Fibroblast growth factors (FGFs), encompasses a family of 22 related polypeptide. There are 18 different biologically active FGF proteins. They influence a wide array of biological and physiological responses such as migration, proliferation, tissue homeostasis, and wound healing. Most FGF are secreted and bind to heparin sulphate binding proteins (HPSG) in the extra cellular matrix (ECM). FGF-binding protein (FGF-BP) is a chaperon protein that binds to FGF releasing from the ECM and chaperoning it to bind to FGF receptors (FGFRs). FGFR dimerize when bound to FGF and starts series of a signal cascade that ultimately leads to the activation of MAPK. FGF2 and FGF4 are selected from the pool of 18 different FGF to be studied in this investigation. Both FGF2 and FGF4 have been reported to be critical for development during embryogenesis. De-regulation of these proteins could lead to various pathologies including different types of cancers. Hence we attempt to investigate the cellular and molecular role of these proteins and their implication on cells in an attempt to set up a foundational understanding for further studies that will include FGF-BPs and FGFRs. To do so, we used HUVECs to examine FGF2 and FGF4 activity through Western Blot analysis. We also investigated their effect on migration using the ECIS Migration Assays, on wound healing by the ECIS Wound Healing Assays and captured wound healing images through Incucyte. Data from these experiments indicated that both, FGF2 and FGF4, have a role in cellular migration and wound healing. They also show they have a dose response on these cells. As a result, we can use these models to further investigate FGF2 and FGF4 modulation by FGF-BP1 and FGF-BP3 and the affects cellular response.

During the last decades the number of biologics increased dramatically and several biopharmaceutical drugs such as peptides, therapeutic proteins, hormones, enzymes, vaccines, monoclonal antibodies and antibody-drug conjugates conquered the market. Moreover, administration and local delivery of growth factors has gained substantial importance in the field of tissue engineering. Despite progress that has been made over the last decades formulation and delivery of therapeutic proteins is still a challenge. Thus, we worked on formulation and delivery strategies of therapeutic proteins to improve their biological performance. Phase I of this work deals with protein stability with the main focus on a liquid protein formulation of the dimeric fusion protein PR-15, a lesion specific platelet adhesion inhibitor. In order to develop an adequate formulation ensuring the stability and bioactivity of PR-15 during storage at 4 °C, a pH screening, a forced degradation and a Design of Experiments (DoE) was performed. First the stability and bioactivity of PR-15 in 50 mM histidine buffer in relation to pH was evaluated in a short-term storage stability study at 25 °C and 40 °C for 4 and 8 weeks using different analytical methods. Additionally, potential degradation pathways of PR-15 were investigated under stressed conditions such as heat treatment, acidic or basic pH, freeze-thaw cycles, light exposure, induced oxidation and induced deamidation during the forced degradation study. Moreover, we were able to identify the main degradation product of PR-15 by performing LC/ESI-MS analysis. Further optimization of the injectable PR 15 formulation concerning pH, the choice of buffer and the addition of excipients was studied in the following DoE and finally an optimal PR-15 formulation was found. The growth factors BMP-2, IGF-I and TGF-β3 were selected for the differentiation of stem cells for tissue engineering of cartilage and bone in order to prepare multifunctionalized osteochondral implants for the regeneration of cartilage defects. Silk fibroin (SF) was chosen as biomaterial because of its biocompatibility, mechanical properties and its opportunity for biofunctionalization. Ideal geometry of SF scaffolds with optimal porosity was found in order to generate both tissues on one scaffold. The growth factors BMP-2 and IGF-I were modified to allow spatially restricted covalent immobilization on the generated porous SF scaffolds. In order to perform site-directed covalent coupling by the usage of click chemistry on two opposite sides of the scaffold, we genetically engineered BMP-2 (not shown in this work; performed by Barbara Tabisz) and IGF-I for the introduction of alkyne or azide bearing artificial amino acids. TGF β3 was immobilized to beads through common EDC/NHS chemistry requiring no modification and distributed in the pores of the entire scaffold. For this reason protein modification, protein engineering, protein immobilization and bioconjugation are investigated in phase II. Beside the synthesis the focus was on the characterization of such modified proteins and its conjugates. The field of protein engineering offers a wide range of possibilities to modify existing proteins or to design new proteins with prolonged serum half-life, increased conformational stability or improved release rates according to their clinical use. Site-directed click chemistry and non-site-directed EDC/NHS chemistry were used for bioconjugation and protein immobilization with the aim to underline the preferences of site-directed coupling. We chose three strategies for the incorporation of alkyne or azide functionality for the performance of click reaction into the protein of interest: diazonium coupling reaction, PEGylation and genetic engineering. Azido groups were successfully introduced into SF by implementation of diazonium coupling and alkyne, amino or acid functionality was incorporated into FGF-2 as model protein by means of thiol PEGylation. The proper folding of FGF-2 after PEGylation was assessed by fluorescence spectroscopy, WST-1 proliferation assay ensured moderate bioactivity and the purity of PEGylated FGF-2 samples was monitored with RP-HPLC. Moreover, the modification of native FGF-2 with 10 kDa PEG chains resulted in enhanced thermal stability. Additionally, we genetically engineered one IGF-I mutant by incorporating the unnatural amino acid propargyl-L-lysine (plk) at position 65 into the IGF-I amino acid sequence and were able to express hardly verifiable amounts of plk-IGF-I. Consequently, plk-IGF-I expression has to be further optimized in future studies in order to generate plk-IGF-I with higher yields. Bioconjugation of PEGylated FGF-2 with functionalized silk was performed in solution and was successful for click as well as EDC/NHS chemistry. However, substantial amounts of unreacted PEG-FGF-2 were adsorbed to SF and could not be removed from the reaction mixture making it impossible to expose the advantages of click chemistry in relation to EDC/NHS chemistry. The immobilization of PEG-FGF-2 to microspheres was a trial to increase product yield and to remove unreacted PEG-FGF-2 from reaction mixture. Bound PEG-FGF-2 was visualized by fluorescence imaging or flow cytometry and bioactivity was assessed by analysis of the proliferation of NIH 3T3 cells. However, immobilization on beads raised the same issue as in solution: adsorption caused by electrostatic interactions of positively charged FGF-2 and negatively charged SF or beads. Finally, we were not able to prove superiority of site-directed click chemistry over non-site-directed EDC/NHS. The skills and knowledge in protein immobilization as well as protein characterization acquired during phase II helped us in phase III to engineer cartilage tissue in biofunctionalized SF scaffolds. The approach of covalent immobilization of the required growth factors is relevant because of their short in vivo half-lives and aimed at controlling their bioavailability. So TGF-β3 was covalently coupled by means of EDC/NHS chemistry to biocompatible and biostable PMMA beads. Herein, we directly compared bioactivity of covalently coupled and adsorbed TGF-β3. During the so-called luciferase assay bioactivity of covalent coupled as well as adsorbed TGF-β3 on PMMA beads was ensured. In order to investigate the real influence of EDC/NHS chemistry on TGF-β3’s bioactivity, the amount of immobilized TGF-β3 on PMMA beads was determined. Therefore, an ELISA method was established. The assessment of total amount of TGF-β3 immobilized on the PMMA beads allowed as to calculate coupling efficiency. A significantly higher coupling efficiency was determined for the coupling of TGF-β3 via EDC/NHS chemistry compared to the reaction without coupling reagents indicating a small amount of adsorbed TGF-β3. These results provide opportunity to determine the consequence of coupling by means of EDC/NHS chemistry for TGF β3 bioactivity. At first sight, no statistically significant difference between covalent immobilized and adsorbed TGF-β3 was observed regarding relative luciferase activities. But during comparison of total and active amount of TGF-β3 on PMMA beads detected by ELISA or luciferase assay, respectively, a decrease of TGF-β3’s bioactivity became apparent. Nevertheless, immobilized TGF β3 was further investigated in combination with SF scaffolds in order to drive BMSCs to the chondrogenic lineage. According to the results obtained through histological and immunohistochemical studies, biochemical assays as well as qRT-PCR of gene expression from BMSCs after 21 days in culture immobilized TGF-β3 was able to engineer cartilage tissue. These findings support the thesis that local presentation of TGF β3 is superior towards exogenous TGF β3 for the development of hyaline cartilage. Furthermore, we conclude that covalent immobilized TGF β3 is not only superior towards exogenously supplemented TGF-β3 but also superior towards adsorbed TGF-β3 for articular hyaline cartilage tissue engineering. Diffusion processes were inhibited through covalent immobilization of TGF-β3 to PMMA beads and thereby a stable and consistent TGF-β3 concentration was maintained in the target area. With the knowledge acquired during phase II and III as well as during the studies of Barbara Tabisz concerning the expression and purification of plk-BMP-2 we made considerable progress towards the formation of multifunctionalized osteochondral implants for the regeneration of cartilage defects. However, further studies are required for the translation of these insights into the development of multifunctionalized osteochondral SF scaffolds. / In den letzten Jahrzehnten stieg die Zahl der Biologika dramatisch an und mehrere biopharmazeutische Arzneimittel wie Peptide, therapeutische Proteine, Hormone, Enzyme, Impfstoffe, monoklonale Antikörper und Antikörper-Wirkstoff-Konjugate eroberten den Markt. Darüber hinaus hat die Applikation und lokale Verabreichung von Wachstumsfaktoren im Bereich des Tissue Engineerings eine wesentliche Bedeutung erlangt. Trotz der in den letzten Jahrzehnten erzielten Fortschritte ist die Formulierung und Verabreichung therapeutischer Proteine noch immer eine Herausforderung. Daher haben wir uns in dieser Arbeit mit der Formulierung und Verabreichung therapeutischer Proteine beschäftigt und Strategien entwickelt, um deren biologische Wirkung zu verbessern. In Phase I dieser Arbeit konzentrieren wir uns auf die Stabilität des dimeren Fusionsproteins PR 15, einem Inhibitor der Adhäsion von Plättchen an arterielle Gefäßläsionen. Um eine geeignete flüssige Formulierung zu entwickeln, welche die Stabilität und Bioaktivität von PR-15 während der Lagerung bei 4 °C sicherstellt, wurde ein pH Screening, eine Forced Degradation Studie und ein Design of Experiments (DoE) durchgeführt. Zuerst wurde die Stabilität und Bioaktivität von PR-15 bei verschiedenen pH Werten in 50 mM Histidinpuffer in einer Kurzzeitstabilitätsstudie bei 25 °C und 40 °C nach 4 und 8 Wochen mit Hilfe verschiedener analytischer Methoden beobachtet. Des Weiteren wurden mögliche Abbauwege von PR-15 unter Stressbedingungen wie erhöhter Temperatur, saurem oder basischem pH-Wert, Einfrier-Auftau-Zyklen, Lichteinwirkung, induzierter Oxidation sowie induzierter Deamidierung während der Forced Degradation Studie untersucht. Darüber hinaus konnten wir das Hauptabbauprodukt von PR-15 durch LC/ESI-MS Analysen identifizieren. Im folgenden DoE wurde die injizierbare PR-15 Formulierung weiter optimiert und bezüglich pH, der Wahl des Puffers sowie der Zugabe von Hilfsstoffen analysiert, bis letztendlich eine optimale PR 15-Formulierung gefunden wurde. Die Wachstumsfaktoren BMP-2, IGF-I und TGF-β3 wurden zur Differenzierung von Stammzellen für das Tissue Engineering von Knochen und Knorpel ausgewählt, um multifunktionalisierte osteochondrale Implantate zur Regeneration von Knorpeldefekten herzustellen. Seidenfibroin (SF) wurde aufgrund seiner Biokompatibilität, seiner mechanischen Eigenschaften und seiner Möglichkeiten zur Biofunktionalisierung als Biomaterial gewählt. Zuerst wurden SF-Scaffolds mit idealer Geometrie und optimaler Porosität erzeugt, um sowohl Knochen also auch Knorpel auf einem Scaffold herzustellen. Um eine räumlich begrenzte kovalente Immobilisierung der Wachstumsfaktoren BMP-2 und IGF-I auf den porösen SF-Scaffolds zu ermöglichen, wurden diese mit unnatürlichen Aminosäuren genetisch modifiziert. Das Einführen von Alkin- bzw. Azidresten in die Aminosäuresequenz von BMP-2 (in dieser Arbeit nicht gezeigt; von Barbara Tabisz durchgeführt) und IGF-I erlaubt unter Verwendung der Click-Chemie eine ortsgerichtete kovalente Kopplung der Wachstumsfaktoren an zwei gegenüberliegenden Seiten der Scaffolds. TGF-β3 wurde durch gewöhnliche EDC/NHS-Chemie, welche keine Modifikation erforderte, kovalent an Mikrosphären immobilisiert und in den Poren des gesamten SF-Scaffolds verteilt. Daher beschäftigen wir uns in Phase II mit der Modifikation von Proteinen, dem Protein Engineering, der Immobilisation von Proteinen und mit Biokonjugation. Neben der Synthese lag der Fokus auf der Charakterisierung modifizierter Proteine und deren Konjugaten. Das Gebiet des Protein Engineerings bietet eine Vielzahl von Möglichkeiten, bestehende Proteine zu modifizieren oder neue Proteine mit verlängerter Serumhalbwertszeit, erhöhter konformativer Stabilität oder verbesserten Freisetzungsraten entsprechend der klinischen Anwendung zu entwickeln. Die ortsspezifische Click-Chemie und die nicht-ortsspezifische EDC/NHS-Chemie wurden für die Biokonjugation und die Immobilisierung von Proteinen verwendet mit dem Ziel, die Vorzüge der ortsgerichteten Kopplung hervorzuheben. Für den Einbau der für die Durchführung der Click-Reaktion erforderlichen Alkin- bzw. Azidfunktionalität in das betreffende Protein wurden drei Strategien ausgewählt: die Azokupplung, die PEGylierung und die gentechnische Modifizierung. Azidgruppen wurden mittels Azokupplung erfolgreich in SF eingebaut und die Alkin-, Amino- oder Säurefunktionalität wurde mittels PEGylierung der Cysteine in das Modellprotein FGF-2 integriert. Die korrekte Faltung von FGF-2 nach erfolgreicher PEGylierung wurde durch Fluoreszenzspektroskopie bestätigt, im WST-1 Proliferationsassay wurde eine angemessene Bioaktivität festgestellt und die Reinheit von PEGylierten FGF-2 wurde mittels RP-HPLC analysiert. Darüber hinaus führte die Modifikation von nativem FGF-2 mit 10 kDa PEG-Ketten zu einer erhöhten thermischen Stabilität. Des Weiteren wurde ein IGF-I-Mutant gentechnisch hergestellt, indem die unnatürliche Aminosäure Propargyl-L-Lysin (Plk) an Position 65 in die IGF-I-Sequenz eingebaut wurde. Da letztendlich lediglich kaum nachweisbare Mengen an Plk-IGF-I exprimiert werden konnten, muss die Plk-IGF-I-Expression in anschließenden Studien weiter optimiert werden, um Plk-IGF-I mit höheren Ausbeuten erzeugen zu können. Die Biokonjugation von PEGyliertem FGF-2 und funktionalisierter Seide wurde sowohl mittels Click- als auch mittels EDC/NHS-Chemie erfolgreich durchgeführt. Allerdings wurden erhebliche Mengen PEG-FGF-2 lediglich an SF adsorbiert und nicht kovalent gekoppelt und konnten schlussendlich nicht aus dem Reaktionsgemisch entfernt werden. Die anschließende Immobilisierung von PEG-FGF-2 an Mikrosphären, war ein Versuch die Ausbeute der Reaktion zu erhöhen und adsorbiertes PEG-FGF-2 leichter zu entfernen. Immobilisiertes PEG-FGF-2 wurde mittels Fluoreszenzmikroskopie und/oder Durchflusszytometrie nachgewiesen und die Bioaktivität wurde durch die Analyse der Proliferation von NIH-3T3-Zellen ermittelt. Die Immobilisierung auf Mikrosphären führte jedoch zu demselben Problem wie in Lösung: Adsorption von positiv geladenem FGF-2 an negativ geladenes SF bzw. negativ geladenen Mikrosphären durch elektrostatische Wechselwirkungen. Schließlich waren wir nicht in der Lage, die Überlegenheit der ortsgerichteten Click-Chemie gegenüber der nicht-ortsgerichteten EDC/ NHS-Chemie zu beweisen. Die während Phase II erworbenen Fähigkeiten und Kenntnisse im Bereich der Immobilisierung und Charakterisierung von Proteinen halfen uns in Phase III Knorpelgewebe in biofunktionalisierten SF-Scaffolds zu erzeugen. Der Ansatz der kovalenten Immobilisierung, der für das Tissue Engineering von Knorpel erforderlichen Wachstumsfaktoren, ist aufgrund ihrer kurzen in vivo Halbwertszeiten von Bedeutung und zielt darauf ab, ihre Bioverfügbarkeit zu kontrollieren. So wurde TGF-β3 mittels EDC/NHS-Chemie kovalent an biokompatible und biostabile PMMA-Mikrosphären gekoppelt. Mit Hilfe des sogenannten Luciferase-Assays wurden die Bioaktivitäten von kovalent gekoppeltem sowie von adsorbiertem TGF-β3 auf PMMA-Mikrosphären ermittelt. Um die Kopplungseffizienz zu berechnen und den tatsächlichen Einfluss der EDC/NHS-Chemie auf die Bioaktivität von TGF-β3 zu untersuchen, wurde die Menge an immobilisiertem TGF-β3 auf PMMA-Mikrosphären mittels ELISA bestimmt. Für die Kopplung von TGF-β3 mittels EDC/NHS-Chemie wurde eine signifikant höhere Kopplungseffizienz im Vergleich zu der Reaktion ohne Kopplungsreagenzien, welche eine geringe Menge an adsorbiertem TGF-β3 zeigte, bestimmt. Bei alleiniger Betrachtung der Ergebnisse des Luciferase-Assays, bei welchem kein statistisch signifikanter Unterschied zwischen kovalent immobilisiertem und adsorbiertem TGF-β3 bezüglich der relativen Luciferase-Aktivität beobachtet wurde, scheint es als hätte die EDC/NHS-Kopplung keinen Einfluss auf die Bioaktivität von TGF β3. Beim Vergleich der mittels ELISA bestimmten TGF β3 Gesamtmenge und der mittels Luciferase-Assay bestimmten Menge an aktivem TGF-β3 auf den PMMA-Mikrosphären, wurde jedoch ein Verlust der Bioaktivität von TGF-β3 durch die EDC/NHS-Kopplung deutlich. Ungeachtet dessen, wurde immobilisiertes TGF-β3 genutzt, um Knorpelgewebe in SF-Scaffolds aus BMSCs zu generieren. Nach den Ergebnissen der histologischen und immunhistochemischen Untersuchungen, der biochemischen Assays sowie der qRT-PCR der Genexpression von BMSCs nach 21 Tagen in Kultur, gelang es uns unter Verwendung von immobilisiertem TGF-β3 Knorpelgewebe aufzubauen. Diese Ergebnisse unterstützen die These, dass die lokale Präsentation von TGF-β3 gegenüber exogen zugegebenem TGF-β3 für die Entwicklung von hyalinem Knorpel überlegen ist. Außerdem schließen wir daraus, dass kovalent immobilisiertes TGF-β3 nicht nur gegenüber exogen zugegebenem TGF-β3 für die Entwicklung von hyalinem Knorpelgewebe überlegen ist, sondern auch gegenüber adsorbiertem TGF-β3. Diffusionsprozesse konnten durch kovalente Immobilisierung von TGF-β3 an PMMA-Mikrosphären verhindert werden und damit eine stabile und gleichmäßige TGF β3-Konzentration am Wirkort aufrechterhalten werden. Mit den in Phase II und III gewonnenen Erkenntnissen und den Untersuchungen von Barbara Tabisz zur Expression und Aufreinigung von plk-BMP-2 haben wir erhebliche Fortschritte bei der Entwicklung multifunktionaler osteochondraler Implantate zur Regeneration von Knorpeldefekten gemacht. Für die Umsetzung dieser Erkenntnisse zur Herstellung multifunktionaler osteochondraler SF-Scaffolds sind jedoch weitere Studien erforderlich.

FGF2 (Fibroblast Growth Factor 2), o fundador da família FGF, tem funções regulatórias na mitogênese, diferenciação, morfogênese e reparo tecidual. Diversas espécies moleculares de FGF2 compartilham uma seqüência C-terminal comum de 155 aminoácidos, pois se originam de diferentes sítios de iniciação de leitura de um único mRNA. O menor, o FGF2-18kDa, é liberado extracelularmente para se ligar a receptores específicos (FGFRs) para disparar as funções parácrinas e autócrinas pelas quais este fator é conhecido. Por outro lado, as espécies maiores (FGF2-21, 22, 22,5 e 34kDa) são intracelulares se ligam a parceiros moleculares desconhecidos para exercer funções intrácrinas ainda indefinidas. O objetivo desta tese foi produzir espécies recombinantes do FGF2-18 e FGF2-22,5, na forma de proteínas de fusão, para analisar funções biológicas e mecanismos de sinalização. Nas células malignas Y1 de camundongo, os recombinantes de FGF2-18kDa (FGF2-18, His-FGF2-18 e His-FGF2-18-ProA) dispararam uma resposta antagônica estimulando as vias de sinalização mitogênica, mas bloqueando o ciclo celular. Nos fibroblastos não tumorigênicos Balb3T3, estes mesmos recombinantes de FGF2-18kDa dispararam apenas a resposta mitogênica clássica. Todos os efeitos biológicos destes recombinantes de FGF2-18kDa foram bloqueados pelo inibidor específico da proteína quinase de tirosina dos FGFRs, PD173074, demonstrando que são respostas intermediadas pelos FGFRs. Portanto, os domínios estruturais adicionados aos recombinantes de FGF2-18kDa não impediram que estas proteínas se ligassem e ativassem os FGFRs. Por outro lado, o recombinante His-FGF2-22,5 dispara apenas as vias de sinalização mitogênica em ambas as células Y1 e 3T3, mas este efeito biológico não é inibido por PD173074. Estes resultados sugerem que a seqüência N-terminal de 55 resíduos, rica em aminoácidos básicos, impede que o FGF2-22,5kDa se ligue e/ou ative os FGFRs. Entretanto, o recombinante His-FGF2-22,5ProA dispara a resposta antagônica característica do FGF2-18kDa. As implicações destes últimos resultados é que o domínio de ProA adicionado ao C-terminal torna o FGF2-22,5kDa um bom ligante dos FGFRs. A interação física entre ligante e receptor das formas recombinantes His-FGF2-18kDa (ou His-FGF2-18ProA) e FGF2-22,5kDa com os putativos FGFRs foi analisada através da técnica de SPR e os resultados mostram KDs aproximados (Kd18=21, 488.10-9 e Kd22,5=20,70393.10-9), enquanto que o número de sítios ligantes em vesículas microssomais das células é significantemente inferior para o FGF2-22,5kDa. Estes resultados são compatíveis com a existência de receptores diferentes para FGF2-18kDa e FGF2-22,5kDa, uma hipótese ainda a ser definitivamente corroborada. Em conclusão, o FGF2-18kDa, mesmo em formas recombinantes como proteína de fusão, dispara todos os efeitos biológicos descritos para FGF2, através dos FGFRs. Diferentemente, o FGF2-22,5kDa, como fator parácrino, só desencadeou a resposta mitogênica clássica de FGF2, provavelmente através de receptores diferentes dos FGFRs. Os resultados e conclusões desta tese têm um potencial indiscutivelmente relevante para a biologia molecular do câncer, com implicações possíveis em terapia oncológica / FGF2 (Fibroblast Growth Factor 2), the founder of the FGF family, has regulatory functions in mitogenesis, differentiation, morphogenesis and tissue repair. Multiple FGF2 molecular species, sharing a C-terminal sequence of 155 amino acids, are translated from different iniciation sites of the same mRNA. The smaller, the FGF2-18kD, is extracellularly released to bind to specific membrane receptors (FGFRs), performing paracrine and autocrine functions. On the other hand, the larger FGF2s (21, 22, 22.5 and 34kDa) are intracellular species that bind to unknown partners to play still undefined intracrine roles. The aim of this thesis was to produce recombinant species of FGF2-18kDa and FGF2-22,5kDa, in the form of fusion proteins, to analyze functions and signaling mechanisms. In mouse Y1 malignant cells, FGF2-18kD recombinants (FGF2-18kDa and His-FGF2-18kDaProA) triggered an antagonistic response activating mitogenic signaling pathways, but blocking the cell cycle. However, in non tumorigenic Balb3T3 fibroblasts, these same FGF2-18kD recombinants only elicited the classical mitogenic response. All biological effects of these FGF2-18kD recombinants were blocked by the specific inhibitor of FGFR-protein-tyrosine-kinases, PD173074, demonstrating that these responses are mediated by FGFRs. Therefore, the new peptide domains added to FGF2-18kD did not prevent these recombinant fusion proteins to bind and activate FGFRs. Conversely, the recombinant His-FGF2-22,5kDa triggered only mitogenic signaling pathways in both Y1 and Balb3T3 cells, a biological effect not inhibited by PD173074. These results suggested that the additional basic-rich N-terminal sequence of 55 amino acid residues, found in FGF2-22,5kDa, prevents this FGF2 species from binding and / or activate FGFRs. However, surprisingly, the recombinant His-FGF2-22kDaProA triggered the antagonistic response characteristic of FGF2-18kDa. These results imply that the ProA-domain added to the C-terminal end rendered the FGF2-22,5kDaProA a good ligand of FGFRs. The physical interaction between recombinants of both His-FGF2-18kD and His-FGF2-22kDa with putative FGFRs, analyzed by SPR, yielded close KD values (KD18=21, 5.10-9 e K D22,5=20,7.10-9), while the number of binding sites in cell microsomal vesicles were significantly lower for the His-FGF2-22,5kDa. These results are consistent with the existence of different receptors for FGF2 and FGF2-18kD-22,5kDa, a hypothesis that has yet to be definitively confirmed. In conclusion, FGF2-18kD, even as recombinant fusion proteins, triggered all biological effects of FGF2, through FGFRs. Conversely, the FGF2-22, 5kDa only triggered the classical mitogenic response, probably via receptors other than FGFRs. The results and conclusions of this thesis are potentially of great interest in cancer molecular biology, with implications in oncologic therapy.

FGFs compreendem um grande família de 24 proteínas, participando de processos chaves nos mais variados tecidos, tendo funções parácrina, autócrina e intrácrina, regulando mitogênese, diferenciação celular, morfogênese e cicatrização. Mas, a relação estrutura-função dos FGFs é pobremente entendida. O membro protótipo desta família é o FGF-2, que apresenta quatro isoformas moleculares incluindo a forma de 18 kDa que é secretada e se liga aos receptores específicos (FGFRs) e dispara uma complexa sinalização. As outras isoformas, de alto peso molecular (21, 22 e 22,5 kDa) são expressas por códons alternativos (CUG) e permanecem no interior da célula interagindo com parceiros moleculares desconhecidos. Para antecipar mecanismos e parceiros do FGF-2 HMW foi realizada modelagem molecular desta isoforma que mostrou: uma estrutura do N-terminal da proteína com motivo β→α&#8594β e manutenção do barril β. A busca por parceiros intracelulares, foi realizada através da técnica do duplo hibrido de levedura, usando um biblioteca de cDNA de cérebro de rato. Foram encontrados 4 possíveis parceiros: BRD2, UBE2I, BRPF1, PC4. Todas essas interações foram confirmadas através do crescimento da levedura em meio sem histidina, produção de β-galactosidase e ensaios de \"pull-down\" com GST. Analises por FACS confirmam que FGF2 não causa apoptose em células adrenais tumorais Y1 de camundongo, mas promovem um acumulo de células na fase S com bloqueio do ciclo celular e da proliferação, configurando uma forma de senescência. Resultados com as células humanas HEK-ER:Ras permitem fazer a seguinte generalização: FGF2 induz senescência em células malignas transformadas pelos oncogenes raso A superexpressão da proteína de fusão FGF-2(18kDa):protA, mas não a da FGF-2(22,5 kDa):protA, protege a célula Y1 da senescência induzida por FGF-2. Por outro lado, a superexpressão destas mesmas isoformas de FGF-2 fusionadas à proteína A em células imortalizadas Balb3T3 não causou transformação celular e nem alterou a resposta mitogênica destas células ao FGF-2 recombinante adicionado ao meio de cultura. Células Y1 quando tratadas com FGF-2 recombinante produz ROS intracelular e libera anions superóxido no meio extracelular. Além disso, o anti-oxidante NAC protege estas células da indução de senescência induzida por FGF-2, sugerindo que ROS pode ser intermediário no disparo de senescência por FGF-2. / FGFs comprise a large fami1y of 24 proteins that play key roles in a number of tissues as local paracrine, autocrine and intracrine regulators of mitogenesis, cellular differentiation, organ morphogenesis and tissue repair. Structure-function relationship among FGFs is still poorly understood. FGF-2, the fami1y prototype member, exists as four molecular species. The 18 kDa form is released to the extracellular milieu and binds to specific receptors (FGFR), initiating a complex array of signals. Other isoforms of higher molecular weights (21, 22 and 22,5 kDa) are translated from alternative codons (CUG) and remain inside of the cell interacting with unknown partners. Aiming to anticipate mechanisms and partners, we modeled the FGF2-HMW molecule, showing that the protein displays β→α&#8594β motif in the N-terminal region and maintains the β-barrel structure common to ali FGFs. By the yeast two-hybrid method, using a cDNA rat brain library, we found four possible partners for FGF2-HMW: BRD2, UBE2I, BRP1 and PC4. Ali partners were confirmed by yeast growth without histidine, production of β-galactosidase and \"pull-down\" assays with GST. FACS analyses confirmed that FGF2 does not cause apoptosis in mouse Y1 adrenal tumor cells. But, FGF2 inhibited S phase progression blocking cell cycle and proliferation, characterizing a form of senescence. In addition, results obtained with the human HEK-ER:Ras cells support the following general statement: FGF2 triggers senescence in malignant cells transformed by ras oncogenes. Ectopic expression of the fusion protein FGF-2(18 kDa):protA, but not of FGF-2(22,s kDa):protA, protected Y1 cells senescence induced by FGF-2. On the other hand, ectopic expression of FGF-2 isoforms fusioned to protA in Balb3T3 immortalized cells did not cause transformation and neither modified the mitogenic response of this cell to recombinant FGF2. Recombinant FGF-2 stimules Y1 cells to produce intracellular ROS and to release superoxide anions into intracellular medium. Moreover, the ROS scavenger NAC protect Y1 cells from senescence induced by FGF-2, suggesting that ROS may be mediate senescence triggering induced by FGF-2.

FGF2 (Fibroblast Growth Factor 2), o fundador da família FGF, tem funções regulatórias na mitogênese, diferenciação, morfogênese e reparo tecidual. Diversas espécies moleculares de FGF2 compartilham uma seqüência C-terminal comum de 155 aminoácidos, pois se originam de diferentes sítios de iniciação de leitura de um único mRNA. O menor, o FGF2-18kDa, é liberado extracelularmente para se ligar a receptores específicos (FGFRs) para disparar as funções parácrinas e autócrinas pelas quais este fator é conhecido. Por outro lado, as espécies maiores (FGF2-21, 22, 22,5 e 34kDa) são intracelulares se ligam a parceiros moleculares desconhecidos para exercer funções intrácrinas ainda indefinidas. O objetivo desta tese foi produzir espécies recombinantes do FGF2-18 e FGF2-22,5, na forma de proteínas de fusão, para analisar funções biológicas e mecanismos de sinalização. Nas células malignas Y1 de camundongo, os recombinantes de FGF2-18kDa (FGF2-18, His-FGF2-18 e His-FGF2-18-ProA) dispararam uma resposta antagônica estimulando as vias de sinalização mitogênica, mas bloqueando o ciclo celular. Nos fibroblastos não tumorigênicos Balb3T3, estes mesmos recombinantes de FGF2-18kDa dispararam apenas a resposta mitogênica clássica. Todos os efeitos biológicos destes recombinantes de FGF2-18kDa foram bloqueados pelo inibidor específico da proteína quinase de tirosina dos FGFRs, PD173074, demonstrando que são respostas intermediadas pelos FGFRs. Portanto, os domínios estruturais adicionados aos recombinantes de FGF2-18kDa não impediram que estas proteínas se ligassem e ativassem os FGFRs. Por outro lado, o recombinante His-FGF2-22,5 dispara apenas as vias de sinalização mitogênica em ambas as células Y1 e 3T3, mas este efeito biológico não é inibido por PD173074. Estes resultados sugerem que a seqüência N-terminal de 55 resíduos, rica em aminoácidos básicos, impede que o FGF2-22,5kDa se ligue e/ou ative os FGFRs. Entretanto, o recombinante His-FGF2-22,5ProA dispara a resposta antagônica característica do FGF2-18kDa. As implicações destes últimos resultados é que o domínio de ProA adicionado ao C-terminal torna o FGF2-22,5kDa um bom ligante dos FGFRs. A interação física entre ligante e receptor das formas recombinantes His-FGF2-18kDa (ou His-FGF2-18ProA) e FGF2-22,5kDa com os putativos FGFRs foi analisada através da técnica de SPR e os resultados mostram KDs aproximados (Kd18=21, 488.10-9 e Kd22,5=20,70393.10-9), enquanto que o número de sítios ligantes em vesículas microssomais das células é significantemente inferior para o FGF2-22,5kDa. Estes resultados são compatíveis com a existência de receptores diferentes para FGF2-18kDa e FGF2-22,5kDa, uma hipótese ainda a ser definitivamente corroborada. Em conclusão, o FGF2-18kDa, mesmo em formas recombinantes como proteína de fusão, dispara todos os efeitos biológicos descritos para FGF2, através dos FGFRs. Diferentemente, o FGF2-22,5kDa, como fator parácrino, só desencadeou a resposta mitogênica clássica de FGF2, provavelmente através de receptores diferentes dos FGFRs. Os resultados e conclusões desta tese têm um potencial indiscutivelmente relevante para a biologia molecular do câncer, com implicações possíveis em terapia oncológica / FGF2 (Fibroblast Growth Factor 2), the founder of the FGF family, has regulatory functions in mitogenesis, differentiation, morphogenesis and tissue repair. Multiple FGF2 molecular species, sharing a C-terminal sequence of 155 amino acids, are translated from different iniciation sites of the same mRNA. The smaller, the FGF2-18kD, is extracellularly released to bind to specific membrane receptors (FGFRs), performing paracrine and autocrine functions. On the other hand, the larger FGF2s (21, 22, 22.5 and 34kDa) are intracellular species that bind to unknown partners to play still undefined intracrine roles. The aim of this thesis was to produce recombinant species of FGF2-18kDa and FGF2-22,5kDa, in the form of fusion proteins, to analyze functions and signaling mechanisms. In mouse Y1 malignant cells, FGF2-18kD recombinants (FGF2-18kDa and His-FGF2-18kDaProA) triggered an antagonistic response activating mitogenic signaling pathways, but blocking the cell cycle. However, in non tumorigenic Balb3T3 fibroblasts, these same FGF2-18kD recombinants only elicited the classical mitogenic response. All biological effects of these FGF2-18kD recombinants were blocked by the specific inhibitor of FGFR-protein-tyrosine-kinases, PD173074, demonstrating that these responses are mediated by FGFRs. Therefore, the new peptide domains added to FGF2-18kD did not prevent these recombinant fusion proteins to bind and activate FGFRs. Conversely, the recombinant His-FGF2-22,5kDa triggered only mitogenic signaling pathways in both Y1 and Balb3T3 cells, a biological effect not inhibited by PD173074. These results suggested that the additional basic-rich N-terminal sequence of 55 amino acid residues, found in FGF2-22,5kDa, prevents this FGF2 species from binding and / or activate FGFRs. However, surprisingly, the recombinant His-FGF2-22kDaProA triggered the antagonistic response characteristic of FGF2-18kDa. These results imply that the ProA-domain added to the C-terminal end rendered the FGF2-22,5kDaProA a good ligand of FGFRs. The physical interaction between recombinants of both His-FGF2-18kD and His-FGF2-22kDa with putative FGFRs, analyzed by SPR, yielded close KD values (KD18=21, 5.10-9 e K D22,5=20,7.10-9), while the number of binding sites in cell microsomal vesicles were significantly lower for the His-FGF2-22,5kDa. These results are consistent with the existence of different receptors for FGF2 and FGF2-18kD-22,5kDa, a hypothesis that has yet to be definitively confirmed. In conclusion, FGF2-18kD, even as recombinant fusion proteins, triggered all biological effects of FGF2, through FGFRs. Conversely, the FGF2-22, 5kDa only triggered the classical mitogenic response, probably via receptors other than FGFRs. The results and conclusions of this thesis are potentially of great interest in cancer molecular biology, with implications in oncologic therapy.

FGFs compreendem um grande família de 24 proteínas, participando de processos chaves nos mais variados tecidos, tendo funções parácrina, autócrina e intrácrina, regulando mitogênese, diferenciação celular, morfogênese e cicatrização. Mas, a relação estrutura-função dos FGFs é pobremente entendida. O membro protótipo desta família é o FGF-2, que apresenta quatro isoformas moleculares incluindo a forma de 18 kDa que é secretada e se liga aos receptores específicos (FGFRs) e dispara uma complexa sinalização. As outras isoformas, de alto peso molecular (21, 22 e 22,5 kDa) são expressas por códons alternativos (CUG) e permanecem no interior da célula interagindo com parceiros moleculares desconhecidos. Para antecipar mecanismos e parceiros do FGF-2 HMW foi realizada modelagem molecular desta isoforma que mostrou: uma estrutura do N-terminal da proteína com motivo β→α&#8594β e manutenção do barril β. A busca por parceiros intracelulares, foi realizada através da técnica do duplo hibrido de levedura, usando um biblioteca de cDNA de cérebro de rato. Foram encontrados 4 possíveis parceiros: BRD2, UBE2I, BRPF1, PC4. Todas essas interações foram confirmadas através do crescimento da levedura em meio sem histidina, produção de β-galactosidase e ensaios de \"pull-down\" com GST. Analises por FACS confirmam que FGF2 não causa apoptose em células adrenais tumorais Y1 de camundongo, mas promovem um acumulo de células na fase S com bloqueio do ciclo celular e da proliferação, configurando uma forma de senescência. Resultados com as células humanas HEK-ER:Ras permitem fazer a seguinte generalização: FGF2 induz senescência em células malignas transformadas pelos oncogenes raso A superexpressão da proteína de fusão FGF-2(18kDa):protA, mas não a da FGF-2(22,5 kDa):protA, protege a célula Y1 da senescência induzida por FGF-2. Por outro lado, a superexpressão destas mesmas isoformas de FGF-2 fusionadas à proteína A em células imortalizadas Balb3T3 não causou transformação celular e nem alterou a resposta mitogênica destas células ao FGF-2 recombinante adicionado ao meio de cultura. Células Y1 quando tratadas com FGF-2 recombinante produz ROS intracelular e libera anions superóxido no meio extracelular. Além disso, o anti-oxidante NAC protege estas células da indução de senescência induzida por FGF-2, sugerindo que ROS pode ser intermediário no disparo de senescência por FGF-2. / FGFs comprise a large fami1y of 24 proteins that play key roles in a number of tissues as local paracrine, autocrine and intracrine regulators of mitogenesis, cellular differentiation, organ morphogenesis and tissue repair. Structure-function relationship among FGFs is still poorly understood. FGF-2, the fami1y prototype member, exists as four molecular species. The 18 kDa form is released to the extracellular milieu and binds to specific receptors (FGFR), initiating a complex array of signals. Other isoforms of higher molecular weights (21, 22 and 22,5 kDa) are translated from alternative codons (CUG) and remain inside of the cell interacting with unknown partners. Aiming to anticipate mechanisms and partners, we modeled the FGF2-HMW molecule, showing that the protein displays β→α&#8594β motif in the N-terminal region and maintains the β-barrel structure common to ali FGFs. By the yeast two-hybrid method, using a cDNA rat brain library, we found four possible partners for FGF2-HMW: BRD2, UBE2I, BRP1 and PC4. Ali partners were confirmed by yeast growth without histidine, production of β-galactosidase and \"pull-down\" assays with GST. FACS analyses confirmed that FGF2 does not cause apoptosis in mouse Y1 adrenal tumor cells. But, FGF2 inhibited S phase progression blocking cell cycle and proliferation, characterizing a form of senescence. In addition, results obtained with the human HEK-ER:Ras cells support the following general statement: FGF2 triggers senescence in malignant cells transformed by ras oncogenes. Ectopic expression of the fusion protein FGF-2(18 kDa):protA, but not of FGF-2(22,s kDa):protA, protected Y1 cells senescence induced by FGF-2. On the other hand, ectopic expression of FGF-2 isoforms fusioned to protA in Balb3T3 immortalized cells did not cause transformation and neither modified the mitogenic response of this cell to recombinant FGF2. Recombinant FGF-2 stimules Y1 cells to produce intracellular ROS and to release superoxide anions into intracellular medium. Moreover, the ROS scavenger NAC protect Y1 cells from senescence induced by FGF-2, suggesting that ROS may be mediate senescence triggering induced by FGF-2.

The tumour suppressor p53 actually exists as 9 protein isoforms. Among them, D133p53a, b and g result from the use of an alternative promoter and lack the N-terminal transactivation domain. In addition to its multiple functions maintaining cell integrity, p53 is also able to block angiogenesis, a process strongly contributing in tumour development. Here I have examined the role of p5 isoforms in the regulation of angiogenesis and tumor progression. I also focused my work on FGF-2 regulation by p53. In a first part, full length p53 (p53) and/or D133p53 isoforms were selectively knocked-down with siRNAs in human glioblastoma cells U87. Conditioned medium produced by tumour cells knocked- down for D133p53 inhibited endothelial cell - EC - migration and tubulogenesis. Furthermore, in the chicken chorioallantoïd membrane CAM, D133p53 knockdown gave rise to smaller tumours devoid of vessels, whereas, in mice, it strongly inhibited tumour growth. Interestingly, the double knockdown of p53 and D133p53 also slowed town tumour growth in mice. Taqman Low Density Array revealed distinct gene expression profiles of pro and anti-angiogenic factors regulation following D133p53 and/or p53 knockdown. In particular, D133p53 knockdown resulted in specific down-regulation of Angiogenin and hepatocyte growth factor, whereas the main angiogenic factors FGF-2 and VEGF-A were not significantly affected. Secondly we investigated the regulation of FGF-2 by p53 and its isoforms D133p53 in a human osteosarcoma cell line U2OS, at translational, transcriptional and secretion levels. It resulted in a sophisticated mode of regulation mediated by a transient IRES-dependent translation inhibition of FGF-2. Our data reveal D133p53 isoforms as activators of angiogenesis and tumour progression, through a specific modulation of the angiogenic balance. These isoforms exhibit dominant-negative effect towards p53 but also intrinsic activities, while underlining the importance of considering D133p53 expression in cancers, as well as the potential antitumoural interest of drugs targeting this p53 isoform.

<p>Angiogenesis can be defined as the formation of new blood vessels from pre-existing ones. Angiogenesis is required for development and maintenance of our vascular system and thus of fundamental importance to our existence. The endothelial cells that line the inside of the vessels de-differentiate, migrate, proliferate and re-differentiate during angiogenesis. Angiogenesis is tightly regulated, controlled by several angiogenic factors of various classes that promote angiogenesis but also by anti-angiogenic factors that counteract the effect of the pro-angiogenic factors. We have examined three factors involved in angiogenesis regulation, Vascular endotelial growth factor (VEGFR) -3, the matrix protein laminin-1 and the collagen XVIII derived fragment endostatin. </p><p>Five tyrosine phosphorylation sites in the cytoplasmic tail of VEGFR-3 were identified by phosphopeptide mapping (PPM). The data was confirmed by PPM using point-mutated receptors generated by site-directed mutagenesis.</p><p>Laminin-1 was found to promote angiogenesis in the chicken chorioallantoic membrane assay and in a synergistic fashion together with suboptimal levels of fibroblast growth factor 2 (FGF-2) in embryoid bodies. Laminin-1 also promoted endothelial tubular morphogenesis in vitro, and upregulated the expression of the endothelial differentiation marker Jagged-1. </p><p>Endostatin was shown to affect endothelial FGF-2-induced cell survival and morphogenesis. This was a result of direct binding to endothelial cells and induction of tyrosine phosphorylation of many proteins including the adaptor protein Shb. The apoptotic and morphogenic responses induced by endostatin was shown to be dependent on Shb. Further, endostatin inhibited endothelial migration and affected molecules implicated in migration. In particular, FGF-2 induced actin reorganization, and β-catenin regulation was modulated by endostatin. </p>

Angiogenesis can be defined as the formation of new blood vessels from pre-existing ones. Angiogenesis is required for development and maintenance of our vascular system and thus of fundamental importance to our existence. The endothelial cells that line the inside of the vessels de-differentiate, migrate, proliferate and re-differentiate during angiogenesis. Angiogenesis is tightly regulated, controlled by several angiogenic factors of various classes that promote angiogenesis but also by anti-angiogenic factors that counteract the effect of the pro-angiogenic factors. We have examined three factors involved in angiogenesis regulation, Vascular endotelial growth factor (VEGFR) -3, the matrix protein laminin-1 and the collagen XVIII derived fragment endostatin. Five tyrosine phosphorylation sites in the cytoplasmic tail of VEGFR-3 were identified by phosphopeptide mapping (PPM). The data was confirmed by PPM using point-mutated receptors generated by site-directed mutagenesis. Laminin-1 was found to promote angiogenesis in the chicken chorioallantoic membrane assay and in a synergistic fashion together with suboptimal levels of fibroblast growth factor 2 (FGF-2) in embryoid bodies. Laminin-1 also promoted endothelial tubular morphogenesis in vitro, and upregulated the expression of the endothelial differentiation marker Jagged-1. Endostatin was shown to affect endothelial FGF-2-induced cell survival and morphogenesis. This was a result of direct binding to endothelial cells and induction of tyrosine phosphorylation of many proteins including the adaptor protein Shb. The apoptotic and morphogenic responses induced by endostatin was shown to be dependent on Shb. Further, endostatin inhibited endothelial migration and affected molecules implicated in migration. In particular, FGF-2 induced actin reorganization, and β-catenin regulation was modulated by endostatin.

The regenerative capacity of mature mammalian CNS neurons after axonal injury is severely limited by a variety of mechanisms. Retrograde repression is the continuous inhibition of the expression of growth phenotypes by tonic signals produced by target tissues and transmitted to the neuron cell body via retrograde axonal transport. Loss of target contact through axonal injury is thought to interrupt this retrograde signal and allow the up-regulation of growth-associated proteins. Most CNS neurons, however, possess many widespread axon collaterals, such that retrograde repression is maintained by intact sustaining collaterals even if some axons are injured.<p>In this project we investigated whether or not retrograde repression plays a role in limiting the expression of GAP-43 in transcallosal neurons. Because TCNs possess local axon collaterals to nearby cortex and project distal axons to homologous areas of contralateral cortex, we hypothesized that the simultaneous interruption of retrograde repressive signals from both ipsilateral and contralateral cortex would result in an up-regulation of GAP-43 expression in at least some TCNs.<p>We found that a bilateral infusion of a function blocking antibody to FGF-2 into the parietal cortex of rats using implanted osmotic mini-pumps resulted in a significant increase in the level of expression of GAP-43 mRNA in TCNs identified by retrograde fluorescent labeling. In contrast, neither ipsilateral or contralateral antibody infusions alone increased GAP-43 expression significantly compared to controls. The level of expression of GAP-43 in TCNs did not significantly increase after stereotactic callosotomy alone, but callosotomized animals treated with an ipsilateral infusion of anti-FGF-2 had levels of increased GAP-43 expression equivalent to those seen in animals that had received bilateral antibody infusions.<p>We conclude that FGF-2 provides a retrograde repressive signal for at least some mature mammalian TCNs, and that the expression of growth-associated proteins can be up-regulated in CNS neurons by simultaneously blocking retrograde repressive signals from all existing axon collaterals. The ability to alter the gene expression of mature CNS neurons in both normal and injured states through the targeted infusion of a pharmacological agent may have potential clinical implications in the future.