Searching for the engram

Scientists are getting close to finding the engram, where memories are stored in the brain
14 December 2014
Presented by Kat Arney


For many years scientists have been searching for the mysterious engram - the place in the brain where memories are kept. And thanks to advances in genetics and neurobiology, it looks like they're now getting close to finding it. Plus, contagious cancers in dogs and devils, and our gene of the month prefers the cold.

In this episode

Makes the hairs on the back of your neck stand up...

01:02 - Sheena Josselyn - Making fear memories

Sheena Josselyn is searching for the engram - the mysterious place in the brain where memories are stored, focusing on fear memories.

Sheena Josselyn - Making fear memories
with Sheena Josselyn, Sick Kids Toronto

Kat - This month I'm reporting back from the Genetics Society autumn meeting, held at the Royal Society at the end of November, which focused on the genetics and neurobiology of learning and memory. Several of the talks focused on the hunt for something called the 'engram' - the exact part of the brain where memories are stored. Opening the meeting was Sheena Josselyn, from the Hospital for Sick Children in Toronto, who told me about the search.

Sheena -  So, it really starts with, how do we store and use information? Everyone knows the brain is involved, but exactly, where in the brain?  People have been looking where in the brain a memory has been stored for centuries.  I think we're starting to corner where exactly a memory traces in the brain.

Kat -  In your talk, you used the wonderful word 'engram'.  What's the search for the 'engram'?

Sheena -  So, engram is a really old word, first proposed by Richard Semon who was a German scientist.  He thought engram was those bits in the brain that store a particular memory.  So engraphy was the process of writing down a memory in the brain.

Kat -  So say, I was to go somewhere, it would be written into a little particular bit of my brain.

Sheena -  Absolutely.  I mean, it's not exactly a snapshot, but you can sort of think of it as a little snapshot in your brain.

Kat -  Like a photo album.

Sheena -  Exactly.  Just that our memory is not like a photo album, but for this sort of analogy, that's fair enough.

Kat -  So, in the intervening time, where have we come in understanding this and how are you trying to figure out where in the brain our memories are made.

Sheena -  Well, it turns out that the sort of history of looking for the engram or the memory traces is pretty colourful has a wonderful past and a lot of people have looked.  It's very elusive.  People can't seem to find exactly where in the brain a memory is stored and it's been sort of frustrating scientists for generations.  But now, with some more modern techniques, we can sort of look at turning on and off populations of cells in the brain to really try and corner the engram.  I don't know if we've caught it, but it's certainly cornered.

Kat -  What do we know so far about, I guess, what a memory is, what does it look like in the brain?

Sheena -  So, it's sort of like describing what an elephant is by grabbing little parts of it.  We know parts about the memory, but we can't describe the entire thing.  We know memories are stored in collections of brain cells or neurons in the brain.  We know that they can become tightly bound together so that they become an ensemble or group of neurons.  We know that these neurons are pretty sparse.  So, very few of them across the brain can hold the memory.  Other than that, we're still trying to find out more about the memory.  We haven't exactly figured out what it is yet and how exactly it's formed.

Kat -  So, tell me a bit about your work and you're working on the response to fear.  How are you trying to dissect that and figure out where these fear memories are stored?

Sheena -  We're really interested in how a fear memory is encoded in the brain.  We know that it's encoded primarily in one particular area of the brain called the amygdala.  It's been known for a while that that area is really important in fear.  What we're trying to do is turn on and turn off populations of cells in the amygdala to try and turn on a memory or turn off a memory.  In that way, really get how the memory is formed in the brain.

Kat -  So, how are you actually experimentally doing this?  How do you test for fear and fear memories?

Sheena -  So, it's really easy in people.  You can ask them, you can look at them, "Are you afraid?"

Kat -  Aargh!

Sheena -  Exactly!  Just looking at their face, it's amazing.  But we use experimental rodents and it's really hard when you ask them, "Are you afraid?"  They just sort of look at you blankly.  So, what we do is try and tap into their responses, what they normally do when they're afraid.  It turns out if you're a small rodent like a mouse, when you're afraid you display freezing.  So, if you're being predated by a cat, you sort of stop and you pretend like you're not moving and then hopefully, the cat will walk on past you.  So, we use this response called freezing to find out when a mouse is afraid.

Kat -  And by understanding this, what have you found out so far about how this freezing response to fear works?

Sheena -  Well, it's really interesting.  We've been sort of playing around with how freezing works and how we can turn it on and turn it off by just playing around with a very, very small population of cells in the amygdala.  What we found is that a memory is very sparsely encoded.  So, very few cells in the brain can hold a memory and they can hold multiple memories.  So now, what we're trying to do is figure out how different memories interact in the brain, how they can be linked in the brain, just by looking at how we can manipulate cells.

Kat -  And in your talk, you presented some really nifty techniques that are finally enabling researchers to really manipulate at the genetic level, turn genes on and off, turn cells on and off really specifically?  How are you employing some of those new techniques?

Sheena -  The research world has really been opened up with techniques.  That's sort of the decade of techniques in neuroscience and in genetics.  We're trying to take advantage of the different tools that are available.  So, we use viral vectors to express normal genes or overexpress genes.  We can also express optogenetics that make a cell responsive to light.  We can use chemical genetics to make a cell responsive to different chemicals.  By using, really taking advantage and pulling in from all different areas, we can try and use different techniques to manipulate our system.

Kat -  And so, using all these things together, can you paint me a picture I guess of how you think a fear memory gets written into the brain from what we know so far?

Sheena -  So, what we do is we pair a specific tone with a mild electric shock.  Not enough to injure an animal, but certainly enough to make them afraid.  So, we think that when these two stimuli come together that neurons that express either the tone memory or express the fear memory are co-activated.  This co-activation binds these neurons together so that any time these neurons are active together in a group, an animal will display fear.

Kat -  So, they hear the sound, you give them a little electric shock, these two groups of cells that switch on and then forever more, they're entwined.

Sheena -  Exactly.  So, any time you excite any part of this trace, presumably, the entire trace becomes active and the animal becomes afraid.  It's like, one little thing can trigger your memory and then bring the whole memory to life.  We think this is how that happens.

Kat -  So, when they hear the sound again, they'll freeze and feel afraid.

Sheena -  Absolutely.

Kat -  And so, on a kind of a genetic or molecular level, are we getting to close to understanding where the engram is, where the fear is encoded in the brain?

Sheena -  Absolutely.  We're really getting close.  I think this is not only important to understanding how a fear memory is made, but understanding how memories are made, how we can restore memories in patients that have too much memory like posttraumatic stress disorder, or patients that have too little memory like in Alzheimer's disease and really understanding the fundamentals will help us translate these findings into helping people.

Kat -  The work that you do is in mouse.  Where is the field coming to in trying to understand these processes in humans?

Sheena -  The mouse work in the basic research work, I think it's critical to informing what we do with people.  I work in a children's hospital and I don't see us doing any of these techniques in people right away.  But it tells us how we should approach the problem, how can we make drugs, target a specific area of the brain rather than systemically treating a person.  I think that these sort of basic research findings will really inform how we go forward with people.

Kat - That was Sheena Josselyn, from the Hospital for Sick Children in Toronto.

This image shows a 0.1 x 0.03 inch (2.5 x 0.8 mm) small Drosophila melanogaster fly.

08:30 - Scott Waddell - How fruit flies choose

Scott Waddell is figuring out how fruit flies decide what to do when faced with a range of choices, such as eating or drinking.

Scott Waddell - How fruit flies choose
with Scott Waddell, University of Oxford

Kat - Also at the Genetics Society autumn meeting we heard from Scott Waddell from the University of Oxford, one of the scientific organisers of the conference. He's studying tiny fruit flies in order to figure out how they decide what to do when faced with conflicting choices. I asked him to explain what he's up to.

Scott -  The general idea is to try and understand what the neural mechanisms are in the brain that allow an animal to do the right thing at the right time.  We, more often than not, focus on hunger on thirst-directed behaviour.  So, how does an animal approach food when it's hungry and not approach food when it's satiated.  How does it look for a drink of water when it's thirsty rather than look for a piece of food.  So, how does it select and prioritise one behaviour over another?

Kat -  So, I guess when it boils down to it, we kind of eat, drink, sleep, reproduce.  How do we decide what to do?

Scott -  Exactly.  So obviously, if all of these possibilities provide the animal with a conflict, they have to decide which of these potential things should I bother engaging at any given time.  Of course, if they're more sleepy than hungry, then the thing they should do to promote their survival is to sleep rather feed and vice versa.

Kat -  So, how are you trying to understand what's going on here?  What organisms are you looking at?

Scott -  So, we specifically use the fruit fly and the reason for that is that the genetics and the small nervous system allow us to bring these mechanisms down to a cellular resolution.  The genetics in turn allows us to manipulate molecules within these cells.

Kat -  So, what are you actually looking at?  How do you test what's going on when a fly is making a decision what to do?

Scott -  So typically, we train flies with an odour accompanied with a food reward or an odour accompanied with a water reward.  And then the thing that we actually measure is when given a choice between two odours, we ask which one do they prefer.  Obviously, in an experiment where we've rewarded one of the odours with food, we would expect them to approach that food associated odour in the later test.  The same kind of thing is true for water associated odours.

Kat -  It seems incredible to me that something as small as a fruit fly and as simple I guess as an insect like that, you can train it to do something.

Scott -  You can train them very well and the memory that you form is incredibly robust, lasting a few days.

Kat -  So, how do you then try to go in at a molecular level and work out what's going on when they've learned to distinguish and to go to the smell that they like?

Scott -  So, the molecular level was first approached firstly with a classical mutagenesis approach like everything else in flies, people just mutagenize them.  But instead of looking for morphological defects on the exterior - a damaged wing, slightly strangely shaped eye or so forth - people screen for flies that couldn't remember.  People don't so often use that kind of approach now.  Instead, we use genetic based approaches that allows to manipulate specific neurons.  These tools allow us to switch neurons on and off.

Kat -  What does a fly's brain actually look like?  How do you study it and look inside what's going on?

Scott -  So, we can just pop out the head capsule, kind of like shelling a pea and look at it under the microscope or in principle, we can also image neural activity in the brain of a live fly just by peeling off a little piece of the head capsule and looking at the brain with a microscope in the live animal.

Kat -  So, you kind of pin it down, for want of a better word, prise open its skull and open what the nerves are doing.

Scott -  Pretty much, yeah.

Kat -  That seems quite fiddly to me.

Scott -  It requires some skill.  It's true.

Kat -  So, when you do this, when you're actually looking at the brain and seeing what's lighting up, what's it telling you about how flies are making these kinds of decisions in their life?

Scott -  That's a difficult question.  It depends which neurons you're looking at.  So, up until now, we've mostly looked at neurons that represent values.  So, good or bad event in a fly's life and then we can see that certain neurons are activated by food rewards for example.  In other experiments, we've looked at neurons that required for the flies to do the appropriate thing with its memory.  Their assumption is the activation of that neuron is part of the process where the brain is making the appropriate decision to either run away from something or run towards something.

Kat -  In terms of actually starting to understand more general neural mechanisms, how organisms learn, do you think some of the things that you found out will apply in higher organisms like humans, mammals?

Scott -  I'm quite sure.  I mean, the conservation of genes tells us that many of the processes at least use conserved molecular mechanisms and certainly, some of the work we've done has uncovered conserved molecules - so the fly equivalent of neuropeptide Y which is involved in energy homeostasis in mammals is clearly involved in food-seeking related behaviours in the fruit fly dopamine which is the known reward signal in mammalian brain that is clearly the reward signal in the fly brain and so on.  So, I think it's going to be generally informative.

Kat -  We've heard at the meeting, a lot of people talking about the idea of the engram, like where in the brain is this knowledge encoded.  How close do you think we've got so far in uncovering it and how long do you think it will be before we really understand how it works?

Scott -  I think I can be pretty optimistic here.  So, I think we have at least located a synaptic junction, a junction between two sets of neurons where the memory is probably represented.  So, I think that is essentially the engram or part of the engram in the fly brain.  But I think we've already gotten that.  The question is, at what resolution can we actually get to, can we say it's a specific synaptic connection.  So yeah, I think it's only a matter of time.

Kat -  Kind of philosophically, it's a bit strange to think that all the memories we have, everyone we've ever known is just somehow written into the junctions between our nerve cells.

Scott -  I find that quite reassuring that you can explain it rather than not.  But yes, some people...

Kat -  It's kind of a spooky thing.

Scott -  No, but some people are very uncomfortable.  We've been able to explain their life in some sort of physical process in the brain, but it doesn't affect me that way at all.

Kat -  What are you now trying to figure out?

Scott -  We would like to know what the mechanisms are that are involved in changing the efficiency of these synapses.  The same synaptic connection may be bidirectionally changed by either a pleasant learning event or an unpleasant learning event.  And so, we'd like to know how that difference is generated.  As I said, we're ultimately involved in trying to understand how the fly chooses whether to approach a food relevant cue rather than a water relevant cue.  So, that kind of higher interpretetive kind of mechanism and we're very interested in potential individual differences between animals and how that's represented in the neural circuit properties.

Kat - Scott Waddell from Oxford University. 

A Tasmanian Devil

15:40 - Elizabeth Murchison - Dogs and Devils

Elizabeth Murchison studies two rare transmissible cancers, affecting dogs and Tasmanian devils, that can be spread between animals.

Elizabeth Murchison - Dogs and Devils
with Elizabeth Murchison, University of Cambridge

Kat - Also at the Genetics Society autumn meeting, we were treated to a lecture from Cambridge University's Elizabeth Murchison, winner of the society's 2014 Balfour Prize. She works on two unusual cancers: Tasmanian devil facial tumours, and CTVT - canine transmissible venereal tumours. But unlike all other cancers we know of, these diseases are contagious and can be transferred between animals. And unfortunately for Tasmanian devils, it could be driving them to extinction. I asked her to tell me the story.

Elizabeth -  We normally think of cancer as a disease that arises when a cell in our body acquires mutations that causes it to grow into a tumour and sometimes that can spread inside the body.  However, we normally think that cancer doesn't survive beyond the body of its hosts.  So, these two cancers in dogs and in Tasmanian devils have actually acquired adaptations which have allowed them to survive beyond the deaths of the original dog and devil that first gave rise to them and to spread through the population as a transmissible cancer.

Kat -  It sounds like pretty scary stuff.

Elizabeth -  It's pretty scary, yeah, and it's lucky that this has only happened twice that we know of in nature.  But it's really a type of disease that we didn't really know much about previously.

Kat -  So, let's start with the devils.  Tell me a little bit about the Tasmanian devils and this cancer that affects them and what you found out about it.

Elizabeth -  Yeah, so the Tasmanian devil is a carnivorous marsupial, actually, the largest living marsupial carnivore.  It lives only in Tasmania which is just to the south of the mainland of Australia.  The Tasmanian devils are actually threatened with extinction due to the emergence of this transmissible cancer.  The cancer was first observed in 1996 and since then it has spread rapidly through Tasmania and is now affecting, I think more than 60% of their habitat and it's caused massive population declines in the most severely affected areas.  The disease is threatening to cause extinction of the species within only 20 to 30 years.  So, it's a really serious situation for the devils.

Kat -  And seems quite strange for a disease that only appears to have emerged in the past couple of decades.

Elizabeth -  Yes, it's a highly virulent disease.  It's spread by biting between devils.  Devils get tumours on their face or inside their mouth.  When they bite another devil, they actually physically implant the living cancer cells and somehow, they're able to escape the immune system and grow into a new tumour in that next devil.  It seems to be highly transmissible and it causes death of affected devils within only a few months of the appearance of symptoms.  So, it's a really nasty disease.

Kat -  How did people figure out that this was not just a virus that was being transmitted from devil to devil and was actually the cancer cells?

Elizabeth -  The key experiments came from looking at genetics actually.  We found that the genetic patterns that you see in the DNA of the devil cancer are all very similar to other cancers and very different to the genetics of the hosts.  So normally, if a cancer arises from a cell in the body, it would retain most of the genetic variants which are present within that person's normal DNA.  What we see in the devil is that the cancer actually has genetic variants which are more similar to another devil than they are to the cancer zone host.  So, it's really clear from the genetics that it's actually not arising from the devils carrying the cancer, but rather, where it's long time ago from a different devil and has spread this population by biting.

Kat -  This must be pretty exceptional cancer cells to be able to be transmitted and survive like this.  What sort of things do we know about that enable them to survive and that they don't just get rejected by the new host's immune system?

Elizabeth - Yeah.  Well, this is a very good question because everything that we know about immunology suggests that transmissible cancers are impossible because they come from a different individual, they should be rejected.  Somehow, the devil's immune system is not able to reject this cancer.  This is a really active area of research.  Hannah Siddle at the University of South Hampton has recently been doing very ground-breaking work where she's found that molecular markers which are normally expressed on the surface of all cells seem to be downregulated in devil cancer cells, meaning that the immune system doesn't really see the markers at all so is unable to grab on to anything.  In a way, these cancer cells, they're kind of wearing an immune invisibility cloak because they're not expressing the correct markers.

Kat -  Given how rapidly the disease is spreading, how threatened the Tasmanian devils are now, are there any hopes realistically for something like a treatment or a vaccine or just to contain the disease?

Elizabeth -  Of course!  Well, at the moment, there's a lot of effort going into captive breeding programmes.  So, keeping devils alive and breeding in captivity, just in case the disease does cause extinction in the wild.  And this is a really important programme and it's essential to try to keep devils alive and breeding even if they do disappear in the wild.  Of course, we have a lot of hope that there will be eventually a vaccine or a cure for the disease or perhaps the devils themselves might acquire some resistance to the disease.  But unfortunately at the moment, we still haven't got any of these things and it's something that we're really working on very hard.

Kat -  And now, to move from devils to dogs, tell me about the dog tumour that you've been working on?

Elizabeth -  Just like the devils, it's spread by the transfer of living cancer cells between dogs.  But this dog cancer is actually sexually transmitted.  So, the way that cells get transmitted is when dogs are mating with each other.  The tumours which normally appear in the genitalia, tumour cells get transmitted from one dog to another.  And in the same way as we've seen with the devils, these tumour cells somehow don't get rejected and have been able to spread in dogs.  Actually, this tumour is now found all around the world.  Free-roaming dogs which are mating randomly are the reservoir for the disease.  So, this is why we don't see the disease normally in the UK.  However, there are a number of cases of CTVT, this dog cancer in the UK in dogs which have been imported from abroad.  From Romania for instance, there's often a number of cases seen from there.

Kat -  So, when you talk about the devil cancer that's a few decades old, what do we know about the age of this dog cancer and where it came from?

Elizabeth -  Well, as I mentioned, this dog cancer is found all around the world.  So, that suggests that it must be quite old to have spread so widely.  But we were interested in trying to use genetics to try to actually figure out how old it might be.  What we did was to sequence the entire genome of this dog cancer and then to look at a type of mutation which is known to be acquired at a consistent rate throughout the lifetime of a cancer.  Actually, this has been found in human cancers.  Doing this, we were able to discover that this dog cancer may have first arisen in a dog that lived about 11,000 years ago.  So, it's a pretty extraordinary cancer that it arose in that original dog 11,000 years ago.  Rather than dying, when that dog died, its cancer is still alive today, having spread through the population of dogs by the transfer of cancer cells.

Kat -  How did it travel around the world?  What do we know about its routes around the world?

Elizabeth -  It's a really fascinating question.  We don't know exactly where that original dog lived.  It could've been anywhere in the world.  However, we know that it probably only spread around the world about 500 years ago.  So, that original dog probably lived about 11,000 years ago.  Probably, the cancer was confined in some area of the world - we don't know where - for most of its history.  And then about 500 years ago, it seems to have escaped from there and spread widely around the world.  500 years is interesting because it's around the time of the age of exploration where seafarers were spreading globally much more rapidly.  It's interesting to speculate that perhaps some people went to this place where CTVT first arose and took a dog from there and took it around the world and the tumour spread very rapidly.

Kat -  And the rest is genetic history.

Elizabeth -  That's right.

Kat -  Do we know anything about what this original dog who originally had the cancer might have look like from its genetics?

Elizabeth -  If you think about it, the dog that first gave rise to CTVT, its genome, its DNA is actually still alive in the cancer that it gave rise to even though that dog itself probably died 11,000 years ago.  So, we've been able to piece together what that original dog might have been like by looking at its DNA, the DNA of its cancer.  We've been able to determine that probably, it was most closely related to a group of dogs known as the ancient breed dogs.  These are dogs which have a more ancient genetic signature.  They include the east Asian breeds like akitas and African breeds like basenji, and northern breeds like the Alaskan malamute, and the husky.  We don't know exactly which breed it was most closely related to, but it's those types of dogs.  In addition to that, we were able to look at specific genetic regions which are associated with certain traits in dogs like coat colour and morphology, and behaviour.  We were able to determine that the dog that first got CTVT probably had an agouti coat which is a kind of browny greyish coloured coat.  They probably had pointy ears, probably had a straight or wavy coat.  We've been able to determine some of its potential behavioural characteristics.  So for instance, it doesn't seem to have appeared to have had a higher risk of obsessive compulsive disorder.  So, it's really interesting that we've been able to piece together what the dog was like from that cancer that it gave rise to.

Kat -  Obviously, the dog and the devil cancers seem pretty exceptional in the natural world, but are there any lessons we could take about cancer on a wider basis and also, is there a risk that these types of transmissible cancers could arise again?

Elizabeth -  Cancer itself is obviously quite common in humans and in other animals.  It's interesting then to see that transmissible cancers are so rare.  We only know of two of them in nature.  So, that suggests that the process of going from being a cancer in one host to being a transmissible cancer which is spreading through the population is very unlikely to occur.  I think understanding what changes of being acquired by the genomes of these two transmissible cancers in dogs and Tasmanian devils could potentially teach us a lot about how this process has come about. 

In addition I think though, studying transmissible cancers and the genetics of transmissible cancers does give us an opportunity to learn more about the evolution of cancers more generally, and how cancers interact with or escape the immune system more generally including in humans.  Whether or not transmissible cancers could occur in humans, I think it's extremely unlikely because we've only seen transmissible cancers twice in nature ever.  However, I guess it's a possibility we can't discount and it is certainly possible that it could occur one day in the future.  Given that we know of two transmissible cancers in nature, it's highly likely in fact, almost certain that transmissible cancers have occurred many times in the evolutionary path, possibly also, caused the extinction of species in the past.  However, I do think that probably, they have always been a rare occurrence and I guess it's going to be very difficult to prove that any particular species might have gone extinct because of a transmissible cancer.

Kat -  And finally, what next for you?  What are you going to be looking at next?

Elizabeth -  My lab is very interested in understanding the evolution of transmissible cancers.  So, what we're doing now is looking at the genetic changes that have occurred through the population of Tasmanian devil cancers and through the population of dog cancers to understand the mutations that have been acquired so that we can piece together the evolutionary path that these two cancers have taken in their journeys through their respective host populations.  I think this is going to teach us a lot more about how cancers evolve in general and how cancers can escape the immune system and how cancers become transmissible.

Kat - That was Elizabeth Murchison, from the University of Cambridge. 

Dynamin assembling into spirals

27:58 - Gene of the Month - Shibire

Our gene of the month is Shibire, named after the Japanese word for “limbs going to sleep”.

Gene of the Month - Shibire
with Kat Arney

Kat - And finally it's time for our gene of the month, and this time it's Shibire, named after the Japanese word for "limbs going to sleep". Shibire is the name given to an unusual mutation in the fruit fly Drosophila's dynamin gene, which normally makes a protein that helps to release tiny little packets of molecules from nerve cells enabling them to transmit information. In Shibire mutant flies, this protein works normally at regular lab temperature, but stops functioning if it gets over 29 degrees centigrade. As a result, the flies immediately become paralysed as their nerves stop working properly, and drop to the ground. But lower the temperature again and they spring right back as if nothing had happened.

Since it was discovered, researchers have used the temperature-sensitive Shibire mutation to study how fruit flies brains work - it cropped up several times during the Genetics society autumn meeting from scientists trying to figure out how memories are made and stored. You can see a fascinating video of this effect in action by following the link here:


Add a comment