Chimie et Modélisation pour la Biologie du Cancer

Abstract Radiotherapy is a cornerstone of cancer management. The improvement of spatial dose distribution in the tumor volume by minimizing the dose deposited in the healthy tissues have been a major concern during the last decades. Temporal aspects of dose deposition are yet to be investigated. Laser-plasma-based particle accelerators are able to emit pulsed-proton beams at extremely high peak dose rates (~10 9 Gy/s) during several nanoseconds. The impact of such dose rates on resistant glioblastoma cell lines, SF763 and U87-MG, was compared to conventionally accelerated protons and X-rays. No difference was observed in DNA double-strand breaks generation and cells killing. The variation of the repetition rate of the proton bunches produced an oscillation of the radio-induced cell susceptibility in human colon carcinoma HCT116 cells, which appeared to be related to the presence of the PARP1 protein and an efficient parylation process. Interestingly, when laser-driven proton bunches were applied at 0.5 Hz, survival of the radioresistant HCT116 p53 −/− cells equaled that of its radiosensitive counterpart, HCT116 WT, which was also similar to cells treated with the PARP1 inhibitor Olaparib. Altogether, these results suggest that the application modality of ultrashort bunches of particles could provide a great therapeutic potential in radiotherapy.

A library of 52 distyryl and 9 mono-styryl cationic dyes was synthesized and investigated with respect to their optical properties, propensity to aggregation in aqueous medium, and capacity to serve as fluorescence “light-up” probes for G-quadruplex (G4) DNA and RNA structures. Among the 61 compounds, 57 dyes showed preferential enhancement of fluorescence intensity in the presence of one or another G4-DNA or RNA structure, while no dye displayed preferential response to double-stranded DNA or single-stranded RNA analytes employed at equivalent nucleotide concentration. Thus, preferential fluorimetric response towards G4 structures appears to be a common feature of mono- and distyryl dyes, including long-known mono-styryl dyes used as mitochondrial probes or protein stains. However, the magnitude of the G4-induced “light-up” effect varies drastically, as a function of both the molecular structure of the dyes and the nature or topology of G4 analytes. Although our results do not allow to formulate comprehensive structure–properties relationships, we identified several structural motifs, such as indole- or pyrrole-substituted distyryl dyes, as well as simple mono-stryryl dyes such as DASPMI [2-(4-(dimethylamino)styryl)-1-methylpyridinium iodide] or its 4-isomer, as optimal fluorescent light-up probes characterized by high fluorimetric response ( I / I 0 of up to 550-fold), excellent selectivity with respect to double-stranded DNA or single-stranded RNA controls, high quantum yield in the presence of G4 analytes (up to 0.32), large Stokes shift (up to 150 nm) and, in certain cases, structural selectivity with respect to one or another G4 folding topology. These dyes can be considered as promising G4-responsive sensors for in vitro or imaging applications. As a possible application, we implemented a simple two-dye fluorimetric assay allowing rapid topological classification of G4-DNA structures.

Objective: The management of giant pituitary tumors is complex, with few publications and recommendations. Consequently, patient's care mainly relies on clinical experience. We report here a first large series of patients with giant pituitary tumors managed by a multidisciplinary expert team, focusing on treatments and outcome. Methods: A retrospective cohort study was conducted. Giant pituitary tumors were defined by a main diameter > 40mm. Macroprolactinomas sensitive to dopamine agonists were excluded. All patients were operated by a single neurosurgical team. After surgery, multimodal management was proposed, including hormone replacement, radiotherapy and anti-tumor medical therapies. Outcome was modeled using Kaplan-Meyer representation. A logistic regression model was built to identify the risk factors associated with surgical complications. Results: 63 consecutive patients presented a giant adenoma, most often with visual defects. Patients were operated once, twice or three times in 59%, 40% and 1% of cases respectively, mainly through endoscopic endonasal approach. Giant adenomas included gonadotroph, corticotroph, somatotroph, lactotroph and mixed GH-PRL subtypes in 67%, 14%, 11%, 6% and 2% of patients respectively. Vision improved in 89% of patients with prior visual defects. Severe surgical complications occurred in 11% of patients, mainly for tumors > 50 mm requiring microscopic transcranial approach. Additional radiotherapy was needed for 29% of patients, 3 to 56 months after first surgery. For 6% of patients, Temozolomide treatment was required, 19 to 66 months after first surgery. Conclusions: Giant pituitary tumors require multimodal management, with a central role of surgery. Most often, tumor control can be achieved by expert multidisciplinary teams.

Triphenylamines (TPAs) were previously shown to trigger cell death under prolonged one- or two-photon illumination. Their initial subcellular localization, before prolonged illumination, is exclusively cytoplasmic and they translocate to the nucleus upon photoactivation. However, depending on their structure, they display significant differences in terms of precise initial localization and subsequent photoinduced cell death mechanism. Here, we investigated the structural features of TPAs that influence cell death by studying a series of molecules differing by the number and chemical nature of vinyl branches. All compounds triggered cell death upon one-photon excitation, however to different extents, the nature of the electron acceptor group being determinant for the overall cell death efficiency. Photobleaching susceptibility was also an important parameter for discriminating efficient/inefficient compounds in two-photon experiments. Furthermore, the number of branches, but not their chemical nature, was crucial for determining the cellular uptake mechanism of TPAs and their intracellular fate. The uptake of all TPAs is an active endocytic process but two- and three-branch compounds are taken up via distinct endocytosis pathways, clathrin-dependent or -independent (predominantly caveolae-dependent), respectively. Two-branch TPAs preferentially target mitochondria and photoinduce both apoptosis and a proper necrotic process, whereas three-branch TPAs preferentially target late endosomes and photoinduce apoptosis only.

Abstract The TAM kinase family arises as a promising therapeutic target for cancer therapy, as well as auto‐immune and viral diseases. In this study, we report the first photoactivatable caged inhibitors of Tyro3 and Mer. This strategy enables spatial and temporal control of the biological activity of the inhibitor upon irradiation with UV light. We describe the design, synthesis, photocleavage properties, and inhibitory activities of four Tyro3 and Mer photoactivatable small molecules. The proof of concept on the TAM kinase family was achieved in vitro, since irradiation by UV light restored the full inhibitory activity of two prodrugs.

Abstract G‐quadruplexes (G4s) are non‐canonical DNA structures implicated in a number of biological processes. Small‐molecule ligands can alter stability and folding of G4s, which can potentially be exploited for therapeutic purposes. In this work, we investigate the interaction of telomeric DNA fragment from Tetrahymena thermophila (TET25, 5′‐G(TTGGGG) 4‐ 3′) with a G4 ligand PyDH2 belonging to the bisquinolinium family. When alone, TET25 adopts a mixture of three conformations, with the most abundant being a four‐tetrad hybrid G4. In the presence of PyDH2, surprisingly, TET25 folds into an antiparallel chair G4, with PyDH2 intercalated between G‐tetrads 2 and 3, according to our crystal structure. The structure represents the second example, and the first crystallographic evidence, of ligand intercalation into a G4. In solution, the interaction of PyDH2 and TET25 leads to a number of complexes differing by G4 topology and binding stoichiometry, strong stabilization of G4 (∆ T m = 12.4 °C in the presence of one equiv. of PyDH2) and large hysteresis of ∼10 °C, suggesting that ligand binding and G4 folding processes are complex.

Abstract G‐quadruplexes (G4s) are non‐canonical DNA structures implicated in a number of biological processes. Small‐molecule ligands can alter stability and folding of G4s, which can potentially be exploited for therapeutic purposes. In this work, we investigate the interaction of telomeric DNA fragment from Tetrahymena thermophila (TET25, 5′‐G(TTGGGG) 4‐ 3′) with a G4 ligand PyDH2 belonging to the bisquinolinium family. When alone, TET25 adopts a mixture of three conformations, with the most abundant being a four‐tetrad hybrid G4. In the presence of PyDH2, surprisingly, TET25 folds into an antiparallel chair G4, with PyDH2 intercalated between G‐tetrads 2 and 3, according to our crystal structure. The structure represents the second example, and the first crystallographic evidence, of ligand intercalation into a G4. In solution, the interaction of PyDH2 and TET25 leads to a number of complexes differing by G4 topology and binding stoichiometry, strong stabilization of G4 (∆ T m = 12.4 °C in the presence of one equiv. of PyDH2) and large hysteresis of ∼10 °C, suggesting that ligand binding and G4 folding processes are complex.

In this work, the linear and nonlinear optical properties of three Imidazo[4-5,b]pyridine derivatives were investigated The results show an increase in the fluorescence quantum efficiency and 2PA cross-section due to substituent groups.

In this work, the spectroscopic properties of Imidazo[1,2-a]pyridine derivatives were investigated, as well as the two-photon absorption process. An increase in the 2PA cross-section was shown with the different structures of the molecules.

The global pharmaceutical industry has experienced significant growth, accompanied by a worrying increase in counterfeit medicines. Substandard medicines pose serious health risks, including poisoning, exacerbation of untreated conditions and mortality, particularly in regions with weak regulatory systems. Anaesthetics, essential to medical care, are particularly vulnerable to supply chain shortages and counterfeiting, highlighting the need for effective authentication strategies. Herein, we propose a colorimetric sensor-array based on a competitive displacement mechanism exploiting the supramolecular host-guest interactions between cucurbit[n]uril macrocycles and cationic multibranched chromogenic dyes. This approach exploits the competition between anaesthetic molecules and host-dye complexes to induce analyte-specific dye colorimetric changes. The resulting optical fingerprints, being a combination of all the signals of the host/dye complexes, processed by machine learning algorithms, allows for the detection, discrimination, and quantification of six anaesthetic molecules. This newly developed array thereby provides a potential tool for addressing issues related to drug counterfeiting and enhancing the reliability of pharmaceutical quality control. • Counterfeiting medicines pose serious health risks in regions with weak regulatory systems • High-throughput and ready-to-use sensor should be developed to detect easily counterfeit medicines • Colorimetric sensor array are versatile tools to address this issue • Supramolecular host-guest-based sensor array are particularly well suited for drugs classification

A study of two-photon absorption cross-section (2PACS) in imidazo[4,5-b]pyridines grafted to different groups was done. A 2PACS of around 150 GM and 10 GM was observed at 540 nm and 700 nm, respectively.

Deoxyribonucleic acid has different structures in human beings. The most known is the double helix but a lot of secondary structures exist and particularly G-quadruplex. It consists of guanine-rich nucleic acid sequences. The association of four guanines through hydrogen bonds forms a plan called G-quartet. This set of hydrogen bonds is called Hoogsteen base pairs. The stacking of at least two quartets around a monovalent cation like potassium or sodium establishes the G-quadruplex. These structures have been much studied over the past twenty years. They are involved in numerous biological mechanisms like replication, transcription, translation and also telomere maintenance. G-quadruplex presence can cause an important genetic as well as epigenetic instability. That is why many methods have been developed in order to localize these structures and to understand their role in vivo. To this end, a broad panel of molecular tools has been used. However, it is still difficult to bring an answer to all the questions about the involvement of G-quadruplex at the genomic level with this panel. In this thesis work, we developed new molecular tools able to target selectively G-quadruplex in a complex biological medium from two benchmark ligands, PhenDC3 and PDC, which have very good affinity and selectivity for G-quadruplex.On the one hand, functionalized ligands have been synthetized with a biotin and/or a photoactivatable group in order to trap and pull-down G-quadruplex in various cellular contexts. On the other hand, derivative compounds which are able to be functionalized in cellulo by bioorthogonal reactions have been obtained. Once the compound interacts with its cellular target, a function (fluorophore or biotin) can be added through an orthogonal reaction. The new panel of compounds has been evaluated by biophysical techniques, FRET-melting experiment and FID assay, in order to determine their affinity to G-quadruplex and their selectivity. We proposed a relation between the two biophysical experiments in order to have a good ranking of ligands for G-quadruplex structures.One of the most important objectives of this work was to localize G-quadruplex ligands in human cancer cells. First, a complete study in fixed cells has been performed using two reactions of click chemistry: reaction of copper-catalyzed-alkyne-azide-cycloaddition (CuAAC) and reaction of strain-promoted alkyne-azide cycloaddition (SPAAC). Secondly, the study has been pursued in living cells using SPAAC reaction because of the toxicity of copper in cells.These compounds have also been used to extract G-quadruplex from biological systems with cyclooctyne-coated magnetic beads. However, results obtained in this preliminary study are not decisive so it could be interesting to optimize the system before concluding.

International audience

I-Motifs are tetrameric DNA structures constituted by mutually intercalated CH+:C base pairs that can form in cytosine-rich DNA sequences under slightly acidic conditions (pH ≈ 6.0) where some of cytosines are protonated. They have been known for over thirty years, but due to their strong pH dependence, these structures were long considered to exist only in vitro. Quite recently, in 2018, studies renewed interest in the i-motif suggesting its presence at the cellular level, especially due to the development of the iMab antibody, raised against i-motifs. Numerous sequences capable of forming i-motifs have been identified in active regions of the genome, such as at telomeres or oncogene promoters, suggesting potentially important biological roles for this structure. To determine these effects, tools complementary to antibodies need to be developed to control i-motif formation and stability. These chemical compounds capable of stabilizing the structure and thus controlling its formation are called ligands. Many compounds have been described over the years, but none has yet been unanimously accepted as a ligand for DNA motifs. This absence of a reference ligand is partly explained by the methods used so far to observe small-molecule interactions with the i-motif structure. Indeed, the techniques employed were inherited from studies of other DNA structures, and many biases introduced by them have been identified, calling previous observations into question. To clarify the subject and conclude on the various effects of compounds, we proposed a specific method for studying i-motifs that allows the evaluation of ligands' effects on these structures. Toward this end, a potentiometric titration method, based on circular dichroism studies at different pH, was developed and validated for the i-motif. This technique allows the observation of DNA conformation without labelling at different pH, thereby detecting stabilization caused by a ligand by measuring the transition pH between folded and unfolded DNA, in isothermal conditions. This method was then coupled with thermal studies, and both were used to test different compounds previously described in the literature and observe their effects on the i-motif's stability in relation to pH. The results show that among all the tested compounds, none were capable of stabilizing native i-motifs. While three compounds (TMPyP4, BRACO-19, and Mitoxantrone) did indeed show a destabilizing effect on the structure, only the complex [Ru(phen)2dppz]2+ demonstrated some stabilizing capacity. However, this stabilization was only observed with a specific sequence with long loops, suggesting that this interaction occurs through interactions with DNA loops. The effects of other compounds on these loops were also studied, suggesting that the effects previously observed in the literature resulted from secondary, non-specific interactions with the i-motif loops. In the second part of this thesis, the selectivity of the iMab antibody was also assessed, and the results show that this antibody targets cytosine-rich DNA sequences rather than the i-motif specifically. Other results even indicate that this antibody can unfold the i-motif, further suggesting that it binds to unfolded C-rich sequences rather than to i-motifs. Overall, the results presented in this study show that none of the tested compounds could stabilize the i-motif by directly interacting with it, highlighting the difficulties of targeting this DNA structure. Additionally, the results concerning the iMab antibody call into question the interpretation of nuclear foci observed in other studies, which currently represent the only evidence of native i-motifs at the cellular level.

Transmission Electron Microscopy is a major tool for performing structural studies in biology. Some methods used for image sampling and analysis need to be improved in order to observe electron dose sensitive samples with good contrast and good signal to noise ratio. During this thesis, various methodological and computational approaches have been studied which aim to improve image quality. First, I evaluated the relevance of combining energy filtered imaging with the STEM mode. I show that this allows an improvement of the signal to noise ratio of images. Then, I devised an algorithm that generates an image from phase data. This approach allows improving the image contrast over direct imaging. The use of a phase plate and focal tilt series are both efficient tools to achieve this goal. While working on the software approach for processing of tilt series, we found that a qualitative result can be obtained from a single image. I developped the SPCI plugin for the ImageJ software. It allows processing between one and three focal images. My work involves optimization of the tomographic reconstruction process, including working with both alignment algorithms and reconstruction algorithms. I expose my studies on image alignment methods used on tilt series. These methods do rely on the use of key points and associated local descriptors. They have proved to be efficient to process images lacking fiducial markers. Finally, I propose a new unified algorithmic approach for 3D reconstruction of tomographic tilt series acquired with sparse sampling. I then derived another novel method that integrates the image alignment step in the process. Studies and developments will continue on both methods in futur work.

Les i-motifs sont des structures tétramériques d'ADN constituées de paires de bases CH+:C intercalées et qui peuvent se former dans des séquences d'ADN riches en cytosine, notamment dans les conditions légèrement acides (pH≈6,0) où certaines cytosines seront protonées. Bien qu'elles soient connues depuis plus de 30 ans, du fait de leur forte dépendance au pH ces structures ont longuement été considérées comme uniquement présentes in vitro. Plus récemment, en 2018, un anticorps iMab a été développé pour cibler les i-motifs et les études ont suggéré leur présence au niveau cellulaire. De nombreuses séquences pouvant former des i-motifs ont été identifiées dans des zones actives du génome, au niveau de télomères ou de promoteurs d'oncogènes, suggérant alors de potentiels rôles biologiques de cette structure. Afin de déterminer ces effets, des outils complémentaires à l'anticorps doivent être développés afin de pouvoir contrôler la formation et la stabilité du i-motif. Ces composés chimiques, capables de stabiliser la structure et ainsi de contrôler sa formation, sont appelés ligands. De nombreux composés ont été décrits au fil des années mais aucun composé n'a aujourd'hui été unanimement admis comme ligand de i-motif de l'ADN. Cette absence de ligand de référence s'explique notamment par les méthodes utilisées jusqu'à présent pour observer les interactions des molécules avec la structure du i-motif. En effet, les techniques employées ont été héritées des études d'autres structures d'ADN et de nombreux biais introduits par celles-ci ont été identifiés, remettant en question les effets précédemment observés. Afin d'éclaircir le sujet et de conclure sur les différents effets de composés, nous avons alors proposé une méthode d'étude spécifique aux i-motifs permettant d'évaluer les effets de ligands sur la structure. Pour cela, une méthode dite de titration potentiométrique, basée sur des études de dichroïsme circulaire à différents pH à température constante, a été développée et validée sur le i-motif. Cette technique permet d'observer sans marquage de l'ADN l'état de ce dernier à différents pH, et ainsi de détecter une stabilisation causée par un ligand. Cette méthode a ensuite été couplée à des études thermiques et les deux ont alors été utilisées pour tester différents composés issus de la littérature et observer leurs effets sur la stabilité du i-motif vis-à-vis du pH. Nos résultats montrent que parmi tous les composés testés, aucune des petites molécules testées n'est capable de stabiliser un i-motif natif. Si trois composés (TMPyP4, BRACO-19 et Mitoxantrone) ont montré un effet déstabilisant sur la structure, seul le complexe [Ru(phen)2dppz]2+ a montré une certaine capacité de stabilisation. Néanmoins, cette stabilisation n'a été observée que sur une séquence spécifique à longues boucles, suggérant que cette interaction ne se fait que par des interactions avec des boucles d'ADN. Les effets des autres composés ont également été étudiées sur ces boucles, suggérant que les effets précédemment observés dans la littérature ne résultent que d'interactions à longues boucles, non spécifiques au i-motif. Dans la seconde partie, la spécificité de l'anticorps iMab a également été testée et les résultats démontrent que cet anticorps cible des séquences d'ADN riches en cytosines et non spécifiquement le i-motif. D'autres résultats indiquent même que cet anticorps est capable de déplier le i-motif suggérant là encore que cet anticorps se lie aux séquences C-riches dépliées. Dans l'ensemble, les résultats présentés dans cette thèse ne montrent qu'aucun des composés testés n'est capable de stabiliser le i-motif en interagissant directement avec le i-motif, mettant en lumière les difficultés de cibler la structure. De plus, les résultats concernant l'anticorps remettent en question l'interprétation des foyers nucléaires observés dans d'autres études et qui représentent aujourd'hui les seules observations de i-motifs natifs au niveau cellulaire.

La famille TAM regroupe trois tyrosines kinases transmembranaires : Tyro3, Axl et Mer. Ces trois protéines sont impliquées dans de nombreux mécanismes de signalisation et de régulation cellulaire, notamment de croissance et d’apoptose. La famille TAM a été identifiée comme nouvelle cible thérapeutique prometteuse dans le traitement de certains cancers, de maladies auto-immunes et d’infections virales. Toutefois, il n’existe que peu de molécules ayant été spécialement conçues comme inhibiteurs de TAM, la plupart s’intéressant à Axl ou Mer. La première partie de ce manuscrit est consacrée à l’optimisation d’inhibiteurs de TAM basés sur le noyau imidazo[4,5-b]pyridine. La conception par modélisation moléculaire, la synthèse et l’évaluation biologique seront discutées. Pour contourner les problèmes de sélectivité inhérents aux protéines kinases, nous avons choisi d’appliquer le concept de la photopharmacologie. Cette stratégie permet de contrôler spatialement et temporellement l'activité de la drogue uniquement dans la zone à traiter par l’utilisation de la lumière, et donc d’éviter les effets secondaires. L’approche utilisant les groupements photo-labiles (GPL) consiste à masquer une fonction chimique importante pour l’activité thérapeutique, ce qui rend la molécule temporairement inactive. L’irradiation du composé induit le clivage du groupement et ainsi la restauration de l’activité inhibitrice. Nous avons introduit différents groupements photo-labiles (GPL) sur l’azote N3 de nos inhibiteurs. Le choix du groupement, la synthèse, l’étude du photo-clivage et l’évaluation biologique des composés seront présentés dans le deuxième chapitre. Enfin, nous avons développé une méthodologie de sulfénylation directe d’imidazopyridines utilisant le DABCO.(SO₂)₂, un précurseur de dioxyde de soufre comme source de soufre.

The TAM family consists in 3 tyrosine kinases : Tyro3, Axl and Mer. These proteins are involved in many cellular processes and pathways. The TAM family has been identified as a new promising target for cancer therapy, autoimmune diseases and viral infections. However, only a few inhibitors have been developped for this family. The first chapter of this manuscript is dedicated to the conception, synthesis and biological evaluation of new inhibitors for the TAM family. To bypass the maindrawback of protein kinase inhibitors, selectivity, we chose to apply the concept of photopharmacology. This strategy enable spatial and temporal control of the drug activity upon irradiation. By blocking a key position of the inhibitor with a photoremovable group, we can inactivate the molecule, and restore the activity upon irradiation. We introduced different photoremovable protecting groups on our inhibitors. The choice of the groups, the synthesis, the photo-cleavage and the biological evaluation will be discussed in chapter 2.Finally, we also developped a methodology for direct sulfenylation of imidazopyridines using DABCO.(SO₂)₂ as sulfur source.

Dans le contexte de la chimiothérapie, la réparation de l’ADN réduit les dommages induits par les agents alkylants de l’ADN dont le témozolomide (TMZ), conduisant à la chimiorésistance. Une des voies principales de réparation de l’ADN est la voie par excision de base (BER) au sein de laquelle une enzyme clée, APE1 (endonucléase AP 1), clive les sites abasiques générés suite aux traitements par les agents alkylants et initie la réparation de la coupure simple-brin. Ce mécanisme représente une source majeure de chimiorésistance dans certains cancers. Plusieurs études ont ainsi validé la voie BER et plus particulièrement APE1 comme une cible importante dans le but d’améliorer l’efficacité des agents anticancéreux; pour ces raisons, de nombreux inhibiteurs d’APE1 ont été développés. Cependant, à la place d’une inhibition directe de l’enzyme, une stratégie alternative consiste à cibler le substrat de cette dernière : les sites abasiques. Les composés macrocycliques de type naphtalénophane ont montré la capacité à se lier fortement et sélectivement aux sites abasiques. Ce processus interfère avec la reconnaissance de ces derniers par APE1 et conduit in vitro à deux effets : l’inhibition du clivage enzymatique d’APE1 et le clivage du site AP par les macrocycles par un mécanisme différent de celui d'APE1, de type β-élimination. Ainsi, une nouvelle série de naphtalénophanes fonctionnalisés, composée de neuf nouveaux dérivés, a été synthétisée et étudiée. La plupart des macrocycles démontre la capacité à se lier fortement et sélectivement aux sites abasiques de l’ADN ainsi qu’à inhiber l’activité d’APE1 in vitro, avec des constantes d’inhibition s'étalant de 39 nM à 25 µM. De plus, l’activité d’inhibition d’APE1 par les ligands, caractérisée par les valeurs de Kı, a pu être corrélée avec leur affinité et leur sélectivité pour les sites abasiques. La structure moléculaire des macrocycles montre une forte influence sur l’activité de clivage de ces derniers pouvant conduire à une abolition ou à une très haute activité de clivage des sites abasiques. De façon intéressante, la formation d’un adduit covalent ADN – ligand avec un des macrocycles a été caractérisée. Enfin, l’activité biologique des naphtalénophanes sur la lignée cellulaire de glioblastome T98G résistante au TMZ a été étudiée. La plupart des ligands affiche une cytotoxicité élevée, avec des GI₅₀ de l’ordre du micromolaire. De plus, un remarquable effet synergique lors du traitement des cellules avec le TMZ et le MMS en combinaison avec un ligand (2,7-BisNP-O4Me) a été démontré. Ce macrocycle augmente également le nombre de sites abasiques et le nombre de coupures double-brins après un co-traitement cellulaire avec les agents alkylants suggérant ainsi l'inhibition d'APE1 attendue. Ces résultats mettent ainsi en évidence le fort intérêt thérapeutique de ce composé.

Guanine-rich nucleic acid sequences can generate four-stranded, noncanonical secondary structures called G-quadruplexes (G4). G4 structures have been identified in the genomes of different species and are involved in physiological processes. Numerous putative G4 sequences (PQS) have been mined by G4 predicting algorithms and G4 chromatin immunoprecipitation sequencing (ChIP-seq). However, finding a sequence relatively rich in guanines does not necessarily mean it can form a G4 structure. The main aim of this PhD thesis is to develop novel in vitro high-throughput G4 characterization assays, suitable for validating a vast amount of PQS identified by in silico G4 prediction methods and ChIP-seq. Förster Resonance Energy Transfer (FRET) is a powerful tool in characterizing intramolecular G4 structures that were dual-labeled by a chromophore (i.e. FAM) and a corresponding quencher (i.e. Tamra): at the low temperature, the two ends of the G4 are in close proximity and the FAM fluorescence is quenched via a FRET mechanism; with the temperature increasing, G4 unfolds and two ends are split apart, restoring gradually FAM fluorescence. Melting temperature (Tm) of a G4 structure could be calculated based on the FRET-melting curve. In this context, a new methodology called the FRET-melting competition (FRET-MC) assay has been developed, which allows to follow 48 duplicated samples within 2 hours. FRET-MC is based on the competitive binding of a selective G4 ligand (PhenDC3) between a labeled G4 reporter (FAM-Tel21-TAMRA, F21T) and an unknown competitor present in a large excess. PhenDC3 stabilizes the G4 structure of F21T and leads to an increase of the Tm of F21T of about 23 °C. An excess of G4 competitors can trap PhenDC3, leading to a decrease of the Tm of F21T back to the Tm value of F21T alone. In contrast non-G4 competitors have little influence on the F21T-PhenDC3 interaction. FRET-MC works well in most cases, but cannot be used to pick up G4s with low thermal stability, as these weak G4 behave as ssDNA at the temperature where F21T starts to melt. For this reason, an alternative isothermal FRET assay compatible with weakly stable G4s was developed. Iso-FRET exploits two labeled RNA probe strands: one of them 37Q (37-quencher) form a G4 thanks to the binding of PhenDC3, and the second one F22 (FAM-22) is partially complementary to 37Q. To this system an unlabeled competitor sequence is added. When the competitor is not a G4, PhenDC3 remains bound to 37Q and stabilizes its G4 structure, preventing duplex formation between 37Q and F22, allowing a high fluorescence signal of F22. On the contrary, an excess of a G4 competitor traps PhenDC3, which is no longer available to bind to 37Q, which in turns allows this oligonucleotide to hybridize to F22 resulting in fluorescence quenching. Iso-FRET can be performed at 25 °C and 37 °C (physiological temperature), which are acceptable to thermolabile G4 competitors. In parallel, a study focusing on the conformation of G4-prone GC-rich sequences which are rich both in cytosines and guanines was conducted. The structures of GC-rich sequences were dependent on cytosine content and [K⁺] / [Na⁺] ratio in buffer. Spectroscopic results shown that the G4 structure (CEB25) tolerates two three continuous cytosines tracks in potassium / sodium mixed buffers containing 40 mM or higher KCl concentration, implying that cytosine contained G4-forming sequences can still adopt a G4 structure in the intracellular environment. UV-melting results evidenced that the presence of cytosines in G4 loops does not decrease G4s stability, but rather increases the probability of forming a competing structure, either a hairpin or a duplex. As bioinformatics analysis indicates that the majority of cytosine-contained PQS located at pre-mRNAs, it will be interesting to transpose our experiments on DNA sequences to RNA motifs, in order to determine if the latter is more or less prone to hairpin / G4 competition.