The regenerative potential of adult stem cell populations within the human body bears great promises for their use in regenerative medicine. The bone marrow (BM) harbours two different types of adult stem cells, haematopoietic stem and progneitor cells (HSPCs) and multipotent mesenchymal stromal cells (MSCs), which are tightly regulated in their distinct anatomically defined niches by multiple cues such as cytokines, cell-cell contacts, the extracellular matrix (ECM) and the physical microenvironment. The ex vivo expansion of these cells for applications in regenerative therapies is of great interest and several biomaterial approaches attempt to mimic the natural BM niche and its components to control stem cell maintenance and differentiation. However, as of now the complexity of such stem cell niches is hard to recapitulate. Towards this goal, this work was focussing on the ECM environment of BM stem cells and was set out to engineer improved in vitro culture systems. MSC themselves are one of the most important cell types within the BM that secrete and construct ECM-networks and thereby shape the microenvironment of the residing cells. The potential of primary human BM-MSC to secrete ECM in vitro has been exploited to generate niche-like ECM surrogates in a robust and versatile format. Application of decellularisation regimes allowed the fabrication of complex matrices which demonstrated suprastructural, compositional and physicochemical properties compareable to those of the native BM-ECM environment. Reliable stability and reproduciblity was achieved by a dedicated procedure of maleic anhydride co-polymer-mediated covalent binding of fibronectin and subsequent anchorage of cell-secreted ECM molecules. As a result of the high reproducibility, a complete proteomic register of ECM molecules was obtained in combination with determining the complex fibrillar and soft gel-like characteristics of MSC-derived matrices. Based on the established BM niche-like substrate, the impact of extracellular matrices on MSC and HSPC ex vivo behavior has been explored. Both cell types demonstrated strong adhesion to ECM substrates and depicted a changed cellular morphology upon contact with native ECM structures compared to standard culture substrates or simple ECM protein coatings, indicating an intense interplay between the cell and the microenvironment. MSC that re-grew into their own matrices have shown advantageous proliferation and cytokine secretion levels as well as enhanced differentiation intensity (upon differentiation induction) compared to MSC that were cultured on less complex substrates. Similarly, HSPC were also instructed for enhanced expansion on MSC-derived matrices without exhaustion of stem cell-marker expressing progenitor cells. The efficiency of these matrices was related to their ability to mimic the native composite suprastructure, ligand nano-topography, molecular composition and physical properties of natural BM ECM environments. The data obtained within this thesis set the ground for a more rational design of artificial stem cell niches with defined and distinct properties, offering exciting options for the in-depth analysis and understanding of stem cell regulation by exogenous cues.

Neural stem (progenitor) cells (NPCs) from fetal tissue are an ideal transplantable cell source. They divide rapidly, are able to generate cells of all three neural lineages and do not divide uncontrolled once transplanted into a host organism. To obtain large quantities of cells for transplantation strategies and to eliminate primary cell contaminations, long periods of in vitro cultivation are necessary. Mouse NPCs are a crucial tool for further investigations of neural stem cells because they make the employment of transgenic animals in vivo and cells in vitro possible. So far only short-term expanded fetal mouse NPCs have been shown to generate dopaminergic neurons and it is not clear whether this was due to differentiation or a result of increased survival of primary dopaminergic neurons. The aims of the thesis were to characterize mouse fetal NPCs, to establish the long-term expansion of fetal mouse NPCs and the generation of dopaminergic neurons in long-term expanded fetal mouse NPCs, to investigate the signaling mechanisms involved in the differentiation of mouse fetal NPCs towards the dopaminergic phenotype and to compare short and long-term expanded NPCs. Long-term expanded fetal mesencephalic NPCs could be grown under suspension and adherent culture conditions and showed self- renewing capacity as well as markers typical for NPCs. They could be differentiated into the three major cell types of the nervous system, but suspension NPCs had a larger potential to generate neurons than adherently grown NPCs. Signaling cascades involved in this process were p38 and Erk1/2 mediated. Long-term expanded NPCs did not have the potential to generate neuronal sub-types. Importantly, they did not generate dopaminergic neurons. Mouse fetal NPCs from three different developmental stages (E10, E12, and E14) were employed but were not able to differentiate into dopaminergic neurons using factors known to stimulate in vitro dopaminergic specification. When cultivated in vitro for short periods, fetal mesencephalic NPCs were able to generate dopaminergic neurons. By eliminating all primary Th- positive neurons, FACS-sorting of NPCs proved a de novo generation of dopaminergic neurons, because after cultivation and differentiation of Th- depleted cell solutions dopaminergic neurons were present in the culture. However, these newly generated neurons failed to incorporate BrdU, making a generation without cell division from precursors probable. The precursor population of short cultures differed from long-term expanded cultures suggesting an ‘aging’ effect of in vitro conditions. IL-1 was a potent inducer of the dopaminergic neuronal phenotype in short-term expanded in vitro cultures and was expressed in vitro as well as in vivo at E14. Several important conclusions concerning fetal mouse stem cell behavior could be drawn from the results of this work: Firstly, the results showed for the first time that in fetal mouse mesencephalic NPCs dopaminergic neurons differentiate from precursors without cell division, therefore consuming those progenitors. Therein fetal mouse NPCs differ significantly from rat and human NPCs or respond differently to the same in vitro conditions that need to be optimized for fetal mouse NPCs. Secondly, less committed precursors find appropriate conditions to proliferate but not to generate the more committed DA precursors that are able to generate dopaminergic neurons. The hallmarks of stem cells, self-renewal and multipotentiality, seem to be part of a delicate balance, that, when unsettled, goes in favor of one side without the possibility of returning to the previous status. Further research should focus on two coherent issues: the isolation of more pure populations of progenitors and the more precise characterization of progenitor populations to find out which in vitro conditions need to be provided to keep the balance between proliferation and differentiation potential. The knowledge gained about stem cells this way would help establish cell sources for transplantation strategies. / Stammzellen sind ein wichtiges Werkzeug für regenerative Therapien im Bereich der neurodegenerativen Erkrankungen wie der Parkinson’schen Erkrankung. Ein besonderer Vorteil von Stammzellen gegenüber dem bereits zur Transplantation verwendeten Primärgewebe, ist ihre Fähigkeit zur fortlaufenden Zellteilung, so dass ausreichende Mengen zur Transplantation zur Verfügung stehen. Der Vorteil von fetalen neuralen Stammzellen (fNSZ) ist ihre genomische Stabilität, die dazu führt, dass bei Transplantationen keine Tumore entstehen. Dennoch ist der Großteil ihrer Eigenschaften und Potentiale noch unbekannt und die optimalen Wachstumsbedingungen für eine lange in vitro Kultur und optimale Differenzierung in dopaminerge Neuronen müssen erforscht werden, um bessere Transplantate herzustellen. Insbesondere Stammzellen der Maus sind für die Forschung von immenser Wichtigkeit, da sie die Arbeit mit transgenen Tieren ermöglichen. Die Zielsetzungen dieser Arbeit waren die Charakterisierung der fNSZ der Maus, die Langzeitexpansion und die anschließende Differenzierung in dopaminerge Neurone. Die Signalkaskaden der frühen Differenzierung und die Unterschiede von kurz- und langzeitkultivierten Stammzellen wurden untersucht. Es konnte gezeigt werden, dass fNSZ der Maus nach Langzeitkultivierung in alle Zelltypen des zentralen Nervensystems, also Neuronen und Glia differenzieren und die dabei aktivierten Signalkaskaden p38 und Erk1/2 vermittelt sind. Das Differenzierungspotential zu neuronalen Subtypen (also auch zu dopaminergen Nervenzellen) verloren diese fetalen Stammzellen unter Kulturbedingungen schnell. Das steht im Gegensatz zu fetalen Stammzellen aus Ratte oder dem Menschen, die auch nach langer Kultivierung ihr dopaminerge Potential erhalten. Nur nach Kurzzeitkultivierung waren dopaminerge Neurone nachzuweisen, die jedoch nicht durch Zellteilung aus Vorläuferzellen hervorgegangen waren. Die Eliminierung aller primären Neurone aus der Mittelhirnisolation durch FACS-sorting von Th-Gfp transgenen Mäusen bewies die de novo Generation der dopaminergen Neurone aus Vorläuferzellen ohne Zellteilung während der Kultivierung der Stammzellen. Diese Ergebnisse zeigten, dass in fetalen mesenzephalen NSZ der Maus dopaminerge Neurone von spezialisierten Vorläuferzellen differenzieren, wodurch diese der Kultur verloren gehen. Weniger spezialisierte Vorläuferzellen finden Bedingungen, die ihre Kultivierung ermöglichen, sind aber nicht in der Lage, spezifischere Vorläuferzellen zu bilden. Die Markenzeichen von Stammzellen, Selbsterneuerung (durch Zellteilung) und das Potential, die Zelltypen des Nervensystems zu generieren, scheinen fein balancierte Zustände zu sein, die bei einer Störung nicht wiederherzustellen sind. Die Ergebnisse dieses Projektes sind von großer Bedeutung für die Forschung zur Zellersatztherapie der Parkinson’schen Erkrankung, deren ultimatives Ziel es ist, eine sichere und verlässlich expandierbare Zellquelle zu etablieren, die fähig ist, in dopaminerge Neurone zu differenzieren. Solche Stammzellen würden Bemühungen um Transplantationsstrategien für neurodegenerative Erkrankungen unterstützen und vorantreiben.

Mesenchymale Stammzellen bieten eine interessante Option in der regenerativen Medizin, da sie praktisch unlimitiert verfügbar sind. Um das Verhalten von humanen MSC zu studieren, werden Untersuchungen momentan an immundefizienten Mäusen durchgeführt, deren Verwendung kostenintensiv und aufwendig ist. Fra-gestellung war, ob durch Immunisolation (Alginat, Dialyseschlauch, Diffusionskammer) die Knorpel erhaltenden -, bzw. bildenden Eigenschaften von MSC-Konstrukten ebenso gut in immunkompetenten Mäusen untersucht werden können. Gleichzeitig sollte geprüft werden, ob die mit einer Immunabschirmung einhergehende Reduktion der Zellversorgung und damit die Annäherung an die Gelenksituation ihre Mineralisierung vermindern kann und ob Mauszellen für eine Veränderung der vordifferenzierten Knorpelpellets verantwortlich sind. Hierzu wurden hBMSC chondrogen differenziert. Die Zellpellets wurden mit Alginat, dem Dialyseschlauch oder der Diffusionskammer verkapselt und parallel zu unver-kapselten Kontrollpellets subkutan in immundefiziente SCID-Mäuse sowie in immunkompetente BDF1-Mäuse implantiert. Die Explantate wurden mit Alzianblau-, Alizarinrot-, Kollagen Typ II-Färbungen, sowie einer ALU in-situ Hybridisierung mar-kiert und mittels Histologiescore doppelt blind bewertet (MannWhitneyU). Überra-schenderweise zeigten die unverkapselten Kontrollen in den BDF1-Mäusen weder Zeichen von Inflammation noch von Destruktion und 4/5 der Pellets waren auf Kol-lagen Typ-II und Alzianblau positiv. Gleichzeitig war der Grad der Mineralisierung in den BDF1-Mäusen gegenüber SCID-Mäusen reduziert (p = 0,03). Durch Alginat wurde die Mineralisierung in den BDF1 Mäusen (0/8) völlig verhindert, während in den SCID-Mäusen noch 7/8 der Pellets Kalzifizierung zeigten (p = 0,001). Die Verkapselung mit Alginat verglichen mit der Kontrolle führte in beiden Mausstämmen zu höheren Scores für Kollagen Typ II (SCID: p = 0,013, BDF1: p = 0,042) und zeigte gleichzeitig eine Reduktion der Mineralisierung (SCID: p = 0,018, BDF1: p = 0,031). In SCID-Mäusen war außerdem der Alzianblau-Wert gegenüber den Kontrollen erhöht (p = 0,003). Die Diffusionskammer erwies sich als ungeeignet, da die Pellets ihre knorpeligen Eigenschaften verloren. Durch die Verwendung des Dialyseschlauches konnte lediglich in der SCID-Maus eine Erhöhung der Kollagen Typ II (p = 0,03) und eine Reduktion der Kalzifizierung (p = 0,004) erreicht werden. Sowohl im Alginatbead in der BDF1-Maus (1/3 Spendern), als auch im Dialyseschlauch mit Kollagenmembran (2/3 Spendern) konnte eine erfolgreiche in vivo Chondrogenese durchgeführt werden. Zur Untersuchung der in vivo Stabilität knorpeliger MSC-basierter Konstrukte stellt die BDF1-Maus eine attraktive, kostengünstige Alternative mit einer gegenüber der SCID-Maus verringerten Mineralisierungsrate dar. Die in vitro gebildete knorpelige Extrazellulärmatrix erzeugt dabei bereits eine Immunisolation, welche die Transplantatdestruktion verhindert. Ob ein intaktes lymphozytäres System die Knorpelstabilität gegenüber defizienten Immunsystemen begünstigt, muss durch die Untersuchung weiterer Ansätze belegt werden. Im Gegensatz zur Diffusionskammer bietet Alginat das richtige Maß an Versorgungsreduktion, um die Stabilisierung des Knorpelphänotyps der Konstrukte zu ermöglichen.

Articular cartilage is a highly specialized, hierarchically organized tissue covering the articular surfaces of diarthrodial joints that absorbs and distributes forces upon mechanical loading and enables low-friction movement between opposing bone ends. Despite a strong resilience towards mechanical stress, once damaged cartilage is generally not regenerated due to a limited repair potential of the residing cells (chondrocytes) and the local absence of vascularized blood vessels and nerves. Eventually, this may lead to osteoarthritis, a chronic degenerative disorder of the synovial joints which has a strongly growing prevalence worldwide. Modern regenerative therapies that aim to rebuild cartilage tissue in vivo and in vitro using chondrocyte- and stem cell-based methods are still not able to produce tissue constructs with desired biomechanical properties and organization for long-term repair. Therefore, cartilage tissue engineering seeks for new ways to solve these problems. In this regard, the application of hydrogel-based scaffolding materials as artificial matrix environments to support the chondrogenesis of embedded cells and the implementation of appropriate biofabrication techniques that help to reconstitute the zonal structure of articular cartilage are considered as promising strategies for sophisticated cartilage regeneration approaches. In this thesis, a modular starPEG-heparin hydrogel platform as cell-instructive hydrogel scaffold was used in combination with a custom-designed 3D inkjet bioprinting method with the intention to develop a printable 3D in vitro culture system that promotes the chondrogenic differentiation of human mesenchymal stromal cells (hMSC) in printed cell-laden hydrogels with layered architectures in order to fabricate cartilage-like tissue constructs with hierarchical organization. Firstly, the successful bioprinting of horizontally and vertically structured, cell-free and -laden hydrogel scaffolds that exhibit layer thicknesses in the range of the superficial zone, the thinnest articular cartilage layer is demonstrated. The long-term integrity of the printed constructs and the cellular functionality of the plotted cells that generally had a high viability after the printing process are shown by a successful PDGF-BB-mediated hMSC migration assay in a printed multilayered hydrogel construct over a culture period of 4 weeks. Secondly, when the established printing procedures were applied for the chondrogenic differentiation of hMSCs, it was found that the printed cell-laden constructs showed a limited potential for in vitro chondrogenesis as indicated by a weaker immunostaining for cartilage-specific markers compared to casted hydrogel controls. In order to increase the post-printing cell density to tackle the limited printable cell concentration which was regarded as the primary reason for the impaired performance of the printed scaffolds, different conditions with varying culture medium and hydrogel compositions were tested to stimulate 3D cell proliferation. However, a significant 3D cell number increase could not be achieved which ultimately resulted in shifting the further focus to casted hMSC-laden starPEG-heparin hydrogels. Thirdly, the chondrogenic differentiation of hMSCs in casted hydrogels proved to be successful which was indicated by a uniform deposition of cartilage-specific ECM molecules comparable with the outcomes of scaffold-free MSC micromass cultures used as reference system. However, the quantitative analysis of biochemical and physical properties of the engineered hydrogel constructs yielded still significant lower values in relation to native articular cartilage tissue. Fourthly, in order to improve these properties and to enhance the chondrogenesis in starPEGheparin hydrogels, a dualistic strategy was followed. In the first part, specific externally supplied stimulatory cues including a triple growth factor supply strategy and macromolecular crowding were applied. As second part, intrinsic properties of the modular hydrogel system such as the crosslinking degree, the enzymatic degradability and the heparin content were systematically and independently altered. It was found that while the external cues showed no supportive benefits for the chondrogenic differentiation, the reduction of the heparin content in the hydrogel proved to be a key trigger that resulted in a significantly increased cartilage-like ECM deposition and gel stiffness of engineered constructs with low and no heparin content. In conclusion, this work yielded important experiences with regards to the application of inkjet bioprinting for hMSC-based cartilage tissue engineering approaches. Furthermore, the obtained data provided valuable insights into the interaction of MSCs and a surrounding hydrogel-based microenvironment that can be used for the further development of chondrosupportive scaffolding materials which may facilitate the fabrication of cartilage-like tissue constructs.

Introduction The adult brain exhibits low regenerative ability. Stem cell-based transplantation approaches have been largely unsuccessful, due to the difficulty to recapitulate the complex cytoarchitecture of the central nervous system (CNS). eNSCs are a new therapeutic option as pharmacological activation and increase of their number in vivo is accompanied by powerful neuroprotection in various disease models. Hes3 is expressed in both proliferating and quiescent NSCs, which makes it a useful biomarker for NSC identification. Direct injections of insulin in the adult brain increase the number of eNSCs and promote rescue of injured neurons via a novel molecular mechanism, the STAT3-Ser/Hes3 Signaling Axis. This molecular pathway with the STAT3-Ser phosphorylation at its core regulates Hes3 and together they form a merging point for several signals including insulin receptor activation. Main aim and Hypothesis Beyond the brain, STAT3-Ser/Hes3 signaling regulates various plastic cell populations in other organs of the endocrine/neuroendocrine system. In the pancreas, Hes3 is expressed in islets cells and regulates their growth, regeneration, and insulin release. Hes3 is also expressed in mouse hypothalamic tanycytes, which are diet responsive cells and play a very crucial role for the communication between the brain and the endocrine system. Also, Hes3 is expressed in the adrenal gland (both in the cortex and medulla); cultured adrenal progenitors express Hes3 and various treatments that induce Hes3 expression promote their growth. Therefore, STAT3-Ser/Hes3 Signaling may be involved in tissue problems that result from metabolic dysfunction. Metabolic syndrome often results in diabetes (Type I, Type II) and insulin resistance, suggesting that eNSCs may be affected by the condition. There is evidence that obesity induces inflammatory reactions in the hypothalamus, leading to NSC loss. However, it is not clear if damage to NSCs is also directly linked to insulin signaling disruption. Results Our results show that various parameters affect Hes3 levels in the brain. Aging decreased Hes3 mRNA expression. Type I diabetes increased Hes3 expression. Type II diabetes decreased Hes3 expression. Thus, we conclude that eNSCs are modulated by diabetes in an age-dependent manner. We also investigated whether common medication for metabolic related dysfunction also affects Hes3 expression in the adult brain. Indeed, our results show that metformin decreases Hes3 expression in the mouse hypothalamus. To address whether metformin has a direct effect on NSCs we treated primary mouse fNSCs with metformin. Metformin decreases cell number, proliferation and affects cell morphology, giving a more differentiated appearance (large, flat cell body with wider projections). Hes3 expression increases significantly at 72 hours of treatment. The metformin result opens the question if the increase in the Hes3 expression is a direct effect of the signal transduction pathways activated by metformin or due to a stress reaction. To address this we treated NSCs with exendin-4, another diabetes drug that we previously showed to both elevate Hes3 expression and cell number using a mouse insulinoma cell line (MIN6). Exendin-4 increases fNSC cell number but it did not affect the morphology. Similar to metformin proliferation was not affected. Hes3 expression increased significantly at 72 hours of treatment as well. This result indicates the distinctive action of the drugs on the STAT3-Ser/Hes3 signaling pathway. Specifically it dissociates Hes3 levels from other cellular parameters. Importantly it shows that two common diabetes medications have very different effects on NSCs. Because Hes3 is strongly regulated by metabolic parameters and medication we addressed potential roles of Hes3 using an established Hes3 null mouse line. Hes3 null mice exhibit no obvious phenotypes under normal conditions. However, we previously showed that when stressed by chemical induced damage, they exhibit low regenerative potential in the pancreas and brain. To identify additional phenotypes, we performed a phenotypic analysis of the Hes3 null mouse line under normal diet and HFD conditions (which induced type II diabetes). We found mild phenotypes that relate to the nervous system, the immune system and metabolism. At the molecular level, Hes3 deletion affects the expression of other genes within the Hes superfamily in the adult mouse brain. However, we did not observe these molecular differences in the HFD condition, suggesting an interplay between metabolic parameters (possibly, circulating insulin) and the regulation of Hes/Hey genes in the brain. Our data suggest a broad range of roles for Hes3, particularly under abnormal conditions. Conclusions Our work establishes that multiple parameters of metabolic state as well as diabetes medication affect Hes3 expression in the brain. Metabolic syndrome is a risk factor for many neurological disorders such as Alzheimer’s disease, Parkinson’s disease and Multiple Sclerosis. It is important to understand at the molecular and cellular level how metabolic dysfunction affects the brain. Here, we introduced a new cellular biomarker and signaling component that is greatly regulated in metabolic dysfunction.:1 Introduction 18 1.1 The ''plastic brain'': Neural Stem Cells, progenitors and precursors 19 1.2 Functional adult neurogenesis 19 1.3 NSCs in conventional and nonconventional regions of the adult brain 20 1.4 Neurodegenerative diseases, cell replacement and endogenous NSCs 21 1.5 The STAT3-Ser/Hes3 signaling axis in NSCs 24 1.6 Beyond the brain: The STAT3-Ser/Hes3 signaling axis operates in plastic cells 27 1.6.1 STAT3-Ser/Hes3 Signaling Axis in the pancreatic islet 27 1.6.2 STAT3-Ser/Hes3 Signaling Axis in the adrenal cortex and medulla 28 1.6.3 STAT3-Ser/Hes3 Signaling Axis in tanycytes of the hypothalamus? 28 1.6.4 STAT3-Ser/Hes3 Signaling: A new molecular component of the neuroendocrine system? 29 1.7 Metabolic syndrome and neurological disease 31 1.7.1 Metabolic dysfunction and Alzheimer's disease 31 1.7.2 Metabolic dysfunction and Parkinson's disease 31 1.7.3 Metabolic dysfunction and Multiple Sclerosis 32 1.7.4 Metabolism and neurodegenerative disease: Are they connected? 32 1.8 Main Aim – Hypothesis 33 2 Materials and Methods 34 2.1 Animal experiments 34 2.1.1 Animal use authorization 34 2.1.2 Genotyping 34 2.1.3 In vivo models 36 2.1.4 In vivo metabolic Analyses 36 2.1.5 Nociception 37 2.1.6 Histology 38 2.1.7 PCR and Real-Time quantitative PCR (qPCR) 39 2.1.8 Western Blot 41 2.2 Mouse phenotyping 42 2.3 Neural stem cell cultures 43 2.3.1 Preparation – Coatings 43 2.3.2 Cell Isolation and Cell Culture 43 2.3.3 Pharmacological Manipulation (Metformin – Exendin-4) 43 2.4 Heat maps 44 2.5 Statistical analyses 44 3 Results 45 3.1 Hes3 is expressed in the mouse brain 46 3.2 Aging and diabetes models alter Hes3 in the brain 48 3.2.1 Hes3 expression decreases with age 48 3.2.2 Pancreatic islet damage by streptozotocin increases Hes3 expression in the brain 48 3.2.3 High Fat Diet reduces Hes3 expression in the brain 49 3.3 Common diabetes medication affect neural stem cells (NSCs) in the brain 53 3.3.1 Metformin decreases Hes3 expression in the brain 53 3.3.2 Metformin opposes growth but increases Hes3 expression in cultured NSCs 54 3.3.3 Exendin-4 promotes growth and increases Hes3 expression in cultured NSCs 54 3.3.4 Metformin and Exendin-4 affect the STAT3-Ser/Hes3 signaling axis 59 3.4 Hes3 null mice exhibit a quasi-normal phenotype 60 3.4.1 Phenotypic Analysis - Normal Diet (ND) 60 3.4.2 Metabolism Relevant Phenotypes – HFD challenge 63 3.4.3 Phenotypic Analysis – Molecular 67 4 DISCUSSION 70 4.1 Diabetes affects the brain 71 4.2 STAT3-Ser/Hes3: a putative mediator 71 4.3 Hes3 is a special member of the Hes/Hey gene family 72 4.4 Patterns of Hes3 expression may be specific to cell type and microenvironment 72 4.5 Metabolic dysfunction and diabetes medication affect brain Hes3 73 4.5.1 Age regulates Hes3 73 4.5.2 Diabetes models regulate Hes3 expression in the brain 74 4.5.3 Metformin regulates Hes3 expression in the brain 74 4.6 Hes3 phenotyping provides clues to Hes3 functions 76 4.7 Hes3 and metabolic dysfunction: Are they connected? 77 5 Conclusions and Future Remarks 79 References 81

Hämatopoetische Stammzellen besitzen die Fähigkeit, die dauerhafte Erhaltung ihrer eigenen Population im Knochenmark zu gewährleisten und gleichzeitig zur Neubildung der verschiedenen Zelltypen des peripheren Blutes beizutragen. Die Sequenz von Entscheidungsprozessen, die den Übergang einer undifferenzierten Stammzelle in eine funktionale ausgereifte Zelle beschreibt, wird als Linienspezifikation bezeichnet. Obwohl viele Details zu den molekularen Mechanismen dieser Entscheidungsprozesse mittlerweile erforscht sind, bestehen noch immer große Unklarheiten, wie die komplexen phänotypischen Veränderungen hervorgerufen und reguliert werden. Im Rahmen dieser Dissertation wird ein geeignetes mathematisches Modell der Linienspezifikation hämatopoetischer Stammzellen entwickelt, welches dann in ein bestehendes Modell der hämatopoetischen Stammzellorganisation auf Gewebsebene integriert wird. Zur Verifizierung des theoretischen Modells werden Simulationsergebnisse mit verschiedenen experimentellen Daten verglichen. Der zweite Teil dieser Arbeit konzentriert sich auf die Beschreibung und Analyse der Entwick- lungsprozesse von Einzelzellen, die aus diesem integrierten Modell hervorgehen. Aufbauend auf den entsprechenden Modellsimulationen wird dazu eine topologische Charakterisierung der resultierenden zellulären Genealogien etabliert, welche durch verschiedener Maße für deren Quantifizierung ergänzt wird. Das vorgestellte mathematische Modell stellt eine neuartige Verknüpfung der intrazellulären Linienspezifikation mit der Beschreibung der hämatopoetischen Stammzellorganisation auf Populationsebene her. Dadurch wird das Stammzellm- odell von Röder und Löffler um die wichtige Dimension der Linienspezifikation ergänzt und damit in seinem Anwendungsbereich deutlich ausgedehnt. Durch die Analyse von Einzelzellverläufen trägt das Modell zu einem grundlegenden Verständnis der inhärenten Heterogenität hämatopoetischer Stammzellen bei.

The discovery of resetting human somatic cells via introduction of four transcription factors into an embryonic stem cell-like state that enables the generation of any cell type of the human body has revolutionized the field of medical science. The generation of patient-derived iPSCs and the subsequent differentiation into the cells of interest has been, nowadays, widely used as model system for various inherited diseases. The aim of this thesis was to generate iPSCs and to subsequently derive NPCs which can be differentiated into neurons in order to model the two most common forms of the NCLs: LINCL which is caused by mutations within the TPP1 gene, encoding a lysosomal enzyme, and JNCL which is caused by mutations within the CLN3 gene, affecting a lysosomal transmembrane protein. It was shown that patient-derived fibroblasts can be successfully reprogrammed into iPSCs by using retroviral vectors that introduced the four transcription factors POU5F1, SOX2, KLF4 and MYC. The generated iPSCs were subsequently differentiated into expandable NPCs and finally into mature neurons. Phenotype analysis during the different stages, namely pluripotent iPSCs, multipotent NPCs and finally differentiated neurons, revealed a genotype-specific progression of the disease. The earliest events were observed in organelle disruption such as mitochondria, Golgi and ER which preceded the accumulation of subunit c of the mitochondrial ATPase complex that was only apparent in neurons. However, none of these events led to neurodegeneration in vitro. The established disease models recapitulate phenotypes reported in other NCL disease models such as mouse, dog and sheep model systems. More importantly, the hallmark of the NCLs, accumulation of subunit c in neurons, could be reproduced during the course of disease modeling which demonstrates the suitability of the established system. Moreover, the derived expandable NPC populations can be used for further applications in drug screenings. Their robust phenotypes such as low levels of TPP1 activity in LINCL patient-derived NPCs or cytoplasmic vacuoles, containing storage material, observed in CLN3 mutant NPCs, should serve as possible phenotypic read-outs.

All multicellular organisms constantly need to replace aged or damaged cells. This vital task of tissue homeostasis is fulfilled by stem cells. The balance between self-renewal and differentiation of the stem cell is crucial for this task and tightly regulated by a signaling microenvironment termed the niche. A widely used model for studying stem cell niche biology is the Drosophila testis, where two stem cell populations, the germline stem cells (GSCs) and the somatic cyst stem cells (CySCs), reside in a niche located at the apical tip. A lot is known about the signals regulating GSC maintenance in the testicular niche. It is, however, unknown how the spatial regulation of these signals defines the range of the niche. Here I show, that Bone Morphogenetic Protein (BMP) signaling is specifically activated at the interface of niche and stem cells. This local activation is achieved by coupling the transport of adhesion and signaling molecules in the niche cells and directing their transport to contact sites of niche and stem cells. Localized niche signaling at junctions underlies the so called stem-cell-niche synapse hypothesis proposed for the mammalian hematopoietic stem cell niche. I have shown that disrupting the localized transport causes premature differentiation and stem cell loss. BMP signaling between niche and GSCs therefore provides the first description of a stem-cell-niche synapse and will yield valuable insights into mammalian stem cell biology. The CySCs reside in the niche of the testis together with the GSCs. To understand how the niche maintains both stem cell types in a concerted way, it is essential to know the pathways regulating both stem cell types. Here I show that Hedgehog (Hh) signaling is a key stem cell factor of CySCs, while only indirectly affecting GSCs. Loss of Hh signaling in CySCs results in premature differentiation and consequent loss of the cells. Overactivation of the pathway leads to an increased proliferation and an expansion of the cyst stem cell compartment. As Hh signaling is also a regulator of the somatic cells in the mammalian testis and the Drosophila ovary this may reflect a higher degree of homology between these systems than previously expected.

Photoreceptors are sensory neurons in our eyes’ retinas that convert light into electrochemical signals thus allowing us to see the world around us. Human retinas have two types of photoreceptors: rods important for night vision and cones important for high-acuity daylight vision. In some retinal diseases, photoreceptors degenerate leaving patients visually impaired or even blind. One of the promising therapeutic approaches is cell therapy, which acts by supporting surviving or by substituting lost photoreceptors by transplanting donor photoreceptors produced in vitro. Human induced pluripotent stem cells (hiPSCs) represent a favorable donor cell source for transplantation, as they are patient-specific and have the ability to self-renew. While photoreceptors isolated from human retinal organoids represent bona fide cells for cell therapy, long time needed for their production, which coincides with developmental time, hampers their clinical application. Another approach for hiPSC differentiation is by overexpressing different transcription factors (TFs), the so-called forward programming. Although proved fast and efficient in producing multiple neuronal cell types, efficient forward programming protocols for engineering photoreceptors were so far not established. Therefore, aim of my thesis was to find TFs that drive in vitro photoreceptor differentiation from hiPSCs and to establish a fast forward programming protocol for producing photoreceptors in high yields. To find TFs that drive photoreceptor differentiation, I performed a TF-library-on-library screening in a reporter hiPSC line. The reporter hiPSCs expressed fluorescent markers only if synthetic photoreceptor-specific promoters were activated, i.e. in case of photoreceptor differentiation. The specificity of the reporter was confirmed in human retinal organoids. For the screening, I transduced the reporter cell line with two lentiviral libraries: a biased one consisting of 16 TFs known from in vivo photoreceptor development and an unbiased one consisting of 1756 TFs. After overexpressing TFs, cells that activated photoreceptor promoters were fluorescently sorted and analyzed. As 80 % of the sorted cells were positive for photoreceptor-specific genes, the screening was highly specific. Furthermore, the screening identified TFs that I validated in the reporter cell line as single, double and triple combinations to find the most efficient one in driving photoreceptor differentiation. The double combination of OTX2 and NEUROD1, known players in photoreceptor development, activated the cone reporter in 10 % of the cells, while additional overexpression of GON4L increased the activation to 25 %. GON4L was never before associated with photoreceptor development showing that in vitro differentiation might be uncoupled from its in vivo counterpart. The cone differentiation efficiency was increased to 50% by treating the cells with AraC, a cell cycle inhibitor that removes all proliferating cells from the cultures. Whether the cell will activate the cone reporter depended on expression levels of individual TFs. Higher and unequal levels, with NEUROD1 having the highest expression, were favorable for obtaining cells with activated cone reporter. Thus, by producing monoclonal cell lines, I identified competent clones with differentiation efficiencies higher than that of a polyclonal cell line and going up to 58%. Except activating the cone reporter, the cone precursor-like cells differentiated from hiPSCs by overexpressing the three TFs OTX2, NEUROD1 and GON4L acquired neuronal morphology and expressed photoreceptor precursor markers. As precursors are the optimal developmental time point for transplantation studies, the engineered cells were transplanted into mouse model of cone degeneration to assess their possible therapeutic potential. Some of the transplanted cells survived in vivo in the subretinal space but did not show any maturation or integration into the remaining retinal circuitry. Thus, further maturation of the cells in vitro is needed before the transplantation. So far, cone precursor-like cells showed ability to mature in vitro when co-cultured with retinal pigment epithelial cells derived from hPSCs and when cultured in presence of additional previously published growth factors. Therefore, changing culturing conditions from stem cell to photoreceptor-specific showed beneficial for further in vitro maturation and paved the way for further research. In conclusion, this study advanced TF-mediated cone photoreceptor engineering. It showed that overexpressing the three TFs OTX2, NEUROD1 and GON4L is enough to push differentiation of more than 50% of hiPSCs into cone precursor-like cells in only 10 days. Additional research to improve maturity and homogeneity of engineered cells – overexpressing additional TFs and changing culturing conditions is ongoing. Fast and efficient protocol established in this study is beneficial for bringing in vitro differentiated cone photoreceptors closer to their commercial application. Such engineered cones could be used as biomedical testbeds for drug discovery and research and represent a promising donor material for cell transplantation to treat blindness.

Hematopoietic stem cells (HSCs) are rare cells that continuously regenerate the entire hematopoietic system by producing billions of blood cells. Dormant HSCs (dHSCs) represent a distinct subpopulation of HSCs characterized by deep quiescence and very low overall biosynthetic activity. Despite this, dHSCs have the highest reconstitution and self-renewal potential and reside at the apex of the hematopoietic hierarchy. While dHSCs do not significantly contribute to daily blood cell production under steady-state conditions, they can be reversibly activated in response to inflammatory signals or blood loss. Thus, dHSCs serve as a reserve pool of HSCs during the entire life. Insufficient dormancy can subject dHSCs to replication stress and promote the accumulation of somatic mutations, increasing the risk of their exhaustion or malignant transformation. Conversely, excessive dormancy can limit normal blood cell production, potentially resulting in bone marrow failure. Therefore, further investigations exploring the mechanisms controlling HSC dormancy are required, as this knowledge is essential for developing novel therapeutic interventions for supporting blood production following chemotherapy or HSC transplantation. Progress in dHSC research has been impeded by technical difficulties associated with isolating these cells. Current methods include either label retention assay, which is very time-consuming, or the use specialized reporter mouse strains that are not readily available. Herein, we utilized single-cell RNA sequencing of HSCs to identify potential surface markers which would facilitate the direct isolation of dHSCs using fluorescence-activated cell sorting (FACS). We selected CD38 as a candidate gene and confirmed that its high expression levels in LT-HSCs, the most functional HSCs identified by the latest surface phenotype, correspond to the dormant subpopulation. Namely, we employed four techniques (staining for cell cycle, label incorporation assay, label retention assay, and single-cell division tracking assay) and demonstrated that CD38+ HSCs are the most quiescent among LT-HSCs. Additionally, through serial competitive transplantation into lethally irradiated mice, we compared CD38+ and CD38- LT-HSCs and discovered that CD38+ LT-HSCs have superior repopulation and self-renewal capacities compared to CD38- LT-HSCs. Thus, we concluded that CD38 is a marker for dHSCs in mice. Besides, we applied the models of hematopoietic stress – acute thrombocytopenia, injection of viral mimetic polyinosinic:polycytidylic acid, and the chemotherapeutic agent 5-fluorouracil, and showed that high expression levels of CD38 on LT-HSCs define dHSCs both in homeostasis and under stress conditions. Notably, we showed that CD38 is not only a marker but also has a functional role in mediating HSC dormancy. Using CD38 knock-out mice, small molecule inhibitor for CD38 enzymatic activity, in vitro assays, bulk RNA sequencing, and confocal microscopy, we discovered a previously unknown signaling axis that promotes HSC dormancy via CD38 enzymatic activity. We demonstrated that second messenger cADPR, synthesized by CD38 through the conversion of nicotinamide adenine dinucleotide - NAD+, elevates free cytoplasmic Ca2+ in dHSCs. This elevation induces the expression of the transcription factor c-Fos. Subsequently, c-Fos forms complexes with the Smad2/3, ultimately promoting dHSC quiescence in p57Kip2-dependent manner. Thus, we revealed that CD38/cADPR/Ca2+/c-Fos-Smad2/3/p57kip2 axis supports dHSCs. Human HSCs (hHSC) are defined as CD38lo/- ; however, CD38 is expressed by various immune cell types present in human bone marrow, such as B-lymphocytes, T-lymphocytes, NK-cells, neutrophils and monocytes. Our co-culture experiments of hHSCs with CD38+ cells and human bone marrow imaging suggest that CD38 promotes hHSC quiescence, however indirectly, in a paracrine manner. Besides, several hematological malignancies (e.g. multiple myeloma, chronic lymphocytic leukemia, acute myeloid leukemia) express CD38 at a high level. We hypothesize that tumor microenvironment enriched in the products of CD38 ecto-enzymatic activity may suppress the cell cycle of healthy hHSCs leading to cancer-related pancytopenia. Therefore, inhibiting CD38-mediated cADPR production might support healthy hematopoiesis in patients with hematologic malignancies. In summary, while CD38/Ca2+ and c-Fos have individually been implicated in proliferation in other cell types, our study for the first time reveals their role in promoting HSC dormancy in collaboration with well-known mediators of HSC quiescence Smad2/3 and p57Kip2. Therefore, we gathered the pieces of the puzzle together and discovered the novel CD38 enzymatic activity-driven signaling pathway controlling HSC dormancy. Manipulation of this axis can potentially stimulate an efficient dHSC response to hematopoietic stress. / Hämatopoetische Stammzellen (HSZ) sind seltene Zellen, die das gesamte blutbildende System kontinuierlich regenerieren, indem sie Milliarden von Blutzellen produzieren. Ruhende HSZ (rHSZ) stellen eine besondere Subpopulation von HSZ dar, die sich durch tiefe Ruhephasen und eine sehr geringe Biosyntheseaktivität auszeichnet. Trotzdem haben rHSZ das höchste Rekonstitutions- und Selbsterneuerungspotenzial und stehen an der Spitze der hämatopoetischen Hierarchie. Während rHSZs unter Normalbedingungen nicht wesentlich zur täglichen Blutzellproduktion beitragen, können sie als Reaktion auf Entzündungssignale oder Blutverlust reversibel aktiviert werden. Somit dienen rHSZ während des gesamten Lebens als HSZ-Reservoir. Eine gestörte Ruhe der rHSZs kann die Zellen einem Replikationsstress aussetzen und die Anhäufung somatischer Mutationen fördern, was das Risiko für Zellerschöpfung oder maligne Transformation erhöht. Umgekehrt kann eine übermäßige Ruhephase die normale Blutzellproduktion einschränken und möglicherweise zu Knochenmarksversagen führen. Daher ist eine weitere Erforschung der Mechanismen erforderlich, die die HSZ-Ruhephase steuern, da dieses Wissen für die Entwicklung neuartiger therapeutischer Maßnahmen zur Unterstützung der Blutproduktion nach einer Chemotherapie oder HSZ-Transplantation unerlässlich ist. Der Fortschritt in der rHSZ-Forschung wurde durch technische Schwierigkeiten bei der Isolierung dieser Zellen behindert. Zu den derzeitigen Methoden gehören entweder der sehr zeitaufwändige Label-Retentionstest oder die Verwendung spezieller Reportermausstämme, die nicht ohne Weiteres verfügbar sind. In dieser Arbeit haben wir die Einzelzell-RNA-Sequenzierung von HSZs genutzt, um potenzielle Oberflächenmarker zu identifizieren, die die direkte Isolierung von rHSZs mittels fluoreszenzaktivierter Zellsortierung (FACS) erleichtern würden. Wir wählten CD38 als Kandidatengen aus und überprüften, dass dessen hohe Expression auf den funktionellsten HSZ (LT-HSZ), welche mit dem neuesten Oberflächenphänotyp isoliert wurden, die ruhenden Subpopulation identifizieren kann. Wir haben vier Techniken angewandt (Färbung für den Zellzyklus, Label-Inkorporationstest, Label-Retentionstest und Einzelzellteilungstest) und gezeigt, dass CD38+ HSZ die am tiefsten ruhenden LT-HSZs sind. Darüber hinaus haben wir durch serielle konkurrierende Transplantation in letal bestrahlte Mäuse CD38+ und CD38- LT-HSZs verglichen und festgestellt, dass CD38+ LT-HSZs im Vergleich zu CD38- LT-HSZs eine höhere Wiederbesiedlungs- und Selbsterneuerungskapazität haben. Daraus schlossen wir, dass CD38 ein Marker für rHSZs in Mäusen ist. Außerdem wendeten wir Modelle für hämatopoetischen Stress an - akute Thrombozytopenie, Injektion des viralen Mimetikums Polyinosin:Polycytidylsäure und des Chemotherapeutikums 5-Fluorouracil - und zeigten, dass eine hohe Expressionsrate von CD38 auf LT-HSZs rHSZs sowohl in Homöostase als auch unter Stressbedingungen definiert. Besonders bemerkenswert ist, dass wir zeigen konnten, dass CD38 nicht nur ein Marker ist, sondern auch eine funktionelle Rolle bei der Vermittlung der HSZ-Ruhe spielt. Mithilfe von CD38-Knock-out-Mäusen, kleinen Molekül-Inhibitoren für die CD38-Enzymaktivität, In-vitro-Tests, Massen-RNA-Sequenzierung und konfokaler Mikroskopie entdeckten wir eine bisher unbekannte Signalachse, die die HSZ-Ruhe über die enzymatische Aktivität von CD38 fördert. Wir konnten nachweisen, dass der sekundäre Botenstoff cADPR, der von CD38 durch die Umwandlung von Nikotinamid-Adenin-Dinukleotid (NAD+) synthetisiert wird, das freie zytoplasmatische Ca2+ in rHSZs erhöht. Diese Erhöhung induziert die Expression des Transkriptionsfaktors c-Fos. Anschließend bildet c-Fos Komplexe mit Smad2/3 und fördert schließlich die Ruhe der rHSZ in Abhängigkeit von p57Kip2. Wir konnten also zeigen, dass die CD38/cADPR/Ca2+/c-Fos-Smad2/3/p57kip2-Achse rHSZs unterstützt. Humane HSZ (hHSZ) werden als CD38lo/- definiert; CD38 wird jedoch von verschiedenen Immunzellarten im menschlichen Knochenmark exprimiert, z. B. von B-Lymphozyten, T-Lymphozyten, NK-Zellen, Neutrophilen und Monozyten. Unsere Co-Kulturexperimente von hHSZs mit CD38+-Zellen und die Bildgebung des menschlichen Knochenmarks deuten darauf hin, dass CD38 die Ruhe von hHSZs fördert, wenn auch indirekt auf parakrine Weise. Außerdem exprimieren mehrere hämatologische Malignome (z. B. multiples Myelom, chronische lymphatische Leukämie, akute myeloische Leukämie) CD38 in hohem Maße. Wir stellen die Hypothese auf, dass die Mikroumgebung des Tumors, die mit den Produkten der ektoenzymatischen Aktivität von CD38 angereichert ist, den Zellzyklus gesunder hHSZs unterdrücken kann, was zu krebsbedingter Panzytopenie führt. Daher könnte die Hemmung der CD38-vermittelten cADPR-Produktion die gesunde Hämatopoese bei Patienten mit hämatologischen Malignomen unterstützen. Zusammenfassend lässt sich sagen, dass CD38/Ca2+ und c-Fos zwar bereits einzeln für die Proliferation in anderen Zelltypen verantwortlich gemacht wurden, unsere Studie jedoch zum ersten Mal ihre Rolle bei der Förderung der HSZ-Ruhe in Zusammenarbeit mit den bekannten Mediatoren der HSZ-Ruhe Smad2/3 und p57Kip2 aufzeigt. Wir haben also die Teile des Puzzles zusammengefügt und den neuartigen, von der enzymatischen Aktivität von CD38 gesteuerten Signalweg entdeckt, der die HSZ-Ruhephase kontrolliert. Die Manipulation dieser Achse kann möglicherweise eine effiziente rHSZ-Reaktion auf hämatopoetischen Stress stimulieren.