Mon travail de thèse fut de développer différents projets en vue de mieux comprendre la dynamique et l'organisation des réseaux d'actine, ainsi que les mécanismes moléculaires à l'origine de la production de force grâce à différents systèmes reconstitués biomimétiques. Dans un premier temps, je me suis intéressée à l'étude de l'organisation spatiotemporelle des réseaux dynamiques d'actine et de ses protéines associées durant la propulsion de particules recouvertes de promoteurs de nucléation des filaments d'actine (Achard et al, Current Biology, 2010 et Reymann et al, accepté à MBoC). J'ai notamment suivi en temps réel l'incorporation de deux régulateurs de l'actine (Capping protein, protéine de coiffe et ADF/cofilin, protéine de fragmentation) et montré que leurs actions conjuguées assurent un contrôle biochimique de l'assemblage d'un réseau complexe d'actine, mais gouvernent également les propriétés mécaniques de ce réseau. Par ailleurs, afin de mieux caractériser les propriétés mécaniques de ces réseaux d'actine en expansion, j'ai développé un système biomimétique novateur utilisant la procédure de micropatrons ou "micropatterning" qui permet un contrôle spatial reproductible des sites de nucléation d'actine. Cela m'a permis de montrer comment des barrières géométriques, semblables à celles trouvées dans les cellules, peuvent influencer la formation dynamique de réseaux organisés d'actine et ainsi contrôler la localisation de la production de forces. (Reymann et al, Nature Materials, 2010). De plus, l'incorporation de moteurs moléculaires dans ce système versatile, nous a permis d'étudier la contraction induite par des myosines. En particulier, j'ai pu montrer que les myosines VI HMM interagissent de manière sélective avec différentes architectures d'actine (organisation parallèle ou antiparallèle, réseau enchevêtré), aboutissant à un processus en trois phases: tension, puis déformation des réseaux d'actine fortement couplée à un désassemblage massif des filaments. Aussi, ce phénomène de désassemblage massif induit par la myosine est intimement dépendant de l'architecture du réseau d'actine et pourrait, de ce fait, jouer un rôle essentiel dans la régulation spatiale des zones d'expansion et de contraction du cytosquelette in vivo.

The cell surface is an active area where extracellular molecules meet their receptors and affect the cellular fate by inducing for example cell proliferation and adhesion. Syndecans and integrins are two transmembrane molecules that have been suggested to fine-tune these activities, possibly in cooperation. Syndecans are proteoglycans, i.e. proteins with specific types of carbohydrate chains attached. These chains are glycosaminoglycans and either heparan sulfate (HS) or chondroitin sulfate (CS). Syndecans are known to influence cell adhesion and signaling. Integrins in turn, are important adhesion molecules that connect the extracellular matrix with the cytoskeleton, and hence can regulate cell motility. In an attempt to study how the two types of glycosaminoglycans attached to syndecan-1 can interact with integrins, a cell based model system was used and functional motility assays were performed. The results showed that HS, but not CS, on the cell surface was capable of regulating integrin-mediated cell motility. Regulation of intracellular signaling is crucial to prevent abnormal cellular behavior. In the second part of this thesis, the aim was to see how the presentation of glycosaminoglycan chains to the FGF signaling complex could affect the cellular response. When attached to the plasma membrane via syndecan-1, CS chains could support the intracellular signaling, although not promoting as strong signals as HS. When glycosaminoglycans were attached to free ectodomains of syndecan-1, both types of chains sequestered FGF2 from the receptors to the same extent, pointing towards functional overlap between CS and HS. To further study the interplay between HS and CS, their roles in the formation of pharyngeal cartilage in zebrafish were established. HS was important during chondrocyte intercalation and CS in the formation of the surrounding extracellular matrix. Further, the balance between the biosynthetic enzymes determined the ratio of HS and CS, and HS biosynthesis was prioritized over CS biosynthesis. The results presented in this thesis provide further insight into the regulation of HS biosynthesis, as well as the roles of both HS and CS on the cell surface. It is evident, that in certain situations there is a strict requirement for a certain HS structure, albeit in other situations there is a functional overlap between HS and CS.

Cell motility has been found to play a role in many important body functions as well as the embryogenenis of mulitcellular organisms like vertebrates. From a physics point of view the interesting questions behind every motion are: Why is it moving? Where do the forces come from? Today we know that the motility of many cells is dependent on an active polymer network. Actin, one of the most abundant proteins in the body, is constantly polymerized, being moved around and depolymerized in motile cells. Until now, only forces outside the cell like traction forces could be measured. The direct measurement of the force generated by polymerizing actin filaments has only been measured by our lab and the lab of M. Radmacher. In these measurements fish keratocytes were used. Whereas I did these experiments, for the first time, on mammalian cells. To measure forward forces on neuronal growth cones I stabilized the SFM, as measurement times went up from minutes to hours. Furthermore measurements had to be performed at 37°C instead of room temerature, this induced drifts of the substrate. I incorporated an optical trap into the microscope to track the motion of the substrate. A feedback loop moved the SFM cantilever to minimize relative motion of substrate and cantilever. For keratocytes I directly measured the forces produced by actin polymerization and, to my knowledge for the first time, the forces associated with the retrograde actin flow using a SFM. The result was that both actin and myosin play important but different roles in motility. For actin it turned out that considering just the polymerization was not enough. Actin depolymerization and the resulting entropic forces are a completely new physical effect in actin based cell motility. With this new force in the force balance I can explain all effects observed in my experiments without introducing any new biochemical feedback loops. Finally I showed that neuronal growth cones are very soft and weak structures. They are at least one order of magnitude softer and weaker as for example fibroblasts or cells forming the blood vessel walls. As neurons are usually located in soft environments this does not impede their normal outgrowth. It could even serve as a safety mechanism that prevents cell from growing into wrong areas like breaching the blood-brain-barrier, on a physical level. For a neuron the wall of a blood vessel feels like a brick wall for us.

A cell's ability to sense and respond to mechanical signals highlights the significance of physical forces in biology; however, to date most biomedical research has focused on genetics and biochemical signaling. We sought to further understand the physical mechanisms that guide the cellular migrations that occur in a number of biological processes, such as tissue development and regeneration, bacterial infections and cancer metastasis. We investigated the migration of single cells and determined whether the biomechanics of these cells could be used to elucidate multi-cellular mechanisms. We first studied Borrelia burgdorferi (Bb), the bacterium that causes Lyme disease. We created a mathematical model based on the mechanical interactions between the flagella and cell body that explained the rotation and undulation of the cell body that occurs as the bacterium swims. This model further predicts how the swimming dynamics could be affected by alterations in flagellar or cell wall stiffnesses. Fitting the model to experimental data allowed us to calculate the flagellar torque and drag for Bb, and showed that Treponema pallidum (Tp), the syphilis pathogen, is biomechanically similar to Bb. Next, we used experimentally-determined parameters of Bb's motility to develop a population-level model that accounts for the morphology and spreading of the "bulls-eye" rash that is typically the first indicator of Lyme disease. This work supported clinical findings on the efficacy of antibiotic treatment regimes. Finally, we investigated the dynamics of epithelial monolayers. We found that intracellular contractile stress is the primary driving force behind collective dynamics in epithelial layers, a result previously predicted from a biophysical model. Taken together, these findings identify the relevance of physics in cellular migration and a role of mechanical signaling in biomedical science.

Dictyostelium discoideum is a social amoebozoan that is used a model system for studying a variety of cell and developmental processes, especially cell motility and chemotaxis. Genome analyses suggest that this model organism possesses a higher percentage of Epidermal Growth Factor (EGF)-like (EGFL) repeats than any other sequenced eukaryote, including humans. EGFL repeats share strong sequence similarity with EGF. In mammals, EGF binds to an EGF receptor (EGFR) to initiate intracellular signalling that regulates a diversity of cellular processes including cell motility and chemotaxis. Some EGFL repeats, like EGF, have also been shown to increase the rate of cell motility by binding to the EGFR and activating EGFR-dependent signalling. Despite their abundance in Dictyostelium, a function for EGFL repeats in this model eukaryote had not previously been studied. This thesis presents a collection of studies that investigated the function of a specific EGFL repeat from the extracellular, cysteine-rich, calmodulin (CaM)-binding protein CyrA. A synthetic peptide (DdEGFL1), equivalent in sequence to the first 18 amino acids of the first EGFL repeat (EGFL1) of CyrA, was shown to increase random cell motility and cAMP-mediated chemotaxis via a novel signalling pathway that did not require either of the two cAMP receptors that are active during early development of Dictyostelium. Several intracellular signalling components were identified and then incorporated into a model detailing the signal transduction regulating EGFL repeat-enhanced cell movement in Dictyostelium. Finally the expression, secretion, and localization of CyrA are presented to couple the findings from studies on DdEGFL1 function with those for the full-length protein. In mammals, a protein that localizes to the extracellular matrix (ECM) and modulates cellular processes by binding to a cell surface receptor and initiating intracellular signalling is termed a ‘matricellular’ protein. The research presented in this thesis suggests that CyrA is the first matricellular protein identified in Dictyostelium.

Dictyostelium discoideum is a social amoebozoan that is used a model system for studying a variety of cell and developmental processes, especially cell motility and chemotaxis. Genome analyses suggest that this model organism possesses a higher percentage of Epidermal Growth Factor (EGF)-like (EGFL) repeats than any other sequenced eukaryote, including humans. EGFL repeats share strong sequence similarity with EGF. In mammals, EGF binds to an EGF receptor (EGFR) to initiate intracellular signalling that regulates a diversity of cellular processes including cell motility and chemotaxis. Some EGFL repeats, like EGF, have also been shown to increase the rate of cell motility by binding to the EGFR and activating EGFR-dependent signalling. Despite their abundance in Dictyostelium, a function for EGFL repeats in this model eukaryote had not previously been studied. This thesis presents a collection of studies that investigated the function of a specific EGFL repeat from the extracellular, cysteine-rich, calmodulin (CaM)-binding protein CyrA. A synthetic peptide (DdEGFL1), equivalent in sequence to the first 18 amino acids of the first EGFL repeat (EGFL1) of CyrA, was shown to increase random cell motility and cAMP-mediated chemotaxis via a novel signalling pathway that did not require either of the two cAMP receptors that are active during early development of Dictyostelium. Several intracellular signalling components were identified and then incorporated into a model detailing the signal transduction regulating EGFL repeat-enhanced cell movement in Dictyostelium. Finally the expression, secretion, and localization of CyrA are presented to couple the findings from studies on DdEGFL1 function with those for the full-length protein. In mammals, a protein that localizes to the extracellular matrix (ECM) and modulates cellular processes by binding to a cell surface receptor and initiating intracellular signalling is termed a ‘matricellular’ protein. The research presented in this thesis suggests that CyrA is the first matricellular protein identified in Dictyostelium.

Die selbständige, gerichtete Bewegung von biologischen Zellen ist eine der grundlegendsten und komplexesten Erscheinungen der Natur. In höher entwickelten Lebewesen spielt die Zellbewegung eine wichtige Rolle, z.B. bei der Entwicklung des Organismus, bei der Funktion des Immunsystems aber auch bei der Metastase von Krebszellen. Die physikalischen Prozesse die dieser Fähigkeit zugrunde liegen, sind im Fokus dieser Arbeit. Um besser zu verstehen welche Prozesse im Einzelnen und in welcher Kombination den Zellen erlauben sich gerichtet fortzubewegen, wurde in der vorliegenden Arbeit ein representatives Modellsystem von motilen Zellen untersucht. Fischkeratozyten bewegen sich in vitro regelmäßig und gleichförmig, relativ schnell über die Substratfläche, und stellen aus physikalischer Sicht eine optimierte, sich selbständig bewegende Polymermaschine dar. Um Kräfte in der Bewegungsebene der Zellen zu untersuchen, wurde in der vorliegenden Arbeit eine neuartige, auf dem Rasterkraftmikroskop (RKM) basierende Methode entwickelt. Zusätzlich wurden hochaufgelöste, mit dem Phasenkontrastmikroskop aufgenommene Bilderserien analysiert und die Geschwindigkeitsverteilung in der Zelle durch Korrelationsalgorithmen bestimmt. Die Struktur des Polymernetzwerkes wurde in mit Fluoreszenzfarbstoff markierten Zellen untersucht, und elastische Eigenschaften wurden mit rheologischen RKM-Messungen bestimmt. Traktionskraftmessungen an elastischen Substraten runden das umfassende Bild ab. Durch Veränderung der molekularen Strukturen mit verschiedenen Chemikalien, die unterschiedliche Prozesse im Gesamtsystem stören, konnte nun ein Phasenraum der Kraftgenerierungsprozesse untersucht und unterschiedliche Effekte verschiedenen Prozessen eindeutig zugeordnet werden. Es wurde somit erstmalig experimentell bewiesen, dass die Polymerisation von Aktin die treibende Kraft am vorderen Rand der Zelle ist. Darüber hinaus wurde das Verhalten des Kraftaufbaus mit einem Model beschrieben, das Aufschluss über die Funktionsweise der darunterliegenden Aktinpolymerstrukturens gibt. Desweiteren wurde in der Mitte der Zelle, zwischen vorderem Rand und Zellkörper, erstmalig eine rückwärtsgerichtete Kraft gemessen, die wichtig ist um ein Kräftegleichgewicht zu erstellen. Ein Model das auf entropischen Kräften im Polymersystem basiert, beschreibt diese kontraktilen Kräfte und ordnet sie der Depolymerisation von Aktin zu. Die Bewegung des Zellkörpers wiederum basiert auf dem Zusammenspiel dieser beiden Mechanismen, sowie der Kontraktion von Aktin und Aktinbündeln durch molekulare Motoren. Eine umfassendes Charakterisierung über verschiedene lokale Mechanismen und ihrer Wechselwirkungen konnte somit erstellt werden, und damit das Verständnis der Kraftgenerierung zur Zellbewegung vertieft.

Collective cell migration is ubiquitous in biology, from development to cancer; it is influenced by heterogeneous cell types, signals and matrix properties, and requires large scale regulation in space and time. Understanding how cells achieve organized collective motility is crucial to addressing cellular and tissue function and disease progression. While current two-dimensional model systems recapitulate the dynamic properties of collective cell migration, quantitative three-dimensional equivalent model systems have proved elusive. The overarching hypothesis of this work is that cell collectives are heterogeneous in nature; and that the influence of biochemical, physical, and mechanical factors combined leads to diverse physical behaviors. The central goal of this work is to establish standard tools for the understanding of 3D collective cell motility by treating individual cell-collectives as independent entities. An experimental model studies cell collectives by tracking individual cells within cell cohorts embedded in three dimensional collagen scaffolding. A computational model of 3-dimensional multi-scale self-propelled particles recreates experimental data and accounts for intercellular adhesion dynamics. A custom algorithm identifies cellular cohorts from experimental and simulated data so these may be treated as independent entities. A second custom algorithm quantifies the temporal and spatial heterogeneity of motion in cell cohorts during ‘motility events’ observed in experiments and simulations. The results show that cell-cohorts in 3D are dynamic with spatial and temporal heterogeneity; cohesive motility events can emerge without an external driving agent. Simulated cohorts are able to recreate experimental motility event signatures. Together these model systems and analytical techniques are some of the first to address collective motility of adhesive cellular cohorts in 3-dimensions.

Polysialic acid (PSA) is a developmentally regulated glycan made of repeating sialic acid monomers with α2-8 linkages. PSA has very limited expression in adults, and modifies only a few cell-surface proteins. However, PSA is overexpressed in several human cancers and is associated with metastasis and poor prognosis. We have described a derivative of PSA containing a mixture of de-N-acetyl and N-acetyl neuraminic acid residues (dPSA) found intracellularly in many normal human tissues but expressed at much higher levels on the cell surface of many human cancer cell lines. The proteins modified with dPSA and dPSA function in normal and abnormal human biology are unknown. The purpose of this study was to identify protein(s) modified with PSA and possible dPSA-dependent functions in cancer cell lines that express dPSA antigens. Using co-immunoprecipitation with the anti-dPSA monoclonal antibody SEAM 2 and mass spectroscopy, we identified membrane-associated nucleolin that is either directly modified or associated with dPSA. In addition, knocking down expression of the polysialyltransferase ST8SiaII (STX) in SK-MEL-28 human melanoma cells nearly eliminated dPSA and nucleolin from membranes but had no effect on the levels of nuclear nucleolin, and resulted in aberrant cell morphology, cell adhesion, and motility. The data suggest that cell-surface nucleolin depends on modification with dPSA, and that dPSA-modified nucleolin has an important role in cell adhesion and migration.

La formation des métastases est un processus multi étapes à travers lequel les cellules cancéreuses se détachent de la tumeur primaire pour envahir à distance dans un site secondaire. L’acquisition de capacités migratoires et invasives des cellules tumorales est cruciale dans la cascade métastatique. L'activation mutationnelle des protéines Ras favorise l'oncogenèse en perturbant une multitude de molécules et de voies qui sont impliquées dans la régulation de plusieurs processus, y compris l'invasion cellulaire et la motilité. Les petites Rho GTPases (Rac1, Cdc42 et RhoA)jouent un rôle central en contrôlant la migration cellulaire via l'assemblage des fibres d'actine, la contractilité de l'actomyosine et des microtubules.Rac1 stimule la motilité de type mésenchymateuse en favorisant la formation de lamellipodes via la formation du complexe régulateur Wave(WRC), un promoteur clé de la polymérisation de l'actine. Les protéines Ral,une autre famille de petites GTPases agissant en aval de Ras, a récemment été impliquée dans la régulation de la migration cellulaire. En particulier,RalB joue un rôle essentiel dans la motilité cellulaire en mobilisant le complexe Exocyst, son principal effecteur. Durant mon projet de thèse,nous avons investigué les mécanismes moléculaires qui contrôlent la motilité cellulaire et l'invasion en aval de la voie oncogénique Ras via le complexe RalB / Exocyst.Dans la première partie de ce manuscrit, nous avons identifié et caractérisé que le complexe WRC est à la fois un nouveau partenaire et mais aussi acteur du complexe Exocyst. En outre, nous démontrons que le complexe Exocyst dirige le complexe WRC à l’extrémité des cellules4mobiles. Cette hypothèse a été caractérisée dans la deuxième partie du manuscrit. En effet, en utilisant la technique d’optogénétique nous avonsmis en évidence le mécanisme moléculaire impliqué dans l'invasion.L’activation de RalB par Ras via les facteurs d'échange Rgl1 et Rgl2,mobilise le complexe Exocyst qui recrute ainsi le complexe WRC à l’extrémité des cellules. Cette cascade d’activation favorise la formation de protrusions, la migration et l'invasion. De manière surprenante, nous montrons que la GTPase Rac1, considérée comme la GTPase clée dans la formation de protrusions cellulaires, n'est pas impliquée dans ce processus.Enfin, nous avons analysé le niveau des protéines Ral dans une cohorte de patientes atteintes de cancer du sein. Nos résultats montrent pour la première fois une accumulation de la protéine RalB dans les compartemets invasif et métastatique suggérant un rôle potentiel de RalB dans l'invasion et la propagation métastatique des cancers du sein humain. Pour conclure,notre travail met en évidence un rôle crucial de la voie Ral, souvent sousestimée,dans le contexte de l'invasion cancéreuse. / Metastasis is a multistep process by which cancer cells migrate awayfrom the primary neoplastic mass to give rise to secondary tumors at distantsites. Thus, the acquisition of motility and invasive traits by tumor cells is acrucial step for metastasis to occur. Mutational activation of Ras proteinspromotes oncogenesis by disturbing a multitude of molecules andpathways that participate to the regulation of several processes includingalso cell invasion and motility. Among them a central role is played by Rhosmall GTPases (Rac1, Cdc42 and RhoA) which control cell migrationthrough their actions on actin assembly, actomyosin contractility andmicrotubules. Rac1 drives mesenchymal-type motility by promotinglamellipodia formation via the Wave Regulator Complex (WRC), a keypromoter of actin polymerization. Another family of small GTPases that actdownstream Ras, the Ral proteins, has been recently involved in theregulation of cell migration. RalB, through the mobilization of its maineffector the Exocyst complex, was shown to play an essential role in cellmotility. In this work of thesis, we investigated the molecular mechanismsthrough which RalB/Exocyst pathway controls cell motility and invasiondownstream oncogenic Ras.In the first part of this manuscript we describe the identification andcharacterization of the WRC complex as a novel interactor of the Exocyst.Furthermore, we provide evidences for Exocyst to be involved in drivingthe WRC to the leading edge of motile cells. This hypothesis, was finallydemonstrated in the second part of the manuscript. We were able to definethe mechanisms underlying the function of RalB in invasion by exploitingan optogenetic approach. We found that RalB, activated by Ras via the2Rgl1 and Rgl2 exchange factors, mobilizes the Exocyst complex whichrecruits the Wave Regulatory Complex (WRC) at cell edge, promotingprotrusions, migration and invasion. Even more, we show that the Rac1GTPase, usually considered the master of cell protrusions, is not involvedin this process. Finally, we analyzed Ral proteins expression in a cohort ofbreast cancer samples, pointing out for the first time an accumulation ofRalB in the invasive and metastasis compartments, suggesting a role ofRalB in invasiveness and metastatic spread of human breast cancers. Takentogether our work contribute to light up the role of the underestimated Ralpathway in the context of cancer invasion.