Jeff Lichtman (Harvard) Part 1: Connectomics: seeking neural circuit motifs

Just another WordPress site

Jeff Lichtman (Harvard) Part 1: Connectomics: seeking neural circuit motifs

Hi My name is Jeff Lichtman and I’d like to talk to you about the relation between the structure and the function of the nervous system As many of you probably know, a long historically-important theme in biology is the relation between the physical structure of biological entities and their function In fact, the state of the art in biology right now is a field called structural biology, which seems to have a very broad name, but it actually is a very narrow field It’s a field designed to talk about the relation between the structure of molecules, the atomic structure of molecules, and how that helps us understand their function Of course, there was a time when the structure and the function of the nervous system was not understood at all, as well as the rest of the body In fact, structure and function questions did not revolve around atoms, but around much bigger structures There was a time, for example, when humans first began dissecting cadavers and opening up animals they had hunted and they noticed, I’m sure, that there were collections, masses of tissue in various parts of the abdomen and other parts of the body that had distinctive appearances For example, the liver has a particular color and it’s always in a particular position in the body of a wide variety of vertebrate animals And for a long time I think these organs like the lungs, the liver, the kidneys, each had their own appearance that separated one from the other, but there wasn’t a great deal to allow us to understand what they did Everything changed however, with the invention of microscopes, because for most organs, the heart maybe is the one exception where you could see it pumping with your naked eye, but for all the other organs, you basically have to look at it at higher resolution to finally understand what the structure was, and what that structure meant And the discovery of the microscope allowed early biologists to notice that there were structural motifs in each of these organs that were recognizable from one animal to another So even though a zebra and a frog look completely different, if you look at their kidneys or you look at their liver, you will find striking similarities, because a small number of cells are put together into a motif, and these motifs explain the normal function of these organs For example, in the kidney there’s a collection of cells that form this impressive tubule system known as the nephron, and once you understand the filtering purpose of a nephron, you basically understand the kidney, because the kidney’s just made up on many hundreds of thousands of nephrons And the same is true in the lung, where you have these vessels that have air in it next to vessels that have blood in it, and this is for gas exchange between air and your blood So for all these organ systems, this has been really a tremendous advance in the way we think about not only the normal function, by knowing the cellular motif, but also it has provided deep insight into what’s wrong with these organs in disease states So for virtually every disease — not every disease, but virtually every disease — there is some physical underpinning, either in the biochemistry, but more often in the physical structure of these cellular motifs So every disease of lung, whether it’s bronchitis, or asthma, whatever the disease is, if you look in the lung you’ll find an abnormality And similarly in the liver, for all the various liver diseases, or the various kidney diseases, there are abnormalities in the physical structure And this has been a useful advance because once you know what’s wrong, physicians and clinical scientists can begin to think about ways of curing these diseases This has worked great for all the organ systems of the body, with one notable exception, and that is the brain, where for reasons that are interesting, that I’d like to begin by talking about with you, for one reason or another we have still a pretty rudimentary view of the physical underpinnings of brain function That is, the structure of the brain is much less well understood So, compared to all the other organ systems, the relationship between the structure and the function of the nervous system is much more complicated,

and this has led not only to a problem in understanding the normal way in which the structure of the brain gives rise to the function of the brain, but perhaps more importantly for human health, there are a wide range of diseases of the nervous system where we have no idea what is actually wrong, and trying to cure a disease without knowing what is wrong is a very difficult step So there are a number of psychiatric diseases and even pain syndromes like migraine, where all we can go on is the behavior or the complaint of a patient, there isn’t much more one can do other than hope that some accidental discovery of a medicine will help these patients I believe that the problem here is that we have to somehow break through this barrier and seek the equivalent cellular motifs in the nervous system So I’d like to begin by explaining why it’s so much harder to get those cellular motifs in the nervous system than in other organs systems The first reason is that the brain’s diversity of cell types far exceeds the diversity of cell types in the rest of the body, combined For example, in the liver there are about five different kinds of cells that make up the portal triad, which is the main cellular motif in the kidney, whereas in the nervous system we don’t know, actually, how many cell types there are A state of the art question right now is, in the retina, perhaps one of the most organized pieces of neural tissue, where people have been trying over the last few years to get a definitive answer to the number of different kinds of cells there are in the retina The retina is almost certainly going to exceed 50, but it may exceed 50 by 2- or 3-fold or more, no one really knows, and someday we will know how many cells there are in the retina, but this isn’t gonna help us with the cerebral cortex, where there’s another menagerie or cells, or help us in the brain stem, or in the amygdala, or the thalamus, or the spinal cord Each of these parts of the nervous system has its own vast array of cell types, which just makes a mockery of the complexity of other organ systems What goes along with this extraordinary diversity of cell types is, not surprisingly, a diversity of function You could probably write a chapter in a book, maybe 100 pages long, that would describe everything a normal kidney does But how long a chapter would you have to write if you wanted to say everything a human brain could do? You couldn’t write such a book, because not only we do so many things already, but every year, every new child learns something new, and does things that have never been done before The diversity of function in the nervous system is extraordinary So this is reason number one why it’s very hard to understand the nervous system, but this is not the only problem There’s a second problem, which goes along with this, and that is that in organs like the kidney, once you understand the cellular motif, you basically understand the whole organ, because it’s iterated over and over again In fact, a human being can lose a kidney and still function pretty well The same cannot — and you could lose a lung, or you could lose part of your liver and still function fine — but you can’t lose half your brain No matter which half I take, you’ll notice a difference and that’s because, in distinction to this iterated, repetitive function used over and over again to build the whole organ, presumably this multiplication of the same function over and over again gives opportunities for greater capacity, the nervous system does it completely differently Every part of the brain has a unique function, so you lose any part of the brain, you lose a particular function So there’s organizational principles in the brain that span much larger areas than small cellular motifs And so this organization at many size scales is a tremendous challenge for people who are trying to understand the relation between the physical structure of the brain and its function I’m gonna show you a picture which is a series of shots of pieces of brain at orders of magnitude smaller resolution to give you a sense of how much information one has to know at each of these levels The brain is shown here in cross-section, that’s a human brain, approximately 10 cm from the front, which is this part up here, to the back, which is over here And each part of the brain has its own particular function Some of you may know that the occipital lobe is in the back of the brain If you have a stroke or damage to this part of the brain

vision will be impaired, but many other functions will be normal Conversely, if you have a damage or any other kind of stroke in the front part of the brain, executive functions, planning can be damaged, but you’ll see fine The cerebellum, over here, is important for balance Breathing and sleeping are in the brain stem Your emotions are in the cingulate cortex Walking and other rhythmic activities of your body are done in the spinal cord So at the centimeter scale, there’s many important differences from one part of the brain to another If we move down to panel B here, now zooming in to a centimeter of brain, we’re looking at a little slab of the cerebral cortex, and you see that the cortex itself is divided into an outer rind where the cells are, that’s called the gray matter And this stuff, pink, is actually the white matter where cables, axons that leave nerve cells, connect one part of the brain to another So even at the centimeter scale, there’s very different functions between the outer rind and the inner part If we zoom in on a little part of this cortex and then zoom up another order of magnitude, now to a millimeter, one can see that in the cortex there are nerve cells, and these nerves are quite large and ridiculously complicated compared to the cells in the rest of the body They are decorated, each of these cells have a bunch of little things that look like tentacles coming out of it These are the dendrites of the cell, and this little cell body here has a bunch of basal dendrites near the cell body, an apical dendrite, and then a bunch of dendrites way up at the top, which are right at the surface of the cerebral cortex The is a pyramidal cell, and the dendrites are the site where this cell receives information from other cells This cell also has an axon, it’s shown here in this line, and then it breaks off here because this axon can go extraordinarily far distances It can go to the other cortex, it can go down the spinal cord In a giraffe this axon could be several meters long, actually Quite an amazing thing for a single cell So there is this division of labor between the input of the cell that comes onto the dendrites and the output of the cell, which is where the axon is, it’s the output of the cell If we zoom up another order of magnitude here, go to a tenth of a millimeter, a hundred microns, and look at one of these dendrites, we see the dendrites themselves are kind of complicated affairs They have a shaft and then sticking out are these little thorns, which are known as dendritic spines The dendritic spines are interesting because when other cells wanna talk to this cell they do so by making connections, synaptic connections that are exciting the cell, at each of these dendritic spines So the excitatory synapses come onto the dendritic spines, but in general the inputs to these cells that try to keep these cells quiet, the inhibitory inputs, come onto the shafts of these dendrites, not where the spines are So you have, again, a division of labor so that little parts of each cell have different functions And if one wants to really understand what’s going on, one has to look at these synapses at high resolution And for that we’ve gotta zoom up way, way more, down 100x, to a single micron And now you can see in this picture here it’s a cross-section from an electron microscope, we’re below the resolution of light imaging, and this is cross-section of a synapse filled with little synaptic vesicles And they’re making a synapse, this axon terminal from another cell is contacting a dendritic spine, and that’s its spine apparatus, that’s how we know it’s a spine just from this picture, this object down here And this is where information is transferred from this cell to this cell through the release of neurotransmitter than binds to receptors on the target cell Now this is a wide range of scale, 6 orders of magnitude, and to make this point a little stronger, I can tell you that a lot of techniques that people are familiar with are ones that image the brain to see the whole brain’s function This technique is called functional Magnetic Resonance Imaging, fMRI And it looks for local increases in blood flow in parts of the brain that are highly active The resolution of an fMRI image is such that the smallest spot, the smallest 3-dimensional spot, which is called a voxel – the equivalent of a pixel in three dimensions, is a cubic millimeter So this image of the brain’s function that you might see in scientific journals or popular press about parts of the brain that are activated usually has a resolution in the range of cubic millimeters, whereas this image,

if you took a bunch of serial sections through the brain, would have a resolution of 100 cubic nanometers – 10^-9 meters And therefore a voxel of an fMRI is a trillion times larger than an electron microscopy voxel that’s 100 cubic nanometers A trillion is a big number and the problem with the brain is one has to understand everything from this level all the way to that level, and everything in between This is an extraordinary challenge So these are two important challenges when it comes to understanding the structure-function relation of the brain, but it’s not for me the hardest The hardest and perhaps most intriguing challenge is the last one I want to mention to you And this is related to a fact that the brain, like other organ systems, is generated by an instruction book in the genes that build through a remarkable process that’s not completely understood, to put it mildly, the structure of the brain, and that structure generates function That is true for all organ systems: genes generate physical structures, and the physical structures underlie function But in the brain of mammals and especially humans, there are lots of functions we have that could not have possibly have come our genes, and yet they’re there and they’re stably maintained, and I would argue probably have a structural underpinning For example, if you learned to ride a bicycle as a child, you can ride a bicycle for the rest of your life, at least until you’re quite old, without having to relearn how to ride a bicycle This was the case with me, I learned to ride a bicycle as a child and then I moved to St. Louis where I didn’t use a bicycle for a long period of time, and then I moved back East and my wife suggested because I was growing, but not up, but out, that I needed exercise She said we needed exercise, but she meant I needed exercise And I bought a bicycle and it was after maybe about 10 years of not riding a bicycle, I got on the bicycle and no problem After a few moments of being shaky, my new bicycle I could ride perfectly well It all kind of came back to me I had a neighbor who was trying to learn to ride a bicycle as an adult, who had no training with bicycle riding as a child, it was just a different kettle of fish altogether It’s very hard for adults who haven’t learned things in development to learn new things, whereas once you’ve learned something you don’t have to keep practicing it to keep it in your mind Somehow, the experience of learning to ride a bicycle, the use of the nervous system, the activity of the brain in learning to balance, in sensing the world, in changing my motor actions to match the change in the tip of my body as I’m moving forward, changed me in a way that for the rest of my life I have that ability And that must mean that somehow structurally I am changed by virtue of learning to ride a bicycle So rather than genes generating the structure of bicycle riding, one might have to think somehow that experience generated the structure of the nervous system And this is something that’s quite different from most organ systems, where it is more a linear path from genes, through structure, to function The brain you have this loop, that is, physical structure generates functional capability, and the functional capability allows you to experience the world in a way that changes the physical structure, which changes the functions of the brain And this loop, like all loops in biology, are messy and horrible and hard to study, because the causal linkage between a gene and a structure and a function is straight forward, relative to a loop Now, what you might call this loop is a loop that is related to making memories, whether the memory is how to do something, related to an experience as a child, or memory of something that you hold in your mind for long periods of time, ultimately those physical traces of experience might be called engrams, that is the word that Richard Semon came up with in the early 1900s for the physical instantiation of a memory, so this process might be called engramitization That’s a made-up word, that’s not really the word, but that is actually what I am interested in in my laboratory and my own career I’ve studied this question of how experience physically turns into something in your brain So how might we go about studying that? My suspicion is that the brain, like other organs, is a very complicated structure and that to study this one, one’s gonna have to get down deep into

the central way in which nerve cells are connected together I don’t think one’s going to find a big bump of the side of the head of people who ride bicycles, other than when they fell off their bicycle, to show us where it is that bicycle riding is I want to point out one important and somewhat controversial idea, which is that this way of sort of taking information about the world and making it part of yourself is I think a relatively new adaptation in a subset of animal species, and I wanna just quickly compare for you this remarkably species, a young human, each of which is learning about the world and ending up being quite a unique specimen based on the particular experience it has, and this organism, which is a preying mantis, which is also an amazing and remarkable creature, but who depends much more on its genetic heritage for its behavioral repertoire than a human baby And I wanna emphasize this because this is something that keeps humans changing over time in ways that other organisms can only change based on genetic drift, humans can change based on culture So let me just go through this for you briefly In humans, there is a long period of development where we are much less adapted to the world we find ourselves in than a preying mantis We come into the world and we’re unable to walk There are very few animals that aren’t mobile shortly after they’re free from their parents In fact, for us it takes about a year to walk, there’s no other animal that takes so long to walk It takes another 15 years before we get our driver’s licenses, and another 3 years, we’re 18 or so, before many of us leave the nest This is just an extraordinarily slow development In the case of other animals, they’re kind of on their own and they have to grow up very quickly, and they basically grow because they immediately, because they know what they have to do They don’t have to learn it in order to do their behavioral repertoires We depend on experience for our development in ways other animals don’t If you take a human baby and you put it in a dark closet for the first couple years of life, you’ll go to jail, because when you take the baby out, besides being very weird for not seeing the world, it can’t actually understand the world, it can’t understand the visual world without visual experience My suspicion is that raising insects without light has no effect on their visual ability, suggesting that, for humans, experience is not just useful, it’s absolutely essential for them to do their normal development We also have to learn things in order to survive The language we speak we are not given, we learn our language We go to school This animal doesn’t have to go to school, it knows, it can interpret the sounds of other animals, largely based on its genes Many of our behaviors, I would say for humans it’s the majority of our behaviors, are based on culture, whether we’re talking about typing on your computer, using your iPad with your finger, or buttoning your shirt, all of those behaviors are culture-based In these animals, the behaviors are not based on culture, they’re based on genes And finally, one of the best cultures humans have is the culture of science And the culture of science has allowed us to I wouldn’t say master, but perhaps tame the genome So now we know enough about genomic and genetics that we can take genetic material and put it into animals, and ultimately into ourselves, just based on curiosity This is the intelligent design of heritable traits, and that word intelligent design is slightly loaded, humans are not that intelligent, but we can do things I, for example, take the genes from jellyfish and put them into mammals so that we can look at cells that we’re interested in looking at, which I’ll show you in a few moments There’s just no precedent for that in other animals doing that And finally, if you think in the future humans may eventually leave this planet, and may have to leave this planet if it doesn’t survive our insults to it, perhaps And if we leave the planet in rocket ships going in many different directions, our species will be basically immortal, because there will be no single thing that could wipe it out And all the other animals on the planet, it’ll be up to us to put them on these new Noah’s Arcs of the future, perhaps,

if they wanna stay alive as well These are profound changes in biology And there’s only one organism that can do it, and it’s humans, and I think they can do it because of this engramitization, a fact that we take information about the world and make it our own, and then pass it on to the next generation as if it were genetic information, even though it’s just culture So what I’d like to talk about now is how one would actually go about studying the way this information is embedded in our brain And in order to do that we have to map out these connections It’s a lot like what geneticists did when they decided to do genomics, which is to map out the atomic structure of the genes that underlie inheritance If one wants to map out the cultural things that pass from one generation to another, we would have to map them out at the resolution at which they function, at the resolution of the synaptic connections So instead of genomics, we would be talking about connectomics This is for many of you probably a new term, so I’ve brought the dictionary definition here Connectomics is sometimes pronounced “connectOmics”, sometimes “connectamics” It’s a plural noun, but singular in construction, sort of like economics or physics have an “s” at the end, but they’re singular It’s a branch of biotechnology — important to realize that technology is an essential part of doing these kind of mappings — concerned with applying the techniques of computer-assisted image acquisition and analysis to the structural mapping of neural circuits or to even the complete nervous system of selected organisms using high-speed methods — you’ve gotta go fast or you’ll never finish, I’ll make that point in a bit — with organizing the results in databases, and with applications of the data as in neurology or psychiatry — that is there are probably diseases of connections, they might be called connectopathies, where the problem is that the wiring is in some way mis-wired or screwed up And also there are fundamental neuroscience questions like, where is bicycle riding? What does it look like? That would come about by understanding connectomics In this dictionary, one also might find the world connectome This, by the way, is from Merriam-Webster’s Unabridged Dictionary in 2019, so that’s a little bit of wishful thinking, but sooner or later I think humans have to face the fact that if we want to understand the brain, we’re gonna have to get in and look at the brain at the level at which it’s actually functioning, which is the level of individual wires Now, the history of neuroscience is such that it began with scientists who, in the end of the 1800s, in the 19th century, made a good start on connectomics, basically, by using tools that allowed them to see small numbers of nerve cells and see how they were connected I’m speaking particularly of the brilliant discovery of Camillo Golgi, of a stain that allows one to see single nerve cells in their entirety The Golgi Technique, as it’s called, he had to call it the “black reaction” product because I think it would have been bragging or something for him to call it the Golgi Technique He invented this in 1873 when he was 30 years old, working in his kitchen And it was a remarkable technique, not so clear how he figured this out, some thought it that he was playing around with photographic emulsions and dipping pieces of brain in them to see if he could get labeling, but basically you take a piece, a block of brain, and sitting right here is such a block in which he just dunks it into a mixture of silver nitrate and potassium dichromate, and then very rarely, an occasional crystal of silver dichromate forms, and the crystal is the nydus for further crystallization, but the crystals cannot break through the membranes of cells So if a nerve cell happens to have a crystal in it, the crystal will expand, but fill up the cell and not break through And so you see the entire, in this case a pyramidal neuron again, the basal dendrites, the apical dendrite, and the axon coming down here, in it’s beauty and completeness The success of the technique was largely because it was so inefficient If all the brain cells labeled like this, the brain would just be a big brown blob, but it was that only, you know, 1-5% of the brain actually labeled So this was a great discovery and he got the Nobel prize for this work, but the person who used this technique to really inform us on the way the brain works,

that it is actually made of nerve cells that are in a circuit, was another scientist who was alive at the same time Golgi was from Italy, and Cajal, Ramón y Cajal, was from Spain, and they shared the Nobel prize Unfortunately, they didn’t like each other very much and that’s a long and interesting story I’m not gonna go into now, but it’s worth… if you want to read a kind of negative view of Cajal, just read Golgi’s Nobel lecture, which is an entire lecture pooping on all the great things Cajal did So what Cajal did is he used the Golgi Technique to look very closely at nerve cells And Cajal was a great artist, his pictures are really beautiful, but what made him famous was a discovery he made in these pictures as shown in this picture here The picture is on a kind of paper that yellowed, but he used a white-out which stood the test of time, so there’s little bit of white and yellow in the background, but that’s not what made him famous What made him famous were these arrows that you can see all the way along His view, which has stood up, has held the test of time, is that the axons coming out of cells, and here is such an axon, an axon at its ending branches, as it does here, and overlaps with the dendrites of another cell, and at the point of overlap, information from the axon is conveyed into the target cell — this is before the invention and the discovery of synapses, so this was just an amazing idea — and that the target cell then takes that information and then passes it out its own axon, and that axon then branches into the dendritic field of another cell And at the place of overlap of those two cells, information passes from the axon into the dendrite of this example, this pyramidal cell, and passes down towards the cell body, and at the same time here’s another axon of another neuron, that happens to be converging on the same dendrite, the same neuron, and together those inputs collaborate in some way to generate a signal that leaves that cell out its axon, and then that axon then makes contact with the dendrite of this cell, and that information is passed down its axon, and that axon goes on and also contacts the dendrite of this cell, and that information goes to the cell body and this is passed out here And here, in short order, is Cajal’s view of the way the cerebral cortex might send information from one place to another It was a tremendous intellectual feat to be able to do this, and to a first approximation he got it right We still, when we think about neural circuits today, most people think in terms of these arrows connecting one cell to another One of the pleasing things about this is that the sparse labeling approach he used presented us with something simple enough that virtually anyone could understand the way the brain might work But one has to keep in mind, and Cajal understood this but many of us forget it I think because it’s so painful to realize the truth, is that this is all based on very, very limited labeling of the Golgi Technique When it’s simple enough to understand because it’s so sparse When Cajal used the Golgi Technique in ways that he could see many more cells in one place, and I’m just going to show you a blow-up of this little region here, it’s a little more complicated In fact, you can look very carefully, there’s no arrows in this diagram, Cajal was a genius He knew it was impossible, even for him, to figure out exactly how these cells were interconnected And so we’re left with the fact, oh yeah there are a lot of cells there, but how they’re interconnected, well that’s just going to be very hard to get the entire wiring diagram Now, if each of these cells were labeled, maybe, a different color, let’s say you had a different kind of crystal in each of these cells, maybe with his genius he could have disambiguated the wires enough that he could get wiring diagrams that were dense or saturated, but he had the Golgi Technique to use For people now, we have opportunities to use other techniques, but you can see where I’m going with this That there’s some inadequacy, perhaps, with using sparse labeling techniques So what about color? Could we use colors as a means of going forward? Well, one way one could take advantage of color would be to make each cell a different color Is that possible? Many of you may know that there has been a revolution in colors related to the fact that jellyfish are bioluminescent, in part, because of the fluorescent protein in them known as Green Fluorescent Protein

This was a great discovery that has modified lots of laboratories in cell and neurobiology and immunology as well In many fields the Green Fluorescent Protein has been a revolution And one of the discoveries that came along with the Green Fluorescent Protein was the realization that the Green Fluorescent Protein could be mutated to make other color fluorescent proteins, each of which have a genetic equivalent, so there’s a gene for each color And many other marine organisms have fluorescent proteins that are in the same family And so there is a mutations of the original Green Fluorescent Protein from jellyfish and other marine animals have now given us a wide range of fluorescent proteins And with those we now have a strategy where one could label with many different colors to saturate out the wiring diagram, rather than having a sparse wiring diagram I realized, and I think this is pretty obvious, that if you had just three colors, if you had a Green Fluorescent Protein gene and a Red Fluorescent Protein gene, and a Blue Fluorescent Protein gene, you basically have enough colors to generate all the colors a human being can see, in part because we’re trichromic, we only have three photoreceptor types Every color we see in the world is just a mixture of the activation of those three primary photoreceptors, in fact the monitor you’re looking at this film on has almost certainly got and R,G, and a B panel If you zoom up with a magnifying glass at your panel, you’ll see that there are little red, green, and blue pixels And that is basically because that’s all you need to see the entire rainbow of colors that humans can see Jean Livet, who was a postdoc when he started this work, worked with me and Josh Sanes on building a tool that would allow us to use three colors, ultimately, to randomize the color in every single neuron in the brain His strategy, which we called “Brainbow”, was based on a recombination event That is, he made a genetic construct under control of a regulator from a gene that’s widely distributed in the nervous system and so should drive expression well in nervous system cells, it’s the Thy1 promoter, and then put after the Thy1 promoter regulator region, three fluorescent proteins, a red one, a green one, and blue one And he interspersed between them sites that could be allow a piece of that DNA to be excised by an enzyme called CRE And he had these lox sites came in two categories in his original construct, the ones that are blue and the ones that are yellowish here are the two sites he started with, that are incompatible That is, you can cut out where number 1 is, between the two blue, OR you can cut out between the two yellow ones, number 2, but you can’t do both You do one or the other because if you do the first one, one of the yellow sites disappears with it as well So this is a construct that if you do nothing to it, if there’s no CRE, it has a default color, which is the first color of the three, there’s a stop signal at the end of each of these fluorescent proteins So Thy1 will cause Cherry to be expressed, but if you do nothing, if there’s no CRE present, this construct would be red However, if there was some CRE present, the CRE has a choice It either can cut out the Cherry piece, in which case the Thy1 will drive the second color’s expression, or it can cut out the first and second pieces, in which case the Thy1 will cause the expression of the third color So if you had an animal that just had one copy of this gene in it, then the cells should either be red, green, or blue, and especially if there isn’t a saturating amount of CRE, some of the cells won’t recombine at all And indeed, that is what Jean Livet found So here for example, I’m gonna show you a picture of a brain in which the Thy1 construct ended up in a piece of the DNA where for reasons unclear, only the glial cells in the brain expressed the fluorescent protein So this is Astrocyte Brainbow, or Astrobow, and this kind of paintball coloring that you see here is showing the hippocampus down here and the cerebral cortex here I’ll zoom up on a little piece of this, you can see it a little bit more clearly You see first of all you’re not seeing any neurons, the neurons’ cell bodies, those pyramidal cells I’ve been talking about, are these black dots, so you’re only seeing the astrocytes, and you see the total inadequacy of three colors This is not one gigantic red astrocyte, and this is not one big yellow astrocyte

surrounding a blue one In fact, the astrocytes are probably this size, the size of this blue one Three colors are not enough to disambiguate every cell in the nervous system However, then Jean made this construct, in many of the animals, rather than a single copy of the construct being inserted into the genome, many copies in tandem were inserted right next to each other And each of them would randomly either stay red, turn green, or turn blue So you might have a cell, and this neuron has three copies that don’t recombine and one copy that turned green, and one that turned blue, that cell might be a brownish-orange Another cell might have a blue, a red, and three green, and be a sort of spearmint color And in most of the animals he made, that was what he had He had multiple tandem copies, and he generated images as shown in the following picture This is again the hippocampus and the cerebral cortex You can already see there’s a lot more color here If we zoom up on the dentate gyrus you see just this huge number of colors So this has been really useful as a screensaver, lots of people have been interested in these pictures ’cause they’re so pretty And in certain animals, in Drosophila and in zebrafish, for example, and in other parts of the body like in skin and in gut, these kinds of tools have been quite useful, but in the central nervous system, so far, of mammals, they haven’t been ideal and we’ve had to work harder to make these tools better I’m gonna tell you a little bit about the way we’ve made them better in a second But one thing I want to point out is that despite this diversity of colors, really this is only three images You only need a Blue Fluorescent Protein image, and you superimpose on top of that a Yellow Fluorescent Protein, and in this case a green image, and we superimpose on top of that a Red Fluorescent Protein image of the same dataset, and one gets this wide range of colors These colors are quite beautiful, especially when you’re looking at axons that travel long distances This provides us with a way of seeing these axons over a very long distance, and in fact these are images, this is an image from one of these new second-generation Brainbow mice, where the labels are more intense and more varied in color than in the previous ones I was very impressed when we saw this, I saw we had invented something tremendous until someone reminded me that we did not invent this idea This idea had been invented previously, by computer scientists who were trying to trace out the wires in thinking machines One of the features in the new Brainbow mice that’s different from the original is that the three fluorescent proteins now are no longer variants and mutations of a single fluorescent protein, which were nicely distinct as colors, but quite difficult to discriminate in terms of antibodies, so we couldn’t enhance with antibodies Whereas the new Brainbow’s we’re making, we’re using three fluorescent proteins from three different marine animals, and we have then distinct antibodies against each of those each of those from a different mammal, and then a secondary antibody with Alexa dyes to each of those, so that generates wide ranges of color So that’s one big advantage, is the fact that we have the ability to use antibodies to enhance the color A second advantage is that these are membrane-labeled fluorescent proteins, rather than labeling the cytoplasm And this is, for example, a reconstruction of those same axons where they get to muscle, and each of these things at the end, these pretzel-shaped objects, are neuromuscular junctions Because they are farnesylated fluorescent proteins, they’re only in the membranes And what that does is it means that the cell bodies are not very bright, whereas before the cell bodies would swamp the intensity of the finest processes Now all the processes have basically the fluorescence per square micron This in fact is a cross-section of this little region in here, showing that the fluorescent proteins are just in the membranes of these cells Another interesting difference from the original Brainbow is that there is no default color anymore In these, that Dawen Cai made, you can see that the first color is actually a non-fluorescent color,

and then the three fluorescent colors each only occur after recombination In this way, cells that do not recombine are black and the cells that recombine are going to me some mixture, depending on how many copies of this transgene, of these three different fluorescent proteins So this is a confocal reconstruction through the cerebellum, and you see just this wide range of beautiful colors and you can see probably that the membrane proteins are mainly in the membranes of these Purkinje cell dendrites, not in the cytoplasm as much So, these are tools that would allow us now to do some kinds of connectomics studies, and I’d like to end at this point with these techniques and say that what one can do is take advantage of color in certain parts of the nervous system to understand the way neural connections are put together in development, and perhaps get a hint of what an engram is like And I’d like to talk about that in a subsequent lecture Thank you