DIVISION OF NEUROSCIENCE
RESEARCH ESSAYS
Introduction
Researchers in the Division of Neuroscience are studying the fundamental processes of the brain, spinal cord and peripheral nervous system. The knowledge gained from such fundamental research is essential to our understanding of human thought, emotion and behaviour, and provides the basis for treatment of many devastating neurological and psychiatric disorders.
Individual research groups within the Division are investigating the neuronal mechanisms underlying vision, hearing, movement, cardiovascular function, emotion and memory.An understanding of the nervous system demands a wide-ranging knowledge, from the basic molecular events which occur within a single neuron, to the highly complex synaptic interactions between thousands, or tens of thousands, of neurons. Over the past year, the Division has strengthened its emphasis on the study of fundamental mechanisms at the level of single neurons, with major foci on synaptic transmission in different regions of the brain and spinal cord, and on sub-cellular molecular signalling pathways in neurons and their target cells.
The field of neuroscience has advanced rapidly over the past decade, accelerated by the development and use of new technologies. Of note is the increasing use in the Division of molecular, genetic and advanced imaging techniques. Collaborations with the JCSMR gene targetting facility are enabling researchers to develop transgenic mice with altered neuronal expression of particular molecules, or with selective neuronal expression of fluorescent labels for imaging studies in living neurons. The acquisition in the year 2000 of a two-photon and confocal imaging facility, funded by the Wellcome Trust, will greatly enhance the research capabilities of the Division.
Dr Bruce Walmsley, Head of Division
Recent studies have highlighted the importance of structural variation as a determinant of heterogeneity in vascular responses. The anatomical relationship between the smooth muscle cells in the walls of arteries and veins is critical to the coordination of vasoconstriction and vasodilation. Since stimuli operate at the outer smooth muscle or inner endothelial layer, there is a requirement for efficient coupling between the smooth muscle cells, between the endothelial cells and also between the smooth muscle cells and the endothelial cells. This is achieved through structures called gap junctions which are the sites of electrical and chemical coupling through their constituent molecules called connexins. While our physiological studies have shown that the smooth muscle cells of the vascular wall are electrically coupled, our comparative immunohistochemical studies have not detected any of the known vascular connexins in the media of small muscular arteries; the vessels most important to the development of blood pressure. Existing members of the connexin family can be detected in the muscle layers of the larger elastic arteries, which serve more as conduits for blood to flow to the various bodily organs. These results suggest the existence of a novel connexin within the vascular system. Cloning of cDNA from muscular and elastic arteries has, however, failed to provide evidence for such a new family member. Thus, the anatomical form of the gap junctions must vary between elastic and muscular arteries so that the ones in the muscular arteries are unable to be detected using conventional immunohistochemical techniques. For example, the junctions may be very small in the muscular arteries and hence below the limit of detection of the light microscope. Ultrastructural studies to test this hypothesis are in progress.
Both vasoconstrictor and vasodilatory mechanisms vary between different vascular beds. Variations in vasodilatory mechanisms lie in the degree of involvement of the endothelial factors, nitric oxide and endothelium-dependent hyperpolarizing factor (EDHF), the precise nature and mechanism of action of the latter remaining elusive to date. Recent studies in our laboratory have shown that the incidence of myoendothelial gap junctions (MEGJs), or the gap junctions which exist between the smooth muscle and endothelial cells of blood vessels, varies between different parts of the mesenteric vascular bed. More importantly, the incidence of MEGJs is positively correlated with the prevalence of EDHF over nitric oxide as vasodilator mechanisms. These results raise the possibility that vasodilation attributed to EDHF could result from the passage of small molecules through MEGJs as well as from direct electrical coupling through MEGJs. We are currently attempting to identify the nature of the connexins present at these MEGJs, as this will provide important information regarding the conductance properties of the channels formed between the muscle cells and the endothelial cells.
An important influence on vascular function are the nerves of the autonomic nervous system which lie at the outer edge of the smooth muscle layer of blood vessels. Our previous serial section electron microscopic studies have demonstrated that autonomic nerves come into close apposition with the membranes of arterial smooth muscle cells in arterioles of the iris. In concert with other similar studies, these results provide evidence for a common structure for neuroeffector associations in arteries, analogous to the neuromuscular junction in skeletal muscle. We have now completed a developmental study which has shown that these specialised sites form early during development before the maturation of nervous responses. These exciting results suggest that modification of physiological responses during development may result from close range interactions and the possible exchange of growth factors.
We are currently testing our hypothesis that structural variation may underlie many of the physiological variations between different vascular beds. Experiments are underway to produce mice which have significant changes in structural parameters. Physiological experiments will determine whether these changes have an effect on vascular function.
The human brain contains an immense number of neurones that are wired together in extremely complex ways. The activity of these neurones is responsible for many functions such as movement, hearing and vision. While we have some understanding of these functions, the creation of our emotions stands as one of the most amazing and puzzling aspects of brain function. We all know what emotions are. At some time we have all experienced love, hate or fear. While life without emotions is hard to imagine, where they originate often seems a mystery. Once a particular emotion is triggered controlling it is also difficult. Activities during which emotional reactions are strong are often our most enduring memories. Disorders of the storage or expression of emotional memories are thought to underlie such mental disorders as panic attacks, anxiety and post traumatic stress disorder. Thus understanding the brain mechanisms that give rise to and control emotions is one of the most important goals in trying to understand these disorders.
The amygdala is a small part of the brain that plays a key role in the control of emotional behaviour. It is particularly important in the storage of emotional memories. Much is understood about the anatomy of the amygdala and which parts of the brain communicate with it. However little is understood about the physiology of this structure. It is not known how information entering the amygdala is processed and what receptors and transmitters are involved. The Cellular Neurophysiology group is involved in studying the properties of cells and circuits within the amygdala. We expect that understanding the function of the amygdala will help us to understand some of the mechanisms underlying emotional behaviour. In addition, understanding the roles of different neurotransmitters and receptors present in this structure might help with the development of new and useful therapeutic strategies in the treatment of disorders affecting the amygdala.
Sensory and cortical inputs enter the amygdala via the lateral and basal nuclei. These inputs are processed by neurones in these nuclei and this information is then sent to the central nucleus which is one of the main output stations of the amygdala. To understand how the information entering the amygdala is processed, we need to know which types of cell are activated by these inputs, and how these inputs modulate the properties of these cells. Thus, in one current project we are recording from cells in the lateral amygdala in coronal brain slices maintained in vitro. We examine the firing properties of these cells and how these properties are affected when synaptic inputs are activated. Cells are also filled with a marker which following fixation of the slices allows us to reconstruct the morphology of the cells. These experiments are showing us that there are four types of neurones in the lateral amygdala, each of which has distinct firing properties. These four cell types are divided into interneurons and three types of pyramidal neurone. Another project in the lab involves examining the properties of the intrinsic currents that control the electrophysiological properties of these different neurones. We have found there these cells have quite different intrinsic currents which account for their different properties. These data suggest that inputs arising from different regions of the cortex may be activating different neurones within the lateral and basal amygdala. If this were the case then each type of inputs may be processed differently.
We have shown previously that cortical inputs that enter the lateral amygdala innervate both pyramidal cells and interneurons in this nucleus. These inputs use glutamate as the primary neurotransmitter. The synapses made by these inputs onto pyramidal cells activate both AMPA and NMDA receptors. Interestingly the synapses onto GABAergic interneurons activate only AMPA receptors. Furthermore these AMPA receptors have a different subunit composition than those present on pyramidal neurons. The outputs of interneurons form inhibitory GABAergic synapses onto pyramidal cells. These synapses play a key role in processing the information reaching the lateral amygdala. However, the morphology and properties of synapses between interneurons and pyramidal cells are not well understood. Interneurons form approximately 5-7% of the total cell population of the lateral amygdala and for this reason these cells are difficult to identify in acute brain slices. In order to identify these cells in brain slices another project aims to produce mice in which interneurons can be easily identified. Interneurons are known to express the calcium binding protein parvalbumin. We have begun to make a transgenic mouse which has a novel gene inserted into its genome. This gene is for the protein green flourescent protein (GFP) under the control of the parvalbumin promoter. Thus, GPF will be expressed by interneurons in the lateral amygdala. GPF is a protein that flouresces when illuminated by 488 nm light. Thus we expect that in these transgenic mice interneurons will be able to be identified by their green flourescence when they are illuminated. Once we have identified interneurones in the amygdala we will be able to examine the projections of interneurones and how they communicate with other neurones.
Nerve cells are dependent on information they receive from the cell they are to make contact with in order to survive and mature. As they are extremely long, sometimes over a meter in length from the cell body along the axon to the terminal, there is a major problem for signal transduction for molecules acting on receptors at the nerve terminal that need to convey the survival signal to the cell body. Survival of neurons during development of the nervous system is dependent upon both soluble and insoluble trophic factors signalling from the nerve terminal to the cell body. Understanding the processes that control the flow of information from the nerve terminal to the cell body is essential if we are to understand neuronal development and apply this understanding to nerve regeneration.
Soluble neurotrophic factors
Soluble neurotrophic factors such as NGF promote the survival of sympathetic and sensory neuronal populations by binding to receptors present on the nerve terminal, and the complex internalised and transported to the cell body. This transport occurs in vesicles where the neurotrophic factor is in the lumen of the vesicle and its receptor is in the wall. In this way the machinery required for the signal transduction is physically moved from the terminal to the cell body where it can then interact with the signal cascade leading to nuclear responses. We are studying the interaction of the neurotrophin receptors by examining the p75 low affinity neurotrophic factor receptor deficient animals and have shown that BDNF, NT3 and NT4 are completely dependent on the presence of this molecule for their retrograde transport while NGF is not. With the aid of video and confocal microscopy, we have studied cultures of dissociated sensory and sympathetic ganglia to which rhodamine-labelled NGF was added. We found rapid binding of label, apparently to surface membrane receptors, followed by a more gradual accumulation of labelled vesicles in the growth cone. Incubation of these cultures with unlabelled NGF one hour later led to a rapid loss of label in the growth cones. These results suggest that there is a pool of internalised neurotrophin, in the nerve terminal, which is in rapid equilibrium with the external environment. It is from this pool that a small fraction of the neurotrophin containing vesicles is targeted for retrograde transport. Our current research aims to examine this targeting to try to find a way to increase the availability of neurotrophins during regeneration.
Second messengers regulating retrograde axonal transport
Other molecules that are not transported but generate second messengers in the nerve terminal may influence neurotrophin endocytosis or transport. We have been investigating the mechanisms of control of the retrograde axonal transport of 125I-b-NGF, 125I-NT-3 , 125I-NT-4/5 125I-BDNF to determine the additional factors that may regulate this pathway of signal transduction and thus be able to act in synergy with the neurotrophins. These molecular mechanisms have been investigated by identifying potential signalling molecules using specific pharmacological inhibitors. Our studies suggest that retrograde axonal transport of these proteins relies on the dynein motor protein in vivo and that a wortmannin-sensitive isoform of PI3-kinase or PI4-kinase has an important function in the intracellular transport of neurotrophic factors. The results of this study will suggest alternate ways to enhance neuronal regeneration by perturbing the second messenger cascades promoting axonal transport.
Non-transported neurotrophic factors
Molecules that are not themselves retrogradely transported may generate stable second messenger molecules at the nerve terminal, which are stable enough to survive the long time for transport to the cell body. In an attempt to describe such messengers we examined the role of the GTP-binding proteins as potential stable second messengers and demonstrated the retrograde transport of Gi and Gz.
The Gza deficient mouse
In order to study the function of Gza in more detail, in collaboration with the Gene Targeting Group, we have made mice deficient in this protein. The animals had a prolonged bleeding time with no obvious platelet abnormality. The animals fail to thrive over the first 3 weeks when they are about 25% lighter than littermates postnatal but recover to reach control weights by 3 months of age. Behavioural tests on the Gza deficient mouse showed no abnormalities in the rotarod, maze, righting reflex and cortical placing reaction. The activity of choline acetyltransferase (CAT) and tryosine hydroxylase (TH) was examined in the Gza deficient mouse in the adrenal gland, superior cervical ganglia and the submaxillary gland. There was a decrease in the levels of TH in the adrenal gland of Gza deficient animals.
Morphine Tolerance
Heterotrimeric G-proteins couple to m and d opioid receptors. Two of these G-proteins, Gz and Gi, have been shown to be involved in m-mediated supraspinal analgesia. We therefore tested the Gza deficient animals for morphine induced spinal analgesia using the tail flick test and supraspinal analgesia using the hot plate test. Morphine treatment over five days led to the the Gza deficient mice becoming significantly more tolerant to morphine than the wild type. These results indicate that Gza deficient mice develop tolerance to morphine more rapidly and to a greater extent than control mice. In these mice receptor binding assays show a small decrease in the affinity of tolerant animals but the increased tolerance in the Gza deficient is not due to receptor down regulation as the maximum binding and receptor affinity do not differ between control and Gza deficient mice. Thus it appears that the alteration in tolerance observed in these experiments is due to a perturbation of G-protein coupled second messenger cascades. Thus Gz may normally play a role in the prevention of the development of opioid tolerance and the presence of Gza delays the development of morphine tolerance. Conversely, preliminary data suggests that during naloxone precipitated withdrawal the Gza deficient animals respond as if they were naive rather than more tolerant than the wild type. These results provide insights as to the mechanisms of morphine tolerance and dependence and may lead to new approaches to the control of morphine addiction.
Leader: Dr M. Crouch
The Molecular Signalling group has continued its work on several areas of cell biology, particularly relating to hormonal stimulation of cell proliferation and cell migration. Summaries of the results of this work are given below.
Cell Proliferation
Mitosis:
We have shown that stimulation of proliferation of Swiss 3T3 fibroblasts is dependent on the activity of a trimeric GTP-binding protein, Gi. Importantly, this dependence is found in response not only to hormones known to use G-protein-coupled receptors, such as thrombin, but also tyrosine kinase receptor agonists, such as epidermal growth factor. Gi has been found to translocate from the cell periphery to the nucleus during G2/M of the cell cycle, where it binds to chromatin. The nuclear Gi colocalizes with kinetochores. Inhibition of Gi function has no immediate effect on induction of S-phase, but retards mitosis, consistent with the timing of migration of Gi to the nucleus. This is associated with inhibition of cyclin-dependent kinase (cdk) activity of cyclin B (mitotic) but not cyclin E (S-phase). Our most recent work in this area has shown that Gi regulates a nuclear MAP kinase pathway that appears to be involved in cytokinesis. We have shown that MEKK1, MEK and ERK 1 and 2, and the p90 RSK enzymes are associated with the spindle and midbody of dividing cells and that MEK, ERKs and the RSKs are activated during cell division. Inhibition of Gi function has no effect on these enzymes after acute growth factor stimulation, but during mitosis their activation is largely reliant on Gi. Therefore, we propose that Gi plays a role in initiation / progression of mitosis in part via a stimulation of the MEK / MAPK / RSK enzyme cascade. S-phase:
We have also examined the events underlying stimulation of S-phase of Swiss 3T3 cells. Our data supports an essential role of phospholipase C in stimulation of DNA synthesis, and a less active role for the MAP kinase pathway. Thrombin is the most effective single stimulus of S-phase in our Swiss 3T3 cells, and the response to this hormone is dependent on both PKC and a tyrosine kinase activity. The latter is activated by Ca2+, which is also mobilized strongly by thrombin. EGF is a weak mitogen and a poor activator of Ca2+ release. Conversely, EGF is a potent activator of MAP kinase whereas thrombin is weak.
The tyrosine kinase activated by thrombin is not the EGF receptor, nor another receptor-like tyrosine kinase, as there was no binding of SH2-containing adaptor molecules to the phosphotyrosines. EGF, however, caused substantial tyrosine phosphorylation and binding of both shc and grb2 to phosphotyrosines. An EGF receptor kinase inhibitor also had little effect on the ability of thrombin to stimulate DNA synthesis, while totally inhibiting the response to EGF. Thus, tyrosine kinase receptor transactivation cannot account for the initiation of the cell cycle by thrombin. Such transactivation, which has recently been strongly advanced as a component of several G-protein-coupled receptor signalling systems, therefore, cannot be claimed as a universal mechanism.
Cell Migration
(a) p70 S6 kinase: Our work on this enzyme in relation to cell migration stemmed from the observation that immunohistochemical localization showed it to be associated with actin stress fibres. Inhibition of p70 S6 kinase activity with rapamycin disrupted stress fibre formation. Consistent with other reports, we have found that p70 S6 kinase is dependent on PI 3-kinase activity. Importantly, we have also found that a specific subset of the cellular PI 3-kinase colocalizes with p70 S6 kinase on actin stress fibres. This is a p110a PI 3-kinase that is present without association of the p85 subunit. Inhibition of either PI 3-kinase or p70 S6 kinase has no effect on the ability of hormones to induce S-phase (unlike T-lymphocytes) showing that the signalling pathways for stimulation of migration and cell cycle initiation are totally separable. Our most recent work on p70 S6 kinase has shown that a major cellular role is in its regulation of migration stimulated by hormones. This is likely to account for its cellular location and effects on the actin cytoskeleton. Migration in response to either thrombin or EGF is greatly attenuated with rapamycin treatment.
(b) Effects of Nitric Oxide: We have also examined the effect of nitric oxide donors on this pathway. This was initially studied, as NO has been proposed to be anti-mitogenic in some cells. However, we found no effect of NO donors on proliferation, but there was a marked enhancement by NO donors of p70 S6 kinase stimulated by either EGF, serum or thrombin. This was a highly selective potentiation, with no effect on hormone-induced Ca2+ release, DNA synthesis, cell number or MAP kinase activation. Associated with the enhanced activation of p70 S6 kinase was an increased cellular migration response.
(c) Flightless I protein: In collaboration with Drs. Hugh Campbell and Klaus Matthaei we have made considerable progress on characterizing the dynamics and potential cellular role of the flightless I (Fli-I) protein. Dr. Campbell cloned the mouse and human genes for this protein, which was first identified in Drosophila as a protein involved in formation of flight muscles. It was since shown to be essential for embryonic fly development, particularly during cellularization of the blastoderm. We have localized the protein to the invaginating membranes of the Drosophila embyro, fully consistent with its genetically-determined role. We have made anti-peptide antibodies to this protein, which is composed of a gelsolin-like domain and an LRR domain. We have shown this protein to be localized to the nucleus of quiescent mouse fibroblasts, and to migrate to the leading edge of motile cells. This process is regulated by both p70 S6 kinase and PI 3-kinase. We are presently making a GFP-Fli-I construct for transfection of fibroblasts to examine the real-time migration of the molecule.
Synapses provide the basis of communication between neurons in the brain. The properties of each synaptic connection are precisely matched to a particular brain function. Synaptic strength may be modified during critical developmental periods, and in the adult brain during processes such as learning. Synaptic modifications may involve both structural and molecular changes. At the structural level, the total number of synaptic contacts between individual neurons may be altered, and the structural features of individual synapses may be modified. At the molecular level, many protein-protein interactions control the mobilization, docking, priming and fusion of synaptic vesicles at neurotransmitter release sites in the presynaptic nerve terminal, and the postsynaptic response to vesicular transmitter release via receptors in the membrane of the target neuron. Research in our Group is aimed at understanding the fundamental structural and molecular mechanisms regulating the strength of synaptic transmission between central neurons.
Our work is focussed on synapses in the mammalian central auditory system, which offer a number of key experimental advantages for studying basic synaptic signalling mechanisms. Synaptic connections between auditory nerve fibers and neurons in the anteroventral cochlear nucleus of the brainstem are large and powerful. These excitatory synapses are optimized to transmit precise temporal auditory information, and generate some of the largest and fastest synaptic signals in the brain. Our previous studies of this connection have provided the most direct observations of the quantal nature of synaptic transmission in the mammalian brain, and of the changes which occur in quantal transmission during normal development. In addition to receiving excitatory inputs from the auditory nerve, neurons in the anteroventral cochlear nucleus also receive inhibitory synaptic inputs, which modify the transfer of auditory information through the cochlear nucleus. Our current experiments are designed to examine the mechanisms underlying both excitatory and inhibitory transmission in the cochlear nucleus, and the results serve as a model of central sensory processing:
Postsynaptic receptor clustering and synaptic strength
Our recent studies have provided the first direct experimental evidence of a link between the structure of a synapse in the brain and the strength of the synaptic signal generated at that synapse. This new evidence has provided a basis for understanding how synaptic strength may be altered at inhibitory synapses, and was achieved in collaboration with an expert immuno-histochemist in the USA, Dr F.J. Alvarez. Our current experiments are designed to examine the hypothesis that patterns of presynaptic nerve activity regulate synaptic structure at these synapses, as this will then establish a direct link between nerve activity and the strength of synaptic signals. Using a population of congenitally deaf mice, we are examining how a lack of auditory nerve activity during development interferes with the normal regulation of synaptic structure and function. One of the major conclusions form our recent work is that there is a very large site to site variability in the strength of synaptic signals generated at different inhibitory synapses in the cochlear nucleus. Our initial results support this hypothesis, and in further experiments we will make direct recordings from individual synapses using a combination of electrophysiological and optical imaging techniques.
The role of inhibitory neurotransmitters in central auditory processing
The two major inhibitory neurotransmitter chemicals in the brain are glycine and GABA. Although most central pathways use only one or the other of these neurotransmitters, previous experimental evidence has demonstrated surprisingly that most nerve terminals in the anteroventral cochlear nucleus contain both glycine and GABA. Our investigations have now revealed distinct roles for these neurotransmitters in the cochlear nucleus. Glycine has a major postsynaptic inhibitory role, acting directly through receptors in the postsynaptic neurons. GABA is also released by these nerve terminals, but acts on receptors in the membrane of presynaptic nerve terminals located over the surface of the same neurons. Further experiments will investigate how these different pre- and post-synaptic inhibitory mechanisms modify the transmission of auditory information through the anteroventral cochlear nucleus.
Molecular mechanisms underlying synaptic transmission
The arrival of a nerve impulse causes calcium ions to enter the presynaptic terminal and release quanta of neurotransmitter contained in membrane-bound vesicles from the synaptic terminal. The release of a quantum of neurotransmitter is not certain, but is a probabilistic event. There are many factors involved in the regulation of release probability, which represents a major determinant of synaptic strength. The large size of auditory nerve terminals in the brainstem allows the possibility of access to the nerve terminal via microelectrodes to directly study the presynaptic mechanisms regulating transmitter release. Such experiments are not possible at most central synapses because of their small size (in the order of several microns). In addition, the large postsynaptic signals allow the quantal nature of transmission to be studied in great detail.
At the large auditory synapse in the anteroventral cochlear nucleus, we have revealed a novel mechanism which dramatically inhibits neurotransmitter release on a rapid (millisecond) time scale. Although transmitter release is triggered by a rise in the intraterminal calcium level, our results indicate that there is also an inhibitory mechanism triggered by calcium, which rapidly suppresses release. Current experiments are investigating the molecular basis of this regulatory mechanism.
We have used a new experimental and analytical method, in collaboration with Dr. John Clements (BAMBI, ANU), to study the relationship between release probability and a form of short-term synaptic plasticity, called paired-pulse depression (or facilitation). These results have provided direct evidence for a precise relationship between release probability and the relative amount of transmitter release following a second, closely spaced nerve impulse. Further experiments have examined transmitter release during trains of nerve impulses, and the recovery of release following synaptic depression. These results are functionally important, as the auditory nerves naturally generate trains of nerve impulses, some at very high frequencies.
The activity of many of the molecules involved in the mobilization, docking, priming and fusion of synaptic vesicles is regulated at specific phosphorylation sites. Our experimental results indicate that phosphorylation of presynaptic proteins is an important regulatory mechanism at brainstem auditory synapses. Direct injection of molecules into the presynaptic terminal is being used to study the role of particular regulatory pathways underlying the modulation of transmitter release by phosphorylation events.
The ultrastructure of central synapses The interpretation of the results of physiological experiments on synaptic transmission is often made in terms of the structural features of a synaptic connection. For example, release probability and depression of transmitter release during trains of impulses may be related to the average number of available docked vesicles at individual synapses, and the total number of quanta released may be determined by the total number of synapses in a connection. Our physiological results at the large auditory synapses have indicated that release probability may be very different between connections, and our theoretical models indicate that this may be due to differences in the ratio of docked to undocked vesicles at auditory terminal release sites. Serial section electron-microscopy of auditory synapses is being used to investigate this issue. In a further study, the ultrastructural differences between auditory synapses in normal and congenitally deaf animals will provide an important basis for interpreting the results of our physiological experiments on the role of nerve activity during development in the auditory system.
The induction of long-term potentiation (LTP) at CA1 hippocampal slices in dependent on the conditioning stimulation increasing the cytoplasmic calcium concentration above a threshold. One potential source of calcium is cytoplasmic calcium stores, of which there are two distinct types. One releases calcium in response to raised cytoplasmic calcium, a process known as calcium induced calcium release, or CICR. The other releases calcium in the presence of IP3 and raised cytoplasmic calcium. The contribution of calcium released from these stores to the induction of LTP has remained controversial. We have titrated the amount of LTP by using theta burst stimulation and varying the number of bursts. We have blocked release of calcium by use of specific pharmacological blockers for each type of release process. In this way we have shown that each type of calcium release process contributes to the induction of LTP up to a certain magnitude of potentiation. When the conditioning stimulation is increased to induce further LTP, blockade of either type of calcium store makes no difference to the amount of LTP induction. It appears that calcium released from these stores is an effective contributor to total cytoplasmic calcium up to a certain level of synaptic activation, beyond which other sources of calcium are presumed to dominate the induction process.
Transmission at synapses on lumbar motoneurones in vitro has been investigated by recording spontaneous miniature excitatory post-synaptic potentials (mEPSPs). All impulse activity and all inhibitory transmission was blocked. The response of each motoneurone to a brief intracellular current pulse was recorded. The motoneurones were labelled with biocytin and their morphology reconstructed. The morphology and the current pulse response will be used to calculate the passive transmission properties of the dendrites. These data, together with the time course of the mEPSP, will be used to calculate a synaptic site of origin for the mEPSP. The objective is to investigate the relationship between the amplitudes of mEPSPs and the distance from the soma at which they are generated. If the dendrites are passive, and synaptic conductances are similar at all dendritic locations, amplitudes should decrease with greater distances from the soma.
|
| Vision is often called the primary sense, justifiably so when one considers that a large fraction of our brain is devoted to processing visual input and generating our internal representation of the external world. Primates, human and non-human, are intensely visual creatures, and the extent to which this sensory system pervades our thought processes is reflected in the preponderance of visual metaphors in our language. |
Visual processing begins in the retina, embodied in three neural layers lining the back of the eyeball. The first layer, the photoreceptors, is connected to the intervening layer, the bipolar cells, via a layer of chemical connections or synapses. At a chemical synapse, a voltage change in the first neuron causes it to secrete a chemical neurotransmitter, which binds to the second neuron and generates a corresponding voltage change. One area of our research focuses on the molecular mechanisms by which the light signals in the photoreceptors pass through this synaptic layer to the bipolar cells. The bipolar cells in turn connect to the ganglion cells which send their axons to the brain via the optic nerve. Unlike the photoreceptors or bipolar cells, the ganglion cells come in a large variety of different types, some displaying quite complex response properties. For example, one class of ganglion cells respond only when the image moves in a particular direction. A second line of research attempts to understand how these ganglion cells are able to compute the direction of image motion.
The connections between photoreceptors and bipolar cells are of particular interest because the mechanism of signal transmission is significantly different from that used in the rest of the brain. Photoreceptors release the excitatory neurotransmitter, glutamate, in a continuous stream. Glutamate release is only interrupted to signal detection of light. Although excitatory synapses in the rest of the brain also release glutamate they can only do so in very brief pulses lasting one thousandth of a second. Our aim is to elucidate the specific molecular differences between tonically and transiently releasing synapses which underlie the functional differences observed. This work will provide insights into the molecular mechanisms of neurotransmitter release at all synapses. The biochemical and molecular biological aspects of this work are being undertaken in Dr Catherine Morgans' laboratory as part of a collaboration. In this lab laboratory we are concentrating our efforts on performing the electrophysiological measurements. Amy Berntson, a PhD student, has succeeded in characterising, for the first time, the light-evoked synaptic responses from bipolar cells of mouse retina. This ground-breaking work establishes the basis for further examination of the mechanisms of transmitter release from the photoreceptors.
A collaboration, with Dr Robert Duvoisin of the Dyson Vision Research Institute at Cornell Medical School in New York, involves the application of transgenic techniques to study photoreceptor-to-bipolar cell synaptic transmission. Dr Duvoisin has produced a transgenic mouse deficient in a glutamate autoreceptor. This receptor is normally present in the synaptic terminals of photoreceptors, and in this location enables the glutamate released from the photoreceptor to feed back and modulate its own release. Such feedback control is common throughout biological systems, but its role in this case is enigmatic. A physiological analysis of the receptor-deficient transgenic mouse is helping to clarify this question. Dr Duvoisin was a Visiting Fellow in the School for two months early in the year during which time we successfully performed preliminary experiments.
Motion detection is fundamental for any visual system and in many vertebrate retinas there are specific classes of ganglion cells which are tuned to respond to stimuli moving in particular directions. These direction selective ganglion cells will respond, for example, to a bar of light moving in a preferred direction, and remain quiescent for the same bar moving in the opposite direction. The main thrust of this work has been to find the locus of the underlying synaptic interaction. Is direction tuning due to synaptic integration within the dendrites of the ganglion cells, or is it a property of the preceding neural circuitry which is simply registered and reported by the ganglion cells? This year we have made a significant breakthrough in being able to demonstrate that the direction selectivity is due to postsynaptic interactions. The next step will be to identify the mysterious inhibitory interneuron (amacrine cell) which appears to be a key element in generating direction selective responses. Three other researchers have collaborated on this project: Professor Bill Levick, a former faculty member of the School but now in the Psychology Department at ANU and Dr Shigang He and Dr David Vaney at the Vision, Touch and Hearing Research Centre in the University of Queensland.
Of the many kinds of illness that afflict humankind, few incur greater social and economic cost than disorders of the nervous system. Paralysis, epilepsy, depression and schizophrenia are all diseases which result from malfunctions of the brain or spinal cord. Not only do these diseases often rob the sufferer of those qualities that define his or her personality, they also in many cases persist for decades, placing enormous and continuing demands on society.
Given this situation, it is not surprising that neuroscience - the study of the nervous system in both health and disease - has undergone rapid growth during the closing years of the twentieth century. During this time a great deal of basic neuroscientific knowledge has been gathered. Tentative advances have also been made in the treatment of a number of neurological disorders. However, there is still a very long way to go. The nervous system is exceedingly complicated, and neuroscience is still a young field. Much fundamental research remains to be done before we can turn the tables on brain disease.
The Synaptic Biophysics Group is engaged in fundamental research of this kind, research which will hopefully underpin medical advances of the future. During 1999 the Group conducted research on two important areas of brain function. One project studied the behaviour of nerve cells (neurons) in a part of the brain called the nucleus accumbens, known to be involved in drug addiction. Another project examined the properties of neurons in the neocortex (the wrinkled grey matter visible on the surface of the brain) which mediates high-level functions like sensation and language.
The first project addressed basic questions about how mammals become addicted to drugs of abuse, such as heroin. Heroin contains compounds called opioids which bind (react chemically) with molecules called opioid receptors that are found in the membranes of neurons. After this binding occurs, something changes in the neuron which eventually gives rise to the sensations of drug intoxication. Neurons in the nucleus accumbens are particularly involved in this process. The nucleus accumbens forms the core of a neural system that is responsible for behavioural manifestations of opioid addiction, such as reward, sensitisation, craving and withdrawal. We studied single neurons isolated from the nucleus accumbens of rat brains, measuring some of the consequences of opioid binding for the electrical properties of these cells.
This work yielded information about how opioids modify the behaviour of neurons, and also about how opioid receptors respond to drugs that may have clinical applications in alleviating withdrawal symptoms. The next step is to follow these phenomena to the next highest level of complexity, by studying networks of neurons in brain slices containing the nucleus accumbens.
The second project studied neurons in the neocortex, which is a brain region that is the repository of many higher-level functions and which is unusually elaborated in humans. The fundamental question being addressed in this project was, how do neurons in the neocortex summate the signals they receive from other neurons? A typical neocortical neuron receives signals from thousands of other neurons, and somehow all of this information must be combined and processed before being passed on to the next neuron in the chain. This processing ability of neurons is obviously fundamental to the way the brain works.
By making electrical recordings from single neurons in rat neocortex, we were able to map out the distributions of different kinds of membrane proteins involved in sculpting the electrical signals that flow through these cells. We also developed mathematical equations to describe the behaviour of the proteins. The next step is to incorporate the equations into a complete computer model of the electrical behaviour of this class of cell. Ultimately the model neuron can be embedded in synthetic networks of many similar model neurons in order to study how the brain integrates and processes information.
The brain is made up of billions of neurons connected to each other via synapses to form specific neuronal networks. The main objective of my group is to understand how neurons within these networks convert inputs into an output signal. Most neurons in the mammalian central nervous system receive thousands of synaptic inputs largely onto their dendritic tree (the major receiving part of a neuron). These inputs summate with each other and if a threshold voltage is reached an electrical impulse, or action potential, is initiated. Neuroscientists call this process "synaptic integration". It is the most fundamental computation a neuron performs.
In my laboratory we use an acute "brain slice" preparation to study synaptic integration in single neurons. To do this a slice of living brain tissue is placed under a high-power light microscope, allowing individual neurons and their processes (dendrites and axon) to be visualized. Electrical recordings using the patch-clamp technique are then made from neurons under visual control.
There are two main types of synaptic inputs to neurons: excitatory and inhibitory. Excitatory inputs increase the likelihood of action potential initiation, whereas inhibitory inputs make the initiation of action potentials less likely. One project in my laboratory has investigated how inhibitory synaptic inputs in the cortex are effected by a set of proteins called voltage-activated sodium channels. When activated these channels are the main proteins responsible of the generation and propagation of action potentials. As a consequence these channels are classically thought to play an excitatory role in synaptic integration. To our surprise we found that these channels can also amplify inhibitory inputs, increasing their ability to block action potential firing.
Another project has investigated the mechanisms responsible for the generation of bursts of action potentials in cortical neurons. Some neurons respond to synaptic inputs by the generation of a single action potential, other neurons generate bursts of action potentials at high frequency (> 100 Hz). Action potential bursts are thought to improve information transfer from one neuron to the next, and in the cortex may be important for the generation of synchronized activity during both physiological (e.g. sleep) and pathological states (e.g. epilepsy). We have found that action potential bursts are generated due to activation of dendritic voltage-activated calcium channels. Like voltage-activated sodium channels, voltage-activated calcium channels are proteins which, among other things, increase neuronal excitability. Previous studies have shown that action potentials actively propagate into the dendrites of cortical neurons. We now show that this action potential "backpropagation" leads to activation of dendritic calcium channels. This increase in dendritic excitability feeds back to the soma and axon leading to a burst of action potentials at high frequency.
A third project has investigated action potential initiation and backpropagation in thalamocortical neurons. These are the main neurons involved in the transfer of sensory information to the cortex. Action potentials evoked by sensory or cortical inputs were initiated near the soma of these neurons and backpropagated into the dendrites. The dendrites of thalamocortical neurons were found to contain a non-uniform distribution of voltage-activated sodium but roughly uniform density of voltage-activated potassium channels. Dendritic action potential backpropagation was found to be reliant on the activation of dendritic voltage-activated sodium channels, but was compromised by dendritic branching, and so can fail to invade the distal dendrites. Calcium channels were found to be concentrated at proximal dendritic locations, suggesting that they play a key role in the amplification of sensory inputs to these neurons.
Finally, in collaboration with Michael Häusser (University College London), we are investigating the interaction of action potentials and synaptic inputs. We have found that action potentials which occur simultaneously with synaptic inputs can substantially reduced them. This effect is greater for inputs with rapid kinetics and located near the cell body, and is caused primarily by the action potential itself. These findings indicates that action potentials can reset synaptic integration in a spatially and temporally precise manner.
In conclusion, this work extends our knowledge of the complex processes that occur in neurons during synaptic integration. The picture that is evolving is that individual neurons do not act as simple integrators. Rather, they perform complex computations which presumably play an important role in brain function. Defining precisely what role this is will be one of the big challenges of the next century.
Research in the Synaptic Biochemistry Group is aimed at understanding how the signaling properties of neurons are influenced by the biochemical and structural design of the synapse. Neurons transmit signals from one to another at synapses, specialized contacts where a chemical transmitter is secreted by the presynaptic neuron and detected by the post-synaptic neuron. At conventional synapses in the brain, transmitter release is triggered by brief (~1ms) voltage signals called action potentials. The release of transmitter is correspondingly short, and the neuron signals in an "all or none" fashion. In contrast, some sensory neurons form specialized synapses, called ribbon synapses, at which neurotransmitter is released continuously (or tonically). In mammals, these neurons are photoreceptors and bipolar cells in the retina, and saccular and vestibular hair cells in the inner ear. These neurons send information to the post-synaptic cell by modulating the rate of tonic release in response to small graded changes in the membrane potential. Photoreceptors, for example, are depolarized in darkness and secrete the neurotransmitter, glutamate, at a high rate. Detection of a photon leads to a tiny hyperpolarization of the plasma membrane and a brief reduction in the rate of glutamate release.
Neurotransmitter release occurs by the calcium dependent fusion of transmitter-filled synaptic vesicles with the plasma membrane. Depolarization of the nerve terminal opens voltage gated calcium channels allowing calcium ions to flow down their electrochemical gradient into the terminal. Fusion is initiated upon the detection of calcium by a synaptic vesicle-associated calcium sensor. Following fusion, the synaptic vesicle membrane is retrieved by endocytosis and a transmitter-loaded synaptic vesicle is regenerated within the terminal. The basic molecular machinery governing neurotransmitter release and the synaptic vesicle cycle is conserved between ribbon synapses and conventional synapses.
We have found the retina to be an excellent experimental system for understanding the combination of molecular and structural differences that determines whether a synapse releases neurotransmitter tonically or phasically. It is a highly organised, laminar structure with discrete synaptic and nuclear layers. The outermost synaptic layer is composed entirely of photoreceptor ribbon synapses, and the inner synaptic layer contains both ribbon and conventional synapses. It is easy to isolate as an intact neural system and it retains its response properties to its natural stimulus, light. Our group has identified three major differences between ribbon and conventional synapses in the retina:
1. The presynaptic calcium channels at ribbon synapses are different from those at conventional synapses. The electrophysiological properties of the presynaptic calcium channels, their interaction with the proteins catalyzing synaptic vesicle fusion and their distribution within the nerve terminal influence the time course and probability of transmitter release. One aim of the group is to characterize the presynaptic calcium channels of photoreceptors and bipolar cells in the retina
2. The presynaptic plasma membrane protein, syntaxin 1, is a critical component of the protein complex that catalyzes the fusion reaction between the synaptic vesicle and plasma membrane at conventional synapses. We discovered that ribbon synapses express a different member of the syntaxin gene family, syntaxin 3, rather than syntaxin 1. A second aim of the group is to compare the effects of syntaxins 1 and 3 on neurotransmitter release
3. Electron micrographs of ribbon synapses in the retina reveal an electron dense bar, the synaptic ribbon, anchored over the active zone and extending into the cytoplasm by about 0.5 micrometers. The synaptic ribbons are surrounded by a halo of synaptic vesicles that can be seen to be attached to the ribbons by short filaments. The appearance of the ribbons is highly suggestive of a role in the delivery of synaptic vesicles to the active zone, although little is known about them biochemically. A third aim is to identify the constituent proteins of synaptic ribbons and to reconstitute the interaction of the ribbons and synaptic vesicles in vitro.
STAFF - DIVISION OF NEUROSCIENCE
Professor and Head: (until March)