Biological information processing in living systems like cells, tissues and organs critically depends on the physical interactions of molecular signaling components in time and space. How endocytic transport of signaling molecules contributes to the regulation of developmental signaling in the complex in vivo environment of a developing organism is not well understood. In a previously performed genome-wide screen on endocytosis, several genes have been identified, that selectively regulate transport of signaling molecules to different types of endosomes, without disrupting endocytosis. My PhD thesis work provides the first functional in vivo characterization of one of these candidate genes, the novel, highly conserved Rab5 effector protein P95 (PPP1R21). Cell culture studies suggest that P95 is a novel endocytic protein important to maintain the balance of distinct endosomal sub-populations and potentially regulates the sorting of signaling molecules between them (unpublished work, Zerial lab). The scientific evidence presented in this study demonstrates that zebrafish P95 is essential for early zebrafish embryogenesis. Both, knockdown and overexpression of zebrafish P95 compromise accurate morphogenetic movements and patterning of the zebrafish gastrula, showing that P95 functions during zebrafish gastrulation. P95 is functionally required to maintain signaling activity of signaling pathways that control embryonic patterning, in particular for WNT/β-catenin signaling activity. Knockdown of zebrafish P95 amplifies the recruitment of β-catenin to early endosomes, which correlates with the limitation of β-catenin to translocate to the nucleus and function as transcriptional activator. The obtained results suggest that zebrafish P95 modulates the cytoplasmic pools of β-catenin in vivo, via endosomal transport of β-catenin. In conclusion, the data presented in this thesis work provides evidence that the cytoplasm-to-nucleus shuttling of β-catenin is modulated by endocytic trafficking of β-catenin in vivo. We propose the endocytic modulation of β-catenin cytoplasm-to-nucleus trafficking as potential new mechanism to fine-tune the functional output of WNT/β-catenin signaling during vertebrate gastrulation.

Mutations in the leucine-rich repeat kinase 2 (LRRK2) gene are a major cause of Parkinson’s disease (PD), which is why modelling PD by replicating effects in animal models attracts great interest. However, the exact mechanisms of pathogenesis are still unclear. While a gain-of-function hypothesis generally receives consensus, there is evidence supporting an alternative loss-of-function explanation. Yet, neither overexpression of the human wild-type LRRK2 protein or its pathogenic variants, nor Lrrk2 knockout recapitulates key aspects of human PD in rodent models. Furthermore, there is conflicting evidence from morpholino knockdown studies in zebrafish regarding the extent of zygotic developmental abnormalities. Because reliable null mutants may be useful to infer gene function, and because the zebrafish is a more tractable laboratory vertebrate system than rodents to study disease mechanisms in vivo, clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein-9 nuclease (Cas9) genomic editing was used to delete the ~60-kbp-long zebrafish lrrk2 locus containing the entire open reading frame. Constitutive removal of both the maternal and the zygotic lrrk2 function (mzLrrk2 individuals) causes a pleomorphic phenotype in the larval brain at 5 days post-fertilisation (dpf), including increased cell death, delayed myelination, and reduced and morphologically abnormal microglia/leukocytes. However, the phenotype is transient, spontaneously attenuating or resolving by 10 dpf, and the mutants are viable and fertile as adults. These observations are mirrored by whole-larva transcriptome data, revealing a more than eighteen-fold drop in the number of differentially expressed genes in mzLrrk2 larvae from 5 to 10 dpf. Additionally, analysis of spontaneous swimming activity shows hypokinesia as a predictor of Lrrk2 protein deficiency in larvae, but not in adult fish. Because the catecholaminergic (CA) neurons are the main clinically relevant target of PD in humans, the CA system of larvae and adult fish was analysed on both cellular and metabolic level. Despite an initial developmental delay at 5 dpf, the CA system is structurally intact at 10 dpf and later on in adult fish aged 6 and 11 months. However, monoamine oxidase (Mao)-dependent degradation of biogenic amines, including dopamine, is increased in older fish, possibly suggesting impaired synaptic transmission or a leading cause of cell damage in the long term. Furthermore, decreased mitosis rate in the larval brain was found, in the anterior portion only at 5 dpf, strongly and throughout the whole organ at 10 dpf. Conceivably, lrrk2 may have a more general role in the control of cell proliferation during early development and a more specialised one in the adult stage, possibly conditional, for example upon brain damage. Because the zebrafish can regenerate lost neurons, it represents a unique opportunity to elucidate the endogenous processes that may counteract neurodegeneration in a predisposing genetic background. To this aim, the regenerative potential of the adult telencephalon upon stab injury was tested in mzLrrk2 fish. Indeed, neuronal proliferation was reduced, suggesting that a complete understanding of Lrrk2 biology may not be fully appreciated without recreating challenging scenarios. To summarise, the present results demonstrate that loss of lrrk2 has an early effect on zebrafish brain development that is later often compensated. Nonetheless, perturbed aminergic catabolism, and specifically increased Mao-dependent aminergic degradation, is reported for the first time in a LRRK2 knockout model. Furthermore, a link between Lrrk2 and the control of basal cell proliferation in the brain, which may become critical under challenging circumstances such as brain injury, is proposed. Future directions should aim at exploring which brain cell types are specifically affected by the mzLrrk2 hypoproliferative phenotype and the resulting consequences on a circuitry level, particularly in very old fish (i.e., over 2 years of age).

Introduction: 23 million patients worldwide suffer from heart failure. These patients depend on cardiac research, because cardiac research enables the development of new therapeutic strategies and –targets. In cardiomyocytes, the compartmentalization of cAMP and cGMP depends on many factors. T-tubuli and PDEs are responsible for the division of cells in microdomains in which localized and specific cAMP and cGMP-signaling occurs. The aim of this thesis was to develop a method to answer the open questions that remain about the physiological and pathophysiological significance of cAMP/cGMP compartmentalization. Methods: I used the zebrafish as a model, because the transparency of zebrafish larvae enabled non-invasive fluorescent imaging in cardiomyocytes in the living animal. I cloned the Fluorescence Resonance Energy Transfer (FRET) sensors EPAC1-camps for cAMP and cGi500 for cGMP and injected them into zebrafish fertilized embryos. Then I used the F0 generation for Fluorescence Lifetime Imaging (FLIM) -FRET-measurements of cAMP and cGMP. Ca2+ is an important downstream mediator of cAMP and cGMP, because Ca2+ regulates cardiac contraction. Therefore, I also cloned the Ca2+ sensor GCaMP6 and used the dye Fluo-4 AM to include intracellular Ca2+ in the imaging. Results: The cloned sensors for cAMP, cGMP and Ca2+ were successfully injected into the zebrafish and showed expression in individual cardiomyocytes. I developed a protocol to mount the living zebrafish embryos and to measure intracellular cAMP and cGMP with FLIM-FRET in vivo with high spatial resolution. I characterized the sensors in their functionality by showing that the sensors react to changes in intracellular concentrations of cAMP and cGMP. The results of this study include evidence that zebrafish have mechanisms that lead to cAMP/cGMP compartmentalization in the absence of T-tubuli, and these mechanisms keep compartmentalization constant even under extreme cAMP or cGMP increasing drug treatment. Furthermore, I imaged intracellular Ca2+ by confocal microscopy and developed a protocol to use Fluo-4 AM for Ca2+ imaging. Conclusion: The method used in this thesis should allow the investigation of subcellular cAMP/cGMP compartmentalization and Ca2+ and to subsequently answer open questions in the field, for example whether a change of cAMP compartmentalization leads to the pathological phenotypes of cardiac disease or if a changed compartmentalization of cAMP in cardiac disease influences Ca2+ concentrations and therefore contraction. Additionally, this method can be used to learn more about cAMP, cGMP und Ca2+ during regeneration in the heart, because the zebrafish cardiomyocytes can regenerate. / Einleitung: Weltweit sind mehr als 23 Millionen unter Herzinsuffizienz leidende Patienten auf die kardiologische Grundlagenforschung angewiesen, da diese die Voraussetzung für eine bessere Versorgung durch adaptierte und neue Behandlungswege schafft. In Kardiomyozyten hängt die Kompartimentierung von cAMP und cGMP von vielen Faktoren ab. T-Tubuli und PDEs werden unter anderem für die Aufteilung der Zellen in Mikrodomänen, in denen lokalisierte und spezifische cAMP- und cGMP-Signalgebung stattfinden kann, verantwortlich gemacht. Das Ziel dieser Arbeit war die Etablierung einer Methode, mithilfe derer offene Fragen bezüglich der physiologischen und insbesondere der pathophysiologischen Relevanz der cAMP- und cGMP Kompartimentierung beantwortet werden können. Methode: Als Modell diente der Zebrafisch, da die Transparenz von Zebrafisch Embryonen eine nicht-invasive Bildgebung von Fluoreszenz in Kardiomyozyten im lebenden Tier ermöglicht. Dafür klonierte ich die Förster Resonance Energy Transfer (FRET) -Sensoren EPAC1-camps als cAMP-Sensor und cGi500 als cGMP-Sensor und injizierte diese in befruchtete Zebrafisch Embryonen. Anschließend benutzte ich die F0-Generation für Fluorescence Lifetime Imaging (FLIM) -FRET-Messungen von cAMP und cGMP. Da Ca2+ als wichtiger downstream Mediator von cAMP und cGMP die kardiale Kontraktion reguliert, klonierte ich außerdem den Ca2+-Sensor GCaMP6 und benutzte den Farbstoff Fluo-4 AM, um intrazelluläres Ca2+ darzustellen. Ergebnisse: Die klonierten Sensoren für cAMP, cGMP und Ca2+ konnten erfolgreich in den Zebrafisch injiziert werden und zeigten alle Expression in einzelnen Kardiomyozyten. Ich entwickelte ein Protokoll, dass die Fixierung von lebenden Zebrafisch Embryonen und nachfolgender Bildgebung von cAMP und cGMP mit hoher zellulärer Auflösung mit FLIM-FRET in vivo erlaubte. Ich konnte eine funktionelle Charakterisierung der Sensoren durchführen, indem ich zeigte, dass sie auf Konzentrationsänderungen von intrazellulärem cAMP und cGMP reagieren sowie zeigen, dass Zebrafische trotz fehlender T-Tubuli eine signifikante cAMP- und cGMP Kompartimentierung aufweisen, auch unter extremen Bedingungen nach Gabe von cAMP/cGMP stimulierenden Substanzen in hoher Dosierung. Ich konnte zudem subzelluläres Ca2+ durch konfokale Mikroskopie bildgebend darstellen und entwickelte ein Protokoll, um mit Fluo-4 AM eine schnelle Möglichkeit zu haben, Ca2+ mit in die Messungen einzubeziehen. Ausblick: Die in dieser Arbeit benutzte Methode bietet eine gute Möglichkeit, subzelluläre cAMP- und cGMP-Kompartimentierung und Ca2+ zu untersuchen und damit zum Beispiel die Fragen zu beantworten, ob eine veränderte cAMP/cGMP Kompartimentierung zu Herzkrankheiten wie Hypertrophie führt oder ob eine veränderte cAMP Kompartimentierung den zellulären Ca2+ Haushalt und damit die kardiale Kontraktion beeinflusst. Darüber hinaus kann das von mir etablierte Protokoll dazu genutzt werden, mehr über cAMP, cGMP und Ca2+ während der Regeneration im Herzen zu lernen, da der Zebrafisch über ausgeprägte Regenerationsfähigkeiten verfügt.

Wnt11 is a key signalling molecule that regulates cell polarity/migration during vertebrate development and also promotes the invasive behaviour of adult cancer cells. It is therefore essential to understand the mechanisms by which Wnt11 signalling regulates cell behaviour. The process of vertebrate gastrulation provides an excellent developmental system to study Wnt11 function in vivo. It is known that Wnt11 mediates coordinated cell migration during gastrulation via the non-canonical Wnt pathway that shares several components with a the planar cell polarity pathway (PCP) in Drosophila. However, the mechanisms by which these PCP components facilitate Wnt11 function in vertebrates is still unclear. While in Drosophila, the asymmetric localization of PCP components is crucial for the establishment of cell polarity, no asymmetric localization of Wnt11 pathway components have so far been observed in vertebrates. To shed light on the cellular and molecular mechanisms underlying Wnt11 signalling, I developed an assay to visualize Wnt11 activity in vivo using live imaging of Wnt11 pathway components tagged to fluorescent proteins. This allowed me to determine the sub-cellular distribution of these components and to correlate the effect of Wnt11 activity with the behaviour of living embryonic cells. I found that Wnt11 locally accumulates together with its receptor Frizzled7 (Fz7) at sites of cell-cell contacts and locally recruits the intra-cellular signalling mediator Dishevelled (Dsh) to those sites. Monitoring these apparent Wnt11 signalling centres through time-lapse confocal microscopy revealed, that Wnt11 activity locally increases the persistency of cell-cell contacts. In addition, I found that the atypical cadherin Flamingo (Fmi) is required for this process. Fmi accumulates together with Wnt11/Fz7 at sites of cell-cell contact and locally increased cell adhesion, via a mechanism that appears to be independent of known downstream effectors of Wnt11 signalling such as RhoA and Rok2. This study indicates that Wnt11 locally interacts with Fmi and Fz7 to control cell-contact persistency and to facilitate coherent and coordinated cell migration. This provides a novel mechanism of non-canonical Wnt signalling in mediating cell behaviour, which is likely relevant to other developmental systems. (Die Druckexemplare enthalten jeweils eine CD-ROM als Anlagenteil: 50 MB: Movies - Nutzung: Referat Informationsvermittlung der SLUB)

The correct development of multicellular organisms is tightly regulated by intrinsic and extrinsic factors at specific time points. Disturbance on any level of these multiple processes may result in drastic phenotypes or eventually death of the organism. The midbrain-hindbrain boundary (also termed isthmic organizer) is a region of high interest as well in early as also in later development. The isthmic region carries organizer identity by the expression and subsequent release of FGF8. False patterning events of this region in early developmental stages would therefore display dramatic results over time. As it has been shown that the midbrain-hindbrain boundary (mhb) in the zebrafish is a compartment (or lineage restriction) boundary I tried to understand the underlying molecular mechanism for its correct establishment. In this work I focused both on embryological, molecular and genetic means to characterize involved molecules and mechanisms. In the first part of the thesis I followed in vivo cell transplantation assays, having started with an unbiased one. Cells of either side the mhb were challenged with this boundary by bringing them into direct cell contact with their ectopic counterpart. In a biased approach, cells overexpressing mRNA of specific candidate genes were transplanted and their clonal distribution in host embryos was analyzed. In the second part of the thesis I started interfering with specific candidate genes by transiently knocking down their protein translation. The adhesion molecules of the Eph/ephrin class had been shown to restrict cell mixing and thereby creating compartment boundaries in other tissues, such as the hindbrain, in the zebrafish and other organisms. Additionally, we generated several stable genetic mutant lines in cooperation with the Tilling facility at the Max-Planck-Institute. The only acquired potential null mutant ephrinB2bhu2971 was analyzed and characterized further. I observed that a knock down or knock out of only one of the ephrinB2 ligands does not seem to be sufficient for a loss of compartment boundary formation. The combinatory approach of blocking translation of EphrinB2a in ephrinB2bhu2971 mutants gave very complex and interesting phenotypes, which need to be investigated further.

Gastrulation leads to the formation of the embryonic germ layers, ectoderm, mesoderm and endoderm, and is the first key morphogenetic process that occurs in development. Gastrulation provides a unique developmental assay system in which to study cellular movements and rearrangements in vivo. The different cell movements occurring during gastrulation take place in a highly coordinated spatial and temporal manner, indicating that they must be controlled by a complex interplay of morphogenetic and inductive events. Generally, cell movement constitutes a highly integrated program of different cellular behaviors including sensing, polarization, cytoskeletal reorganization, and changes in adhesion and cell shape. During migration, these different behaviors require a continuous regulation and feedback control to direct and coordinate them. In this work, we analyze the cellular and molecular mechanisms underlying the different types of cell behaviors during gastrulation in zebrafish. Specifically, we focus on the role of the adhesive and mechanical properties of germ layer progenitors in the regulation of gastrulation movements. In the first part of the project, we investigated the role of the adhesive and mechanical properties of the different germ layer progenitor cell types for germ layer separation and stratification. In the second part of this study, we applied the same methodology to determine the function of germ layer progenitor cell adhesion in collective cell migration. Tissue organization is thought to depend on the adhesive and mechanical properties of the constituent cells. However, it has been difficult to determine the precise contribution of these different properties due to the lack of tools to measure them. Here we use atomic force microscopy (AFM) to quantify the adhesive and mechanical properties of the different germ layer progenitor cell types. Applying this methodology, we demonstrate that mesoderm and endoderm progenitors are more adhesive than ectoderm cells and that E-cadherin is the main adhesion molecule regulating this differential adhesion. In contrast, ectoderm progenitors exhibit a higher actomyosin-dependent cell cortex tension than mesoderm and endoderm progenitors. Combining these data with tissue self-assembly in vitro and in vivo, we provide evidence that the combinatorial activities of cell adhesion and cell cortex tension direct germ layer separation and stratification. It has been hypothesized that the directionality of cell movement during collective migration results from a collective property. Using a single cell transplantation assay, we show that individual progenitor cells are capable of normal directed migration when moving as single cells, but require cell-cell adhesion to participate in coordinated and directed migration when moving collectively. These findings contribute to the understanding of the gastrulation process. Cell-cell adhesion is required for collective germ layer progenitor cell migration, and cell cortex tension is critical for germ layer separation and stratification. However, many questions still have to be solved. Future studies will have to explore the interaction between the adhesive and mechanical progenitor cell properties, as well as the role of these properties for cell protrusion formation, cell polarization, interaction with extracellular matrix, and their regulation by different signaling pathways.

Oscillations are present at many different levels of biological organization. The cell cycle that directs the division of individual cells, the regular depolarization of neurons in the sinu-atrial node which underlies the regular beating of the heart, the circadian rhythms that govern the daily activity cycles of virtually all organisms, and the clocks that make entire populations of fireflies flash on and off in unison feature as prominent examples of biological clocks. During development, biological clocks regulate the patterning of growing tissues, as is the case in vertebrate somitogenesis, and potentially also in vertebrate limb outgrowth and axial segmentation of invertebrate embryos. During vertebrate segmentation, the embryonic axis is subdivided along its anterior-posterior axis into epithelial spheres of cells called somites. This rhythmic process is thought to be driven by a multicellular oscillatory gene network, the so-called segmentation clock. Oscillations of hairy and enhancer of split gene products have been proposed to constitute the core clockwork in individual cells, and these oscillators are coupled to each other by Delta-Notch intercellular signaling. The interaction of the segmentation clock with a posteriorly-moving arrest wavefront then translates the temporal information encoded by the clock into a spatial pattern of segments. In the framework of this Clock and Wavefront model, segment length is determined by both clock period and arrest wavefront velocity. How the period of the segmentation clock is regulated is presently unknown, and understanding the mechanism of period setting might yield insight into the nature and function of the segmentation clock. In this study, two different but complementary approaches were pursued to investigate how period is regulated in the zebrafish segmentation clock. First, it has been reported that zebrafish mind bomb (mib) mutant embryos form somites more slowly than their wt siblings, suggesting that Mib might be implicated in period setting. Mib is an E3 ubiquitin ligase required for ubiquitination and endocytosis of the Notch ligand Delta, and Notch signaling is impaired in mutants with defective Mib. It has been suggested that the mechanistic basis for the requirement of Delta endocytosis in Notch signaling is a need for Delta to enter a particular endocytic compartment, potentially a recycling endosome, in a ubiquitin-dependent manner, where its signaling ability might be established or amplified by an as yet unknown posttranslational modification. In the present study, Delta trafficking through the endocytic pathway was analyzed in the PSM of wt and mib embryos through colocalization studies with endocytic markers. The rationale of this approach was that if Delta gained access to a particular endocytic compartment through Mib-dependent endocytosis, the presence of Delta in this compartment would be expected to be reduced in mutants with defective Mib, thereby revealing the compartment’s identity. However, no qualitative changes in colocalization with different endocytic markers could be detected in mib mutants, and the methods available did not allow for quantification of colocalization in wt or mutant backgrounds. However, Delta colocalized with 13 markers of recycling endosomes, consistent with the hypothesis that these are functionally important in Notch signaling. More refined techniques will be necessary for a quantitative analysis of normal as compared to impaired Delta trafficking. A genetic approach to period regulation proved to be successful for the Drosophila circadian clock, where the identification of period mutants advanced the understanding of the clock’s genetic circuitry. This motivated a screen for period mutants of the segmentation clock, which was carried out by measuring somitogenesis period, segment length and arrest wavefront velocity in a pool of candidate mutants. A subset of Delta-Notch mutants, and embryos treated with a small-molecule inhibitor that impairs Notch signaling, displayed correlated increases in somitogenesis period and segment length, while there was no detectable change in arrest wavefront velocity. Combined, these findings suggested that segmentation clock period is increased in experimental conditions with impaired Delta-Notch signaling. Using a theoretical description of the segmentation clock as an array of coupled phase oscillators, the delay in the coupling and the autonomous frequency of individual cells were estimated from the direction and magnitude of the period changes. The mutants presented here are the first candidates for segmentation clock period mutants in any vertebrate. The nature of the molecular lesions in these mutants, all of which affect genes implicated in intercellular Delta-Notch signaling, suggests that communication between oscillating PSM cells is a key factor responsible for setting the period of the segmentation clock.

Fibroblast growth factor (Fgf) signaling plays multiple inductive roles during development of vertebrates (Itoh 2007). Some Fgfs, such as Fgf8, are locally secreted and signal over a long range to provide positional information in the target tissue (Scholpp and Brand 2004). Fgf ligands signal in a receptor-dependent manner via tyrosine kinase receptors, four of which have been so far identified. Fgf8 signaling was shown to depend both on receptor activation as well as endocytosis. The specificity of Fgf ligands and receptors as well as the function of receptors in the control of the Fgf signaling range have been, however, largely unclear. In this study, we show that the putative Fgf8 receptor Fgfr1 is duplicated in zebrafish and that it acts redundantly in the formation of the posterior mesoderm. Also, in overexpression studies we confirm the notion that receptor endocytosis influences Fgf8 signaling range. Through TILLING mutant recovery and morpholino knockdown studies we also show that Fgfr2 is required for growth and skeletal development in zebrafish, whereas Fgfr4 is required for pectoral fin specification and growth.

Glucose homeostasis is fundamental for all living organisms. In vertebrates, the hormone insulin regulates the metabolism of carbohydrates, fats and proteins. In order to sustain the glucose homeostasis, the pancreatic β-cells, which produce and secrete insulin, must coordinate their efforts to secrete the right amounts of insulin required by the organism. In vitro studies, have suggested that a subpopulation of β-cells, referred to as “hub-cells”, coordinate islet Ca2+ dynamics during insulin secretion. However, it is unclear whether the hub-cell model pertains to an in vivo scenario, where the islet is densely vascularized and innervated. In this thesis, we employed the genetically-encoded calcium indicator GCaMP6, confocal imaging and optogenetics, to characterize the Ca2+ dynamics of the zebrafish β-cells in vivo. We found that pancreatic β-cells present endogenous Ca2+ spikes in vivo under basal conditions. These Ca2+ spikes are rapidly suppressed after lowering glucose levels via insulin administration. In addition, the temporal inhibition of blood flow decreases the Ca2+ spikes, suggesting that β-cells are systemically connected. Furthermore, β-cells show a synchronized response to a pericardial glucose injection. Specifically, we found that Ca2+ spikes originate and emanate from a subset of β-cells that are the first to respond to a glucose stimulus. We define these cells as “leader-cells”. We tested if these cells could coordinate the islet in vivo by employing 2-photon laser ablation. Whereas ablation of control cells had no significant effect on the amplitude and duration of the subsequent Ca2+ spikes responses, ablation of leader cells led to a reduction in the Ca2+ response. Furthermore, we developed systems for optogenetic interrogation of β-cells in vivo. We show that the light-gated Cl- ion pump halorhodopsin (NpHR) can be applied to inhibit β-cell depolarization in the zebrafish. We also present the optically orthogonal system of the red Ca2+ indicator K-GECO1 in combination with the blue-shifted channelrhodopsin CheRiff to activate individual β-cell in vivo. Using these new tools, we provide examples where the activation of individual β-cells showed heterogeneous potential to trigger influx of Ca2+ in the rest of the β-cells. Overall, our results led us to propose a hierarchical model of islet coordination. In contrast to the majority of β-cells, which occupy the bottom of the hierarchy since they present low capability to recruit other cells, the leader cells occupy the top levels, being capable to coordinate a majority of the islet’s β-cells.:List of figures xii List of Tables xiii 1. Introduction 1 1.1. Diabetes and insulin 1 1.2. The endocrine pancreas 2 1.3. The diabetes pandemic 4 1.4. β-cell development in zebrafish and mammals 4 1.5. β-cells function and heterogeneity 6 1.6. β-cell coordination 8 1.7. Genetically-encoded calcium indicators 10 1.8. Genetically-encoded optogenetic actuators 13 1.9. Models to study In vivo β-cell coordination 16 2. In vivo β-cell Ca2+ dynamics 19 2.1. β-cells present endogenous Ca2+ spikes in vivo, which are not present ex vivo 19 2.2. Insulin injection reduces endogenous β-cell Ca2+ activity 22 2.3. Pharmacological inhibition of β-cell Ca2+ spikes interferes with glucose control 24 2.4 Transient blood flow interruption decreases β-cell calcium spikes 26 2.5 Glucose bolus leads to a synchronous response of β-cells 29 3. Leader β-cells coordinates Ca2+ dynamics in vivo 32 3.1. High speed 2D and 3D imaging reveals “leader” β-cells 32 3.2. Pan-islet response to glucose is impaired after leader β-cells ablation 41 4. Optically orthogonal toolset for in vivo optogenetics and Ca2+ imaging 46 4.1. Development of optogenetics actuators systems in zebrafish β-cells 46 4.2. Red fluorescent calcium reporters in zebrafish β-cells 47 4.3. In vivo temporal optogenetic silencing of β-cells 50 4.4. In vivo temporal optogenetic silencing of a subset of β-cells can inhibit the islet response 52 4.5. In vivo temporal optogenetic activation of β-cells 55 5. Discussion and future directions 61 5.1. β-cell calcium spikes are systemically influenced 61 5.2. First responder β-cells are present in vivo 64 5.3. Leader β-cells coordinate Ca2+ influx in vivo 66 5.4. β-cell optogenetic interrogation shows heterogeneous potential of individual β-cells for islet coordination 68 6. Materials and methods 75 6.1. Zebrafish strains and husbandry 75 6.2. Transgenic lines generation 76 6.3. Glucose measurements 77 6.4. Pericardial injection of glucose and insulin 77 6.5. Live imaging 77 6.6. Fast whole islet live imaging 78 6.7. Selective two-photon laser ablation of leader cells in the zebrafish islet. 78 6.7. Selective one-photon optogenetic interrogation of β-cells in the zebrafish islet. 79 6.8. Islet blood flow imaging 80 6.9. Mechanical heart stop 80 6.10. Immunostaining 80 6.11. TUNEL assay 81 6.12 Image analysis of GCaMP6s fluorescence intensity from in vivo imaging. 82 6.13 Quantification of GCaMP6s fluorescence intensity 82 6.14 Spatial drift correction images. 83 6.15 Statistical analysis 84 7. References 85 8. Annexes 90 9. Acknowledgments 97 / Die Glukosehomöostase ist für alle lebenden Organismen von grundlegender Bedeutung. Bei Wirbeltieren reguliert das Hormon Insulin den Stoffwechsel von Kohlenhydraten, Fetten und Proteinen. Um die Glukosehomöostase aufrechtzuerhalten, müssen die β-Zellen der Bauchspeicheldrüse, welche Insulin produzieren und absondern, ihre Bemühungen koordinieren, um die richtigen Mengen an Insulin zu sekretieren, die der Organismus benötigt. In-vitro-Studien haben gezeigt, dass eine Subpopulation von β-Zellen, die als „Hub-Zellen“ bezeichnet werden, die Insulinsekretion der Inseln koordiniert. Es ist jedoch unklar, ob sich die Hub-Cell-Theorie auf ein in-vivo-Szenario bezieht, bei dem die Insel dicht vaskularisiert und von Neuronen innerviert ist. In dieser Arbeit verwendeten wir den genetisch kodierten Calcium-Indikator GCaMP6, konfokale Bildgebung und Optogenetik, um die Ca2+-Dynamik der Zebrafisch-β-Zellen in vivo zu charakterisieren. Wir fanden heraus, dass Pankreas-β-Zellen in vivo unter basalen Bedingungen endogene Ca2+-Spitzen aufweisen. Diese Ca2+-Spitzen werden nach Senkung des Glukosespiegels durch Insulinverabreichung schnell unterdrückt. Darüber hinaus verringert die zeitliche Hemmung des Blutflusses die Ca2+-Spitzen, was darauf hindeutet, dass β-Zellen systemisch verbunden sind. Darüber hinaus zeigen β-Zellen eine synchronisierte Reaktion auf die perdikale Glukoseinjektion. Insbesondere fanden wir heraus, dass Ca2+-Spitzen von den β-Zellen hervorgerufen werden, die zuerst auf den Glukosestimulus reagieren. Wir definieren diese Zellen als 'Leader-Zellen'. Wir haben in vivo durch den Einsatz einer 2-Photonen-Laserablation getestet, ob diese Zellen die Insel koordinieren können. Während die Ablation von Kontrollzellen keinen signifikanten Einfluss auf die Amplitude und Dauer der nachfolgenden Ca2+-Spitzenreaktionen hatte, führte die Ablation von Leader-Zellen zu einer signifikanten Verringerung der GCaMP-Reaktion. Darüber hinaus haben wir Systeme für die optogenetische Abfrage von β-Zellen in vivo entwickelt: Wir zeigen, dass die lichtgesteuerte Cl—Ionenpumpe Halorhodopsin (NpHR) angewendet werden kann, um die Depolarisation von β-Zellen in vivo zu hemmen. Wir präsentieren auch das optisch orthogonale System des roten Ca2+-Indikators K-GECO1 in Kombination mit dem blauverschobenen Channelrhodopsin CheRiff, um einzelne β-Zellen in vivo abzufragen. Unter Verwendung dieser neuen Werkzeuge liefern wir Beispiele, bei denen die Aktivierung einzelner β-Zellen ein heterogenes Potenzial für die Auslösung des Ca2+-Einstroms in die übrigen β-Zellen in vivo zeigte. Insgesamt bietet diese Studie Hinweise darauf, dass eine Untergruppe von β-Zellen ein hohes Potenzial zur Koordination der Ca2+-Dynamik der Insel in vivo aufweist.:List of figures xii List of Tables xiii 1. Introduction 1 1.1. Diabetes and insulin 1 1.2. The endocrine pancreas 2 1.3. The diabetes pandemic 4 1.4. β-cell development in zebrafish and mammals 4 1.5. β-cells function and heterogeneity 6 1.6. β-cell coordination 8 1.7. Genetically-encoded calcium indicators 10 1.8. Genetically-encoded optogenetic actuators 13 1.9. Models to study In vivo β-cell coordination 16 2. In vivo β-cell Ca2+ dynamics 19 2.1. β-cells present endogenous Ca2+ spikes in vivo, which are not present ex vivo 19 2.2. Insulin injection reduces endogenous β-cell Ca2+ activity 22 2.3. Pharmacological inhibition of β-cell Ca2+ spikes interferes with glucose control 24 2.4 Transient blood flow interruption decreases β-cell calcium spikes 26 2.5 Glucose bolus leads to a synchronous response of β-cells 29 3. Leader β-cells coordinates Ca2+ dynamics in vivo 32 3.1. High speed 2D and 3D imaging reveals “leader” β-cells 32 3.2. Pan-islet response to glucose is impaired after leader β-cells ablation 41 4. Optically orthogonal toolset for in vivo optogenetics and Ca2+ imaging 46 4.1. Development of optogenetics actuators systems in zebrafish β-cells 46 4.2. Red fluorescent calcium reporters in zebrafish β-cells 47 4.3. In vivo temporal optogenetic silencing of β-cells 50 4.4. In vivo temporal optogenetic silencing of a subset of β-cells can inhibit the islet response 52 4.5. In vivo temporal optogenetic activation of β-cells 55 5. Discussion and future directions 61 5.1. β-cell calcium spikes are systemically influenced 61 5.2. First responder β-cells are present in vivo 64 5.3. Leader β-cells coordinate Ca2+ influx in vivo 66 5.4. β-cell optogenetic interrogation shows heterogeneous potential of individual β-cells for islet coordination 68 6. Materials and methods 75 6.1. Zebrafish strains and husbandry 75 6.2. Transgenic lines generation 76 6.3. Glucose measurements 77 6.4. Pericardial injection of glucose and insulin 77 6.5. Live imaging 77 6.6. Fast whole islet live imaging 78 6.7. Selective two-photon laser ablation of leader cells in the zebrafish islet. 78 6.7. Selective one-photon optogenetic interrogation of β-cells in the zebrafish islet. 79 6.8. Islet blood flow imaging 80 6.9. Mechanical heart stop 80 6.10. Immunostaining 80 6.11. TUNEL assay 81 6.12 Image analysis of GCaMP6s fluorescence intensity from in vivo imaging. 82 6.13 Quantification of GCaMP6s fluorescence intensity 82 6.14 Spatial drift correction images. 83 6.15 Statistical analysis 84 7. References 85 8. Annexes 90 9. Acknowledgments 97

Mammals cannot re-form heavily damaged bones as in large fracture gaps, whereas zebrafish efficiently regenerate bones even after amputation of appendages. However, the source of osteoblasts that mediate appendage regeneration is controversial. Several studies in zebrafish have shown that osteoblasts are generated by dedifferentiation of existing osteoblasts at injured sites, but other observations suggest that de novo production of osteoblasts also occurs. In this study, we found from cell-lineage tracing and ablation experiments that a group of cells reserved in niches serves as osteoblast progenitor cells (OPCs) and has a significant role in fin ray regeneration. Besides regeneration, OPCs also supply osteoblasts for normal bone maintenance. We further showed that OPCs are derived from embryonic somites, as is the case with embryonic osteoblasts, and are replenished from mesenchymal precursors in adult zebrafish. Our findings reveal that reserved progenitors are a significant and complementary source of osteoblasts for zebrafish bone regeneration.