Cortical networks are comprised of a multitude of cell types. To understand sensory processing, the function and interaction of these cell types must be investigated. Neurons can be separated into two main groups: excitatory pyramidal (Pyr) cells and inhibitory interneurons. Inhibitory interneurons make up 20% of the total cortical neuronal population and they exhibit a striking array of molecular, morphological and electrophysiological characteristics. The most numerous are the parvalbumin-expressing (PV+) interneurons, accounting for 35-40% of the interneuron population in adult mouse visual cortex. Somatostatin-expressing (SOM+) neurons are another significant group, comprising 20-25% of the interneuron population. The visual responses of SOM+ and PV+ interneurons were measured using 2-photon targeted cell-attached recordings and compared with Pyr cells in the primary visual cortex of anaesthetized mice. These interneuron populations exhibited higher firing rates than Pyr cells in response to oriented gratings, but were less orientation selective, with PV+ interneurons exhibiting the lowest orientation selectivity. Next, SOM+ interneurons were stimulated optogenetically using channelrhodopsin to measure their effect on Pyr cell and PV+ interneuron responses to visual stimuli. Activating small numbers of SOM+ interneurons in vivo inhibited stimulus- evoked firing in PV+ interneurons but not in Pyr cells. Stimulating a large number of SOM+ interneurons confirmed this differential effect, inhibiting PV+ interneurons twice as effectively as Pyr cells. Moreover, the remaining responses to oriented gratings in PV+ cells were more orientation-tuned and time-modulated. In short, inhibitory SOM+ cell activity does not summate with PV+ cell activity, but suppresses it, reconfiguring the inhibitory input to Pyr cells. These results suggest a new role for SOM+ cells, which are activated more slowly and provide dendritic inhibition to Pyr cells while strongly antagonizing PV+ cells, thereby shifting inhibitory input to Pyr cells from somatic to dendritic inhibition throughout the course of the network's visual response.

The rules governing synaptic connectivity in neural circuits are still poorly understood. In the cerebellar cortex, the general wiring patterns between different cell types are known, but the precise circuit at the level of individual cells is still unclear. In the molecular layer, thought to be the principal site of cerebellar learning, I investigated three fundamental microcircuits and their connectivity: the network of chemically and electrically coupled inhibitory interneurons, their inhibitory connection to Purkinje cells, and their shared parallel fibre input, which together form a feed-forward inhibitory circuit. I used both electrophysiological and anatomical techniques to probe functional and structural connectivity patterns in these circuits. Optimisation was performed to combine electrophysiological recordings and imaging with a focused ion beam scanning electron microscope. The resulting high-resolution three-dimensional imaging reveals the precise connectivity of hundreds of individual parallel fibres. I found that their connectivity is constrained by volume exclusion rules between presynaptic boutons. These spatial constraints are overcome between directly adjacent interneurons and Purkinje cells, receiving input from a common parallel fibre bouton. Using multiple patch-clamp recordings, I found evidence for specific, non-random connectivity amongst molecular layer interneurons. The electrical and chemical networks are both highly clustered and their overlap is non-random. Inhibitory connections between interneurons are preferentially organized in a feed-forward (or transitive) fashion. I show that individual interneurons can both decrease and increase the inhibition received by Purkinje cells, via an indirect connection through electrically coupled interneurons. A highly specific connectivity pattern supports this control of inhibition of Purkinje cells by the coupled network. To summarize, I investigated the rules governing neural connectivity in the molecular layer: from spatial constraints, via formation of specific connectivity motifs, to the implementation of functional roles. This new information will help shape our understanding of the interactions and computations performed in the cerebellar cortex.

Oligodendrocyte progenitors (OLPs) are widely distributed throughout the embryonic and adult central nervous system (CNS). They are the primary, possibly the only source of oligodendrocytes in the developing CNS and in the repair of demyelinated lesions, but it is unclear whether they continue to generate myelinating oligodendrocytes or any other cells in the normal healthy adult. A transgenic mouse line, Pdgfra-CreERT2, allowed temporally-controlled, permanent activation of YFP expression in PDGFRa-expressing OLPs and their differentiated progeny in the Rosa26-YFP reporter background. In vivo BrdU labelling and fate mapping of YFP-positive cells revealed that, in the corpus callosum, approximately half of the OLPs were dividing with a cell cycle time of approximately 8 days. Half of the daughter cells differentiated into myelinating oligodendrocytes, the others remained undifferentiated to maintain the OLP population. In the cortex, the majority of differentiated cells were SOX10+ non-myelinating oligodendrocyte-lineage cells of unknown function. In addition, YFP-labelled projection neurons accumulated in the piriform cortex (primary olfactory cortex). YFP-labelled astrocytes were not found anywhere in the forebrain. In collaboration with RMJ Franklin and colleagues (University of Cambridge) the fates of OLPs during repair of gliotoxin-induced demyelinated spinal cord lesions were investigated. In the lesions, OLPs generated mainly remyelinating oligodendrocytes and Schwann cells as well as a few astrocytes. Taken together, my in vivo Cre/lox fate-mapping studies showed that OLPs can differentiate into myelinating oligodendrocytes, non-myelinating oligodendrocytelineage cells, cortical projection neurons and - during remyelination - Schwann cells and astrocytes. Thus, OLPs possess multipotential stem cell-like properties in the adult mouse CNS.

microRNAs and other small RNAs generated by the ribonuclease Dicer are involved in a variety of physiological processes including immune responses, neuronal function and modulation of synaptic plasticity. This thesis uses the Nav1.8-Cre mouse to ablate Dicer in post-mitotic nociceptors, in order to study the role of small RNAs in nociception. Colleagues have previously identified that loss of small RNAs within these nociceptors leads to deficits in inflammatory pain behaviour, and that the expression of nociceptor transcripts is greatly reduced. In this thesis electrophysiological recordings were used. It was determined that Dicer null nociceptors were not capable of being sensitised using inflammatory mediators. This was found to be due to reduced functional voltage gated sodium channel expression, particularly Nav1.8 and Nav1.9, leading to impaired action potential electrogenesis and loss of repetitive firing properties. Deep sequencing and bioinformatic analysis were used to identify the miRNA and small RNA repertoire in dorsal root ganglia (DRG) tissue. Additionally, nociceptor enriched miRNAs were identified using the conditional Dicer knockout mouse. The observation that the loss of miRNAs leads to reduced expression of nociceptor transcripts does not fit with the canonical RNA interference (RNAi) paradigm. Therefore a hypothesis that small RNAs can induce transcription was tested using bioinformatic approaches and functional studies in DRG neurons.

Neurons of the central nervous system rely on finely tuned integrative properties to perform the computations that allow transformation of sensory input into an output. It has long been a goal of neuroscience to understand this computation, in order to grasp the fundamental function of neuronal circuits. In my thesis, I have focused on cerebellar granule cells to dissect the input-output transformation that takes place within a single neuron. The small size of the granule cell together with its limited number of inputs make it an ideal cell in which to study integration of synaptic input. Using whole-cell patch clamp recordings in awake, head-fixed mice I have characterised the input-output transformations that underlie the flow of sensory information through the cerebellar cortex. I found that in the awake state the cerebellum receives a greatly increased amount of synaptic input compared to the anesthetised state. This high frequency input appears to contribute to distinct integrative properties in granule cells, such as a considerably lower input resistance. Surprisingly, despite the dramatically higher rate of excitatory input, output spiking rates in the resting awake state remain similar to the anesthetised state. However, the onset of locomotion was correlated with an increase in spiking, associated with a further increase in excitatory synaptic input. This suggests that these cells may be ‘primed’ to fire explicitly during motor function, allowing transmission of highly filtered sensory information with a very high signal-to-noise ratio. My experiments suggest that glutamatergic spillover may contribute to synaptic transmission during locomotion. Furthermore, spatial segregation of inhibitory inputs or their modulation may also play a role in this function-related firing, but this remains an open question. Together these results represent the first example of relating synaptic input of single cells with the behavioural state of an awake animal.

Introduction: Mitochondria are considered to be the powerhouse of the cell being the primary generators of ATP, they also have numerous other important functions including; being the main generator of reactive oxygen species (ROS) and a central role in apoptosis. As the main intracellular source of ROS, many people believe that mitochondria play a significant role in ageing. Senescence is associated with ageing and has been associated with atherosclerotic vascular disease. The concept of human cells lacking functional mitochondria (Rho 0 cells) is not new and was first described by Attardi et al in 1989. However, most of this work has been done on immortalised cell lines. Aims: To see if it is possible to generate and characterise Rho 0 human endothelial cells. To use these cells as a tool to investigate the mechanisms by which they respond to stress and whether differences in ROS production and/or antioxidant defences account for any differences observed. Methods: Human Umbilical Vein Endothelial Cells (HUVEC) were grown in media supplemented with glucose and uridine in the presence of low dose ethidium bromide. Rho 0 status of the cells was confirmed by auxotrophy for uridine, quantitative PCR for mitochondrial-encoded gene expression and western blots for mitochondrial-encoded proteins. Results: The Rho 0 status of the cells was confirmed by; auxotrophy for uridine (Rho 0 cells die in medium lacking uridine), absence of mitochondrial-encoded genes (subunit-1 of complex IV and subunit-6 of subunit V) and lack of expression of the mitochondrial-encoded protein subunit-1 of complex IV. Rho 0 cells are resistant to both stress-induced senescence and apoptosis. They produce less ROS and have upregulated antioxidant defences. Conclusions: It is possible to grow Rho 0 HUVEC. These cells are a useful tool for studying the role of mitochondria in senescence and apoptosis in the cardiovascular system.

Nitric oxide (NO) functions as a signalling molecule throughout the brain where, via the intracellular generation of cGMP, it participates in many functions, such as in synaptic plasticity. The initial experiments were based on the finding that, in optic nerve, NO released from blood vessels tonically depolarises axons. The aim was to test the hypothesis that the tonic NO production is maintained by phosphorylation of endothelial NO synthase (eNOS). The results from extracellular recordings of changes in the axonal membrane potential suggested that PI3 kinase-mediated eNOS phosphorylation is partially responsible. The subsequent aim was to determine if blood vessel-neuron communication may be more widespread, by investigating if this mechanism accounts for basal NO production in the developing rat hippocampus. For this purpose, measurements of cGMP were chosen as a sensitive index of the local NO concentration. Contrary to expectations, no clear evidence for a dominant role of either eNOS or the neuronal NO synthase emerged, although the data suggested that NO formation was calcium-dependent. The next step was to characterise the target cells of endogenous and exogenous NO in the hippocampus, particularly in the light of findings that, with a better tool for inhibiting the dominant phosphodiesterase activity (phosphodiesterase-2), much higher cGMP levels could be evoked than previously. Accordingly, instead of a predominant location in astrocytes, cGMP immunocytochemistry showed widespread staining of neuronal elements (somata, dendrites, neuropil) throughout the tissue. The final objective was to begin to analyse NO transduction in cells in real-time, using a newly developed fluorescent cGMP sensor. Cell lines expressing various levels of guanylyl cyclase and phosphodiesterase were selected for study. Cellular responsiveness to extremely low NO concentrations (down to 3 pM) could be detected. Moreover, the findings illustrated how the interplay between guanylyl cyclase and phosphodiesterase activities serves to generate distinct cellular cGMP profiles.

Epithelial ovarian cancer (EOC) is the most common form of gynaecological malignancy in the developed world. Identifying molecular markers of disease may provide novel approaches to screening and could enable targeted treatment and the design of novel therapies. In previous work, 141 primary ovarian tumours were analysed using metaphase comparative genomic hybridization to identify complete or partial chromosome deletions that may harbour tumour and/or metastasis suppressor genes. Chromosome 18 (Ch18) was found to have deletions in 50% of the tumours. Microcell-mediated chromosome transfer (MMCT) of normal Ch18 material into EOC cell lines resulted in hybrids that displayed significant suppression of anchorageindependent growth, invasiveness and reduced tumour growth in nude mice. The major aim of this project was to identify protein changes associated with the tumour suppression observed in the EOC Ch18 MMCT cell models. The project involves a detailed quantitative proteomic comparison of two parental ovarian cancer cell lines (derived from primary endometrioid and clear cell carcinomas) and their MMCT-derived Ch18 hybrid clones. The cellular, secreted and surface proteomes were probed in order to gain comprehensive coverage and to improve the likelihood of useful biomarker identification. A combination of quantitative two-dimensional difference gel electrophoresis (2DDIGE), affinity chromatography and two-dimensional-liquid chromatography and tandem mass spectrometry (2D-LC-MS/MS) have been employed to examine the whole cell, secreted and cell surface proteomes of the parental and hybrid cell models to identify differentially expressed proteins as potential markers of tumour suppression. Proteins of interest have been validated using immune-based detection methods in the parental cell lines, Ch18 hybrids, revertant cell lines, a panel of cancer cell lines and normal ovarian surface epithelium cell lines and in serum from a set of ovarian cancer cases and healthy controls.

Comprehending gnathostome evolution requires insights into key cellular and molecular components of craniofacial and shoulder development. For the work of this PhD, I made use of genetically modified mouse models to study aspects of mammalian head and shoulder morphogenesis by triple fluorescent RNA in situ hybridisation, immunohistochemistry and high resolution imaging. First- I use a genetically defined sentinel cell population labelled by the Hand2-Cre transgene to establish the expansion of the distal-most branchial arch domain and correlate this by triple fluorescent RNA in situ hybridisation with the system controlling proximo-distal branchial arch patterning, the Dlx system. I find that the axis of the Dlx system does not correspond to the proximo-distal but an endodermal-ectodermal axis of the arch and rotates during development; the overall expansion of the arch is explicable by telescopic outgrowth along this new axis. Second- I study the cellular and molecular characteristics of head/ shoulder skeleto-muscular connectivity and the contribution of limb lateral plate mesoderm to the shoulder girdle, which allows me to identify part of the manubrium sterni as the ‘lost’ mammalian procoracoid and to demonstrate that the interaction between lateral plate mesodermal subpopulations is non-random. Third- I establish novel roles for Hand2 in lower incisor ameloblasts and in laminar dermal bone formation, suggesting a fundamental role for Hand2 in epithelial and mesenchymal cell layer arrangements. My detailed study of the murine frontal bone reveals that the establishment of an internal and an external layer initiates dermal bone formation; the latter shows intermediate molecular periosteal/ perichondrial characteristics and generates the intermediate layer by a Hand2-dependent invagination process. For a comparative amphibian data set, I begin to establish genetic lineage labelling as technique in Xenopus tropicalis. I generate and test a Xenopus Hand2-Cre transgene and establish a stable generic Xenopus tropicalis Cre-reporter line by I-SceI mediated transgenesis.

During development oligodendrocyte progenitor cells (OLPs) are responsible for the production of oligodendrocytes. Cells with similar antigenic properties to developmental OLPs persist throughout postnatal life, beyond the cessation of “developmental” myelination. These postnatal cells are often referred to as “NG2 cells” because they (and their perinatal counterparts) express the NG2 proteoglycan, but little is known about their function in the adult brain. Experiments documented in this Thesis use transgenic lineage tracing technology to characterize the in vivo behavior of OLPs in the brain at various ages. Fate mapping of OLPs revealed that they give rise to oligodendrocytes throughout life. In addition, OLPs were shown to generate a small proportion of the projection neurons present in the posterior piriform cortex, while no evidence for astrogliogenesis from OLPs was found. Cumulative in vivo labelling of OLPs with thymidine analogues (BrdU and EdU) showed that they proliferate continuously throughout life with an increasing cell cycle time with age. At all ages examined, there was a proportion of OLPs that never underwent cell division, indicating that there are cycling and non-cycling populations of OLPs in the mouse brain that persist throughout life. The observed contribution of adult-born oligodendrocytes to myelinating the brain was surprisingly large, and raised intriguing questions as to the necessity and function of these new myelinating cells. To investigate this directly, I generated a new transgenic mouse line that when crossed to a transgenic mouse that expresses an inducible form of Cre, can be used to selectively ablate the myelinating oligodendrocytes produced in adult life.