Erwin Neher (2014) - Short-term Synaptic Plasticity

Thank you, Dr. Blatt, for the introduction. It’s a great privilege to speak in front of about 600 hundred of the very best young researchers of the world which I hope a fraction of it at least is interested in a very interesting topic, namely understanding how our brain functions. So Roger Tsien a few minutes ago told you about some interesting ideas and interesting facts about the very long term forms of plasticity which indeed are believed or are apt to store information in the brain to mediate learning and memory. Let me go to the other side, to the very short term forms of plasticity. But before doing so, let me show you in a few slides a little bit about how our understanding of what happens in the brain developed over the last 200 years. And this of course starts with the famous experiments of the Italian scientists Walter and Galvani, Galvani who taught us or showed in spectacular experiments that the frog muscle can be made to twitch when the nerve is stimulated by a shock of electricity. This demonstrated that indeed there was something like electricity, electrical signalling in our body. that our brain is made up of this filigrane network of neurons. Roger Tsien already referred to this. And today we know that our brain is a network of about 10^12 of such neurons which are connected with each other via synapses. And on average a given neuron receives about 1000 or 10,000 inputs from other neurons. So what is it, this strange cell, a neuron? This strange structure? It is nothing else but a quite general cell which however has some special structures. So, even a normal cell has some kind of processes called microvilli. In the neuron such processes are exaggerated in the sense that a neuron possesses two kinds of protrusions some are called dendrites which are the receiving organs of the neuron. One special protrusion is the axon, a tube-like structure which can be very long and which transmits the nerve impulse to cells to which this given neuron is connected to. So each neuron receives inputs from thousands of other neurons primarily on its dendrites and these inputs delivered by proceeding cells can be both excitatory and inhibitory. A given neuron integrates or adds up these signals and whenever the electrical potential inside the cell soma, as a consequence of all these influences converging onto a given cell, whenever the potential inside the cell surpasses a certain threshold an action potential or a nerve impulse is generated at the so-called axon hillock. And this nerve action potential then travels down the axon to the nerve endings exciting or inhibiting other neurons. So there’s a nice movie or animation provided by Dr. Hasan from the Science Bridge in Heidelberg who illustrates, gives an impression on how this firework of axon potentials may happen in parts of our brain. You can see that an input fires an action potential which then spreads, excites other ones and this is going on. The movie is not correct in every detail but I think that doesn’t matter for getting the general idea. So the point of interest of course when you want to study the connectivity between neurons is the synapse. What appears to be here is this little bouton where a preceding neuron signals... or a sending neuron signals onto a receiving neuron. And knowledge about what happens at the synapse started actually not at the synapse between nerve cells but it started at the so-called neuromuscular junction, a synapse which makes a connection between a neuron and the underlying muscle cell. So when the nervous system wants a certain muscle to contract it sends an action potential through the motor nerve to the corresponding muscle cells. There the nerve ending splits up into several endings and what happens there is exactly the same thing which happens between two neurons, namely that the presynaptic ending liberates a transmitter which excites the underlying neuron. So our knowledge about these processes came to a large extent from experiments by Sir Bernard Katz Castillo and Del Castillo studying the synapse. And I show here one of their early registrations of the signal which can be recorded in a muscle fibre You can see two phases, first a so-called excitatory postsynaptic potential which represents the effect of the transmitter being liberated by the nerve and then surpassing the threshold just like in a nerve cell an action potential is being elicited in the muscle cell which then initiates the contraction. What Katz and Del Castillo did is they looked in detail at what happens here before or just during resting periods. And they recognised when they turned up the gain of their amplifier that there were all these little spontaneously occurring little blips. You know, very, very small signals. Now ordinarily many researchers dealing with sensitive amplifiers at high gain would dismiss such signals because there are multiple sources of interference, just influences by some instruments in the vicinity. Katz and Del Castillo did not do so. They actually started to study these small signals because they realised that they saw these little blips only when they recorded at sites close to the neuromuscular junction to where the nerve is. When they recorded a few millimetres away on the same muscle they wouldn't see these characteristic blips. Also some distance away from the muscle the wave form of the global signal would be different. This initial postsynaptic potential would be missing. What they saw is just the electrical excitability spreading in the muscle fibre. Also what they realised is that this amplitude of this excitatory postsynaptic signal is very much dependent on the calcium concentration in the medium. So when they reduced the calcium concentration, this initial signal became smaller so that eventually the action potential failed. What was left was a small signal. And when they reduced this to the extreme, they saw that the stimulating nerve impulse sometimes elicited a signal like this. And sometimes it failed to elicit any signal. So this kind of finding together with electron microscopy data from the De Robertis Laboratory in Argentina which showed that there were in the synaptic terminal some small structures, some vesicles now known as synaptic vesicles lead to this idea that indeed what happens at the nerve, at the synapse is the following. The nerve impulse causes influx of calcium into the nerve terminal. This is why the process is so much dependent on the extracellular calcium concentration. The calcium causes the fusion of such vesicles which then release their contents in neurotransmitter. Now, of course an action potential with calcium influx leads to the simultaneous release of many of such vesicles. The little blips which they saw in the recording at high resolution was interpreted to be spontaneous fusion of such vesicles with the plasma membrane. And then postsynaptically the neurotransmitter diffuses to the postsynapse and opens ion channels in the postsynaptic membrane. So we have a kind of transformation of an electrical signal into a chemical signal released from a presynaptic terminal and back translated into an electrical signal in the postsynaptic membrane. Now, synaptic plasticity. The term plasticity describes the observation that synaptic strength which is the signal being produced in the postsynaptic neuron by a presynaptic action potential is not a fixed quantity like in the electronic computer but changes constantly depending on the use of the synapse. Neuroscientists are convinced that this plasticity is a very important aspect of neuronal signal processing in the nervous system. And that indeed particularly the long term changes, the long term changes in connectivity between neurons underlies learning and memory. And this as I understand was Roger Tsien’s topic. Now, short term plasticity on the other hand I think is not less important because it mediates basic information processing tasks like adaptation, filtering and others. I will mention a few of these tasks in one of my following slides. What's also interesting is that different types of synapses have their own personality with respect to this short term plasticity. Some synapses, when you stimulate them repetitively like those of the climbing fibres, show depression in the sense that after some rest period a first response is very large followed by smaller responses. Other synapses show exactly the opposite, namely at first smaller response followed by subsequently increasing responses. Still other synapses have more complex types of behaviour. So given synapse on average displays a certain type of plasticity. Now here is just an example of what short term depression can be good for in a network in which many of such neurons are put together to subserve some task. I mentioned already a depression is good for providing adaptation to sensory stimuli. It is good for so called gain control, regulating the amine activity in certain brain areas. You can build networks with synapses displaying depression which provide, which generate rhythms. You can do temporal filtering but also more complex tasks of the central nervous system like sound localisation can be implemented in neural networks which use a short term depression. And even the so-called orientation tuning, an individual system which is seen in the primary visual cortex in the sense, that there are certain cells which primarily respond to contrasts in a certain orientation. One can simulate this and obtain insensitivity towards, invariance towards differences in absolute contrast by implying in the network short term depression. Okay, so another aspect which is being discussed now is a manifestation of short term plasticity, is the switching between different brain states. We know that our brain can switch within seconds between different states such as quiet state, stress, arousal, and focusing, attention. There are different sleep states which can be observed with EEG, where within seconds again the pattern of the EEG can change between the well-known REM sleep phase and other phases. Such fast switching of course cannot be due to long term plastic changes because they need time to develop, they need time to be trained and so on. So it is known that such brain states are controlled by a number of diffusely projecting transmitter systems such as dopaminergic, adrenergic, cholinergic, peptidergic and so on. Basically all the kinds of input mechanisms which were discussed by Dr. Kobilka yesterday in his lecture. And it is also known that such diffusely projected transmitter systems short term change, short term plasticity in their target areas. So my conclusion is that studying such short term plasticity, its dynamic nature and its mechanism may reveal some very important aspects of brain function, which may not be less important than the long term changes. Now, although these short forms of plasticity have been described already in the very early work on the neuromuscular junction by Katz and Miledi, we still... I think there are still many things which we don’t know about them and some of the essential features that we don’t probably understand yet. Now, what can happen when a synapse changes its strength, its short term plasticity? These are of course many things because it’s a multi-step process. There may be changes in the action potential wave form and the associated calcium influx. There may be postsynaptic changes, there may be inhibitory so-called auto receptors which feed back so that the transmitter feeds back onto the willingness of the remaining vesicles to release. There is a problem of recycling of vesicles and consumption of vesicles. In general, I think these are processes which molecularly mainly involve changes in second messenger levels, changes in phosphorylation and possibly also cytoskeletal reorganisations due to all the trafficking. Now in the rest of my talk I want to concentrate on basically two aspects, namely the depression mediated by the depletion of vesicles and the changes in the... molecular changes in the release apparatus which modulate the willingness of release-ready vesicles to actually fuse in response to a calcium stimulus. Now, we have been studying this process for a number of years in a very special synapse, the so-called Calyx of Held synapse. This is a synapse in the auditory pathway. What you see here is a slice of the brain stem with the auditory pathway laid out. The auditory fibres make first synapse in the ventral cochlear nucleus. The axon of the receiving neuron transverses the midline and makes a contact here with another cell in the so-called medium MNTB, the medium nucleus of trapezoid body. And both of these synapses here which have this very special shape of a calix or cup-like presynaptic terminal surrounding a compact postsynaptic cell body. And the special shape of the synapses is probably due to the fact that for the sound localisation for directional hearing the signals are coming from the two ears have to be processed very precisely so that at the point where the information from one ear meets the information from the other ear in the lateral superior olive the very fine differences in intensity and timing can be detected. So we have here a synapse which is very special in its geometry but which offers for electrophysiology the advantage that both the pre- and the postsynaptic compartments can be voltage clamped, can be a subject of very precise electrophysiological measurements simultaneously. And this was indeed found in Bert Sakmann’s lab by Gerard Borst and Bert Sakmann now almost twenty years ago. So we studied this synapse and studied its plasticity. Just to point out what we can do since we can voltage clamp the pre- and postsynaptic compartments and record corresponding currents. We can control the ionic milieu inside both compartments for precise analysis of the current flow. We can load the terminal with fluorescent indicator. Theis and me have been doing so over fifteen years, mainly using Roger Tsien’s fluorescent dyes. Well, we can do many things with this synapse which cannot be done in other synapses because usually synapses are very small bouton-like structures which are just too small to be approached and targeted by electrophysiological electrodes. Now, most of the calyxes of this kind show prominent synaptic depression. This is a response to a 100 Hz stimulus after the synapse was given some rest. We see a very large first excitatory current now, an invert current which is induced by the neurotransmitter glutamate which is released from the presynaptic terminal. A second stimulus 10 ms later gives a smaller response and finally after four or five stimuli the response is quite small. Now the synapse works at very high spike rates. The auditory nerve is highly active with between 10 and 50 action potentials per second even in the completely silent state. So this synapse, at least the one shown here, would work in a permanent state of partial depression. So that’s the physiological state of the synapse. But when we look in detail... This is the average response. When we look in detail at the pattern of short term plasticity, we see that some of the cells here just one particular cell displayed does show this very strong depression. This is shown here for three frequencies: 200 Hz, 100 Hz and 50 Hz. But in the same preparation next to that cell which shows this behaviour maybe we may find another cell which shows exactly the opposite, namely an initial facilitation. The second response is bigger than the first one followed by depression. Now, this is understood in the sense... Or I should point out that compared to this cell that cell has a very much smaller first EPSC namely of only about 1.1 nanoamp, whereas this had 2.5 nanoamp. So the understanding of this pattern is that in this cell the release probability was so high that during the first response already a good fraction of the available vesicles got released so that subsequent responses were smaller. In this cell the release probability during the first stimulus was smaller so that the record reveals that indeed for the second and third stimulus the excitation power of the nerve is stronger that the release probability for the remaining vesicles increases. So this biphasic behaviour comes about by first an increase in the release probability which is seen only when the initial release probability is very small. But then eventually there’s more and more stimuli. The synapse will use up all its vesicles and will fall into depression. Okay, so this consideration already points out that the so-called paired-pulse ratio, the ratio of the second to the first stimulus is a kind of measure which is used by many physiologists to first of all characterise the short term plasticity behaviour of the synapse and then also this measure is used as a kind of indicator of how large initial release probability is. Now when I record from 31 synapses like this and plot this so-called paired-pulse ratio against the amplitude of the first EPSC, I see there is a systematic relationship and that the larger the first EPSCs, the smaller the paired-pulse ratio. This is in line with many observations that were made at other synapses. Okay, but still since we are dealing here with synapses which are supposed to do the same job, which are morphologically all the same, which no obvious differences, one asks the question Why does this have a smaller release probability and a large PPR and this one is the opposite? So the hypothesis is that vesicles which are release-ready can indeed exist in two states. In a state of low release probability, typically the numbers would be that about 5 - 10% of the vesicles would be released by a single action potential. And the so-called super prime state with high release probability. Now, this term “super-prime” has been used in a number of contexts and a number of molecular mechanisms have been proposed. Why a super-prime vesicle has higher release probability than the other one, I will come back to that and try to explain that it’s probably the energy barrier of fusion which is being changed by biochemical modulatory influences. Okay, that’s written here in this last sentence, namely that it’s background modulator influences probably by diffusely projecting transmitting systems which cause diversity. Now, what kind of signalling mechanism might be underlying this? The most prominent one which probably is connected to this is the diacylglycerol mediated signalling pathways, one variant of the g-protein mediated pathways discussed yesterday by Dr. Kobilka. Where a receptor interacts with a g-protein, the g-protein activates phospholipase C splitting inositol phospholipids generating the second messenger diacylglycerol and IP3 which releases calcium. Now, it has been shown before that this diacylglycerol branch of the pathway has a special role in priming of synaptic vesicles and also in the release of synaptic vesicles. Namely there is a protein called MON-13 which has a so-called C1 domain. A domain which responds to a diacylglycerol and previous work has shown that this MON-131 is involved in the priming process and probably also then in the regulation of the energy barrier for release. So, we cannot of course change the brain state by stimulating... by diffusely projecting neurons in the brain slice preparation. I should mention this, all these experiments are done in rat brain slices with the conventional electrophysiological techniques. But we can mimic so to speak the action of this diacylglycerol mediated signalling by applying a phorbol ester, a mimic of diacylglycerol. And what you observe is when you apply this substance all the EPSCs are on average 2-5 times larger than under control conditions. And if we now plot the paired-pulse ratio again against the first EPSCs, we see that these new data points which are the filled symbols nicely follow the trend from the control data point and this more or less seamless continuation also overlapping between some of the points under phorbol ester and some of the points under control conditions. Seems to tell me or provokes the hypothesis that phorbol ester shifts this equilibrium or this dynamic equilibrium between primed and super-primed vesicles to the super-primed site. Now when you look at this property further, you can see that super-priming is a property of rested synapses. At 200 Hz stimulation EPSCs converge to the same steady state value. This is shown here in these recordings. Here is the first response of a relatively large cell or a cell with large EPSC under the influence of phorbol ester. This value is from the same cell without phorbol ester. This is a cell with a relatively small amplitude but under phorbol ester and this is the corresponding first value for the same cell without phorbol ester. Now you see the first responses are very, very different. Nevertheless the second, third, fourth, fifth responses converge to the same level. And the same level is not 0, it is just very similar. So in that sense it seems that super-priming or this differentiation between synapses only occurs... it needs time. It’s only present when the time, when the synapses had time to develop its super-priming. Interpretation of this finding is that as I just mentioned if the synapse has time to develop a super-prime state, then there is this difference. If however during such a drain there is a kind of constant flow of vesicles being released and newly recruited... I mean this steady state is a kind of steady state between vesicle consumption and vesicle delivery. So the hypothesis is that, well, if the cell is not allowed to go into super-prime state, then all these cells behaves the same. Of course this can be readily put into a simple kinetic model, where you assume that in order for vesicles to be released, they have to first undergo some molecular, some step which puts them docked at the membrane and into a state which is ready to be released. And if there is enough time then they can even mature further into super-prime state. So the first priming state will be fast. The second state would be slow. Now, this explains this behaviour. It also explains quantitatively what you observe when you plot the steady state value of release versus stimulation frequency. Here you see two curves, one under control conditions, one under a phorbol ester. These are normalised to the respective first value but you can see that the broken curves which are predictions of this simple model very nicely follow the experimental points. This is also the case for the low end of the relationship connecting steady state release and frequency. And of course from this model you then can extract the relevant and most interesting parameters, namely the data are compatible with the idea that the normally primed state has a release probability of here in this case 0.06, a relatively small release probability, whereas the super-prime state has a probability of 0.42. The so-called priming rate is fast, 4 per second, which means that in about 250 mms a new vesicle is loaded onto a release site whereas the rate for this second super-priming step is very low in the control case and enhanced by the action of the phorbol ester. Okay, now the findings which I described are not singular to the Calyx of Held. The various aspects of the findings are shared by a variety of other synapses. So the heterogeneity between synapses of the same type has been described also for the mouse neuromuscular junction. The prototype synapse, the hippocampus between CA3 and CA1 neurons, has been shown a few years ago to show similar, very similar kind of heterogeneity by a series of papers by Hanse and Gustafsson and some more recent papers. Also this phenomenon of what I call here the collapse of heterogeneity at higher frequency has been described by Hanse and Gustafsson and discussed by Walters and Smith in the context of information processing in neuronal networks. And this slow conversion from prime to super-prime state has indeed also been an element in a model proposed by Stefan Hallermann for cerebellar mossy fibre synapses. So, although I think the action potential wave form and calcium influx is the strongest modulatory influence on the short term plasticity because calcium currents are known to be able to show what's called facilitation, also inactivation, they are regulated themselves by G-protein pathways and so on. And when there are changes in calcium influx there are very large changes as a consequence in the release due to the very steep relationship between calcium influx and release. But I think the experiments and the Calyx of Held make a strong point for this other form of modulation of short term plasticity which is not due to calcium influx. We can tell this is a Calyx of Held because the influence of phorbol ester on calcium influx has been very carefully studied by Ralf Schneggenburger’s laboratory a few years ago. Although these previous papers have shown that there are no major changes in pool size. So the phenomenon I described really seems to be a relatively novel form of change of modulation in the energy barrier for releasing a vesicle. Well, the similarity between modulation by phorbol ester and the heterogeneity at the synapses at rest suggests that the heterogeneity actually comes from different degrees of background modulation. And I think that this concept of two states of release-ready vesicles and the influence of super-priming will contribute to understanding a number of important phenomena of signal processing in our brain. So of course this is not my work, indeed all the data shown were obtained by Holger Taschenberger who until recently was a postdoctoral fellow in my lab. My contribution as an Emeritus is only data analysis and model building. Many of the ideas and background data were provided by Takeshi Sakaba, a long term collaborator who is now in Doshisha University in Kyoto, Ralf Schneggenburger who is a professor at the APFL in Lausanne and Suk-Ho Lee, a colleague from Seoul National University. Thank you for your attention.

Erwin Neher (2014)

Short-term Synaptic Plasticity

Erwin Neher (2014)

Short-term Synaptic Plasticity

Abstract

Our brain is a network of about 1011 neurons, which are connected via synapses. A neuron typically receives input from about 10000 other neurons, which can be either excitatory or inhibitory. The neuron integrates these inputs and generates an action potential, when its membrane potential surpasses a certain threshold. In synaptic transmission neurotransmitter is released upon an increase in intracellular calcium concentration ([Ca++]) in the presynaptic terminal. Neurotransmitter diffuses across the synaptic cleft and opens ion-selective channels in the postsynaptic membrane.
Synaptic strength (the size of the signal in the postsynaptic neuron, elicited by a nerve impulse in the presynaptic one) displays ‘plasticity’ - unlike signal transfer across elements in a digital computer. The term ‘synaptic plasticity’ describes the fact that connection strengths between the neurons of our brain change constantly in a use-dependent manner. These changes occur on many time scales and underlie many of the computational capabilities of our brain. Long-term changes, such as ‘long-term potentiation’ and ‘long-term depression’ are being studied intensely, since they are believed to underlie learning and memory. Short-term changes, on the other hand, are in no way less important, since they subserve basic signal processing tasks, such as adaptation, gain control, filtering, short-term memory, etc. Molecular mechanisms for this, so-called ‘short-term plasticity’, are still a matter of debate. In my laboratory we are studying the dynamics and pharmacology of short-term plasticity in a particular synapse, which has very special features for electrophysiological investigations:
The ‘Calyx of Held‘, a glutamatergic presynaptic terminal in the auditory pathway, is a giant synapse, which is large enough that quantitative biophysical techniques, such as voltage clamp, Ca++ fluorimetry, and Ca++ ion uncaging can be applied. Using these experimental tools, the role of Ca++ and other second messengers in neurotransmitter release can be conveniently studied (see E. Neher and T. Sakaba, 2008, Neuron 59, 861-872 for review). We identified specific roles of Ca++ in the triggering of release and in the maintenance of release during sustained high-frequency stimulation (Lipstein et al., 2013, Neuron 79. 82-96), as well as influences of other signaling pathways, which modulate amount and kinetics of short-term plasticity. A better understanding of these phenomena will not only clarify mechanisms of signal processing in dedicated brain circuits, but also provide explanations regarding the question how our brain is able to rapidly switch between distinct states, such as various forms of sleep and wakefulness.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content