC signaling pathway activation and upregulation of hallmark MYC target genes. The use of JQ1 and OTX015 (BET bromodomains inhibitors) strongly reduced the viability in vitro of retinoblastoma cell lines representatives of the subtype 2, together with a significant MYC/MYCN gene and protein downregulation. We provided preliminary results to explore a new therapeutic avenue of BET protein inhibition in retinoblastoma. Preclinical models widely used in retinoblastoma research has not been characterized or classified at the molecular level. We have used the same approach as for primary human tumor’s classification, and found that most cellular and PDX models studied classed in the molecular subtype 2 and shared many of the molecular, genomic and protein characteristics found in primary tumors of this molecular subtype. Taken together, we have performed a deeper characterization of subtype 2 retinoblastomas, which seems to represent a more aggressive phenotype, and is the represented subtype in the preclinical models analyzed.

Molecular mechanisms of lower-grade (II–III) diffuse gliomas (LGG) are still poorly understood, mainly because of their heterogeneity. They split into astrocytoma- (IDH-A) and oligodendroglioma-like (IDH-O) tumors both carrying mutations(s) at the isocitrate dehydrogenase (IDH) gene and into IDH wild type (IDH-wt) gliomas of glioblastoma resemblance. We generated detailed maps of the transcriptomes and DNA methylomes, revealing that cell functions divided into three major archetypic hallmarks: (i) increased proliferation in IDH-wt and, to a lesser degree, IDH-O; (ii) increased inflammation in IDH-A and IDH-wt; and (iii) the loss of synaptic transmission in all subtypes. Immunogenic properties of IDH-A are diverse, partly resembling signatures observed in grade IV mesenchymal glioblastomas or in grade I pilocytic astrocytomas. We analyzed details of coregulation between gene expression and DNA methylation and of the immunogenic micro-environment presumably driving tumor development and treatment resistance. Our transcriptome and methylome maps support personalized, case-by-case views to decipher the heterogeneity of glioma states in terms of data portraits. Thereby, molecular cartography provides a graphical coordinate system that links gene-level information with glioma subtypes, their phenotypes, and clinical context.

Connor O’Hara July 29, 2019 Inhibition of Cancer Stem Cells by Glycosaminoglycan Mimetics In the United States cancer is the second leading cause of death, with colorectal cancer (CRC) being the third deadliest cancer and expected to cause over 51,000 fatalities in 2019 alone.1 The current standard of care for CRC depends largely on the staging, location, and presence of metastasis.2 As the tumor grows and invades nearby lymph tissue and blood vessels, CRC has the opportunity to invade not only nearby tissue but also metastasize into the liver and lung (most commonly).3 The 5-year survival rate for metastasized CRC is <15%, and standard of care chemotherapy regimens utilizing combination treatments only marginally improve survival.3-5 Additionally, patients who have gone into remission from late-stage CRC have a high risk of recurrence despite advances in treatment.6-7 The Cancer Stem-like Cell (CSC) paradigm has grown over the last 20 years to become a unifying hypothesis to support the growth and relapse of tumors previously regressed from chemotherapy (Figure 1).8 The paradigm emphasizes the heterogeneity of a tumor and its microenvironment, proposing that a small subset of cells in the tumor are the source of tumorigenesis with features akin to normal stem cells.9 The CSCs normally in a quiescent state survive this chemotherapy and “seed” tumor redevelopment.10 First observed in acute myeloid lymphoma models, CSCs have since been identified in various other cancers (to include CRC) by their cell surface antigens and unique properties characterizing them from normal cancer cells.11-12 These include tumor initiation, limitless self-renewal capacity to generate clonal daughter cells, as well as phenotypically diverse, mature, and highly differentiated progeny.13-14 Previously our lab has identified a novel molecule called G2.2 (Figure 2) from a unique library of sulfated compounds showing selective and potent inhibition of colorectal CSCs in-vitro.15 G2.2 is a mimetic of glycosaminoglycans (GAGs) and belongs to a class of molecules called non-saccharide GAG mimetics (NSGMs). Using a novel dual-screening platform, comparisons were made on the potency of G2.2 in bulk monolayer cells, primary 3D tumor spheroids of the same cell line, and subsequent generations of tumor spheroids. This work has shown in-vitro the fold-enhancement of CSCs when culturing as 3D tumor spheroids. Spheroid culture serves as a more accurate model for the physiological conditions of a tumor, as well as the functional importance of upregulating CSCs. Evaluation of G2.2 and other NSGMs was performed in only a few cell lines, developing a need to better understand the ability of G2.2 to inhibit spheroids from a more diverse panel of cancer cells to better understand G2.2’s mechanism. The last few decades have seen the advancement in fundamental biological and biochemical knowledge of tumor cell biology and genetics.16 CRC, in particular, has served as a useful preclinical model in recapitulating patient tumor heterogeneity in-vitro.17 Recent work has characterized the molecular phenotypes of CRC cell lines in a multi-omics analysis, stratifying them into 4 clinically robust and relevant consensus molecular subtypes (CMS).18-19 Our work was directed to screen a panel of cells from each of the molecular subtypes and characterize the action of G2.2 and 2nd generation lipid-modified analogs, synthesized to improve the pharmacokinetic properties of the parent compound. Four NSGMs, namely G2.2, G2C, G5C, and G8C (Figure 2) were studied for their ability to inhibit the growth of primary spheroids across a phenotypically diverse panel. Compound HT-29 IC50 (μM) Panel Average IC50 (μM) G2.2 28 ± 1 185 ± 55 G2C 5 ± 2 16 ± 15 G5C 8 ± 2 63 ± 19 G8C 0.7 ± 0.2 6 ± 3 Primary spheroid inhibition assays were performed comparing the potency of new NSGMs to G2.2. Fifteen cell lines were evaluated in a panel of colorectal adenocarcinoma cell lines with several cell lines representing each CMS. Primary spheroid inhibition assays revealed 3 distinct response with regard to G2.2’s ability to inhibit spheroid growth. Cells from CMS 3 and 4, which display poor clinical prognosis, metabolic dysregulation, and enhanced activation of CSC pathways, showed the most sensitivity to G2.2 (mean IC50 = 89 ± 55 μM). Mesenchymal CMS 4 cell lines were over 3-fold more sensitive to treatment with G2.2 when compared to CMS 1 cell lines. Resistant cell lines were composed entirely of CMS 1 and 2 (mean IC50 = 267 ± 105 μM). In contrast, all lipid-modified analogs showed greater potency than the parent NSGM in almost every CRC cell line. Of the three analogs, G8C showed the greatest potency with a mean IC50 of less than 15 μM. Of the CRC spheroids studied, HT-29 (CMS 3) was most sensitive to G8C (IC50 = 0.73 μM). To evaluate the selectivity of NSGMs for CSC spheroid inhibition, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium) cytotoxicity assays were performed on monolayer cell culture, and the fold-selectivity of NSGM for spheroids was analyzed. Data shows that NSGMs preferentially target CSC-rich spheroids compared with monolayer cellular growth, with G2.2 having over 7-fold selectivity for spheroid conditions. This fold selectivity was enhanced in CMS 3/4, supporting the idea that G2.2 targets a mesenchymal and stem-like phenotype. To further validate this selectivity, limiting dilution assays were performed across the panel to determine the tumor-initiating capacity of each cell line. Cell lines which showed a sensitive response to G2.2 were over 2-fold more likely to develop into spheroids, validating the previous hypothesis. Further characterization was performed analyzing the changes G2.2 induced on CSC markers, as well as the basal expression of a unique pair of cancer cells. Western blots showed a reduction in self-renewal marker across all CMS after treatment with G2.2, and that cell lines sensitive to G2.2-treatment overexpress mesenchymal and stem-like markers. G2.2-resistant cell lines show an epithelial phenotype, lacking this expression. The positive results observed in these studies enhance the understanding of G2.2 and analogs, and further evaluation with additional cell lines of various tissues would improve the knowledge thus far gained. However, all experiments described take valuable time to perform and analyze. Thus, there became a need to develop a high-throughput screening (HTS) platform for our assays that standardized analysis and enhanced productivity. Initial development of the method for this assay are underway, and recent evidence from these evaluations of breast cancer spheroids suggests that G2.2 and analogs may be tissue-specific compounds for the treatment of cancer. Future work entails refining the application of this method for evaluation of the NCI-60 (National Cancer Institute) tumor cell panel. Overall, these results make several suggestions concerning the NSGMs evaluated against the panel. First, G2.2 selectively targets CSCs with limited toxicity to monolayer cells of the same cell line. Further, G2.2 has the greatest potency with CMS 3/4, whose mesenchymal phenotypes are associated with poor clinical prognosis and enrichment of CSCs. Supporting evidence include that sensitive cell lines are highly tumorigenic and show enhanced expression of mesenchymal/CSC markers compared to resistant cell lines. Lipid-modification of G2.2 enhances in-vitro potency against spheroid growth, with nM potency reached in the most sensitive cell lines. Evidence in the development of a HTS platform also suggests these NSGMs show tissue specificity to cancers of the intestine. Further work characterizing the mechanism of NSGMs in a broader multi-tissue panel will enhance our understanding of the compounds as a potential therapy to dramatically improve patient survival through specific targeting of tumorigenesis. References 1. Colorectal Cancer Facts & Figures 2017-2019. American Cancer Society 2017. 2. Compton, C. C.; Byrd, D. R.; Garcia-Aguilar, J.; Kurtzman, S. H.; Olawaiye, A.; Washington, M. K. Colon and rectum. In AJCC Cancer Staging Atlas, 2nd ed.; Ed. Springer Science: New York, 2012; pp 185–201. 3. Van Cutsem, E.; Cervantes, A.; Adam, R.; Sobrero, A.; Van Krieken, J. H.; Aderka, D.; Aranda Aguilar, E.; Bardelli, A.; Benson, A.; Bodoky, G.; et al. ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann. Oncol. 2016, 27, 1386–422. 4. Siegel, R. L.; Miller, K. D.; Fedewa, S. A.; Ahnen, D. J.; Meester, R. G. S.; Barzi, A.; Jemal, A. Colorectal cancer statistics, 2017. CA Cancer J. Clin. 2017, 67, 177–193. 5. Moriarity, A.; O'Sullivan, J.; Kennedy, J.; Mehigan, B.; McCormick, P. Current targeted therapies in the treatment of advanced colorectal cancer: a review. Ther. Adv. Med. Oncol. 2016, 8, 276–293. 6. Seidel, J.; Farber, E.; Baumbach, R.; Cordruwisch, W.; Bohmler, U.; Feyerabend, B.; Faiss, S. Complication and local recurrence rate after endoscopic resection of large high-risk colorectal adenomas of >/=3 cm in size. Int. J. Colorectal Dis. 2016, 31, 603–611. 7. Pugh, S. A.; Shinkins, B.; Fuller, A.; Mellor, J.; Mant, D.; Primrose, J. N. Site and stage of colorectal cancer influence the likelihood and distribution of disease recurrence and postrecurrence survival: data from the FACS randomized controlled trial. Ann. Surg. 2016, 263, 1143–1147. 8. Batlle, E.; Clevers, H. Cancer stem cells revisited. Nat. Med. 2017, 23, 1124–1134. 9. Hanahan, D.; Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 2011, 144, 646–674. 10. Tirino, V.; Desiderio, V.; Paino, F.; De Rosa, A.; Papaccio, F.; La Noce, M.; Laino, L.; De Francesco, F.; Papaccio, G. Cancer stem cells in solid tumors: an overview and new approaches for their isolation and characterization. FASEB J. 2013, 27, 13–24. 11. Bonnet, D.; Dick, J. E. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat. Med. 1997, 3, 730–737. 12. Desai, A.; Yan, Y.; Gerson, S. L. Concise reviews: cancer stem cell targeted therapies: toward clinical success. Stem Cells Transl. Med. 2019, 8, 75–81. 13. Munro, M. J.; Wickremesekera, S. K.; Peng, L.; Tan, S. T.; Itinteang, T. Cancer stem cells in colorectal cancer: a review. J. Clin. Pathol. 2018, 71, 110–116. 14. Zhou, Y.; Xia, L.; Wang, H.; Oyang, L.; Su, M.; Liu, Q.; Lin, J.; Tan, S.; Tian, Y.; Liao, Q.; Cao, D. Cancer stem cells in progression of colorectal cancer. Oncotarget 2018, 9, 33403–33415. 15. Patel, N. J.; Karuturi, R.; Al-Horani, R. A.; Baranwal, S.; Patel, J.; Desai, U. R.; Patel, B. B. Synthetic, non-saccharide, glycosaminoglycan mimetics selectively target colon cancer stem cells. ACS Chem. Biol. 2014, 9, 1826–1833. 16. Punt, C. J.; Koopman, M.; Vermeulen, L. From tumour heterogeneity to advances in precision treatment of colorectal cancer. Nat. Rev. Clin. Oncol. 2017, 14, 235–246. 17. Mouradov, D.; Sloggett, C.; Jorissen, R. N.; Love, C. G.; Li, S.; Burgess, A. W.; Arango, D.; Strausberg, R. L.; Buchanan, D.; Wormald, S.; et al. Colorectal cancer cell lines are representative models of the main molecular subtypes of primary cancer. Cancer Res. 2014, 74, 3238–3247. 18. Guinney, J.; Dienstmann, R.; Wang, X.; de Reynies, A.; Schlicker, A.; Soneson, C.; Marisa, L.; Roepman, P.; Nyamundanda, G.; Angelino, P.; et al. The consensus molecular subtypes of colorectal cancer. Nat. Med. 2015, 21, 1350–1356. 19. Berg, K. C. G.; Eide, P. W.; Eilertsen, I. A.; Johannessen, B.; Bruun, J.; Danielsen, S. A.; Bjornslett, M.; Meza-Zepeda, L. A.; Eknaes, M.; Lind, G. E.; et al. Multi-omics of 34 colorectal cancer cell lines - a resource for biomedical studies. Mol. Cancer 2017, 16, 116–132.