The cell-cell and cell-matrix adhesion molecule CD44 and its numerous splice variants are involved in a multitude of physiological and pathological processes, including tumour progression. Especially variant CD44 has been implicated in metastasis formation. For long term in vivo experiments on metastasis formation, a plasmid based RNAi technique was applied to generate stable splice variant ‘v7’-specific CD44 knock-down clones of a highly metastatic rat pancreatic adenocarcinoma cell line (BSp73ASML). The resulting phenotype was characterized with an emphasis on interactions of CD44v with the tumour surrounding during the course of metastasis formation. Loss of CD44v is accompanied in vivo by a marked reduction in metastatic growth in the lymph nodes and particularly in the lung, which could be reverted by restoring CD44v expression in the knock-down cells. The impaired metastatic growth was not due to a lower proliferative activity or a reduced anchorage-independence of these cells in vitro. Instead, they display several defects, which can be attributed to perturbed interactions of CD44v with the microenvironment. Compared to ASMLwt cells CD44vk.d. cells do not form cell aggregates in stromal surroundings, such as lymph nodes and the lung, due to lost cell-cell adhesion, mediated by interactions of CD44v and hyaluronic acid (HA). Furthermore, CD44vk.d. cells exhibit an impaired matrix production, as CD44v is most likely involved in the assembly of matrix components, containing HA, collagen and laminin. The matrix supports rapid adhesion of ASML cells through ?1 integrin and in addition contributes to survival. Finally, the loss of CD44v is accompanied by a marked decrease in apoptosis resistance. Impaired PI3K-Akt survival signalling, activated by CD44v was identified as the cause of this defect. In conclusion, CD44v contributes to the metastatic phenotype of ASML cells as a multifunctional player interacting with the surrounding in several ways. First, as cell-cell adhesion molecule by mediating cell aggregation, second, as cell-matrix adhesion molecule by organizing matrix generation and last, as signalling molecule supporting survival. This highlights the role of variant CD44 in the metastatic spread of tumour cells through complex interactions with the tumour microenvironment and underlines the important role of a highly regulated interplay between tumour cells and their surrounding for metastasis formation.
Iron is the most abundant trace element in the human body and zinc is the second one. Control of ion homeostasis is of vital importance for mammals and bacteria. Regulation of the ion flux into or out of the cell is a complex and articulated mechanism that still needs to be elucidated. The highly specialized pathogen Mycobacterium tuberculosis has to contend with iron sequestration in order to survive in the human body. Iron metabolism is regulated by controlling transcription of genes involved in iron uptake, transport and storage. Paucity of this metal triggers an extensive response to increase iron acquisition whereas an excess of it can be toxic for the cell. The control of intracellular iron concentration is also linked to other important processes including oxidative stress response and the regulation of virulence factors. Recent studies have shown that, in patients affected by TB/HIV exposed to high level of dietary iron, the risk of active pulmonary tuberculosis increases. In M. tuberculosis the ferric uptake regulator A (FurA) is activated by Fe2+ to bind specifically to its target DNA sequence thereby repressing the downstream genes. Zinc is also an important element for all living organisms and serves as a cofactor in all six classes of enzymes and also in several regulatory proteins. The intracellular concentration of this metal must be carefully regulated due its toxicity. Compared with eukaryotes, little is known about zinc homeostasis in bacteria. In the tuberculosis genome several genes coding for zinc proteins have been classified but curiously no zinc regulator has been yet defined. Surprisingly, instead, two Fur genes were identified: Mtb furA and Mtb furB, but no clear structural or functional distinction has been reported. In this thesis a careful and detailed structural and biological description of FurA and FurB proteins is presented. Using a variety of biochemical and biophysical methods - including electrophoretic mobility shift assay (EMSA), site directed mutations, isothermal calorimetry (ITC), microPIXE analysis, extended X-ray absorption fine spectroscopy (EXAFS) and X-ray crystallography - we investigated the metal binding sites together with the nature and the structure of these proteins. The combination of these results enable us to distinguish between structurally and functionally distinct metal binding sites, provide a meticulous description and qualitative and quantitative characterization of them, propose biological roles and present for the first time a 3D picture of a zinc uptake regulator.
Peroxisomaler Proteinimport erfolgt mit Hilfe einer speziellen Translokalisations-Machinerie an der peroxisomalen Membran. Obwohl die beteiligten Proteine während der letzten Jahre identifiziert wurden, sind Details zum Mechanismus der Translokalisation noch nicht bekannt. Aktuelle Ergebnisse lassen auf ein „cycling receptor“ Model schließen, welches aus „cargo recognition“ (der Aufnahme von peroxisomalen Matrixproteinen im Cytosol), dem „docking“, der Abgabe der aufgenommen Proteine ins peroxisomalen Lumen und dem Rezeptor Recycling besteht. Das membranassozierte Peroxin Pex14p wird im Allgemeinen als eine Hauptkomponente des peroxisomalen „Docking“-Komplexes angesehen. Es interagiert neben einigen membrangebundenen Peroxinen mit den PTS-Rezeptoren Pex5p und Pex7p und besteht aus drei Domänen: einem konservierten N-terminus, einer hydrophoben Region und einer coiled-coil Domäne. Die N-terminale Domäne erkennt sogenannte WxxxF/Y-Motive, konservierte aromatische Sequenzen im PTS1-Rezeptor Pex5p. Obwohl es kein klassisches WxxxF/Y Motiv aufweist bindet Pex19p, ein Protein mit einer Schlüsselfunktion in der peroxisomalen Biogenese, dieselbe N-terminale Domäne von Pex14p wie Pex5p. In der vorliegenden Arbeit wurden funktionelle und strukturelle Studien am N-terminalen Tel von Pex14p durchgeführt. Die erhaltenen 3-dimensionalen Modelle beschreiben ein 3-Helix Bündel, das eine hydrophobe Interaktionsfläche für amphipatische, helikale Liganden darstellt. Der Vergleich zwischen der Komplexstruktur von Pex14p (aa 16-80W)-Pex5p (aa 116-124) mit der von Pex14p (aa 16-80W)-Pex19p (aa 66-77) zeigt dass beide die gleich Bindestelle besetzen, wobei Pex19p eine unerwartete invertierte Orientierung aufweist. Die Strukturdaten wurden durch NMR Titrations- und ITC-Experimente ergänzt, welche die kompetitive Bindung von Pex5p und Pex19p bestätigten und Pex5p als den stärkeren Liganden charakterisieren. Die so gewonnen Ergebnisse erlauben einen Einblick in die molekularen Abläufe während der frühen Schritte des peroxisomalen Imports und implizieren Voraussetzungen die mögliche Pex14p-Interaktionspartner erfüllen sollten.
In der vorliegenden Arbeit wurde die Präsenz spontan generierter, tumorreaktiver Gedächtnis T-Zellen bei Kopf-Hals-Tumorpatienten gezeigt. Mit Hilfe autologer dendritischer Zellen, die zuvor mit autologem Tumorlysat beladen worden waren, wurden tumorspezifische, inflammatorische Gedächtnis T-Zellen nachgewiesen. 25% der Patienten wiesen eine spontane Tumorimmunität auf. CD4 T-Zellen waren zu 16% tumorreaktiv, CD8 T-Zellen zu 28%. Die Frequenz tumorspezifischer Gedächtnis T-Zellen war in der CD4 Subpopulation signifikant erhöht. CD8 T-Zellen demonstrierten ein polyvalentes T-Zellrepertoire. Die Mehrheit der Tumorpatienten ohne detektierbares Anti-Tumorgedächtnis waren im fortgeschrittenen Spätstadium (Stadium IV). Umfangreichere Analysen der zellulären Tumorimmunität ergaben, dass diese zu 50% Th2 polarisiert war, während 25% der Patientenlymphozyten Th1 Polarisierung aufwiesen. Tumorspezifische, inflammatorische Gedächtnis T-Zellen wurden im peripheren Blut und tumordrainierenden, nicht metastasierten Lymphknoten in ähnlicher Frequenz detektiert. Tumor-Immunreaktivität war im lymphatischen Organ auffällig öfter nachweisbar: 19% Immunreaktivität der Lymphknoten T-Zellen und 5% der Blutlymphozyten. Im Lymphknoten waren CD4 und CD8 T-Zellantworten gleich vertreten. Im Blut war Tumor-Immunreaktivität tendenziell öfter in der CD8 Subpopulation detektierbar. Lymphatische Gedächtnis T-Zellen waren doppelt so oft Th1 polarisiert als entsprechende Lymphozyten im Blut. Die Th2 Polarisierung war in beiden Zellkompartimenten gleich. In Tumorgeweben und Blutplasmata von Kopf-Hals-Tumorpatienten wurden immunsuppressive und -stimulatorische Zytokine detektiert. Im Vergleich zu gesunden Donoren waren erstere in der Regel erhöht, letztere erniedrigt. Eine signifikant inverse Korrelation bestand zwischen erhöhtem TGF-β1 Vorkommen von Tumorgeweben und IFN-α Nachweis in solchen, sowie Induktion von Th1 Tumor-Immunreaktivität. Kopf-Hals-Tumorpatienten mit 3-Jahres-Rezidivmanifestation zeigten signifikant erhöhte TGF-β1 Konzentrationen im Tumorgewebe. Malignome von Kopf-Hals-Tumorpatienten ohne Th1 Tumorreaktivität wurden tendenziell von mehr naiven T-Zellen im Vergleich zu reaktiven Patienten infiltriert. In Tumore von nicht Th1 tumorreaktiven Patienten wanderten signifikant mehr CD4 T-Zellen ein. Dieser Befund korrelierte signifikant mit einer erhöhten Stadium IV Zuordnung der Patienten und mit einem erhöhten TGF-β1 Nachweis im Tumorgewebe. Erhöhtes Vorkommen von tumorinfiltrierenden CD4 T-Zellen und/oder viel TGF-β1 beeinflussten die Tumorgröße signifikant. In Anwesenheit von Th1 tumorreaktiven Gedächtnis T-Zellen im Blut wurden die Malignome von signifikant mehr CD8 als CD4 T-Zellen infiltriert. Die CD8 T-Zellen konzentrierten sich peritumoral und drangen nicht tiefer ins Tumorinnere ein. Die Tumore wurden bei einer Th1 Tumorreaktivität von Lymphknoten T-Zellen gleichmäßig von CD4 und CD8 T-Zellen infiltriert. Letztere waren sowohl peri- als auch intratumoral nachweisbar. Die hiesigen Analysen liefern einen wichtigen Beitrag zum Verständnis von Tumor-Immuninteraktionen bei Kopf-Hals-Tumorpatienten.
Insulin-like growth factor-1 (IGF-1) has pleiotropic effects on various tissues during pre- and postnatal development by stimulating proliferation and differentiation, and plays a major role in hypertrophy and tissue remodeling in the fetal, perinatal, and adult organism. In skeletal muscle the role of IGF-1 is well established, but questions remain about the significance of the various isoforms that are produced from the single-copy IGF-1 gene. By the use of different promoters, differential splicing and post-translational modifications, at least six IGF-1 isoforms are generated from the IGF-1 gene, that differ in the N-terminal signal peptide (Class) and the C-terminal E-peptide. The aim of this work was to dissect the different roles of IGF-1 isoforms in skeletal muscle in vitro and in vivo. Cell culture experiments revealed that IGF-1Ea isoforms promoted myogenic differentiation and cell hypertrophy, resulting in enlarged myofibers, while IGF-1Eb isoforms instead did not show an effect on fiber size but on proliferation of myoblasts. Correlating with the results obtained in vitro, transgenic animals over-expressing IGF-1Ea isoforms showed pronounced muscle fiber hypertrophy, accompanied by an increase in force generation and strength, while IGF-1Eb isoforms showed very mild effects on muscle size and no changes in muscle strength, further implicating the Ea-peptide in the hypertrophic response. Analysis of the intracellular signals transduced by the different IGF-1 isoforms revealed a complex regulatory network, excluding certain pathways previously implicated in the induction of skeletal muscle hypertrophy in response to IGF-1. Preliminary analysis of regeneration in response to IGF-1 isoforms demonstrated that each isoform enhanced the regeneration process, suggesting that Eb-peptide-containing isoforms did so by stimulating the proliferation of satellite cells, while IGF-1Ea enhanced the growth of newly forming fibers. Class 2 IGF-1Eb was found to specifically induce a calcineurin isoform (CnAß1) that has been linked to enhanced regeneration. In addition, this thesis describes the cloning of a Class 2 IGF-1En isoform that was previously not described in rodent species, but known to exist in humans. The present work constitutes the first evidence for different functions of IGF-1 isoforms in vitro and in vivo, provides an overview of their variable effects in skeletal muscle and a strong basis for future research into their specific functions.
Die vorliegende Dissertation befasst sich mit der Herstellung und funktionellen Analyse neuartiger, IL-2- bzw. GM-CSF-enthaltender bi- und trispezifischer single-chain Antikörper (Immunzytokine) zur Verbesserung der antitumoralen Immunantworten bei Krebspatienten. Die Arbeit ist thematisch in ein Projekt zur Entwicklung einer aktiv-spezifischen Immuntherapie (ASI) von Tumoren eingebunden. Bei dieser Therapie werden inaktivierte, autologe Tumorzellen (aus dem chirurgisch entfernten Primärtumor des Patienten) mit dem Newcastle Disease Virus (NDV, Stamm Ulster) infiziert und die so hergestellten autologen Tumorvakzinezellen (ATV-NDV) dem Patienten nach einem bestimmten Schema geimpft. Die beiden schon beschriebenen bispezifischen, rekombinant hergestellten Konstrukte bs HN-CD3 und bs HN-CD28 sind so beschaffen, dass sie mit einem Arm an das virale Antigen Hämagglutinin-Neuraminidase (HN) auf der Oberfläche der Vakzinezellen binden. Der zweite Arm hingegen ist entweder gegen das T-Zell-Rezeptor-assoziierte CD3-Molekül oder gegen den kostimulatorisch wirkenden CD28-Rezeptor gerichtet. Diese zwei Signale rekrutieren und aktivieren T-Zellen am Ort der applizierten Vakzine. Die in dieser Arbeit vorgestellten bi- und trispezifischen Antikörper enthalten darüber hinaus die Zytokine IL-2 und/oder GM-CSF und sollen durch das Einbringen zusätzlicher immunstimulierender Signale die Wirkung der Vakzine verbessern. Zunächst konnte demonstriert werden, dass es möglich war, die neuen Antikörper stabil in CHO Zellen zu produzieren. Die aufgereinigten Antikörper, die nicht nur gegen HN, sondern auch gegen das virale Fusionsprotein (F0) gerichtet sein können, wiesen alle Bindungseigenschaften und biologischen Aktivitäten auf, um stimulierend wirken zu können. In Gegenwart eines suboptimalen CD3-Signals induzierten Antikörper, die zusätzlich zum CD28-Signal einen IL-2-Zytokinanteil aufwiesen, eine stärkere und länger anhaltende Aktivierung von T-Zellen, welche aus mononukleären Zellen des peripheren Bluts (PBMZ) gesunder Spender isoliert wurden, als die Verwendung von CD28-Signal oder IL-2 alleine. Diese verstärkte antitumorale Wirkung konnte in dem neu etablierten in vitro Tumor-Neutralisations-Assay (TNA) nachgewiesen werden. V.a. naive T-Zellen, die im Gegensatz zu Gedächtnis-T-Zellen noch keinen antigenspezifischen Kontakt durchlaufen haben, ließen sich durch diese neuartigen Antikörper stark stimulieren und besaßen zytotoxische Aktivität. Neu in diesem Zusammenhang war, dass IL-2 eine vergleichbare kostimulatorische Wirkung auf die Aktivierung naiver T-Zellen ausübte wie das CD28-Signal. Zudem wurde nachgewiesen, dass CD14+ Monozyten durch GM-CSF-enthaltende Fusionsproteine aktiviert werden konnten und im TNA zum Antitumor-Effekt beitrugen. Die Ergebnisse dieser Dissertation lassen insgesamt darauf schließen, dass die klinische Anwendung einer Kombinationstherapie – als eine Erweiterung der Grundvakzine ATV-NDV durch Anbindung von CD3- und CD28-Signalen zusammen mit IL-2- und GM-CSF-enthaltenden Immunzytokinen – vielversprechend ist.
Understanding cellular signaling pathways and their cross talk increasingly depends on the visualization of their spatio-temporal dynamics. This thesis investigates different factors influencing the performance of two FRET based protein kinase C (PKC) probes, KCP-1 and KCP-2. KCP-1 consists of the truncated PKC substrate pleckstrin, sandwiched between two fluorescent proteins, EYFP and GFP2. KCP-2 is a shortened version of KCP-1, missing the last 18 amino acids of the pleckstrin insert (the acidic loop). The EYFP/GFP2 emission ratio of KCP-1 increases upon phosphorylation, while it decreases in KCP-2. Both probes are reversible. We examined the influence of linker length, charge distribution in particular domains of the probe, and fluorophore dimerization on probe performance. Different FRET pairs, including novel fluorescent proteins and novel labeling techniques, were tested in an effort to vary and optimize the spectral properties of KCP probes. Both probes, KCP-1 and KCP-2, were shown to be sensitive to elongation in their N-terminal linker region. The signal amplitude of the probe was diminished with increasing linker length. Shortening of the same linker region reduced probe performance, although not to the same extent as elongation. This demonstrated the importance of linker length for proper orientation of the fluorophores in the probe molecule. Any changes in the N-terminal PH domain of the pleckstrin insert, for example shortening or replacing basic amino acids with uncharged or acidic residues, had severe impact on probe performance. The direction of the KCP-1 signal was reversed, showing a decrease instead of an increase after phosphorylation. The reversal of signal in KCP-1 probes reflects that specific interactions between the PH domain and the acidic loop are crucial for probe performance. In KCP-2, the impact was less striking, although the signal was reduced. Abolishing the fluorescent proteins' ability to dimerize, led to strongly reduced KCP-2-like signals in both probes. This is the first time that fluorophore dimerization was shown to be essential for the mechanism of action of a genetically encoded sensor. Based on this data, we propose the following models for the mechanism of these PKC probes: In unphosphorylated KCP-2, dimerization of fluorophores serves as a clamp, pulling the two fluorophores together to a closed, quasi cyclized conformation. Phosphorylation of the probe leads to a more open conformation, increasing the average distance between the two fluorophores. This results in a decrease of FRET efficiency after phosphorylation. Two interactions determine the mechanism in KCP-1. Again, the dimerization of the two fluorophores serves as a clamp, pulling the fluorophores to close proximity. An additional interaction between the PH domain and the acidic loop changes the relative orientation of the transition dipole moments in the probe, resulting in a decreased initial FRET efficiency, compared to KCP-2. Phosphorylation of the probe imposes a strain on the intramolecular architecture that rearranges the transition dipoles, leading to an increase in FRET efficiency. The need for a molecular clamp, dimerization of fluorescent proteins in KCP-1 and KCP-2 probes, within the molecule explains why probes containing new monomeric fluorescent proteins or other labels based on full sized proteins did not yield functional sensors. Using the small fluorophore FlAsH, however, we were able to create a smaller sized PKC probe. This probe will allow NMR experiments and structural analysis of a FRET probe for the first time.
The formation of functionally distinct cell types from a single zygote is a fundamental aspect of metazoan development. The specification of cell identities usually occurs during early development, long before this information is used during differentiation. The molecular basis of cell fate specification is the activation of characteristic gene expression programs. A prerequisite for the appropriate differentiation of cells according to their fate is therefore the faithful transmission of once established gene expression states throughout phases of cell proliferation and growth. This task is accomplished by the so-called cellular memory. A paradigm to study the molecular basis of the cellular memory is the regulation of homeotic gene expression in Drosophila. A number of studies have shown that the cellular memory is an epigenetic mechanism based on the modification of the chromatin structure. This is accomplished through the antagonizing activities of repressive PcG (Polycomb group) and activating TrxG (Trithorax group) proteins. The recruitment of both PcG as well as TrxG proteins to their target loci depends on the presence of identical cis-regulatory elements termed Polycomb group response elements, or PREs. How is the decision between epigenetic silencing and activation at a PRE taken? Presumably, the default state of a PRE is to function as a PcG-dependent silencer, whereas the conversion into the active mode requires incoming signals. It has previously been shown that the epigenetically active state at a PRE correlates with non-coding transcription through these elements themselves. This raised the question whether the non-coding transcription is of functional significance for the epigenetic activation of a PRE, or if it reflects only a consequence of this process. The major aim of this thesis was to determine whether the transcription through the well-characterized Fab-7 PRE has any function in the regulation of the epigenetic state of this element. In the first part, a transgenic reporter system was used to answer this question. The results obtained show that the transcription through Fab-7 functions as a novel anti-silencing mechanism that counteracts the PcG-mediated repression by default. Depending on the tissue and/or on the locus, transcription through PREs may be required throughout development to prevent the re-establishment of PcG silencing. Importantly, PREs located outside the homeotic gene complexes are also transcribed, suggesting that the anti-silencing function of transcription through PREs may be a fundamental aspect of the cellular memory. In the second part, the transgenic reporters mentioned above were used to analyze the epigenetic consequences of transcription through the Fab-7 PRE. The results suggest that at least on the transgene, both repressive PcG as well as activating TrxG proteins associate with the Fab-7 PRE, irrespective of the epigenetic state of this element. The primary function of transcription through the Fab-7 PRE may therefore be to modulate the activities of PcG/TrxG complexes, rather than regulating their differential recruitment to the chromatin. To get insight into the precise function of non-coding transcription through PREs, the properties Fab-7 RNA were investigated in SF4 tissue culture cells in vitro, as well as in wildtype embryos in vivo. Although a sequence-specific function of Fab-7 RNA is unlikely, the results obtained do not exclude structural function of non-coding PRE transcripts. Futhermore, Fab-7 RNA can be detected on mitotic chromatin, suggesting that the transmission of epigenetically activated states through cell division might depend on the timing of non-coding transcription with respect to the cell cycle and/or the association of non-coding RNAs with the chromatin. Since the functions of PcG/TrxG proteins have been conserved during evolution, the results presented here might have implications on our understanding of epigenetic gene regulation in mammals.
The protozoan parasite Trypanosoma brucei, which causes human sleeping sickness, has to adapt to rather different environments as it cycles between the mammalian host and the tsetse fly vector. This adaptation is mediated by changes in trypanosome gene expression, which are mainly regulated at the post-transcriptional level. Proteins with an RNA-binding "Puf" domain are important for post-transcriptional control by modulation of mRNA stability and regulation of translation in other species. This suggested that Puf domain proteins might also have a similar role in trypanosomes. In addition to the previously characterized TbPUF1 (Hoek, Zanders et al. 2002) I have identified eight more PUF genes in the T. brucei genome. A comparison of the characteristic RNA binding domain of T. brucei PUF proteins suggested that they bind related but distinct targets. Interestingly, each of the PUF protein has an orthologue in Trypanosoma cruzi and Leishmania major. Phylogenetic analysis suggested that there were several kinetoplastid PUF proteins before separation from the eukaryotic lineage. Depletion of each of the nine T. brucei PUF proteins by RNA interference did not result in an obvious phenotypic change. Furthermore, PUF1 knock out procyclic cell lines were viable, indicating that this PUF protein is not essential for in vitro growth. Additionally, double RNAi analyses suggested that the proteins tested did not share redundant functions in T. brucei. Microarray studies comparing wild-type cells with cells where PUF levels have been perturbed (either by RNAi or overexpression) revealed one putative mRNA target for PUF5: CAP17 (corset associated protein 17) mRNA is downregulated upon overexpression of PUF5 in the insect form of the parasite. Interestingly, PUF5 overexpression was also lethal for procyclic cells. Multiple mRNAs which associated with PUF proteins in mRNP complexes were identified. PUF5 for example, selectively binds to mRNAs encoding for amino acid transporters in bloodstream form cells. Attempts to identify PUF interaction partners have so far failed. The effect of PUF proteins on global protein expression level was also investigated using two-dimensional gel electrophoresis approach. A few proteins which were differentially regulated upon RNAi or knockout of PUF proteins were identified. These results indicated that PUF proteins in T. brucei, as is true for PUF proteins in general, are involved in regulating gene expression.