Wiring Up the Brain: Axon Navigation_PDF

So, we're not sure that everyone's in, but nearly everyone's in, and it's full anyway. I'm the Biological Secretary of the Royal Society, John Scahill, and it's my very pleasant duty tonight to introduce the Farrer Lecture. The award is given this year to Professor Christine Holt for pioneering ou...

So, we're not sure that everyone's in, but nearly everyone's in, and it's full anyway. I'm the Biological Secretary of the Royal Society, John Scahill, and it's my very pleasant duty tonight to introduce the Farrer Lecture. The award is given this year to Professor Christine Holt for pioneering our understanding of the key molecular mechanisms involved in nerve growth, guidance and targeting. The lecture is named for Sir David Ferrier, who was a pioneer in experimental physiology of the brain, and who was considered to be the originator of modern cerebral surgery in the late 19th century. And by surgery, he allocated regions of the brain specifically concerned with vision and hearing. Christine is Professor of Developmental Neurobiology at Cambridge, and very appropriately, her work is on the formation of nerve connections between the eye and the brain, and importantly has included the demonstration that growth of nerve fibres involves the local production of proteins at their tips. Her lecture tonight that we're going to hear, as you can see here, is Wiring Up the Brain, How Axons Navigate. And at the end of the lecture, following your discussion, I'll give her the award of the Ferrier Medal and some other thing in an envelope. Thank you. APPLAUSE Well, thank you very much for that kind introduction. It's really a great pleasure and an honour for me to give this special lecture here today and to honour David Ferrier, who did some pioneering work in functional brain mapping. Now, the mature brain is made up of many billions of nerve cells, or neurons, that are wired together in a highly complex and highly complex and organised way. And neurons in one part of the brain make precise connections with target neurons that are often a long distance away. And the precision of this wiring is really key to the animal's ability to interact with the outside world and, indeed, its very survival. So the process that we would like to understand is how these long distance connections are made. And the process, called axon guidance or axon navigation, is the subject of this lecture today. Now, neurons are highly specialised cells with unique morphologies, as shown here. They have a cell body where the nucleus is housed with the genetic material. And they have multiple short processes called dendrites. And they have a single process called an axon. And, as you know, the neurons are electrically excitable cells, so signals come into the neuron through the dendrites and are passed on to other cells through the axon through synapses that they make with the dendrites on their target neurons. So the axon can be very long, indeed. And, in fact, this neuron here, if it were drawn to scale, would have to have many thousands of these loops here that I've indicated. Even though this would be a typical neuron, even though the cell body here is about the same size as any other sort of cell body in our bodies, which is about 10 microns across. So if I were to draw it to scale myself, if I were a neuron, for example, a motor neuron with a long axon that travels for a metre or so, then my axon would have to extend all the way to Birmingham. So this is really quite a formidable kind of navigational feat. Most of the connections, these long distance connections, are made early during development. And our research is done at the earlier stages of development, looking at the embryonic visual system. And we work mostly on this little beastie here, which is a Xenopus frog. This is in its tadpole phase in an aerial view, looking down on its head and its eyes are here. And what's shown here is a diagram as a slice through the visual system here, a sort of typical plan for any vertebrate visual system. You see the eyes in cross-section and the brain. And what's shown here is this cell called a retinal ganglion cell. Now these are the cells that send an axon out of the eye to the brain. They're cells that relay all our visual information into the brain. And if we follow its navigational route of the axon, you can see it exits the eye, enters the optic nerve, crosses the midbrain here and then grows through the brain up into its target here in the midbrain, a structure called the optic tectum. Now in humans, this begins at around six weeks of gestation. And it probably takes the axon in humans, because the distances are quite great, around two or three weeks for the axon to reach its target. In Xenopus, this process is much faster. The whole distance here is about a millimetre. It takes about 18 hours or so. So here what we're showing is a real ganglion cell that we've injected with a dye. And you can see its axon crossing the midline here and growing up the side of the brain here. And this one was caught in the middle of navigating. It hasn't got to its target. And you might notice it's got a little swelling on the end here. This tip is a specialised growth structure that leads the axon through the brain. And here it's illustrated in higher magnification. And it's called a growth cone. It's about five microns across. Now the first person to describe a growth cone was Ramon y Cajal, who was a Spanish, very famous Spanish neuroanatomist. And these are some of his original drawings here. And he said, this curious club, I christened the growth cone, endowed with exquisite chemical sensitivity and rapid amoeboid movement. And this was really a remarkable insight, considering he had only ever seen growth cones in fixed dead material. And here you can see this dynamic structure of the growth cone here. This is a growth cone in culture. And you can see it's constantly putting out these phylopodia. And Cajal imagined that the phylopodia were searching the environment for clues as they were going along. And in fact, sometimes this has been likened to an amoeba on a leash at the end of the axon. Now we can actually image these growth cones in the live brain. And I'm going to show you a movie here. So first off, we're looking at the side of an embryo. All we've done is taken the skin off the brain here. And this is superimposed, the beginning of the movie and the end of the movie. And as we focus in, you'll see two little growth cones from two retinal ganglion cells. And now they're moving through the optic tract, and they're constantly putting out phylopodia. And right about now, they've entered the tectum, and you can see they're beginning to branch. So they're in their target, and they'll branch and form synapses there. So one of the things that we noticed when we started to do these experiments was that the growth is very directed. The axons aren't making a lot of errors. In fact, very rarely do they make errors. So they very accurately select the correct pathway to take. So how do they do this navigational feat? So what I want to do in the talk today is I'll talk about how they're guided in the embryonic visual system. And we'll look briefly at two choice points within the pathway where the axons make decisions. And then for the second part of the talk, we'll look inside the growth cone and talk about a rather surprising mechanism that we've found that's involved in axon guidance. So the growth cone navigates by sensing its environment. And to take the map analogy a little further, the growth cone navigates by following signposts in its environment. So for example, it turns right on Axon Avenue, left on Gleel Boulevard, and then enters its target here, illustrated, the Cerebellar Centre. Well, what we'd like to understand then is sort of what are the signposts in the brain, what are the molecular signposts, and how does the growth cone detect them? Well, over the last 20 years or so, work from a lot of people's labs has shown that actually the early brain is highly patterned molecularly. And so the embryonic visual system then is subdivided into like a patchwork quilt of molecular domains that are kind of arrayed like signposts along the pathway. So as the growth cone advances along the pathway, it encounters a succession of these signposts, such as illustrated here, names such as Netrin, Efrin, Slitz, and Decemers. So how do the growth cones detect these signposts? Well, the guidance cues in the environment, which it's sort of interested in, it has receptors that are able to detect these guidance cues by binding to them. And so, for example, a guidance cue will bind and activate a receptor. And you can think of these rather like telephone receivers that are taking the message from the outside and signalling it into the inside of the cell. And then there's a cascade of events that are initiated inside the cell, leading to cytoskeletal remodelling and the growth cone steering one way versus another. Now, we can look at whether a growth cone is responsive to a guidance cue using an assay in culture that we call the growth cone turning assay. And very simply, we take a little piece of retinal tissue, a little piece of embryonic eye tissue, place it in a dish, and after just a few hours, the neurons will start to send axons along the bottom of the dish. And then what you can do is select a growth cone and present it with a gradient of your favourite molecule that you want to test. And then over the next hour, you can ask whether the growth cone turns towards the gradient or away from the gradient, which we call a traction or repulsion, or indeed ignores the gradient and just grows straight ahead. So this has been a very useful assay for looking at the behaviour of growth cones. And it's nice because it's quantitative. You can measure the angle of the turning. What I'd like to do now is just look at one particular point, decision point in the pathway, the optic nerve head. Now, this is where axons have to make a decision to leave the eye. So this is a ganglion cell. Its axon grows across the surface of the retina and then it takes a 90 degree turn and dives into the retina in order to get out of the eye. And here you can see a stained retina. These are the ganglion cells in red. And you can see the optic nerve head here with the axons going through. This is the optic nerve coming out of the back of the eye. And humans, I should say, the optic nerve sounds like it's one nerve, but it's actually composed of about a million different axons of these sorts of retinal ganglion cell axons. Now, if you look on fast at the retina, you get a really beautiful view where the retina's just been flattened out and a hole mount, as we call it. And you can see these axon bundles that are converging on this central region here, the optic nerve head, which becomes the optic disc. And this sort of bicycle spoke pattern, this sort of bicycle spoke wheel pattern, sort of suggests that the axons are being attracted by something right at the nerve head here. And indeed, it turns out that a molecule called nectrin is expressed right at the nerve head. And this is a molecule that was originally described and isolated by Mark Tessier-Levine, who was a previous Ferrier lecturer in 2008. And he called it nectrin because it's Sanskrit for one who guides. And you can then test if this is attracting axons. You can test it in one of these growth cone turning assays, as illustrated here, you put nectrin in the pipette. And what we found was that the growth cones of the retina are very attracted to nectrin. So that's consistent with the idea that the nectrin guidance cue at the nerve head is pulling them into the nerve head and helping them get guided out. And importantly, you can block the receptor to nectrin and then they don't turn anymore. Now, if you knock out the nectrin gene or its receptor, and you can do this in mouse, then if you look at this bicycle spoke sort of pattern and you label a few of the cells right in this region here, their axons stream towards the nerve head here in a control. But in the mutants, the axons become quite lost. And you can see they just spiral around in this region looking for what to do, but they're unable to get out of the eye. And actually these mutants have severe optic nerve defects and partial or complete blindness. So this is fairly good evidence then that nectrin is a guidance cue that's involved in getting axons out of the eye. Now I'd like to turn to the next choice point in the pathway. And this is the optic chiasm. And the optic chiasm is the part of the brain where the optic nerves cross over or partially cross over to the other side of the brain. Now in humans, each half of the nasal part of the retina, that's right next to the nose here, sends their projection across to the other side of the brain, marked in red or blue here, or in this simple diagram here in green. So they completely cross over. Whereas the ganglion cells from the temporal part of the retina next to the temple, stay on the same side, marked in red here. And this crossover allows our brain to receive information from the same point in visual space. And by superimposing these signals from the two eyes separately, we're able to... This allows the brain then to construct binocular vision or stereoscopic vision and gives us depth perception, which is certainly a useful sort of feature to have if you're a predatory animal. Now animals with laterally positioned eyes, so these are frontal positioned eyes, have very little binocular vision and their axons all completely cross. So frogs are actually really interesting because at the tadpole phase, shown in this kind of mugshot here of a tadpole, you can see that the eyes are placed laterally and if you look at their projections, they completely cross. But when they become a frog, as shown by this mugshot of a little froglet here, their eyes shift to the top of the head and they gain binocularity so that a bug flying above them is now seen by the two eyes. And right around when that happens, their eyes, they begin to develop an ipsilateral projection, a projection that's uncrossed, allowing them depth perception, which really sort of helps with their new bug catching predatory lifestyle. So we were interested to know, well what is directing this divergent choice here at the chiasm? And we screened the chiasm of tadpoles and frogs for molecules that might be different and what we found was that right around metamorphosis, a molecule here called Efrin B pops up right discreetly, right at the midline at the chiasm. Now if I just superimpose what the axons are doing here, in the tadpole the axons are going across, in the frog the axons are veering, going towards the midline and then not crossing the midline. Now Efrin is classed as a repulsive molecule, it often repels axon growth. And so in order to test whether indeed it's the model that Efrin is actually the molecule that's really helping to guide these axons, we precociously introduced it into the tadpole chiasm here. So this is a little clump of Efrin which wouldn't normally be there and what we found was that indeed the axons now, a subset of these axons now were repelled and went to the same side of the brain and failed to cross over. So that seems to indicate that Efrin is having a role here in this divergent choice. And the different crossing behaviour depends very much on receptor expression. And this is just looking at a sort of face on picture of the eye flattened out of the retina and in the lower half of the retina the FB receptor, the receptor for Efrin is expressed by, you can see from this blue staining here, and those cells marked in red are the ones that are turned away from the chiasm if you like, whereas the ones in the upper part of the retina, green here, are oblivious to this and then they just cross to the other side. And this mechanism is very similar in mammals. So in mouse if you knock out the receptors for the Efrin ligand, then the axons now do the wrong sort of thing and they get sort of mixed up and the uncrossed projection is very reduced in these mutant animals that lack the receptor. So I've told you so far about two choices, very briefly. The first one at the nerve head here in the pathway seems to work by mechanism of attraction, whereas at the optic chiasm axons are repelled at this choice point. So just sort of very simplistically then, the growth cone leads the axon to its target by making a series of choices through interacting with repulsive or attractive cues that decorate the pathway. Now as I said, this is a very simplified view really of what's going on because there are many molecules, the growth cone actually sees numerous different cues at the same time and has to integrate these cues and this is something that people are working on trying to understand. But it illustrates the general principle of axon guidance. Now when we first started making movies of the growth cones in the live brain, Bill Harris, Friedrich Bonhoeffer and I, we cut an axon severing the growth cone from its cell body in the eye and we were really surprised to find that first of all the axon continues to grow for about three hours, but importantly it continues to path find correctly and it continues to recognise its target properly. So it's able to make all of these decisions properly, these navigational decisions, without having to consult its cell body. So this indicated that the growth cone is acting rather like an autonomous cellular outpost, able to make its own decisions and we wondered about this autonomy for a number of years and the experiments that I'm going to tell you about now sort of give us some light I think in terms of understanding how they can be so autonomous. So the central dogma in biology is that DNA, the genetic material, gets transcribed into messenger RNA and the messenger RNA, shown in this blue squiggle, gets translated and made into, synthesised into protein. Now proteins do the bulk of the work around any cell. So then the messenger RNA gets synthesised by joining a structure called the ribosome translation machinery where it gets synthesised into protein. Now the old idea was that all of this happens in the cell body, so you get new proteins synthesised in the cell body and those proteins then get taken out to various places in the periphery of the cell. And of course that does, of course that's one of the major mechanisms. However, there's a real problem if you're a neuron and you have a very long axon, because it can take days, even with the fastest transport mechanisms of taking protein around the cell, it can take days to get a protein out to the very tip of an axon. So any local demand for proteins out there can't be met quickly. Now we now know that RNA can be taken out into the peripheral parts of the cell, into the dendrites and into the axons, where it can await instruction silently to then be translated into protein. And a single RNA molecule by multiple rounds of translation can actually give rise to multiple proteins in that spot. So it's actually quite an efficient way of getting protein, if you want a build-up protein in a particular place out in the periphery. Now our work has shown over the last 10 years or so, that the RNA gets transported out into the axon, and that guidance cues and growth factors can trigger this local translation, thereby very quickly supplying any sort of demand for a new protein. And now I'll give you some evidence for this. So the first experiment that we did, that was really a sort of key experiment is showing here. So here we're back to using the turning assay, and what we found was that if you use the turning assay with netron in the pipette, where the axons are usually attracted towards it, if we did that in the presence of a drug that inhibits protein synthesis, a protein synthesis inhibitor, then the axons now lost their ability to turn towards netron. They lose this sort of directional response. And importantly, they can still grow, and they basically grow without showing any directional bias. Now they do this even when you cut them and separate them from the cell body. So shown here is a little explant of an eye with some axons coming out here, and these can be cut away, so you have this little sort of headless growth cone here, and you can do the turning assay on it. And as shown here, these little headless growth cones behave very well in a turning assay. They'll be attracted to attractants and repelled by repellents. But when we add a protein synthesis inhibitor, then a number of different guidance cues are inhibited in their ability to cause these directional responses. And so this indicates that sort of multiple cues exhibit this protein synthesis dependent response, and indicates that the mechanism that we're looking at here is rather a common mechanism that's used by numerous different guidance cues. So we built up a picture really just shown here of the signalling pathway. It looks a little bit complicated, but just briefly, we're looking at two proteins, a guidance cue, netron, sematractant, repellent. And they converge on this protein here, mTOR, which is a kinase that initiates, sets off the initiation of translation, leading to protein synthesis, cytoskeletal remodelling, and then turning. So what we'd really like to know is what proteins are being synthesised here. Now, we focused our attention to begin with on the cytoskeleton, which as its name suggests, it is responsible for shaping the growth cone, shaping cells, rather like our skeleton inside. And if you look at the cytoskeleton in a growth cone, there are two main elements. These are two main elements that are in all cells. There's actin and microtubules. Actin in red here is very abundant in the periphery, and it fills the phylopodia with these long filaments of actin. And the microtubules tend to be along the axon here, and occasionally probe the phylopodia. And this is a real growth cone that's been stained with an antibody to actin and to tubulin here. Now, the cytoskeleton is really important for mediating this steering response. In an attractive gradient, the phylopodia, the actin, is assembled on the near side of the gradient, which results in more additional phylopodia being generated on the side of the gradient and a turning towards. The opposite happens with repulsive turning, though. In a repulsive gradient, the actin gets disassembled on the near side of the gradient and results in the growth cone turning away and being repelled by the gradient and withdrawing phylopodia on this near side and heading off in the opposite direction. So if you look at this messenger RNA for beta actin, we find that it's highly abundant in the growth cones of the retinal ganglion cell axons. And so in red here, these little red puncta are that messenger RNA, is the messenger RNA signal in the growth cone. And I'm going to show you a movie of these because they're incredibly mobile inside the axons. So we can do this now. So in red, you'll see these little puncta, these little granules moving along. There's a growth cone here. There's an axon here. There's other axons in blue here. And you can see these little RNA granules kind of moving into the growth cone, accumulating in the growth cone here, and even some of them moving out. And the green signal here are the mitochondria, which are the energy sort of powerhouses of the cell that we've labelled to look at the sort of interactions that occur between these two sort of substructures. So what we've built up a picture of then so far, the idea is shown here. The idea is that netrin, a guidance Q, binds to the receptor. The mRNA then gets recruited, joins the ribosome, and gets synthesised into protein. Now, if beta-actin is our mRNA here, and if it gets synthesised into protein, we would expect that when we add netrin, the growth cone, if we look at the protein levels, should change in intensity, and the actin should increase over time after adding the netrin. And indeed, that's what we see as shown here. So this is a control growth cone, and we're looking at the protein now. We've stained beta-actin protein, and after just five minutes of netrin addition, you can see a significant increase in the brightness here, indicating that there's higher levels of the protein. And this is completely abolished if you add this drug that inhibits protein synthesis. So it's surprisingly fast. Now, what happens in a gradient, because here what we've done is just add netrin to the whole growth cone, and what we find in a gradient, so here we've set up the gradient coming at 90 degrees here from this side, and what we're looking at is actin in the growth cone, and this intensity scale is the hot colours here are high intensity, cool colours are low intensity. And there's a polarised increase in beta-actin on the side, on the near side of the gradient. And this suggests then that the polarised translation of beta-actin can give this directional bias to the growth cone in a gradient of an attractant, and it's sort of remarkably precise. The scheme here then, so the way we sort of understand it then, is netrin binds to its receptor, DCC in the membrane, and that then triggers the recruitment of the RNA to the ribosome on the near side of the growth cone, which then synthesises new protein that helps to promote actin assembly on the near side of the growth cone, and in so doing helps the axon growth cone turn towards netrin. And if we look at it in this kind of model then, the idea then is that the beta-actin message gets translated to help cytoskeletal assembly on the near side. But what about attraction, is the same sort of thing happening, what about repulsion, does the same thing happen but in the opposite way? In repulsion? Well it turns out no, because in a repulsive gradient, if you can block beta-actin, and I should say that we can block beta-actin translation specifically with an antisense morpholino, and that will block the turning of the gradient of the growth cone towards the gradient. But if we use that technique in a repulsive gradient, then the growth cone still turns away from the gradient, so it doesn't need the translation of beta-actin. What we find instead is that repellents trigger the translation of proteins that disassemble the actin cytoskeleton, so for example, cofilin or rho-A. So then the idea is that you still get this polarised recruitment and translation on the near side of the signal, here, the external signal, but in this case what happens is that the newly synthesised proteins help to disassemble the cytoskeleton on the near side and lead to the withdrawal of phylopodia on the near side. So this gives us the idea then that there's a differential translation model, depending on which signals come in, which receptors are activated, you get different sets of RNAs being translated that then give rise to a different type of response. So one of the things we really wanted to be able to do was to visualise translation live, and a number of years ago we made this translation reporter here so that we could see translation happening in live growth cones. And this relies on this remarkable protein here called Caida, which is from the stony coral, and it has an amazing property that its native colour is green, but if you hit it with UV light it then changes irreversibly to red. So we hooked that up to beta-actin, a 3' UTR regulatory region for beta-actin, and the idea then is that wherever actin is translated you see the Caida fluorescent protein. We introduced this into neurons, cut the axons so we know we're looking at local events, hit the growth cone with UV, and then that gets switched to red, and then we can add netrin and ask whether the green signal comes back in the growth cone. And of course the green signal, if that returns, then that indicates that that's new synthesis of new Caida protein because all the old protein, the pre-existing protein, is now red. And you can see this works very nicely, so here's a growth cone before conversion, it's green, no red. After conversion the green has disappeared and now it's red. And if we add netrin, then over the course of quite a short time period we can start to see the green signal coming back and it even goes out into filopodia. So we've used this live reporter to look inside, to look at axons, growth cones in the living brain. And we hooked up the reporter to a molecule that we're interested in called NFPC, which is an adhesion molecule. And so the idea is that the reporter sees changes in translation of this adhesion molecule NFPC. So we converted to red the signal in growth cones in the lower part of the tract versus the mid part of the tract. And in the mid part of the tract the axons make a bend, they change their orientation, which has to do with this guidance molecule, SAMA3A, which they start to see here. And we found that these growth cones in the lower tract, the green signal doesn't return after we've photo-converted, indicating that they're not translating the NFPC molecule. But amazingly once they hit this region here, right where they encounter this guidance cue, this new guidance cue, SAMA3A, they start to be able to quite quickly return, the green signal quite quickly returns. And this synthesis of this NFPC here is important for helping the growth cone make its decision to turn this bend and is important for the next leg of the journey. So the idea then that this gives us is that as growth cones are growing through the brain, they will potentially, when they meet a new signal, a new signpost, they can potentially switch on the translation of some mRNAs that will then give protein that's useful for the navigation through the pathway. And we've now been able to look at this at a single molecule level, which we're quite excited about. So here we've used a different fluorescent protein, Venus, which matures very quickly. And what you're going to see here, we've used a sort of continuous bleaching method so that as soon as the Venus molecule is made, and this is live over here, it bleaches. And so what you see are these little spots, bright spots happening. You should be able to see those, I hope. And over here, this is the same growth cone and we've plotted it cumulatively over this time. So you can see where all the little flashes have happened. So this is wherever you get a flash, that's telling us that that's where you get a molecule of Venus made and a molecule of the beta actin. So this is something that we are planning to use a lot more in the future. So then which mRNAs are actually present in the growth cone? This was a really challenging question to answer because first of all, you have to get pure growth cones, you can't have any contamination from the nucleus. And so we did this by using a special technique that employs a laser to cut out the growth cones from cultured axons. And for these experiments, we had to cut out a thousand growth cones, one by one, for every single experiment. And then we isolated the mRNA and profiled it to find out which ones were there. And what was surprising to start with was that there is a very large number of RNAs there, almost a thousand in these Xenopus growth cones. And we saw the cytoskeletal category here in red, but the other surprising thing was that they spanned really the whole array really of functional categories, going from metabolic, membrane trafficking, transmembrane, secreted and extracellular matrix molecules. So this sort of suggested that there was, that mRNAs could be involved in a number of other processes locally in axons. However, this was really good, but it didn't tell us that the RNAs are being translated because they could have just been shuttled out there and could just be sort of sitting there. So then we embarked on this experiment sort of fairly recently, which is called Axon Trap, and a method that allows us to identify translating messenger RNAs that are actually translating in the retinal axons in the living mouse brain. And what we did was to use a trick, a genetic trick, to introduce a tag onto the ribosome so that you can then use an antibody to pull down the ribosome. And the ribosome comes with the RNAs that are being translated. And we introduced that into the eyes specifically, and so the only things in the brain that have these tagged ribosomes are the retinal axons that have come from those eyes. And then we were able to isolate the RNAs, take the RNAs, sequence them, and get an idea of the in vivo axonal, what we call translateome, the sort of global analysis of all the RNAs in here. And we did this at several different stages, three developing stages, before birth and two around birth and a week later. And we got around about 2,000 RNAs for each of these stages, so it's a very complex translateome. And interestingly, the mRNAs, the function of the proteins encoded by those mRNAs were, a lot of them were axon specific, and they matched what the function was at the time of the isolation. So, for example, elongation and axon guidance were represented highly before birth when the axons are just growing across the tectum, and during a wiring phase around birth and P7, the axons have arrived and they're branching and pruning and undergoing synaptogenesis, and there's a lot of the mRNAs out there involved in this process. Now, the other surprise, which is really interesting as well, was that we also did the same analysis on mature axons. And this has been, there's been a big debate about whether mature axons actually translate, have mRNAs and whether they undergo, whether they can do protein synthesis. And we found a surprisingly complex translateome out there, over 1,000 mRNAs in these mature adult axons. And these are a number of different categories, but their predominant categories are synaptic transition, transmission, so they seem to have a role in keeping your synapses going, and axon survival. And we also found when we did this sort of bioinformatic analysis, interestingly, that a number of diseases are very strongly linked to the adult translateomes, so Parkinson's, Huntington's, Alzheimer's disease, for example, which suggests that perhaps local translation is needed to prevent the axon from degenerating. So you have to actually actively keep the axon alive. And this makes sense. I mean, these neurons, these axons are made, you know, they have to survive for over 80 years or so in humans, so there has to be mechanisms to keep them alive. And in some cases, for example, in ALS, the axon seems to be the first thing to degenerate and go, and once that's degenerated, then the synaptic transition transmission fails, and there's severe clinical consequences. So just to summarise then, at the end, so we started here looking at the directional guidance, and the retinal ganglion cells I've talked about to you today are now probably the best studied example of a CNS axon navigation. And we saw how the decisions of their growth cones are guided by cues along the pathway and how this involves receptors, cytoskeleton, attractive and repulsive responses. And then we also saw how this new, sort of rather novel RNA-based mechanism helps growth cones to navigate by supplying new protein on demand exactly where it's needed. And in fact, I should have sort of said this earlier, but the idea then is it's extremely localised, this RNA-based mechanism, where you, literally where you sort of tickle a receptor is where underneath you can get new proteins, and some receptors, for example, the nectrin receptor binds has been shown to bind ribosomal proteins directly, kind of linking the machinery directly to the receptor. So I should also say that the RNA-based mechanism seems to be, it's a highly conserved mechanism, it happens in many other cells and it also happens in dendrites, for example. Now, and our new translatome work and other functional studies that we've done recently, actually, suggests that this mechanism is also involved in a number, multiple aspects of axonal function, for example, axon branching, where if we reduce translation, local translation when these are branching, then we get very impoverished branches. And we know that in some neurodevelopmental diseases there is an association with branches that are very poorly formed in the brain. And with axon regeneration, for example, the axonal growth cone, when you cut it off, fails to regenerate in the absence of protein synthesis, local protein synthesis, and we also have some evidence that axon maintenance requires local protein synthesis in the axon. So in trying to understand this whole process of axon guidance, our work actually may offer some fundamental new insights into the future of understanding some neurological diseases and nerve regeneration. So none of this work would have been possible without really a very talented crew of colleagues, both in my lab and also my collaborators over the years. And I would like to give a special thanks to Bill Harris, who we have worked together for 30 years, and he's been involved in all of the projects, really, that I've talked about today. And I also just want to mention to thank, I know many people in the audience have contributed to the work that I've talked about today, and I'm sorry I haven't been able to mention your names, but what I've talked about is a sort of a result of many, many people's work. And I'll stop there, and thank you. APPLAUSE So other questions? There's one there, can you see? I wondered what the mitochondria might be doing. Are they doing anything for protein synthesis? Well, that's an interesting question. Some people are looking at that. What's very interesting is the mitochondrial genes themselves, the nuclear encoded mitochondrial genes, are highly represented across all of the ages in the axons. So we think that actually mitochondrial sort of health, if you like, is possibly replenished locally, but they also may have, you see, I don't know whether you noticed, some of those RNA granules come along and they sort of rub up against the mitochondria and then go off and come back and rub up again, so there could be some role there. Yeah. So in your initial navigation analogy, the streets are already there, so somebody only has to choose left or right, which you explained how the neurons do that. Now, are the streets already there for the inner growing brain? The streets are sort of more or less there by the time the axons grow, but obviously they themselves have to be put in there, and so you could, you know, take the question further back and say, well, how do they get there at the right place and the right time? And that's a whole other sort of area to look at. But a lot of people's work has shown that, you know, there's a lot of, the brain gets patterned through transcription factor activity that happens very early in development, and there's a sort of stepwise increase in the sort of complexity and patterning of the brain that happens over time. I was thinking that the patterning of the beacons wouldn't be enough to give you the channeling that you need. It would give you something that's more spreading and more... Well, I mean, you saw the F and B as a really very, very tiny area of the brain. It's sort of very restricted, so that signpost is very restricted. Maybe we can talk afterwards. I'm not quite sure I catch what you're saying there. Thank you for the talk. I was wondering how much the growth cone is reprovisioned. If you cut a growth cone off, is it regenerated at all quickly, and is it coming from the nucleus far away, or what's the situation there? If you cut a growth cone in these very young growing axons, they will quickly regenerate a growth cone. And there's some work that James Fawcett has done showing in a sort of different kind of preparation that if you cut the growth cone off in the presence of these protein synthesis inhibitors, you don't get a reformed growth cone. In old axons that don't regenerate well, then they just don't reform a growth cone. They're just not good at doing that. So the process gets worse and worse with age. Sorry, just to come back on that. What that means is that the growth cone, in normal circumstances, is being given a load of mRNAs to take with it where it gets to. So that means that there's probably differential gene expression in the nucleus. Oh, absolutely. You're talking about a genetic effect here. Yes. So, I mean, one of the striking things is how dynamic the changes are from age to age. So the repertoire of RNAs changes within hours. So you've got, obviously this is all coming from the nucleus, so you've got the nucleus which is going through this whole programme that's supplying the RNAs in the first place. And we don't understand how that works. And there's a whole lot of alternative splicing and all sorts of other stuff that I haven't got into. But, yes. You can try that one. Which one? Right there now. Sorry. Keep your hands up if you put your hand up. Thank you very much for the talk. Is there any possibility of supplying some drugs or stimulating an ageing axon to better maintain themselves or replenish the ageing, basically, so in order for these axons to function better later in age? Thank you. So if I understand correctly, is there any sort of drug that we can use that would make axons regenerate better? Yes. There's no sort of silver bullet. Certainly there's experiments that have been done where cyclic AMP, which is cyclic nucleotide, if you add drugs that sort of increase, so with time the axons seem to sort of lower their intracellular levels of cyclic AMP. And you can make them behave like young axons again by putting drugs on and making, increasing the levels of cyclic AMP. But this hasn't really been a very successful model to take into the animal. It's just something that seems to work quite well in culture. And I don't really work in that particular area, but there are people who are actively pursuing lines of investigation like that. There was one at the front down here. Oh, you've got one there. What kind of a role do you think that this kind of research will play in research into regeneration, for example, after spinal cord injury, that kind of thing? It's a really important question. I think we just don't have a good handle on how to make axons regenerate. And I think we're approaching it in a quite simplistic way. You can see the sorts of things that I've talked about here, incredibly complex, and we don't understand them yet. And I think first you really have to understand how an axon grows, how it navigates. What are the basic mechanisms that get them there in the first place to then be able to go to the regenerating system and be able to say, okay, let's try this, this and this. So I think, as you can see, we're still, the state of play is still we don't fully understand. We're increasing our knowledge. And so I think I'm hopeful in the future that some of the things that we've talked about here will be helpful when it comes to sort of therapeutic sorts of things down the line. Thank you. I may have missed it, but I was wondering how long it takes an axon to grow to reach its eventual target, because you still have to transport all the ribosomes and the mRNAs and so on, and that has to be done to sort of keep up with the sort of leading edge of the axon. Yes. So the axons that I talked about mostly today in Xenopus take about 20 hours, 24 hours to make that whole journey. But obviously in a mammal, you know, we're talking sort of weeks to get there. So as the growth cone, as soon as the growth cone is formed, it goes out there, it's got ribosomes in it, so it sort of takes everything with it, but then of course it's got to be replenishing those things all the time. Yes. Thanks for the talk. I had two questions. My main question was about probably you answered the gentleman, did genes play the main role of this whole complex process? That's the first question. The second question, you mentioned microtubula. What do you think about Penrose, Roger Penrose's idea of consciousness somehow arises from works of microtubula through quantum mechanical effects? Have you seen any signs of it? Okay. Well, the first question about the genes controlling, yes, I mean, this is a remote form of gene regulation. I mean, this is all coming from, it's all directed from the nucleus out into the periphery. So yes, it totally depends on the sort of nuclear things that are happening in the nuclear programme. Penrose consciousness, microtubules, I don't think I'm the best person, I don't know. Sorry. Yes. Yes, hello there. I have a son who has cerebral palsy. He was born very premature and took him a long time to learn to walk. Eventually, he even taught himself to bounce on a trampoline and remove his muscles and whatever by seeing himself in a window. Now, in some way, it seems to me that he's almost regrown those pathways to understand how the muscle movement should work from different parts of the brain. Have you come across anything like that before? Well, that's very interesting. I haven't because, I mean, I don't work in that area, but I would sort of, yeah, I would believe it. I mean, certainly young nervous systems are very good at adapting and sort of making changes to themselves depending on the sort of demand and the sort of intrinsic environment. So, yeah, I think that's all I can add to that. I'm intrigued by the messenger RNA granules. Are all the messenger RNAs bundled together? Are they specifically bundled in different types of granules, for example, the Sema 3A sensitive messenger RNAs versus the Netrin RNAs? Are they membrane bound? What's known about those guys? Yeah, we don't know too much about the RNA granules. I mean, there was some early evidence that a single RNA is carried on its own, which seems sort of very expensive, but there's more recent evidence suggesting that actually, no, there's actually multiple RNAs that can be carried together. They're not membrane bound. They're granules. The RNAs are bound to RNA binding proteins. There's a whole slew of proteins that their only job is to bind and carry around the RNA and to regulate the RNA, and I didn't tell you anything about those at all today, but yeah, they're sort of fascinating how you kind of regulate who gets into your granule, who gets translated when, and all that sort of thing. It seems to me that you've answered all our questions. So all that remains for me is to make this presentation and to thank you very much for an extremely clear talk. Thank you. This is it. Thank you very much. Thank you very much. Thank you. Very good. Very good. Thank you. Thank you. Thank you. Thank you. Thank you.