Includes bibliographical references (leaves 142-168) Aim was the purification and identification of the early primitive ectoderm-like (EPL) cell induction signals within the medium conditioned by the human hepatocellular carcinoma cell line HepG2 and the localisation of the signals that induce EPL cell and primitive ectoderm formation.

Includes bibliographical references (leaves 142-168) / x, 168 leaves : ill. (chiefly col.) ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Aim was the purification and identification of the early primitive ectoderm-like (EPL) cell induction signals within the medium conditioned by the human hepatocellular carcinoma cell line HepG2 and the localisation of the signals that induce EPL cell and primitive ectoderm formation. / Thesis (Ph.D.)--University of Adelaide, Dept. of Molecular Biosciences (Biochemistry), 2001

Fibrocartilages in the body, including the temporomandibular joint (TMJ) disc and knee meniscus, lack intrinsic healing capacity following trauma or disease. Current treatments only address the symptoms of fibrocartilage damage and do nothing to prevent further degradation of the joint. A tissue engineered replacement, with biochemical and biomechanical properties approaching those of native tissue, could provide a solution. This thesis investigates two components critical to the generation of a tissue engineered TMJ disc: 1) characterization of the native disc to identify a suitable animal model and create design parameters, and 2) development of approaches to use human embryonic stem cells (hESCs) in fibrocartilage tissue engineering. The first step to achieving this goal was to identify an animal model for the human TMJ disc based on quantitative biochemical and biomechanical properties. To this end, rabbit, goat, pig, cow, and human discs were analyzed, and the pig disc was shown to possess properties most similar to the human. The next step was to further characterize the pig TMJ, as many aspects of the joint were still poorly understood. Though the TMJ disc is anchored to the surrounding bony tissue on all sides by discal attachments, little was known about their properties. Biochemical and histological analysis was performed on these attachments and indicated that they are similar to the disc but possess distinct regional matrix content related to joint biomechanics. Finally, though the contribution of collagen to the mechanical properties of the TMJ disc was well characterized, the contribution of the glycosaminoglycans (GAGs) was unknown. By removing sulfated GAGs with chondroitinase ABC, it was found that these molecules contribute to the viscoelastic compressive properties of the disc, but only in regions with the highest native GAG content. The second aspect of this thesis involved producing fibrocartilage tissue from hESCs. The pluripotency and unlimited self-renewal of these cells makes them ideally suited for producing fibrocartilages that contain a spectrum of matrix components. This work began by investigating what factors are necessary for fibrochondrogenic differentiation of hESCs in embryoid bodies (EBs). Growth factors and co-cultures with primary fibrochondrocytes were both shown to be potent modulators of fibrochondrogenesis, although differentiation of hESCs consistently produced a heterogeneous cell population. To purify populations of fibrochondrocytes differentiated form hESCs, two inexpensive and novel techniques were investigated. First, density gradient separation was the first technique attempted. This technique was able to isolate distinct subpopulations of cells, some of which were mechanically similar to native chondrocytes. Second, a chondrogenic tuning technique was applied to differentiated hESCs. Following fibrochondrogenesis in EBs, cells were expanded in monolayer in chondrocyte specific media before being used for tissue engineering. Chondrogenic tuning produced several distinct cell populations during expansion, and, as a result, a spectrum of different cartilaginous tissues was achieved for tissue engineering. Three of the cell populations produced tissues similar to the native TMJ disc, outer meniscus, and inner meniscus. Overall, this thesis identified an animal model for TMJ characterization and in vivo studies, furthered understanding of structure-function relationships of the TMJ disc and its attachments, and developed a technique for producing a spectrum of engineered fibrocartilages from hESCs.

The Hedgehog (Hh) signalling pathway is one of the key signalling pathways orchestrating intricate organogenesis, including the development of neural tube, heart and skeletal muscle. Yet, insufficient mechanistic understanding of its diverse roles is available. Here, we show the molecular mechanisms regulating the neurogenic, cardiogenic and myogenic properties of Hh signalling, via effector protein Gli2, in embryonic and adult stem cells. In Chapter 2, we show that Gli2 induces neurogenesis, whereas a dominant-negative form of Gli2 delays neurogenesis in P19 embryonal carcinoma (EC) cells, a mouse embryonic stem (ES) cell model. Furthermore, we demonstrate that Gli2 associates with Ascl1/Mash1 gene elements in differentiating P19 cells and activates the Ascl1/Mash1 promoter in vitro. Thus, Gli2 mediates neurogenesis in P19 cells at least in part by directly regulating Ascl1/Mash1 expression. In Chapter 3, we demonstrate that Gli2 and MEF2C bind each other’s regulatory elements and regulate each other’s expression while enhancing cardiomyogenesis in P19 cells. Furthermore, dominant-negative Gli2 and MEF2C proteins downregulate each other’s expression while imparing cardiomyogenesis. Lastly, we show that Gli2 and MEF2C form a protein complex, which synergistically activates cardiac muscle related promoters. In Chapter 4, we illustrate that Gli2 associates with MyoD gene elements while enhancing skeletal myogenesis in P19 cells and activates the MyoD promoter in vitro. Furthermore, inhibition of Hh signalling in muscle satellite cells and in proliferating myoblasts leads to reduction in MyoD and MEF2C expression. Finally, we demonstrate that endogenous Hh signalling is important for MyoD transcriptional activity and that Gli2, MEF2C and MyoD form a protein complex capable of inducing skeletal muscle-specific gene expression. Thus, Gli2, MEF2C and MyoD participate in a regulatory loop and form a protein complex capable of inducing skeletal muscle-specific gene expression. Our results provide a link between the regulation of tissue-restricted factors like Mash1, MEF2C and MyoD, and a general signal-regulated Gli2 transcription factor. We therefore provide novel mechanistic insights into the neurogenic, cardiogenic and myogenic properties of Gli2 in vitro, and offer novel plausible explanations for its in vivo functions. These results may also be important for the development of stem cell therapy strategies.

Newly established F1 hybrid Embryonic Stem cells allow the production of ES cell-derived animals at a high enough efficiency to directly make ES cell based genetics feasible. An F1 hybrid ES cell line, G4 was used to generate transgenic over-expressing cell lines. The consequence of the expression of a panel of transgenes was assessed directly from ES cell-derived embryos produced by the tetraploid complementation assay. The generation of ES cell-derived embryos/animals was very efficient. A sufficient number of mutants for initial phenotypic analyses was derived only a few weeks after the establishment of the cell lines. The genes used in the study had either angiogenic/vasculogenic, anti-angiogenic or unknown properties. Of these transgenic mouse lines VEGF-A and Flt-Fc were used to further elucidate the effects of altered VEGF signaling on cell fate decisions in embryonic development and ES differentiation in two experimental systems. A. Early but transient Flk-1 activation led to enhanced generation of blood progenitors, whereas continuous activation of Flk-1 abolished this effect and enhanced endothelial cell generation. Ex vivo analysis of cells derived from E7.5 embryos demonstrated that sFlt-1-mediated control of Flk-1 activity also impacted the fate of hematopoietic and endothelial cells. The Flt-1-Fc transgenic mouse model was used to alter Flk-1 activation in vivo and show the relevance of the in vitro observations. These results demonstrate that sFlt-1 regulates Flk-1 activation in an oxygen responsive manner. Inhibition of Flk-1 activation by sFlt-1 increases the specification of hemangioblasts to blood cells consistent with a VEGF-independent default mechanism. B. Ubiquitous over-expression of VEGF164 isoform led to E8.75 embryonic lethality. The primary cause of lethality was the failure to form an organized cardiovascular system, which was manifested in three ways: the absence of yolk sac blood vessels, the lack of embryonic-maternal circulation due to the failure of allantochorionic fusion and improper cardiac function. The described phenotypes suggest that VEGF does not inhibit embryonic or extra-embryonic mesoderm formation at gastrulation but perturbs the balance amongst the mesodermal components.

Construction of the vertebrate nervous system begins with the decision of a group of ectoderm cells to take on a neural fate. Studies using Xenopus ectodermal explants, or with mouse ectoderm cells or embryonic stem (ES) cells, demonstrated that this process of neural determination occurred by default – the ectoderm cells became neural after the removal of inhibitory signals. Whether ectoderm or ES cells directly differentiate into bona fide neural stem cells is not clear. One model suggests that there is an intermediate stage where “primitive” neural stem cells (pNSC) emerge harbouring properties of both ES cells and neural stem cells. The goal of my research was to address this question by evaluating the role of growth factor signaling pathways and their impact on the function of the zinc-finger homeobox transcription factor, Zfhx1b, during mouse neural stem cell development. I tested whether FGF and Wnt signaling pathways could regulate Zfhx1b expression to control early neural stem cell development. Inhibition of FGF signaling at a time when the ectoderm is acquiring a neural fate resulted in the accumulation of too many pNSCs, at the expense of the definitive neural stem cells. Interestingly, over-expression of Zfhx1b was sufficient to rescue the transition from a pNSC to definitive NSC. These data suggested that definitive NSC fate specification in the mouse ectoderm was facilitated by FGF activation of Zfhx1b, whereas canonical Wnt signaling repressed Zfhx1b expression. Next I sought to determine whether Zfhx1b was similarly required during neural lineage development in ES cells. FGF and Wnt signaling modulated expression of Zfhx1b in ES cell cultures in manner resembling my observations from similar experiments using mouse ectoderm. Knockdown of Zfhx1b in ES cells using siRNA did not affect the initial transition of ES cells to pNSC fate, but did limit the ability of these neural cells to further develop into definitive NSCs. Thus, my findings using ES cells were congruent with evidence from mouse embryos and supported a model whereby intercellular signaling induced Zfhx1b, required for the development of definitive NSCs, subsequent to an initial neural specification event that was independent of this pathway.

Embryonic stem cells (ESC) constitute a promising source of cells for cardiac transplantation strategies. However, complexities associated with in vivo studies have made it difficult to develop a thorough understanding of cell integration. We have engineered an in vitro system that recapitulates the native cardiac environment using 300μm thick collagen scaffolds seeded with neonatal cardiomyocytes (CM) and electrical field stimulation. The injection of undifferentiated ESC served as a baseline to assess the validity of studying cell transplantation in this model. Yfp-ESC survived and proliferated over several days in model tissue. ESC were not observed to significantly differentiate into the cardiac lineage, and did not integrate with the cardiac cell population. While the injection of ESC improved cardiac cell number, tissue functional properties were hindered. The methods developed herein can be readily adapted to study ESC derived progenitor and differentiated cells, to elucidate the optimal cell state for ESC-mediated cell therapy.

Embryogenesis occurs within a complex and dynamic cellular environment that influences cell fate decisions. Pluripotent stem cells (PSCs) are a valuable tool for research into disease models as well as a resource for cell therapy due to their capacity to self-renew and differentiate into all cell types. Mimicking aspects of the embryonic microenvironment in vitro impacts the resultant functional cells. The aim of this work was to develop a controlled and scaleable process for the generation of hematopoietic progenitor cells (HPCs) from embryonic stem cells (ESCs). We demonstrated with bioreactor-grown embryoid bodies (EBs) that increased HPC generation can be elicited by decreasing the oxygen tension by a mechanism where vascular endothelial growth factor receptor 2 (VEGFR2) activation is controlled through competition with the ligand decoy VEGFR1. This is important as it demonstrates the inherent responsiveness of the developing hematopoietic system to external forces and influences. We also established a serum-free system that facilitates directed differentiation, determining 5 ng/ml bone morphogenetic protein-4 (BMP4) with 50 ng/ml thrombopoietin (TPO) could generate 292 ± 42 colony forming cells (CFC)/5 x 10^4 cells with early VEGF treatment (25 ng/ml, day 0-5). We also controlled aggregate size influencing relative endogenous and exogenous growth factor signaling and modulating mesodermal differentiation; CFC output was optimal when initialized with 100 cell aggregates. For the first time, we demonstrated efficacy of local growth factor delivery by producing HPCs with gelatin microparticles (MP). Overall, these design components generate HPCs in a controlled and reproducible manner using a serum-free bioprocess that couples size controlled aggregates containing gelatin MPs for localized growth factor release of BMP4 and TPO with hypoxia to induce endogenous VEGF production. These strategies provide a tunable platform for developing cell therapies and high density growth, within a bioreactor system, can be facilitated by hydrogel encapsulation of the aggregates.

Construction of the vertebrate nervous system begins with the decision of a group of ectoderm cells to take on a neural fate. Studies using Xenopus ectodermal explants, or with mouse ectoderm cells or embryonic stem (ES) cells, demonstrated that this process of neural determination occurred by default – the ectoderm cells became neural after the removal of inhibitory signals. Whether ectoderm or ES cells directly differentiate into bona fide neural stem cells is not clear. One model suggests that there is an intermediate stage where “primitive” neural stem cells (pNSC) emerge harbouring properties of both ES cells and neural stem cells. The goal of my research was to address this question by evaluating the role of growth factor signaling pathways and their impact on the function of the zinc-finger homeobox transcription factor, Zfhx1b, during mouse neural stem cell development. I tested whether FGF and Wnt signaling pathways could regulate Zfhx1b expression to control early neural stem cell development. Inhibition of FGF signaling at a time when the ectoderm is acquiring a neural fate resulted in the accumulation of too many pNSCs, at the expense of the definitive neural stem cells. Interestingly, over-expression of Zfhx1b was sufficient to rescue the transition from a pNSC to definitive NSC. These data suggested that definitive NSC fate specification in the mouse ectoderm was facilitated by FGF activation of Zfhx1b, whereas canonical Wnt signaling repressed Zfhx1b expression. Next I sought to determine whether Zfhx1b was similarly required during neural lineage development in ES cells. FGF and Wnt signaling modulated expression of Zfhx1b in ES cell cultures in manner resembling my observations from similar experiments using mouse ectoderm. Knockdown of Zfhx1b in ES cells using siRNA did not affect the initial transition of ES cells to pNSC fate, but did limit the ability of these neural cells to further develop into definitive NSCs. Thus, my findings using ES cells were congruent with evidence from mouse embryos and supported a model whereby intercellular signaling induced Zfhx1b, required for the development of definitive NSCs, subsequent to an initial neural specification event that was independent of this pathway.

Embryonic stem cells (ESC) constitute a promising source of cells for cardiac transplantation strategies. However, complexities associated with in vivo studies have made it difficult to develop a thorough understanding of cell integration. We have engineered an in vitro system that recapitulates the native cardiac environment using 300μm thick collagen scaffolds seeded with neonatal cardiomyocytes (CM) and electrical field stimulation. The injection of undifferentiated ESC served as a baseline to assess the validity of studying cell transplantation in this model. Yfp-ESC survived and proliferated over several days in model tissue. ESC were not observed to significantly differentiate into the cardiac lineage, and did not integrate with the cardiac cell population. While the injection of ESC improved cardiac cell number, tissue functional properties were hindered. The methods developed herein can be readily adapted to study ESC derived progenitor and differentiated cells, to elucidate the optimal cell state for ESC-mediated cell therapy.