Orientador: Marcos Hikary Toyama / Resumo: A giroxina é a principal serino protease encontrada no veneno total de Crotalus durissus terrificus. As serino proteases de veneno de serpentes são proteínas que apresentam uma grande similaridade estrutural com a trombina humana incluindo a alta conservação do sitio catalítico. Os resμltados já obtidos pelo nosso grupo revelaram a capacidade dos flavonoides glicosilados de Laguncularia racemosa em reduzir de forma significativa a atividade enzimática das trombinas humanas. A identificação destas molécμlas revelou que dois flavonoides glicosilados (quercetina-3-O-arabinosidio (QAra) e quercetina-3-O-ramnosidio (QRham)) respondem pela inibição da trombina (Rodrigues et al., 2015). Sabe-se que a soroterapia antiveneno não consegue neutralizar de forma eficiente a ação farmacológica das serino proteases e consequentemente não é capaz de neutralizar os danos homeostáticos e teciduais. Neste trabalho, uma nova serino protease foi isolada do veneno total de Crotalus durissus terrificus (cdtsp-2) através de quatro estapas de separação, usando a superdex 75 com uma fase móvel de bicarbonato de amônia pH 8, a cdtsp-2 Estas duas frações foram então fracionadas em coluna de troca iônica em DEAE para separação das principais frações e posteriormente fracionadas em CLAE de fase reversa para confirmar o grau de homologia molecular e em todas as etapas cromatográfica foram acompanhadas pelos ensaios enzimáticos em BApNA. A nova serino protease denominada como Cdtsp-2 (Crotalus durissus terr... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Gyroxin is the major serine protease present in Crotalusdurissusterrificus venom. Serine proteases from snake venom are proteins, which exhibit a great structural similarity to hμMan thrombin including the catalytic site conservation. The resμlts already obtained by our group revealed the glycosylated flavonoidsability of Laguncμlariaracemosato significantly reduce the enzymatic activity of hμMan thrombins. The molecμles identification reveals two glycosylated flavonoids (quercetin-3-O-arabinosidiμM (QAra) and quercetin-3-O-rhamnosidiμM(QRham)), which inhibit thrombin (Rodrigues et al., 2015). Antivenomserotherapycan not efficiently neutralize the serine proteases pharmacological action and is not able to neutralize homeostatic and tissue damage. In this work, a new serine protease was isolated from Crotalusdurissusterrificusvenom (cdtsp-2) by throμgh four separation stages , using superdex 75 with a mobile phase of ammoniμM bicarbonate pH 8. These two fractions were fractionated on ion exchange colμMn in DEAE for separation of the major fractions and subsequently fractionated in reverse phase CLAE to confirm the degree of molecμlar homology.All chromatographic steps were monitored by the enzymatic assays in BApNA. The new serine protease called cdtsp-2 (Crotalusdurissusterrificus serine protease 2) is an unpublished protein, with a molecμlar mass of 27 kDa and its N-terminal sequence showed structural differences when compared to the gyroxin of Crotalusdurissusterrificus. Re... (Complete abstract click electronic access below) / Mestre

La thérapie photodynamique (PDT pour Photodynamic Therapy) met en jeu des molécules nommées photosensibilisateurs (PS), de l'oxygène et de la lumière. Les PS, non cytotoxiques à l'obscurité, produisent des espèces réactives de l'oxygène (ROS) lorsqu'ils sont excités avec une longueur d'onde appropriée en présence d'oxygène. Les ROS regroupent les radicaux de l'oxygène et l'oxygène singulet (1O2), qui est la principale forme de ROS formés lors du processus de PDT. En présence de tissus vivants, l'1O2 va conduire à la mort cellulaire par apoptose ou par nécrose. Pour améliorer l'efficacité photodynamique, une des pistes étudiées par la communauté scientifique consiste à améliorer la sélectivité du traitement. Le traitement des tumeurs primaires malignes du cerveau, dont le glioblastome multiforme (GBM ou astrocytome de grade IV) est la forme la plus agressive, reste un challenge. Lorsqu'elle est possible, la chirurgie occupe une place prépondérante. L'exérèse ne concerne que la partie volumineuse centrale de la tumeur, tandis que la zone périphérique infiltrante est, quant à elle ciblée par des traitements supplémentaires. Malgré les progrès de la neurochirurgie et de la radiothérapie, l'espérance de vie à 5 ans ne dépasse pas 10%. La thérapie photodynamique se présente comme une alternative thérapeutique grâce aux améliorations apportées par le contrôle local. Pour traiter le gliobastome par PDT, une première approche a consisté à coupler un peptide, à un photosensibilisateur (la chlorine) via un bras espaceur Ahx (acide aminohexanoïque). Le peptide utilisé (ATWLPPR) est un ligand spécifique du récepteur neuropiline 1 (NRP-1). NRP-1 est lui-même un co-récepteur au récepteur du facteur vasculaire de croissance endothéliale (VEGFR) qui est surexprimé au niveau des néovaisseaux et qui favorise la néoangiogenèse au cours du développement des tumeurs solides. L'assemblage PS-Ahx-ATWLPPR a montré une stabilité peptidique in vivo et in vitro avec une bonne pharmacocinétique et une bonne biodistribution. Ses efficacités anti-tumorales et anti-vasculaires ont notamment été prouvées. Cependant, in vivo, le peptide ATWLPPR montrait une dégradation par le système réticulo-endothélial et l'assemblage présentait une affinité moindre pour NRP-1 par rapport au peptide seul. Afin de résoudre ces problèmes, une nouvelle stratégie décrite dans cette thèse a consisté à développer des nanoparticules multifonctionnelles. Ces nanoparticules sont constituées d'un coeur d'oxyde de gadolinium (Gd2O3) pour permettant un réhaussement de contraste positif en IRM, enrobé d'une couche de polysiloxane biocompatible dans laquelle est greffé le photosensibilisateur par liaison amide. La nanoparticule est ensuite fonctionnalisée en surface avec des agents chélatants (DOTA, DTPA) par l'intermédiaire de fonctions amines libres de la couche de polysiloxane. Les peptides de type ATWLPPR sont greffés sur les agents chélatants, ce qui permet un ciblage spécifique de NRP-1. De cette façon, on obtient des nanoparticules qui offrent à la fois une possibilité de ciblage actif des néovaisseaux tumoraux, de visualisation par IRM et un effet PDT. Dans l'objectif d'obtenir un effet PDT optimal, une augmentation du contraste en imagerie IRM et une sélectivité maximale pour les cellules endothéliales, un plan d'expérience a été élaboré. Chaque lot du plan d'expérience a été synthétisé en faisant varier la composition chimique du coeur, l'épaisseur de la couche de polysiloxane, le nombre de photosensibilisateurs, le type de surfactant, le nombre et le type de peptides. Une fois la synthèse et la purification de ces nanoparticules effectuées, chaque lot a été caractérisé pour vérifier la conservation des propriétés photophysiques, en particulier la formation d'oxygène singulet. Des études biologiques sur des cellules tumorales de type MDA-MB-231 et U87 ont été réalisées, pour étudier la cytototoxicité, la phototoxicité et le réhaussement de contraste IRM de ces nanoparticules / Photodynamic therapy (PDT) involves molecules called photosensitizers (PS), molecular oxygen and light. PS are non-cytotoxic in the dark but produce reactive oxygen species (ROS) when they are excited with light of an appropriate wavelength in the presence of oxygen. ROS include oxygen radicals and singlet oxygen (1O2), which is the main form of ROS formed during PDT processes. In the presence of living tissue, 1O2 leads to cell death by apoptosis or necrosis. To improve photodynamic efficiency, a strategy developed by scientists consists in improving the selectivity of the treatment. The treatment of primary malignant brain tumors, including glioblastoma multiforme (GBM or astrocytoma level IV) which is the most aggressive form, remains a challenge. When it is possible, surgery is performed by removing the central volume of the tumor, while infiltrating peripheral zone is treated by additional treatments. Despite advances in neurosurgery and radiotherapy, the life expectancy at 5 years after the tumor detection does not exceed 10 %. PDT appears as an alternative treatment. In preliminary study a photosensitizer (chlorin) coupled to a peptide (ATWLPPR) through an Ahx linker (aminohexanoic acid) has been designed. The peptide is a specific ligand of neuropilin-1 receptor (NRP-1). NRP-1 is a co-receptor of vascular endothelial growth factor receptor (VEGFR) overexpressed in neovessels and which promotes the formation of new vessels during the development of solid tumors. This targeted photosensitizer presented a peptidic stability in vivo and in vitro with good pharmacokinetic and biodistribution. Its anti-tumor and anti-vascular efficiencies have been proven. However, the ATWLPPR peptide showed degradation in the reticuloendothelial system (RES) and a reduced affinity for NRP-1 compared with peptide alone. To solve these problems, a new strategy using multifunctional nanoparticles has been developed in this thesis. The nanoparticles consist of a core of gadolinium oxide (Gd2O3) for MRI contrast, coated with a layer of biocompatible polysiloxane wherein the photosensitizer is covalenty grafted. The nanoparticle surface is functionalized by chelating agents (DOTA, DTPA) via free amine functions of the polysiloxane layer. ATWLPPR peptides are grafted on chelating agents, which allows specific targeting of NRP-1. Nanoparticles allow a MRI visualization, a PDT effect and an active targeting of the tumor neovasculature. With the aim to obtain an optimal PDT effect, an enhancement of contrast in MRI imaging and a high selectivity for endothelial cells, an experimental design has been developed. Each batch of the experimental design was synthesized with various chemical compositions of the core, the size of the polysiloxane layer, the number of photosensitizers, the number and the type of peptides and the type of surfactant. Once the synthesis and purification of these nanoparticles done, each batch was characterized to ensure the conservation of the photophysical properties, in particular the formation of the singlet oxygen. Biological studies on tumor cell type MDA- MB-231 and U87 were carried out, especially their cytototoxicity and phototoxicity

圖形化模型編輯器是模型編輯器開發長久以來一直努力的方向。Eclipse.org 已提供兩個功能強大的圖形化編輯器設計框架：GEF 圖形化編輯器設計框架，和將 EMF 與 GEF 結合並簡化其設計流程的 GMF 圖形化模型編輯器設計框架，來協助開發者發展圖形化模型編輯器。 / 使用設計框架時所面臨的高學習門檻是開發者選用設計框架上的一大障礙。為降低開發者在 GMF 使用上的學習門檻，我們希望能提供一個以模型開發為導向的圖形編輯器設計專屬模型語言，讓程式設計師能夠以其慣用的程式設計角度來描述其圖形化模型編輯器的組成架構。 / 本篇研究主要是在探討 GMF 圖形化模型編輯器開發流程的簡化。在研究中希望藉由提供 GMF 圖形化模型編輯器設計框架的 GM3 專屬模型語言，讓開發者能夠透過 MDA 的角度來開發圖形化模型編輯器的架構，藉以縮短 GMF 的開發流程。實驗中使用 JavaCC 文法剖析器產生工具來開發 GM3專屬模型語言與 GMF 各項模型定義間的模型轉換。 / The availability of a visual graphical editor for a target domain is the prerequisite of visual graphical modeling, which has been adopted by classical software development for decades and is especially emphasized in today's model-driven engineering. However, compared with traditional textual editors, developing a visual graphical editor from scratch is not an easy work. As a result, there were frameworks developed such as GEF and GMF aimed to simplify the construction of graphical editors. Even so, however, it is still though hard for an average programmer to construct a visual graphical editor by using these frameworks without a long time of learning. / Our result is a modeling langauge of graphcial editors called GM3, serving as a bridge betwen developers of graphical editors and the GMF framework. With GM3, the developer can specify the model of his editing domain, how each model element should be presented by which kind of graphical elements in the editor, and how the tool pallete should be filled with various kind of creation and manipulation tools for model elements. After the GM3 specification of an editor is produced, the GM3 transformation engine developed by us using the JavaCC parser generator can be used to generate all files required of the GMF framework and, finally, a subsequent application of the standard GMF code generation procedure can produce a complete graphical editor on Eclipse platform.

Model-driven software development (MDSD) and component-based software development are both paradigms for reducing complexity and for increasing abstraction and reuse in software development. In this thesis, we aim at combining the advantages of each by introducing methods from component-based development into MDSD. In MDSD, all artefacts that describe a software system are regarded as models of the system and are treated as the central development artefacts. To obtain a system implementation from such models, they are transformed and integrated until implementation code can be generated from them. Models in MDSD can have very different forms: they can be documents, diagrams, or textual specifications defined in different modelling languages. Integrating these models of different formats and abstraction in a consistent way is a central challenge in MDSD. We propose to tackle this challenge by explicitly separating the tasks of defining model components and composing model components, which is also known as distinguishing programming-in-the-small and programming-in-the-large. That is, we promote a separation of models into models for modelling-in-the-small (models that are components) and models for modelling-in-the-large (models that describe compositions of model components). To perform such component-based modelling, we introduce two architectural styles for developing systems with component-based MDSD (CB-MDSD). For CB-MDSD, we require a universal composition technique that can handle models defined in arbitrary modelling languages. A technique that can handle arbitrary textual languages is universal invasive software composition for code fragment composition. We extend this technique to universal invasive software composition for graph fragments (U-ISC/Graph) which can handle arbitrary models, including graphical and textual ones, as components. Such components are called graph fragments, because we treat each model as a typed graph and support reuse of partial models. To put the composition technique into practice, we developed the tool Reuseware that implements U-ISC/Graph. The tool is based on the Eclipse Modelling Framework and can therefore be integrated into existing MDSD development environments based on the framework. To evaluate the applicability of CB-MDSD, we realised for each of our two architectural styles a model-driven architecture with Reuseware. The first style, which we name ModelSoC, is based on the component-based development paradigm of multi-dimensional separation of concerns. The architecture we realised with that style shows how a system that involves multiple modelling languages can be developed with CB-MDSD. The second style, which we name ModelHiC, is based on hierarchical composition. With this style, we developed abstraction and reuse support for a large modelling language for telecommunication networks that implements the Common Information Model industry standard.

Modeliais grįsta sistemų architektūra (MDA) yra „Object Management Group“ (OMG) konsorciumo iniciatyva apibrėžti naują požiūrį į programų sistemų kūrimą remiantis modeliais ir automatizuota jų transformacija į programinį kodą. Siekdama standartizuoti šį požiūrį, OMG patvirtino visą eilę standartų, bet esminiai MDA principai ir praktikos glūdi modeliais grįstame sistemų kūrimo stiliuje, kuris yra fundamentalus programų sistemų inžinerijoje. MDA idėjos, pradžioje sukėlusios didelį entuziazmą IT bendruomenėje, ilgainiui peraugo į skepticizmą ir kai kurie autoriai atvirai pradėjo abejoti modeliais grįsto sistemų kūrimo perspektyva. Šiame darbe analizuojamos praktinio MDA taikymo programų sistemų kūrimo procese aspektai ir galimybės, analizuojami galimi sprendimai ir kliūtys, dėl kurių MDA požiūris gali būti sunkiai pritaikomas. Įvairių autorių įvardijami praktiniai MDA taikymo programų sistemų kūrimo procese sunkumai, didelės tam reikalingos investicijos, pastangos ir resursai, tinkamų instrumentų trūkumas, didelė standartų įvairovė ir sudėtingos technologijos iš dalies lėmė išaugusį skepticizmą MDA paradigmai. MDA apibrėžia naujus sistemų kūrimo principus ir standartais apibrėžia technologijas, kurios pagrindžia šiuos principus. Tai neišvengiamai įtakoja patį programų sistemų kūrimo procesą, kuris turi pasipildyti naujomis veiklomis, praktikomis ir technologijomis, kaip kurti sistemas taikant MDA požiūrį. Šiame kontekste būtinas tam tikras pragmatinis požiūris į MDA ir su juo... [toliau žr. visą tekstą] / Model Driven Architecture (MDA) is an approach to using models in software development, which states that models and model-based transformations are a key part of effective automated software development. The Object Management Group (OMG) has defined standards for representing MDA models, but the principles and practice of MDA are rooted in model-based styles of development that have been fundamental to software engineering from its earliest days. Unfortunately, early enthusiasm for Model Driven Architecture (MDA) has dissipated to the point that many people are openly skeptical of the value of any model-driven approach. This paper examines the practical realities of MDA, difficulties and challenges in adopting an MDA approach to software engineering process. While MDA requires additional efforts and high investment to be adopted in software engineering process, it doesn’t provide any means or guidelines for this. This paper argues that to be successful, a pragmatic MDA approach must be executed in context of a sound Enterprise Architecture providing an integrated business architecture and governance structure that enables an organization to respond to business requirements quickly and appropriately.

導引精靈(Wizard)是一種用於收集用戶端資料的互動式人機介面程式，它是由一系列的對話框組成。使用導引精靈，可以避免資料遺漏、更能確保資料的完整性。由於方便好用，導引精靈不僅大量的使用在既有的電腦程式中，很顯然地，它也應該是許多新應用軟體系統不可或缺的一部分。 為了克服開發導引精靈過程的繁雜，瑣碎與費時，我們曾開發一套遵循MDA架構的多平台導引精靈快速開發系統，稱為 MoDWizII。利用MoDWizII，我們只要輸入高階的平台無關精靈描述，系統即能自動產生各支援平台的對應導引精靈實做。目前系統支援的平台包括：Java、Eclipse 以及Web 應用。 隨著手機與平板電腦的普及與無所不在，其應用系統顯然亦有收集使用者資料的需求。相對於傳統電腦，行動裝置的先天限制並不適合傳統的編輯輸入，因此導引精靈顯然是一種更值得考量的替代選項。此一因素，促使我們考量將MoDWizII的支援對象擴充至行動平台。 本研究的主要任務是擴充MoDWizII的支援對象至行動平台。新系統名稱為MoDWiz3，目前選擇的唯一支援平台是市佔率最高的Android平台。據IDC報告，至2013年第三季時Android的行動裝置佔有率達81%。 為了使得ModWiz3得以產生Android平台的精靈實做，我們在此研究中，貢獻了一套針對Android平台特性而設計的精靈程式Android平台專屬超模型(PSMM)。除此之外，我們亦實做了在MDA架構下，針對Android平台的必要M2M與M2T轉換工具。利用這些工具鏈，我們即能將高階的平台無關精靈描述轉換成Android平台的導引精靈實做。和以往支援平台有一個顯著不同之處是，Android平台本身並無提供立即可用的導引精靈API可供實做時使用，因此本研究的另一貢獻是使用Android fragment機制設計、選用和實做所有系統所需的精靈元件。 / Software wizards are interactive programs consisting of a series of UI dialogs aimed to acquire responded data from the users. Using wizards to collect data has the advantage of ensuring the completeness and integrity of the collected data. They are not only pervasively found in existing computer systems but apparently would also be an indispensable part of many software applications to be developed. In order to overcome the shortage of developing wizards from scratch, which are complicated, tedious and time-consuming, our lab has developed a system called MoDWizII, which can produce multi-platform wizards from a single declarative wizard description by following OMG's MDA approach. Currently the supported platforms include Java, Eclipse and Web applications. With the explosive popularity of mobile phones and tablets, the need of complex data input using mobile devices increases. However, due to the size limitation, it is very inconvenient to use traditional editing approach to input large amount of data on these devices. In contrast, wizards in these use cases seem more suitable. This motivated the inclusion of mobile platforms to the supported platforms of our system. In this research we extend MoDWizII to a new one called MoDWiz3 with the goal of supporting also mobile platforms. Currently the only selected platform is Android, which is used by most people as indicated by a report of International Data Corporation (IDC) that the market share of Android platform was 81% in the 3rd quarter in 2013. To enable the support of Android platform, this research created a definition of PSMM (Platform-Specific MetaModel) for wizards on the Android platform; standard M2M and M2T toolchain was also developed and can be used to generate practical Android wizard applications from a platform independent wizard description. While previous supported platforms have well-established wizard API, it is not the case for Android. We thus included in the research also the design, selection and implementation of all required wizard components for the Android platform.

本研究遵循模型驅動架構的概念，發展一適用於現實商業流程之工作流程模型，採用創新的方式對於工作流程做描述，我們只需要按照固定的步驟，便可以將真實的商業流程發展成可用的模型，以利於後續發展工作流程系統之用。在發展模型時，我們採用了IBM的資訊框架（IFW）做為塑模的方式，以及利用統一塑模語言中的狀態移轉圖和循序圖的概念來做為呈現工作流程模型，解決了傳統流程無法表達流程內部資料變化所帶來的影響，除此之外，我們在模型發展後，採用包含事件-驅動概念的工作流程引擎來產生工作流程系統，我們將商業流程視為一個個的物件，物件收發事件導致狀態的改變而引發另一個事件的發生，如此循環下去，直到整個流程完成。

Απο τις απαρχές της μελέτης του PAR1, είχε βρεθεί οτι βρίσκεται σε στενή συνάφεια με τον καρκίνο, με ποικιλία πειραμάτων που έγιναν σε καρκινικές σειρές και σε διάφορα πειραματικά μοντέλα ζώων. Οι σκοποί της παρούσας εργασίας μπορούν να συνοψιστούν ως εξης: Η διερεύνηση της έκφρασης του υποδοχέα 1 της θρομβίνης (PAR1) σε καρκινικές σειρές προερχόμενες από ανθρώπινους όγκους και συγκεκριμένα: Στις σειρές από καρκίνο του προστάτη PC3 και LNCaP και στις σειρές από καρκίνο του μαστού MDA-231 και MCF-7. Η διερεύνηση της λειτουργικότητας του παραπάνω υποδοχέα στις προαναφερθείσες σειρές και το αποτέλεσμα της αναστολής του υποδοχέα στην επιβίωση και στον πολλαπλασιασμό των μελετώμενων κυττάρων. Η διερεύνηση της ενεργοποίησης της ΜΑΡ κινάσης μέσω του PAR1. Και τελικά, η διερεύνηση της έκφρασης του PAR-1 σε ασθενείς που χειρουργήθηκαν για καρκίνο του πνεύμονα στην Παν/μιακη Καρδιοθωρακοχειρουργική κλινική της Πάτρας, απο τον Καθηγητή Κο Δ. Δουγένη και την ομάδα του. Ο ανταγωνιστής του PAR-1, SCH 79797, προκάλεσε μείωση του κυτταρικού πολλαπλασιασμού των προαναφερθέντων καρκινικών σειρών, όπως μελετήθηκε με τη μέθοδο ΜΤΤ και την ενσωμάτωση ραδιενεργού θυμιδίνης. Αυτή η μη ειδική ανταπόκριση όλων των μελετώμενων σειρών στον SCH, αποδόθηκε κατόπιν στο γεγονός ότι αυτοί οι ανταγωνιστές δεν ήταν απολύτως εκλεκτικοί για τον PAR-1 όπως πιστευόταν, όταν σχεδιάστηκαν. Ο συγκεκριμένος ανταγωνιστής επιλέχθηκε μεταξύ των λίγων, της μοναδικής κατηγορίας που υπήρχε, όταν ξεκίνησαν τα πειράματα. Η θρομβίνη υπερδιπλασίασε τον πολλαπλασιασμό της σειράς PC3 και δεν είχε κανένα αποτέλεσμα στον πολλαπλασιασμό της σειράς MDA-231. Το τελευταίο συμφωνεί και με προηγούμενη έρευνα όπου καταδείχθηκε ότι η θρομβίνη δεν επηρεάζει τον πολλαπλασιασμό, αλλά μειώνει τη μεταστατικότητα της σειράς MDA-231 (Kamath et al., 2001). Η αύξηση του πολλαπλασιασμού των καρκινικών κυττάρων του προστάτη PC3, από τη θρομβίνη γίνεται μέσω ενεργοποίησης του υποδοχέα της PAR-1, και μέσω ενεργοποίησης της ΜΑΡ κινάσης, όπως φάνηκε από τα πειράματα στα οποία χρησιμοποιήθηκε ο ειδικός αγωνιστής του PAR1, το εξαπεπτίδιο SFLLRN. Από κάποια πρώτα ενδεικτικά πειράματα φαίνεται ότι η ενεργοποίηση της ΜΑΡΚ λαμβάνει χώρα μέσω διενεργοποίησης του EGFR, κάτι που έχει αποδειχθεί για άλλους GPCRs. Η έκφραση του PAR1 όπως μελετήθηκε με RT-PCR, σε δείγματα ασθενών που χειρουργήθηκαν για κακοήθεις όγκους στους πνεύμονες, ανιχνεύθηκε σε όλα τα δείγματα καρκινικού ιστού. Η υψηλότερη έκφραση του PAR1 ανιχνεύθηκε στον ασθενή με μελάνωμα και στον ασθενή του υψηλότερου σταδίου. Φυσικά ο αριθμός των ασθενών που μελετήθηκαν δεν αρκεί για εξαγωγή συμπερασμάτων, αλλά τα παραπάνω αποτελέσματα συμφωνούν με τα γνωστά ως σήμερα ευρήματα για τον PAR1. Παραμένει να διευκρινιστεί η σημασία της αυξημένης έκφρασης του PAR1 σε ασθενείς με κακοήθεις όγκους στους πνεύμονες, αφού πρώτα επιβεβαιωθεί αυτή η αυξημένη έκφραση σε μεγαλύτερο αριθμό ασθενών. / From the onset of studies of PAR1, it has been concluded that this receptor is closely related to cancer. This relationship has been established after various experiments in cancer cell lines and in experimental animal models. The purposes of the present study can be summarized as follows: To explore the expression of PAR1 in cell lines established from human solid tumors and specifically PC3 and LNCaP from prostate cancer and MDA-231 and MCF-7 from breast cancer. To explore the suppression of PAR1 to the above cell lines in cell division. To determine if the activation of PAR1 to the above cell lines leads to MAPK phosphorylation. And ultimatilly, to explore the expression of PAR1 in patients that have been operated for tumor in lungs in Patras University Hospitall by Dr. D. Dougenis and colleagues. It was found that PAR1 is strongly expressed in highly metastatic cell lines PC3 and MDA-231, opposite to the cells LNCaP and MCF-7 that have lower metastatic capacity. The finding for the breast cancer cells MDA-231 and MCF-7 was according to published results (Kamath et al., 2001). PAR1 selective antagonist SCH 79797, reduced cell survival and DNA synthesis to all the above mentioned cell lines, independently of PAR1 expression. These non-specific results contributed to the recent fact that these antagonists were not PAR1 selective finally. Thrombin caused more than 100% induction of DNA synthesis in PC3 cells and had no effect in MDA-231 cells in accordance with published results that thrombin reduces the metastatic capacity of MDA-231 cells (Kamath et al., 2001). This effect of thrombin in PC3 cells, is mediated by activation of PAR1 as it was shown with the use of the selective agonist peptide SFLLRN. The activation of PAR1 by thrombin in PC3 cells leads to MAPK activation as it was shown by Western analysis. Furthrmore, preliminary experiments indicate that MAPK phosphorylation after PAR1 activation may be result of EGFR transactivation. In the sample tissues from patients, PAR1 expression was detected in all cancers with different ODs. The number of the samples is not enough to lead to conclusions, but there are some important observations. The highest level of PAR1 expression as was detected by RT-PCR were found to the sample tissues of the patient diagnosed for melanoma and of the patient with the most advanced stage of lung cancer. More patients shoulde be examined and more experiments to be done in order to proceed to conclusions for the significance of PAR1 in lung cancer.

Notre thèse s’inscrit dans le cadre des langages contrôlés pour le génie logiciel. Elle a pour but de faciliter l’adoption de l’approche par règles métier (ARM) par les entreprises en créant un langage contrôlé en vue d’aider à la spécification des règles métier par les experts métier. Notre solution va permettre de réduire la distance sémantique entre les experts métier et les experts système afin de répondre non seulement au besoin d’intercompréhension entre ces derniers mais aussi pour réaliser un transfert automatique de la description des règles métier vers les systèmes d’information (SI). Ce langage contrôlé que nous avons créé permettra d’assurer en plus la consistance et la traçabilité de ces règles avec leur implantation / Our thesis focuses on controlled natural languages (CNL) for software engineering. It aims at facilitating the adoption of the business rule approach (BRA) by companies by creating a CNL in order to help business experts in the specification of their business rules. Our solution will allow reducing the semantic gap between business experts and system experts to meet not only the need for mutual understanding between them but also to achieve an automatic transfer of the description of business rules to information systems (IS). The CNL that we have created will also ensure the consistency and the traceability of these rules together with their implementation

The central theme of the thesis is to study linear and non linear SVM formulations in the presence of uncertain observations. The main contribution of this thesis is to derive robust classfiers from partial knowledge of the underlying uncertainty. In the case of linear classification, a new bounding scheme based on Bernstein inequality has been proposed, which models interval-valued uncertainty in a less conservative fashion and hence is expected to generalize better than the existing methods. Next, potential of partial information such as bounds on second order moments along with support information has been explored. Bounds on second order moments make the resulting classifiers robust to moment estimation errors. Uncertainty in the dataset will lead to uncertainty in the kernel matrices. A novel distribution free large deviation inequality has been proposed which handles uncertainty in kernels through co-positive programming in a chance constraint setting. Although such formulations are NP hard, under several cases of interest the problem reduces to a convex program. However, the independence assumption mentioned above, is restrictive and may not always define a valid uncertain kernel. To alleviate this problem an affine set based alternative is proposed and using a robust optimization framework the resultant problem is posed as a minimax problem. In both the cases of Chance Constraint Program or Robust Optimization (for non-linear SVM), mirror descent algorithm (MDA) like procedures have been applied.