Dans ce travail, nous nous sommes intéressés à l‘étude, par spectroscopie optique, des interactions que peuvent avoir des acides nucléiques entre eux ou avec des surfaces de différentes natures telles que des protéines ou des nanoparticules inorganiques. Dans une première partie, nous présentons une étude structurale, réalisée par spectroscopies UV et IR sur des séquences nucléotidiques contenues dans DsrA, un ARN non-codant. Plusieurs modèles d‘auto-association sont proposés. Dans une deuxième partie, l‘association de la protéine Hfq avec un duplexe dA₂₀-dT₂₀ est étudiée par spectroscopie IR. Cette interaction se produit via une ouverture partielle de la double-hélice et une transition conformationnelle des sucres du brin purine. Dans une troisième et quatrième partie, nous nous intéressons à la vectorisation d‘oligonucléotides et de peptides. Différentes stratégies de fonctionnalisation de surface de nanoparticules de maghémite sont explorées et l‘efficacité des nano-vecteurs est évaluée in vitro. / In this work, we were interested in the study, by optical spectroscopy, of the interactions that may have nucleic acids between them or with surfaces of various types such as proteins, or inorganic nanoparticles. In the first part, we present a structural study by UV and IR spectroscopy on nucleotide sequences contained in DsrA, a non-coding RNA. Several models of self-association are proposed. In the second part, the association of the Hfq protein with a duplex dA₂₀-dT₂₀ is studied by IR spectroscopy. This interaction occurs via a partial opening of the double-helix and a conformational transition of the sugars of the purine strand. In the third and fourth part, we are interested in the vectorization of oligonucleotides and peptides. Different strategies of maghemite nanoparticles surface functionalization are explored and the effectiveness of nano-vectors is assessed in vitro.

A system has been developed for converting raster images into vector images. Raster images are made of pixels, while vector images are made of smoother shapes. The image is first segmented, and the segments are layered. The boundary of each segment is approximated with a periodic B-Spline curve. This curve is then used to create a B-Spline surface to approximate the interior of the segment. The algorithm fits each B-Spline to the colors of the image using least-squares approximation. The color and shape of each B-Spline surface are extrapolated into regions behind other segments. The result is a vector image made of layered B-Spline surfaces.

No description available.

Loop vectorization is a powerful optimization technique that can significantly boost the runtime of loops. This optimization depends on functional equivalence between the original and optimized code versions, a requirement typically established through the compiler's static analysis. When this condition is not met, the compiler will miss the optimization. The process of manually rewriting the source code to pass an already missed compiler optimization is time-consuming, given the multitude of potential code variations, and demands a high level of expertise, making it impractical in many scenarios. In this work, we propose a novel framework that aims to take the code blocks that the compiler failed to optimize and transform them to another code block that passes the compiler optimization. We develop an algorithm to efficiently search for a code structure that automatically passes the compiler optimization (weakly verified through a correctness test). We focus on loop-vectorize optimization inside OpenMP directives, where the introduction of parallelism adds complexity to the compiler's vectorization task and is shown to hinder optimizations. Furthermore, we introduce a modified version of TSVC, a loop vectorization benchmark in which all original loops are executed within OpenMP directives. Our evaluation shows that our framework enables " loop-vectorize" optimizations that the compiler failed to pass, resulting in a speedup up to 340× in the blocks optimized. Furthermore, applying our tool to HPC benchmark applications, where those applications are already built with optimization and performance in mind, demonstrates that our technique successfully enables extended compiler optimization, thereby accelerating the execution time of the optimized blocks in 15 loops and the entire execution time of the three applications by up to 1.58 times. / Master of Science / Loop vectorization is a powerful technique for improving the performance of specific sections in computer programs known as loops. Particularly, it simultaneously executes instructions of different iterations in a loop, providing a considerable speedup on its runtime due to this parallelism. To apply this optimization, the code needs to meet certain conditions, which are usually checked by the compiler. However, sometimes the compiler cannot verify these conditions, and the optimization fails. Our research introduces a new approach to fix these issues automatically. Normally, fixing the code manually to meet these conditions is time-consuming and requires high expertise. To overcome this, we've developed a tool that can efficiently find ways to make the code satisfy the conditions needed for optimization. Our focus is on a specific type of code that uses OpenMP directives to split the loop on multiple processor cores and runs them simultaneously, where adding this parallelism makes the code more complex for the compiler to optimize. Our tests show that our approach successfully improves the speed of computer programs by enabling optimizations initially missed by the compiler. This results in significant speed improvements for specific parts of the code, sometimes up to 340 times faster. We've also applied our method to well-optimized computer programs, and it still managed to make them run up to 1.58 times faster.

O objetivo do presente estudo visa o desenvolvimento de um copolímero em bloco constituído por metacrilato de metila (MMA) e de dimetilaminoetila (DMAEMA), tendo como macroniciador poli--caprolactona dibromada (Br-PCL-Br), e que permite formar nanocápsulas sensíveis ao pH, contendo ou não o tripeptídeo leucina-ácido aspártico-valina (LDV) na superfície para a vetorização ativa de anti-neoplásicos. Os métodos envolveram diferentes abordagens sintéticas testadas, sendo que a técnica de transferência eletrônica por regeneração de ativadores (ATRP-ARGET) permitiu obter o copolímero PCL-P(MMA-DMAEMA)2 de forma mais prática e com rendimentos entre 30 e 70%. Por fim, o tripeptídeo LDV foi conjugado ao copolímero por meio do ligante metacrilato de 2-isocianato de etila (IEM). Um método por cromatografia líquida de alta eficiência (CLAE) foi adaptado para a quantificação da doxorrubicina e as nanopartículas foram preparadas por nanoprecipitação e avaliadas quanto à capacidade de expandir em diferentes pHs e citotoxicidade em células de câncer de mama. Os resultados do copolímero demonstram, por análises de infravermelho (IR-FT), sinais característicos em 2900 cm-1 e 1720 cm-1 correspondentes às funções –CH e –C=O. A análise de ressonância magnética nuclear de hidrogênio (RMN 1H) mostra a caracterização das cadeias hidrocarbônicas do copolímero, sendo que os deslocamentos químicos em 2,8 ppm e 3,8 ppm correspondem aos sinais dos grupamentos –CH2-N do DMAEMA e -OCH3 do MMA. As nanocápsulas preparadas a partir do copolímero expandiram de diâmetro quando expostas à pH ácido. Uma vez que o PMMA foi identificado como componente mais citotóxico, o copolímero foi otimizado por meio da redução da quantia de MMA. A quantificação da doxorrubicina encapsulada nas nanopartículas preparadas a partir dos copolímeros não otimizado (ARGET-A) e otimizado (ARGETB) foi de 61,42% e 64,88%, respectivamente. No estudo de citotoxicidade, as nanopartículas preparadas a partir do copolímero ARGET-B apresentaram-se eficazes no controle da proliferação celular de MCF-7. Conclui-se que o método de síntese ATRP-ARGET-B foi o mais apropriado para a produção do copolímero empregado no desenvolvimento de nanopartículas pH responsivas eficazes no 6 controle da proliferação de células tumorais. Ainda, existe a possibilidade do emprego do copolímero contendo o tripeptídeo LDV para alcançar uma vetorização ativa em células de câncer por meio da interação com integrinas específicas. Entretanto, até o presente, não foi realizada a avaliação das nanopartículas contendo LDV. / The objective of the present study looks for the development of a block copolymer constituted by methyl methacrylate (MMA) and dimethylaminoethyl methacrylate (DMAEMA), having poly--caprolactone dibromated (Br-PCL-Br) as a macroinitiator and, that could form pH sensible nanocapsules with or without the tripeptide leucineaspartic acid-valine (LDV) in its surface for active vectorization of anti-neoplasics. The methods employed different synthetic approaches tested, being that the activator regenerated by eletron transfer technique (ATRP-ARGET) allowed to obtain the copolymer PCL-P(MMA-DMAEMA)2 in a practicle way and with incomes between 30 and 70%. Finally, the tripeptide LDV was linked to the copolymer through the 2- isocyanatoethyl methacrylate (IEM). A high performance liquid chromatography method (HPLC) was adapted to doxorubicin quantification and, the nanopartircles were prepared by nanoprecipitation and evaluated conserning its ability to expand in different environments and citotoxycity in mammary cancer cells. The results from the copolymer demonstrated, by infrared (FT-IR), characteristic signals of 2900 cm-1 and 1720 cm-1 from the functions –CH and –C=O. And hydrogen nuclear magnetic resonance (RMN 1H) analysis allowed the characterization of the hydrogen-carbonic chains of the copolymer, being that the chemical displacement in 2,8 ppm and 3,8 ppm corresponds to the signals of the groups –CH2-N from DMAEMA and –O-CH3 from MMA. The nanocapsules prepared from the copolymer expanded its diameter when exposed to acidic pH. Once PMMA was identified as the most toxic component the copolymer was optimized by the reduction of MMA amount. Doxorubicin quantification in the nanocapsules prepared with the copolymers not optimized (ARGET-A) and optimized (ARGET-B) was 61,42% and 64,88%, respectively. In the cytotoxicity study, the nanocapsules prepared from copolymer ARGET-B showed to be efficient to control the cellular proliferation of MCF-7. It can be concluded that the ATRP-ARGET-B method was the more appropriate one for the copolymer production, which was employed in nanocapsules pH responsive effective to control 8 tumor proliferation. Besides, there is the possibility to use the copolymer functionalized with LDV to achieve an active delivery to cancer cells by it interaction with specific integrins. However, till the present, it was not realized the evaluation of the nanocapsules with LDV.

O objetivo do presente estudo visa o desenvolvimento de um copolímero em bloco constituído por metacrilato de metila (MMA) e de dimetilaminoetila (DMAEMA), tendo como macroniciador poli--caprolactona dibromada (Br-PCL-Br), e que permite formar nanocápsulas sensíveis ao pH, contendo ou não o tripeptídeo leucina-ácido aspártico-valina (LDV) na superfície para a vetorização ativa de anti-neoplásicos. Os métodos envolveram diferentes abordagens sintéticas testadas, sendo que a técnica de transferência eletrônica por regeneração de ativadores (ATRP-ARGET) permitiu obter o copolímero PCL-P(MMA-DMAEMA)2 de forma mais prática e com rendimentos entre 30 e 70%. Por fim, o tripeptídeo LDV foi conjugado ao copolímero por meio do ligante metacrilato de 2-isocianato de etila (IEM). Um método por cromatografia líquida de alta eficiência (CLAE) foi adaptado para a quantificação da doxorrubicina e as nanopartículas foram preparadas por nanoprecipitação e avaliadas quanto à capacidade de expandir em diferentes pHs e citotoxicidade em células de câncer de mama. Os resultados do copolímero demonstram, por análises de infravermelho (IR-FT), sinais característicos em 2900 cm-1 e 1720 cm-1 correspondentes às funções –CH e –C=O. A análise de ressonância magnética nuclear de hidrogênio (RMN 1H) mostra a caracterização das cadeias hidrocarbônicas do copolímero, sendo que os deslocamentos químicos em 2,8 ppm e 3,8 ppm correspondem aos sinais dos grupamentos –CH2-N do DMAEMA e -OCH3 do MMA. As nanocápsulas preparadas a partir do copolímero expandiram de diâmetro quando expostas à pH ácido. Uma vez que o PMMA foi identificado como componente mais citotóxico, o copolímero foi otimizado por meio da redução da quantia de MMA. A quantificação da doxorrubicina encapsulada nas nanopartículas preparadas a partir dos copolímeros não otimizado (ARGET-A) e otimizado (ARGETB) foi de 61,42% e 64,88%, respectivamente. No estudo de citotoxicidade, as nanopartículas preparadas a partir do copolímero ARGET-B apresentaram-se eficazes no controle da proliferação celular de MCF-7. Conclui-se que o método de síntese ATRP-ARGET-B foi o mais apropriado para a produção do copolímero empregado no desenvolvimento de nanopartículas pH responsivas eficazes no 6 controle da proliferação de células tumorais. Ainda, existe a possibilidade do emprego do copolímero contendo o tripeptídeo LDV para alcançar uma vetorização ativa em células de câncer por meio da interação com integrinas específicas. Entretanto, até o presente, não foi realizada a avaliação das nanopartículas contendo LDV. / The objective of the present study looks for the development of a block copolymer constituted by methyl methacrylate (MMA) and dimethylaminoethyl methacrylate (DMAEMA), having poly--caprolactone dibromated (Br-PCL-Br) as a macroinitiator and, that could form pH sensible nanocapsules with or without the tripeptide leucineaspartic acid-valine (LDV) in its surface for active vectorization of anti-neoplasics. The methods employed different synthetic approaches tested, being that the activator regenerated by eletron transfer technique (ATRP-ARGET) allowed to obtain the copolymer PCL-P(MMA-DMAEMA)2 in a practicle way and with incomes between 30 and 70%. Finally, the tripeptide LDV was linked to the copolymer through the 2- isocyanatoethyl methacrylate (IEM). A high performance liquid chromatography method (HPLC) was adapted to doxorubicin quantification and, the nanopartircles were prepared by nanoprecipitation and evaluated conserning its ability to expand in different environments and citotoxycity in mammary cancer cells. The results from the copolymer demonstrated, by infrared (FT-IR), characteristic signals of 2900 cm-1 and 1720 cm-1 from the functions –CH and –C=O. And hydrogen nuclear magnetic resonance (RMN 1H) analysis allowed the characterization of the hydrogen-carbonic chains of the copolymer, being that the chemical displacement in 2,8 ppm and 3,8 ppm corresponds to the signals of the groups –CH2-N from DMAEMA and –O-CH3 from MMA. The nanocapsules prepared from the copolymer expanded its diameter when exposed to acidic pH. Once PMMA was identified as the most toxic component the copolymer was optimized by the reduction of MMA amount. Doxorubicin quantification in the nanocapsules prepared with the copolymers not optimized (ARGET-A) and optimized (ARGET-B) was 61,42% and 64,88%, respectively. In the cytotoxicity study, the nanocapsules prepared from copolymer ARGET-B showed to be efficient to control the cellular proliferation of MCF-7. It can be concluded that the ATRP-ARGET-B method was the more appropriate one for the copolymer production, which was employed in nanocapsules pH responsive effective to control 8 tumor proliferation. Besides, there is the possibility to use the copolymer functionalized with LDV to achieve an active delivery to cancer cells by it interaction with specific integrins. However, till the present, it was not realized the evaluation of the nanocapsules with LDV.

In the context of urban buildings, architectural floor plans describe a building's structure and spatial distribution. These digital documents are usually shared in file formats that discard the semantic information related to walls and rooms. This work proposes a new method to recover the structural information by extracting walls and detecting rooms in 2D floor plan images, aimed at multi-unit floor plans which present challenges of higher complexity than previous works. Our proposed approach is able to handle overlapped floor plan elements, notation variations and defects in the input image, and its speed makes it suitable for real applications on both desktop and mobile devices. We evaluate our methods in terms of precision and recall against our own annotated dataset of multi-unit floor plans. / Graduate

O objetivo do presente estudo visa o desenvolvimento de um copolímero em bloco constituído por metacrilato de metila (MMA) e de dimetilaminoetila (DMAEMA), tendo como macroniciador poli--caprolactona dibromada (Br-PCL-Br), e que permite formar nanocápsulas sensíveis ao pH, contendo ou não o tripeptídeo leucina-ácido aspártico-valina (LDV) na superfície para a vetorização ativa de anti-neoplásicos. Os métodos envolveram diferentes abordagens sintéticas testadas, sendo que a técnica de transferência eletrônica por regeneração de ativadores (ATRP-ARGET) permitiu obter o copolímero PCL-P(MMA-DMAEMA)2 de forma mais prática e com rendimentos entre 30 e 70%. Por fim, o tripeptídeo LDV foi conjugado ao copolímero por meio do ligante metacrilato de 2-isocianato de etila (IEM). Um método por cromatografia líquida de alta eficiência (CLAE) foi adaptado para a quantificação da doxorrubicina e as nanopartículas foram preparadas por nanoprecipitação e avaliadas quanto à capacidade de expandir em diferentes pHs e citotoxicidade em células de câncer de mama. Os resultados do copolímero demonstram, por análises de infravermelho (IR-FT), sinais característicos em 2900 cm-1 e 1720 cm-1 correspondentes às funções –CH e –C=O. A análise de ressonância magnética nuclear de hidrogênio (RMN 1H) mostra a caracterização das cadeias hidrocarbônicas do copolímero, sendo que os deslocamentos químicos em 2,8 ppm e 3,8 ppm correspondem aos sinais dos grupamentos –CH2-N do DMAEMA e -OCH3 do MMA. As nanocápsulas preparadas a partir do copolímero expandiram de diâmetro quando expostas à pH ácido. Uma vez que o PMMA foi identificado como componente mais citotóxico, o copolímero foi otimizado por meio da redução da quantia de MMA. A quantificação da doxorrubicina encapsulada nas nanopartículas preparadas a partir dos copolímeros não otimizado (ARGET-A) e otimizado (ARGETB) foi de 61,42% e 64,88%, respectivamente. No estudo de citotoxicidade, as nanopartículas preparadas a partir do copolímero ARGET-B apresentaram-se eficazes no controle da proliferação celular de MCF-7. Conclui-se que o método de síntese ATRP-ARGET-B foi o mais apropriado para a produção do copolímero empregado no desenvolvimento de nanopartículas pH responsivas eficazes no 6 controle da proliferação de células tumorais. Ainda, existe a possibilidade do emprego do copolímero contendo o tripeptídeo LDV para alcançar uma vetorização ativa em células de câncer por meio da interação com integrinas específicas. Entretanto, até o presente, não foi realizada a avaliação das nanopartículas contendo LDV. / The objective of the present study looks for the development of a block copolymer constituted by methyl methacrylate (MMA) and dimethylaminoethyl methacrylate (DMAEMA), having poly--caprolactone dibromated (Br-PCL-Br) as a macroinitiator and, that could form pH sensible nanocapsules with or without the tripeptide leucineaspartic acid-valine (LDV) in its surface for active vectorization of anti-neoplasics. The methods employed different synthetic approaches tested, being that the activator regenerated by eletron transfer technique (ATRP-ARGET) allowed to obtain the copolymer PCL-P(MMA-DMAEMA)2 in a practicle way and with incomes between 30 and 70%. Finally, the tripeptide LDV was linked to the copolymer through the 2- isocyanatoethyl methacrylate (IEM). A high performance liquid chromatography method (HPLC) was adapted to doxorubicin quantification and, the nanopartircles were prepared by nanoprecipitation and evaluated conserning its ability to expand in different environments and citotoxycity in mammary cancer cells. The results from the copolymer demonstrated, by infrared (FT-IR), characteristic signals of 2900 cm-1 and 1720 cm-1 from the functions –CH and –C=O. And hydrogen nuclear magnetic resonance (RMN 1H) analysis allowed the characterization of the hydrogen-carbonic chains of the copolymer, being that the chemical displacement in 2,8 ppm and 3,8 ppm corresponds to the signals of the groups –CH2-N from DMAEMA and –O-CH3 from MMA. The nanocapsules prepared from the copolymer expanded its diameter when exposed to acidic pH. Once PMMA was identified as the most toxic component the copolymer was optimized by the reduction of MMA amount. Doxorubicin quantification in the nanocapsules prepared with the copolymers not optimized (ARGET-A) and optimized (ARGET-B) was 61,42% and 64,88%, respectively. In the cytotoxicity study, the nanocapsules prepared from copolymer ARGET-B showed to be efficient to control the cellular proliferation of MCF-7. It can be concluded that the ATRP-ARGET-B method was the more appropriate one for the copolymer production, which was employed in nanocapsules pH responsive effective to control 8 tumor proliferation. Besides, there is the possibility to use the copolymer functionalized with LDV to achieve an active delivery to cancer cells by it interaction with specific integrins. However, till the present, it was not realized the evaluation of the nanocapsules with LDV.

La Cathepsine D (CathD) est une protéase lysosomale surexprimée et sécrétée par de nombreuses tumeurs solides. Cette enzyme favorise la prolifération tumorale et le processus métastatique, faisant d’elle une cible intéressante pour la thérapie anticancéreuse. Il existe un très bon inhibiteur de la CathD, la pepstatine, mais celui-ci traverse trop difficilement la membrane cellulaire pour être actif. C’est pourquoi des vecteurs de pénétration cellulaire, basés sur l’oligomérisation de mimes contraints de dipeptide, ont été développés au laboratoire. De ce travail, un bioconjugué, le JMV4463, a été développé. Composé d’un vecteur AMPA4 (tétramère de l’acide 2-(aminométhyl)phénylacétique), de la pepstatine, et d’une partie hydrophile aidant à la solubilisation du conjugué, ce bioconjugué est capable d’entrer dans les cellules et possède une activité antiproliférative sur différentes lignées de cellules cancéreuses. Partant de ce travail, la synthèse de nouveaux vecteurs potentiels de pénétration cellulaire, oligomères de mimes contraints de dipeptide, a été réalisée. Leur capacité d’internalisation a été établie sous la forme de conjugués avec la pepstatine. Les monomères envisagés étaient des analogues des motifs AMPA et acide (S)-2-(3-amino-4-oxo-3,4-dihydrobenzo[b][1,4]thiazépin-5(2H)-yl)acétique (DBT), motif ayant également montré une importante capacité d’internalisation. Ainsi, des analogues de l’AMPA en série indole, pyrrole et cyclohexane ont été préparés. De plus, la synthèse d’analogues de l’AMPA possédant différentes substitutions a été réalisée, tels que les méta- et para-AMPA, un homologue inférieur ou les 4,5-diméthoxy-, 4,5-dihydroxy- et 4-bromo-AMPA. Un analogue en série benzodiazépine du DBT a également été étudié. Tous les vecteurs correspondants sont capables d’internaliser la pepstatine dans les cellules mais, de façon surprenante, aucun des conjugués n’a montré d’activité anti-proliférative, indiquant un rôle essentiel du vecteur AMPA4. Une étude des relations structure / activité du JMV4463 a confirmé que le vecteur AMPA4 possédait un mode d’action unique mais celui-ci n’a pu être identifié à ce jour. Dans cette étude, il a également été montré qu’il était possible de simplifier la structure de la pepstatine, en remplaçant un de ses motifs statine par un motif gamma-alanine, permettant un coût de production réduit. Enfin, une étude in vivo sur un modèle xénogreffé de souris a montré un fort potentiel anti-tumoral du conjugué JMV4463. / Cathepsin D (CathD) is an overexpressed lysosomal protease secreted by several solid tumors. This enzyme is involved in tumor proliferation and metastasis, which makes it a promising target for cancer therapy. There exists a potent CathD inhibitor called Pepstatin, but it presents a too poor ability to cross the plasmic membrane in order to be active. To overcome this drawback, the development of cell penetrating vectors based on the oligomerization of constrained dipeptide mimetics has been undertaken in our lab. From this work, the bioconjugate JMV4463 has been developed. This bioconjugate is made of an AMPA4 (2-(aminomethyl)phenylacetic acid) vector, the pepstatin and a hydrophilic part which increases its solubility. This bioconjugate is able to penetrate the cells and has an antiproliferative effect on different cancer cell lines. Based on this work, synthesis of new potential cell penetrating vectors, oligomers of constrained dipeptide mimetics, was performed. Their internalization ability was established as their conjugate with pepstatin. We developed analogues of AMPA scaffold and (S)-2-(3-amino-4-oxo-3,4-dihydrobenzo[b][1,4]thiazépin-5(2H)-yl)acetic acid (DBT), which showed an important internalization ability. Thus, AMPA derivatives in indole, pyrrole and cyclohexane series were prepared. Furthermore, the synthesis of AMPA analogues having different substitutions on its aromatic part was performed, such as meta- and para- AMPA, an inferior homologue or 4,5-dimethoxy-, 4,5-dihydroxy- and 4-bromo-AMPA. An analogue of DBT in the benzodiazepin series has also been studied. All the corresponding vectors were able to internalize the pepstatin into the cells. Surprisingly, none of the conjugates showed an anti-proliferative activity, indicating the essential role of the AMPA vector. A structure/activity relationships study was performed for JMV4463. The obtained results confirmed that the AMPA vector has a unique mode of action that has not been identified yet. Furthermore, it has been demonstrated that it was possible to simplify the pepstatin structure by replacing one of its statin residues by a gamma alanine, which allows decreasing the production cost. Finally, an in vivo study on a xenografted mouse model showed a high anti-tumour potential for the JMV4463 conjugate.

Les oligonucléotides sont de courtes séquences d’acide nucléique qui ont la capacité d’inhiber ou de moduler l’expression d’un gène cible par différents mécanismes. Cependant, leur potentiel thérapeutique est limité par leur faible internalisation. Pour pallier ce problème de vectorisation, il a été envisagé de conjuguer les oligonucléotides à des molécules biocompatibles. Ce travail de thèse porte sur des bioconjugués composés d’un nucléolipide à l’extrémité 5’ ou 3’ des oligonucléotides. Tout d’abord, les propriétés physico-chimique et d’hybridation de ces nouveaux composés ont été évaluées. Des études biologiques ont ensuite été réalisées in vitro et in vivo. Les résultats obtenus ont permis de mettre en évidence le mécanisme d’internalisation cellulaire mais aussi de prouver l’efficacité de transfection et d’inhibition de ces conjugués. En outre, le caractère amphiphile de ce type de composé rend possible leur auto-assemblage pour la formulation de substances actives. Dans ce cadre, différentes formulations ont été investiguées. Ainsi, dans ce travail de thèse une nouvelle technologie d’oligonucléotides conjugués a été développée. La séquence de ces molécules peut être modulée et maitrisée de manière à l’adapter à la cible thérapeutique visée. Actuellement, ce système est appliqué sur treize projets différents à l’échelle nationale. L’effet thérapeutique est ainsi évalué dans différentes pathologies telles que des cancers hormono-dépendants, des leucémies, des maladies neurologiques chroniques ou encore la résistance aux antibiotiques. / Oligonucleotides are short nucleic acid molecules able to modulate or inhibit gene expression. However the main drawback of oligonucleotides lies in their poor cellular internalization, which limit their therapeutic applications. Herein, to overcome this limitation, oligonucleotides were conjugated to biocompatible molecules as a nucleolipid to either the 5'- or the 3'-end. First, physico-chemical properties and binding behaviour of this newly compound were investigated. Then in vitro and in vivo biological assays were performed to characterize but also understand the cellular internalization pathways and their biological activities. Finally, the amphiphilic nature of the oligonucleotide-nucleolipid confers spontaneously self-assembling properties for drugs loading and vectorization purposes. This Ph.D. thesis focuses on new oligonucleotide bioconjugates for various biological applications. Sequences of nucleotides can also be modulated to specifically bind to the therapeutic target. This tuneable technology is actually used in 13 different projects, including hormone-dependent cancers, leukemia, chronic neurological disorders and antibiotic resistance, for example.