Tumorstammzellen scheinen das Triebwerk für die Initiierung und Progression des Mammakarzinoms zu sein. Durch ihr Potential zur Proliferation von Tumorgewebe, zur Metastasierung und zur Bildung von Rezidiven bestimmen sie maßgeblich die Prognose und Mortalität von Brustkrebspatientinnen. Diese Arbeit demonstriert, welche Mechanismen sich Brustkrebsstammzellen zu Nutze machen, um einer Immunantwort durch NK Zellen zu entkommen. Mittels durchflusszytometrischer Analysen konnte innerhalb der Gesamtpopulation an MCF 7-Brustkrebszellen eine CD44highCD24low-Subpopulation, die dem Tumorstammzellanteil entspricht, abgegrenzt werden. Im Vergleich zur Ausgangspopulation war nach einer Kokultur mit aktivierten NK Zellen gesunder menschlicher Spender eine Anreicherung von Tumorstammzellen in vitro zu verzeichnen. Die Inkubation von Brustkrebszellen mit NK Zell-Überstand führte zu keiner wesentlichen Veränderung der Tumorstammzellpopulation, was die Notwendigkeit eines direkten Zell-Zell-Kontakts impliziert. Diese Tumorstammzellen könnten nach einem Angriff durch NK Zellen einerseits durch Selektion übrig geblieben sein oder andererseits durch epithelial-mesenchymale Transition (EMT) neu entstanden sein. Hinweise auf einen Selektionsprozess ließen sich anhand der verminderten Oberflächenexpression von NK Zell-Liganden auf Tumorstammzellen im Vergleich zu Nichtstammzellen finden. Die untersuchten Brustkrebszelllinien (MCF 7, SKBR 3, BT 474 und MDA MB 231) besaßen ein jeweils individuell reguliertes Muster der aktivierenden NKG2D Liganden (MICA, MICB, ULBP1, ULBP2, ULBP3), DNAM 1-Liganden (CD112, CD155) und von MHC1-Molekülen auf Tumorstammzellen und Nichtstammzellen. Die niedrigere Expression von NK Zell-Liganden auf Tumorstammzellen lässt auf eine verminderte Angreifbarkeit durch NK Zellen schließen. Eine Induktion von Tumorstammzellen aus differenzierten epithelialen Tumorzellen via EMT nach einer Kokultur mit NK Zellen konnten wir beweisen. Aus einer stammzelldepletierten MCF 7-Population gingen nach dem Kontakt zu NK Zellen Tumorzellen mit dem Phänotyp CD44highCD24low de novo hervor. Die Herunterregulation des epithelialen Adhäsionsmoleküls E-Cadherin sowie die Hochregulation mesenchymaler Marker wie des Strukturproteins Vimentin, der EMT-auslösenden Transkriptionsfaktoren Slug, Snail und Twist, und der stammzelltypischen Transkriptionsfaktoren Oct4, KLF4 und cMyc auf mRNA-Ebene sprachen für eine EMT-getriggerte Induktion von Tumorstammzellen nach einer Kokultur von MCF 7-Zellen mit NK Zellen. Desweiteren stellten wir fest, dass der direkte Kontakt zwischen Tumorzellen und NK Zellen für die Induktion von Tumorstammzellen von großer Bedeutung ist, und zwar auch nach Inhibition des zytotoxischen Effektorpotentials der NK Zellen. Diese Zell-Zell-Interaktionen scheinen von NKG2D und DNAM 1 abhängig zu sein und eine konsekutive Stammzellinduktion via EMT zu beinhalten. Da aus einer nativen Population nach dem Kontakt zu NK-Zellen ein doppelt so hoher Anteil an Tumorstammzellen hervorging wie aus einer ebenso mit NK-Zellen behandelten stammzelldepletierten Fraktion, ist davon auszugehen, dass ein überdurchschnittlich gutes Überleben von Tumorstammzellen unter NK-Zell-vermitteltem Selektionsdruck auch zum „Immune Escape“ beitragen kann. Hinsichtlich ihrer Klonogenität gab es zwischen bestehenden und induzierten Tumorstammzellen keinen Unterschied. Beide Fraktionen waren in gleichem Ausmaß in der Lage neue Kolonien zu bilden. Es konnte also gezeigt werden, dass eine EMT-getriggerte Induktion im Sinne eines „Immune Escapes“ von Brustkrebszellen nach dem Kontakt zu NK Zellen maßgeblich zur Tumorstammzellanreicherung beiträgt. Ein zusätzlicher Selektionsprozess bestehender Tumorstammzellen kann als wahrscheinlich angenommen werden. Interaktionen über die NK Zell-Rezeptoren NKG2D und DNAM 1 bzw. deren Liganden auf Tumorzellen scheinen eine Schlüsselrolle zu spielen. Sie könnten als Ansatzpunkt für medizinische Interventionen dienen, die zur Verhinderung einer Tumorstammzellanreicherung im Mammakarzinom beitragen und somit die Prognose von Brustkrebspatientinnen verbessern. / Tumor stem cells seem to be the engine for initiation and progression of breast cancer. By their potential for unlimited proliferation, dissemination and relapse they largely determine the prognosis and mortality of breast cancer patients. This thesis shows mechanisms breast cancer (stem) cells use to escape from an immune response by NK cells. Using flow cytometry analysis, we defined a subpopulation of CD44highCD24low cells corresponding to the tumor stem cell fraction of MCF-7 breast cancer cells. Compared to the native MCF-7 cell population we observed an enrichment of tumor stem cells in vitro after co-culture with activated NK cells from healthy human donors. Incubation of breast cancer cells with NK cell supernatant resulted in no significant change, which implies that the observed NK cell-mediated enrichment of the cancer stem cell population depends on direct cell-cell contact. One possibility is that these tumor stem cells could be enriched by selection, i.e. by preferential killing of their more differentiated counterparts. Alternatively, they could arise de novo via epithelial to mesenchymal transition (EMT). A selection process was supported by data showing reduced surface expression of NK cell ligands on tumor stem cells compared to non-stem cells. The investigated breast cancer cell lines (MCF-7, SKBR 3, BT 474 and MDA MB 231) showed a distinctly regulated pattern of activating NKG2D-ligands (MICA, MICB, ULBP1, ULBP2, ULBP3) and DNAM 1-ligands (CD112, CD155) and of MHC1-molecules on tumor stem cells and non-stem cells. The comparatively lower expression of NK cell ligands suggests a reduced vulnerability of tumor stem cells towards NK cells. However, we also showed the induction of tumor stem cells from differentiated epithelial tumor cells via EMT. Tumor cells with the phenotype CD44highCD24low arose de novo from a stem cell depleted MCF-7 tumor cell population which was then co-incubated with NK cells. Downregulation of the epithelial cell adhesion molecule E-cadherin as well as upregulation of mesenchymal markers such as the structural protein vimentin, the EMT-inducing transcription factors Slug, Snail and Twist and the stem cell-typical transcription factors Oct4, KLF4 and cMyc on mRNA level indicate an EMT-triggered induction of tumor stem cells after co-culture of MCF-7 with NK cells. We also found that the direct contact between tumor cells and non-lytic NK cells is of vital importance for the induction of tumor stem cells. These cell-cell interactions appeared to depend on NKG2D and DNAM-1 and to include a consecutive stem cell induction via EMT. As the proportion of tumor stem cells after co-culture of native MCF-7 cells and NK cells was almost twice the number of stem cells arisen from an equally treated stem cell depleted fraction we can assume that an above-average survival of tumor stem cells due to selection stress can also contribute to immune escape. Regarding their clonogenicity we noticed no difference between pre-existent and induced tumor stem cells. Both fractions possessed the ability to generate new colonies of similar quantity. Thus, we showed that upon exposure to NK cells EMT-triggered induction considerably contributes to the enrichment of breast cancer stem cells. An additional process of stem cell selection seems to be realistic. Interactions of the NK cell receptors NKG2D and DNAM 1 and its ligands on tumor cells respectively seem to play a key role. The recognition that de-differentiation may represent a previously unrecognized immune escape mechanism of breast cancer cells could serve as a starting point for medical interventions, with the aim of preventing stem cell accumulation in breast cancer and thus improving the prognosis of breast cancer patients.

Techn. Universiẗat, Diss., 2004--Darmstadt.

The present thesis is a contribution to unravel the molecular mechanisms that underlie urodele regeneration. Urodele amphibians (newts and salamanders) are among the few vertebrates with the remarkable ability to regenerate lost body appendages, like the limbs and the tail. Urodele tail and limb regeneration occurs via blastemal epimorphic regeneration. A blastema is a mound of progenitor cells that accumulates at the amputation plane and eventually gives rise to the missing structures. It is known today that dedifferentiating muscle fibers at the amputation plane contribute to the blastema cell pool, but how this process occurs on the cellular and molecular level is hardly understood, which is in part due to the lack of molecular methods to test gene function in urodeles. Furthermore, little is known about how coordinated growth and patterning occurs during urodele regeneration, and if the patterning mechanisms in regeneration are related to the ones in development. The goal of this study was to better understand these processes on the molecular level. To address these questions, I first established several methods in our model systems, which are the mexican salamander Ambystoma mexicanum (axolotl) and a cell line derived from the newt Notophthalmus viridescens. In order to monitor gene expression on a cellular level during regeneration, I worked out a good in situ hybridization protocol on axolotl tissue cryosections. To be able to test gene function, I established electroporation conditions to both overexpress genes in the cultured newt cells and to deliver morpholinos into axolotl cells in vivo and newt cells in culture. I demonstrate here that morpholinos are an effective tool to downregulate protein expression in urodele cells in vivo and in culture. Testing the role of two candidate genes in muscle fiber dedifferentiation, the homeobox containing transcription factor Msx1 and Rad, a GTP-binding protein of a new Ras-related protein family, revealed that neither seems to play a major role in muscle dedifferentiation, both in culture and in vivo. In addition to testing gene function I have examined the muscle dedifferentiation process in more detail. I show here that dedifferentiating muscle fiber nuclei undergo morphological changes that are likely due to chromatin remodeling events. I also demonstrate that the axolotl spinal cord expresses embryonic dorsoventral (d/v) patterning markers of the neural tube. The transcription factors Msx1, Pax7 and Pax6 are expressed in their respective d/v domains in both the differentiated and the regenerating axolotl spinal cord. Furthermore, the secreted signaling molecule sonic hedgehog (Shh) is expressed in the floor plate in both the differentiated and the regenerating cord. Using a chemical inhibitor (cyclopamine) and an activator of the hedgehog pathway, I discovered that hedgehog signaling is required for overall tail regeneration. Blocking hedgehog signaling does not only result in d/v patterning defects of the regenerating spinal cord, but it also strongly reduces blastema cell proliferation. In addition, I identified cartilage and putative muscle progenitor cells in the blastema, marked by the expression of the transcription factors Sox9 and Pax7, respectively. Both progenitor populations are reduced in the blastema in the absence of hedgehog signaling. The continuous expression of marker genes for embryonic progenitor cell domains in the mature axolotl may be related to their ability to regenerate.

The present thesis is a contribution to unravel the molecular mechanisms that underlie urodele regeneration. Urodele amphibians (newts and salamanders) are among the few vertebrates with the remarkable ability to regenerate lost body appendages, like the limbs and the tail. Urodele tail and limb regeneration occurs via blastemal epimorphic regeneration. A blastema is a mound of progenitor cells that accumulates at the amputation plane and eventually gives rise to the missing structures. It is known today that dedifferentiating muscle fibers at the amputation plane contribute to the blastema cell pool, but how this process occurs on the cellular and molecular level is hardly understood, which is in part due to the lack of molecular methods to test gene function in urodeles. Furthermore, little is known about how coordinated growth and patterning occurs during urodele regeneration, and if the patterning mechanisms in regeneration are related to the ones in development. The goal of this study was to better understand these processes on the molecular level. To address these questions, I first established several methods in our model systems, which are the mexican salamander Ambystoma mexicanum (axolotl) and a cell line derived from the newt Notophthalmus viridescens. In order to monitor gene expression on a cellular level during regeneration, I worked out a good in situ hybridization protocol on axolotl tissue cryosections. To be able to test gene function, I established electroporation conditions to both overexpress genes in the cultured newt cells and to deliver morpholinos into axolotl cells in vivo and newt cells in culture. I demonstrate here that morpholinos are an effective tool to downregulate protein expression in urodele cells in vivo and in culture. Testing the role of two candidate genes in muscle fiber dedifferentiation, the homeobox containing transcription factor Msx1 and Rad, a GTP-binding protein of a new Ras-related protein family, revealed that neither seems to play a major role in muscle dedifferentiation, both in culture and in vivo. In addition to testing gene function I have examined the muscle dedifferentiation process in more detail. I show here that dedifferentiating muscle fiber nuclei undergo morphological changes that are likely due to chromatin remodeling events. I also demonstrate that the axolotl spinal cord expresses embryonic dorsoventral (d/v) patterning markers of the neural tube. The transcription factors Msx1, Pax7 and Pax6 are expressed in their respective d/v domains in both the differentiated and the regenerating axolotl spinal cord. Furthermore, the secreted signaling molecule sonic hedgehog (Shh) is expressed in the floor plate in both the differentiated and the regenerating cord. Using a chemical inhibitor (cyclopamine) and an activator of the hedgehog pathway, I discovered that hedgehog signaling is required for overall tail regeneration. Blocking hedgehog signaling does not only result in d/v patterning defects of the regenerating spinal cord, but it also strongly reduces blastema cell proliferation. In addition, I identified cartilage and putative muscle progenitor cells in the blastema, marked by the expression of the transcription factors Sox9 and Pax7, respectively. Both progenitor populations are reduced in the blastema in the absence of hedgehog signaling. The continuous expression of marker genes for embryonic progenitor cell domains in the mature axolotl may be related to their ability to regenerate.

Urodele amphibians have the remarkable ability to re-grow lost body parts. This regenerative response after injury in urodeles involves dedifferentiation of fully differentiated cells into proliferative cells. One well-studied example of this is the dedifferentiation of multinucleated muscle cells into mononucleate cells resembling their precursors, the myoblasts. To form these mononucleate cells the differentiated myotubes in vivo must re-enter and complete the cell cycle; they again proliferate and produce progeny. A key question is what factors induce the myotubes to re-enter the cell cycle and proliferate. Early events of cell cycle re-entry can be studied in the A1 cell line, a myogenic cell line isolated from the Notophthalmus viridescens hindlimb, which traverses cell cycle until G2 in response to serum. In particular, it was found that thrombin cleavage induces a factor in serum of all animals tested so far to promote S phase re-entry in A1 myotubes. We have used this S phase re-entry of the A1 cell line to purify the serum activity and developed a 5-step purification protocol that enriches the activity almost 2 000 fold over the starting material, or 40 000 fold over serum. To conveniently produce and test potential candidates for their ability to induce S phase re-entry in A1 myotubes, we also developed an overexpression- and purification system for emerging candidates. Candidates were then tested for this activity with or without prior incubation with thrombin. We identified Bmp proteins as the first pure molecules that were found in fractions across the purification of the activity and that could also induce cell cycle re-entry in a dose-dependent manner when recombinantly added to the A1 myotubes. Furthermore, this response could be blocked in a dose-dependent manner by the known bmp-inhibitor noggin. Finally, we showed that inhibition of Bmp signaling in vivo causes defects in axolotl tail regeneration.

Mastzellen (MZ) reifen zu terminal differenzierten Zellen erst in peripheren Geweben. ATRA ist ein wesentlicher Regulator der Hämatopoese und kann auf positive wie auf negative Weise deren Differenzierung beeinflussen – die Qualität ist dabei häufig abhängig vom Reifegrad. Die Arbeit beschäftigte sich daher mit der Frage, ob die Effekte von ATRA auf MZ ebenfalls abhängig von deren Reifegrad sind. Unreife HMC-1 5C6 Zellen, differenziertere LAD 2 Zellen und reife kutane MZ (KMZ) wurden mit ATRA behandelt und die Effekte auf spezifische Mastzellmarker auf Protein- und mRNA-Ebene untersucht. Die Proteinexpression von c-kit, FceRI, Tryptase und Chymase wurde bei allen drei Zellsystemen herunterreguliert. Änderungen der Proteinexpression wurden qualitativ, wenngleich nicht immer quantitativ auf mRNA-Ebene reflektiert, d.h. es gab Marker, bei denen die Herunterregulation auf der einen oder anderen Ebene überwog. Eine genaue Quantifizierung der ATRA-Effekte auf die Tryptase und Chymase zeigte eine prozentual erheblich stärkere Herunterregulation der mRNA als des Proteins. Invers dazu war der Effekt bei c-kit, ein Effekt, der besonders deutlich bei HMC-1 5C6 auszumachen war. Zur Klärung, welcher Mechanismus der von der mRNA-Ebene unabhängigen Herunterregulation des c-kit-Proteins zugrunde liegt, wurden HMC-1 5C6 zusätzlich zu ATRA mit Inhibitoren fundamentaler Zellfunktionen inkubiert. Cycloheximid allein war dabei in der Lage, die Wirkung von ATRA zu imitieren. Es kann also vermutet werden, dass die starke Herunterregulation des Proteins im Vergleich zur mRNA einem translationellen Mechanismus folgt. Betrachtet man alle Effekte gemeinsam, so zeigte ATRA den stärksten Einfluss auf das am weitesten differenzierte Mastzellsystem, nämlich KMZ. Geringer waren die Effekte bei LAD 2, wohingegen bei HMC-1 5C6 das schwächste Ansprechpotential gegenüber ATRA gefunden wurde. Dies ist von besonderem Interesse, weil ATRA normalerweise wesentlich potenter auf unreif-proliferierende Systeme wirkt. ATRA hat auf humane MZ einen deutlich dedifferenzierenden Effekt, der weitgehend unabhängig vom Reifegrad der Zellen operiert. Die Effekte scheinen dabei in Abhängigkeit vom betrachteten Marker nicht auf die transkriptionelle Ebene beschränkt zu sein, sondern könnten auch einem translationellen Mechanismus unterliegen. / Mast cells (MC) are of hematopoietic origin but complete their differentiation exclusively within tissues. A large number of mediators positively or negatively affect the maturation process of MC. ATRA is a potential master regulator of haematopoiesis, where it primarily affects immature and proliferative leukocytes. Here, the effects of ATRA (3-7d) on MC that span different stages of maturation, i. e. immature-leukemic HMC-1 5C6 cells, intermediately matured LAD 2 cells, and terminally differentiated skin MC were analyzed. The expression of the lineage markers c-kit, FceRI, tryptase, chymase und histidindecarboxylase (HDC) was studied in parallel at protein level by flow-cytometric analysis and at mRNA level by RT-PCR. ATRA exposure led to a down-regulation at protein and at mRNA level of c-kit, FceRI, tryptase and chymase by all MC subtypes. Comparing protein and transcript levels, however, substantial differences between c-kit and the proteases were noted. c-kit down-regulation was more pronounced at protein level, whereas the opposite was found with proteases. Further analysis revealed the existence of a second mechanism of c-kit down modulation that proceeded rapidly and independently of mRNA changes. This pathway was restricted to immature MC and could be mimicked by cycloheximide, suggesting that altered translation may account for the phenomenon. Taken together, the study indicates that MC are significant target cells of ATRA throughout their lifespan, but that the molecular events underlying down-regulation of lineage markers may be shifted in the course of MC differentiation. The strongest effects of ATRA were observed in mature MC, while this accounts to a lesser degree for LAD 2 cells. In contrast, the immature and highly proliferating HMC-1 5C6 cells presented even less sensitive to ATRA. Therefore, ATRA has dedifferentiating potential towards human MC that is most pronounced in mature MC. Depending on the specific marker, protein down regulation can underlie various mechanisms. In this regard, c-kit seems to follow a translational rather than transcriptional inhibitory process.

Experimentelle und neurophysiologische Studien weisen auf eine Spezifität der Gesichterkognition hin. In der differentiellen Psychologie wird ein Schwerpunkt auf die Differenzierbarkeit sozio-kognitiver Leistungen von akademischen Fähigkeiten gelegt. Dabei werden bislang kaum Versuche unternommen, Messmodelle zu etablieren, die in neurokognitiven Modellen verankert sind. Basierend auf neuartigen Versuchen zur Etablierung solcher Modelle ist es das Ziel dieser Dissertation, die Robustheit dieser Modelle aus einer entwicklungspsychologischen Perspektive zu betrachten und diese zu erweitern. Zudem werden altersbedingte Leistungsunterschiede in der Gesichterkognition auf der Ebene latenter Faktoren ermittelt und die Hypothese altersbedingter kognitiver Dedifferenzierung mit modernen Methoden kritisch untersucht. Das Hauptziel ist die Erbringung entwicklungspsychologischer Evidenz für die Spezifität der Gesichterkognition. In einem ersten - primär methodologischen - Manuskript wird erstmalig in der Literatur die Implementierung von Funktionen der Beobachtungsgewichtung aus der nicht-parametrischen Regression für Strukturgleichungsanalysen vorgeschlagen. Diese Methode ergänzt Multigruppenanalysen bei der Untersuchung kognitiver Dedifferenzierung. Weitere vier Manuskripte adressieren Fragestellungen zur Gesichterkognition und zeigen: 1) Gesichterwahrnehmung, Gesichtergedächtnis und die Schnelligkeit der Gesichtererkennung sind separierbare Prozesse über die gesamte erwachsene Lebensspanne; 2) die Schnelligkeit der Gesichtererkennung kann nicht von der Schnelligkeit der Emotions- und Objekterkennung faktoriell getrennt werden; 3) Gesichterwahrnehmung und Gesichtergedächtnis können bis zum späten Alter von allgemeinen kognitiven Fähigkeiten getrennt werden, und 4) eine leichte Dedifferenzierung zwischen Objekt- und Gesichterkognition tritt auf der Ebene von Akkuratheitsmessungen auf. Implikationen sind in den Manuskripten ausführlich diskutiert und im Epilog zusammengefasst. / Cognitive-experimental and neuropsychological studies provided strong evidence for the specificity of face cognition. In individual differences research, face tasks are used within a broader variety of tasks, usually with the intention to measure some social skills. Contemporary individual differences research still focuses on the distinction between social-emotional vs. academic intelligence, rather than establishing measurement models with a solid basis in experimental and neuropsychological work. Building upon recent efforts to establish such measurement models this dissertation aimed to extend available models and assess their robustness across age. Furthermore, it investigates mean age differences for latent factors, critically looks at phenomena of dedifferentiation with novel and innovative analytic methods, and attempts to provide more evidence on the uniqueness and communalities of face cognition throughout adulthood. In a first primarily methodological manuscript, we propose for the first time in the literature an implementation of functions to weight observations used in nonparametric regression approaches into structural equation modeling context, which can fruitfully complement traditionally used multiple-group approaches to investigate factorial dedifferentiation. In the following four manuscripts, we investigated individual and age-differences in face cognition. Results show that: 1). Face perception, face memory and the speed of face cognition remain differentiable throughout adulthood; 2). The speed of face cognition is not differentiable from the speed of perceiving emotional expressions in the face and complex objects, like houses; 3). Face perception and memory are clearly differentiable from abstract cognition throughout adulthood; and 4). A slight dedifferentiation occurs between face and object cognition. Implications are discussed in the manuscripts and the epilogue.

Urodele amphibians have the remarkable ability to re-grow lost body parts. This regenerative response after injury in urodeles involves dedifferentiation of fully differentiated cells into proliferative cells. One well-studied example of this is the dedifferentiation of multinucleated muscle cells into mononucleate cells resembling their precursors, the myoblasts. To form these mononucleate cells the differentiated myotubes in vivo must re-enter and complete the cell cycle; they again proliferate and produce progeny. A key question is what factors induce the myotubes to re-enter the cell cycle and proliferate. Early events of cell cycle re-entry can be studied in the A1 cell line, a myogenic cell line isolated from the Notophthalmus viridescens hindlimb, which traverses cell cycle until G2 in response to serum. In particular, it was found that thrombin cleavage induces a factor in serum of all animals tested so far to promote S phase re-entry in A1 myotubes. We have used this S phase re-entry of the A1 cell line to purify the serum activity and developed a 5-step purification protocol that enriches the activity almost 2 000 fold over the starting material, or 40 000 fold over serum. To conveniently produce and test potential candidates for their ability to induce S phase re-entry in A1 myotubes, we also developed an overexpression- and purification system for emerging candidates. Candidates were then tested for this activity with or without prior incubation with thrombin. We identified Bmp proteins as the first pure molecules that were found in fractions across the purification of the activity and that could also induce cell cycle re-entry in a dose-dependent manner when recombinantly added to the A1 myotubes. Furthermore, this response could be blocked in a dose-dependent manner by the known bmp-inhibitor noggin. Finally, we showed that inhibition of Bmp signaling in vivo causes defects in axolotl tail regeneration.