Research Area: Neurobiology

To distinguish contexts, animals think probabilistically, study suggests
Picower Institute
August 3, 2020

Among the many things rodents have taught neuroscientists is that in a region called the hippocampus, the brain creates a new map for every unique spatial context – for instance, a different room or maze. But scientists have so far struggled to learn how animals decide when a context is novel enough to merit creating, or at least revising, these mental maps. In a study in eLife, MIT and Harvard researchers propose a new understanding: The process of “remapping” can be mathematically modeled as a feat of probabilistic reasoning by the rodents.

The approach offers scientists a new way to interpret many experiments that depend on measuring remapping to investigate learning and memory. Remapping is integral to that pursuit, because animals (and people) associate learning closely with context, and hippocampal maps indicate which context an animal believes itself to be in.

“People have previously asked ‘What changes in the environment cause the hippocampus to create a new map?’ but there haven’t been any clear answers,” said lead author Honi Sanders. “It depends on all sorts of factors, which means that how the animals define context has been shrouded in mystery.”

Sanders is a postdoc in the lab of co-author Matthew Wilson, Sherman Fairchild Professor in The Picower Institute for Learning and Memory and the departments of Biology and Brain and Cognitive Sciences at MIT.  He is also a member of the Center for Brains, Minds and Machines. The pair collaborated with Samuel Gershman, a professor of psychology at Harvard on the study.

Fundamentally a problem with remapping that has frequently led labs to report conflicting, confusing, or surprising results, is that scientists cannot simply assure their rats that they have moved from experimental Context A to Context B, or that they are still in Context A, even if some ambient condition, like temperature or odor, has inadvertently changed. It is up to the rat to explore and infer that conditions like the maze shape, or smell, or lighting, or the position of obstacles, and rewards, or the task they must perform, have or have not changed enough to trigger a full or partial remapping.

So rather than trying to understand remapping measurements based on what the experimental design is supposed to induce, Sanders, Wilson and Gershman argue that scientists should predict remapping by mathematically accounting for the rat’s reasoning using Bayesian statistics, which quantify the process of starting with an uncertain assumption and then updating it as new information emerges.

“You never experience exactly the same situation twice. The second time is always slightly different,” Sanders said. “You need to answer the question: ‘Is this difference just the result of normal variation in this context or is this difference actually a different context?’ The first time you experience the difference you can’t be sure, but after you’ve experienced the context many times and get a sense of what variation is normal and what variation is not, you can pick up immediately when something is out of line.”

The trio call their approach “hidden state inference” because to the animal, the possible change of context is a hidden state that must be inferred.

In the study the authors describe several cases in which hidden state inference can help explain the remapping, or the lack of it, observed in prior studies.

For instance, in many studies it’s been difficult to predict how changing some of cues that a rodent navigates by in a maze (e.g. a light or a buzzer) will influence whether it makes a completely new map or partially remaps the current one and by how much. Mostly the data has showed there isn’t an obvious “one-to-one” relationship of cue change and remapping. But the new model predicts how as more cues change, a rodent can transition from becoming uncertain about whether an environment is novel (and therefore partially remapping) to becoming sure enough of that to fully remap.

In another, the model offers a new prediction to resolve a remapping ambiguity that has arisen when scientists have incrementally “morphed” the shape of rodent enclosures. Multiple labs, for instance, found different results when they familiarized rats with square and round environments and then tried to measure how and whether they remap when placed in intermediate shapes, such as an octagon. Some labs saw complete remapping while others observed only partial remapping. The new model predicts how that could be true: rats exposed to the intermediate environment after longer training would be more likely to fully remap than those exposed to the intermediate shape earlier in training, because with more experience they would be more sure of their original environments and therefore more certain that the intermediate one was a real change.

The math of the model even includes a variable that can account for differences between individual animals. Sanders is looking at whether rethinking old results in this way could allow researchers to understand why different rodents respond so variably to similar experiments.

Ultimately, Sanders said, he hopes the study will help fellow remapping researchers adopt a new way of thinking about surprising results – by considering the challenge their experiments pose to their subjects.

“Animals are not given direct access to context identities, but have to infer them,” he said. “Probabilistic approaches capture the way that uncertainty plays a role when inference occurs. If we correctly characterize the problem the animal is facing, we can make sense of differing results in different situations because the differences should stem from a common cause: the way that hidden state inference works.”

The National Science Foundation funded the research.

Seemingly similar, two neurons show distinct styles as they interact with the same muscle partner
Picower Institute
July 7, 2020

A new study by MIT neuroscientists into how seemingly similar neuronal subtypes drive locomotion in the fruit fly revealed an unexpected diversity as the brain’s commands were relayed to muscle fibers. A sequence of experiments revealed a dramatic difference between the two nerve cells – one neuron scrambled to adjust to different changes by the other, but received no requital in response when circumstances were reversed.

The findings published in the Journal of Neuroscience suggest that these subclasses of neurons, which are also found abundantly in people and many other animals, exhibit a previously unappreciated diversity in their propensity to respond to changes, a key property known as “synaptic plasticity.” Synaptic plasticity is considered an essential mechanism of how learning and memory occur in the brain, and aberrations in of the process are likely central to disorders such as autism.

“By seeing that these two different types of motor neurons actually show very distinct types of plasticity, that’s exciting because it means it’s not just one thing happening,” said senior author Troy Littleton, a member of The Picower Institute for Learning and Memory and Menicon Professor of Neuroscience in MIT’s Departments of Biology and of Brain and Cognitive Sciences. “There’s multiple types of things that can be altered to change connectivity within the neuromuscular system.”

Tonic and phasic neurons

Both of the neurons work in the same way, by emitting the neurotransmitter glutamate onto their connections, or synapses, with the muscles. But these two neurons do so with different styles. The “tonic” neuron, which connects only to a single muscle, emits its glutamate at a constant but low rate while the muscle is active. Meanwhile, the “phasic” neuron connects to a whole group of muscles and jumps in with a strong quick pulse of activity to spring the muscles into action.

Heading into the study Littleton and lead author Nicole Aponte-Santiago were curious to explore whether these different neurons compete or cooperate to drive the muscle fibers, and if they exhibited different plasticity when their functions were altered. To get started on what ultimately became her dissertation research, Aponte-Santiago developed the means to tailor genetic alterations specifically in each of the two neurons.

“The reason we were able to answer these questions in the first place was because we produced tools to start differentially manipulating one neuron versus the other one, or label one versus the other one,” said Aponte-Santiago, who earned her PhD in Littleton’s lab earlier this spring and is now a postdoc at the University of California at San Francisco.

With genetic access to each neuron, Aponte-Santiago distinctly labeled them to watch each one grow in fly larvae as they developed. She saw that the tonic neuron reached the muscle first and that the phasic one connected to the muscle later. She also observed that unlike in mammals, the neurons did not compete to control the muscle but remained side by side, each contributing in its characteristic way to the total electrical activity needed to drive movement.

To study the neurons’ plasticity, Aponte-Santiago employed two manipulations of each neuron. She either wiped them out completely by making them express a lethal protein called “reaper” or she substantially tamped down their glutamate activity via expression of tetanus toxin.

When she wiped out the phasic neuron with reaper, the tonic neuron quickly stepped up its signaling, attempting to compensate as much as it could. But in flies where she wiped out the tonic neuron, the phasic neuron didn’t budge at all, continuing as if nothing had changed.

Similarly when Aponte-Santiago reduced the activity of the phasic neuron with the toxin, the tonic neuron increased the number of boutons and active zone structures in its synapses to respond to the loss of its partner. But when she reduced the activity of the tonic neuron the phasic neuron again didn’t appear to respond.

In all the experiments, the muscle received less overall drive from the neurons than when everything was normal. And while the phasic neuron  apparently didn’t bother to make up for any loss on the part of the tonic neuron, the tonic neuron employed different means to compensate – either increasing its signaling or by enhancing the number of its connections on the muscle – depending on how the phasic neuron was diminished.

“It was quite intriguing that Nicole found that when the phasic input wasn’t there, there was a unique form of plasticity that the tonic neuron showed,” Littleton said, “but if the phasic neuron was there and wasn’t working, the tonic neuron behaved in a very different way.”

Another intriguing aspect of the study is the role of the muscle itself, which may be an active intermediary of the plasticity, Littleton said. The neurons may not sense when each other have been wiped out or inactivated. Instead the muscle appears to call for those changes.

“Even though a muscle has two distinct inputs, it can sort of uniquely control those two,” Littleton said. “When the muscle is getting glutamate, does it know whether it is coming from the tonic or the phasic neuron and does it care? It appears that it does care, that it really needs the tonic more than the phasic. When the phasic is gone it shifts some of the plasticity, but when the tonic is gone the phasic can’t do much about it.”

In new work, the lab is now looking at differences in gene expression between the two neurons to identify which proteins are responsible for the unique properties and plasticity of the tonic and phasic neurons. By defining the genetic underpinnings of their unique properties, the lab hopes to begin to get a handle on the molecular underpinnings of neuronal diversity in the brain.

In addition to Aponte-Santiago and Littleton, the paper’s other authors are Kiel Ormerod and Yulia Akbergenova.

The National Institutes of Health and the JPB Foundation supported the study.

The story behind the science: How discovery develops
Picower Institute
June 29, 2020

Scientific discoveries can sometimes seem like products on a store shelf. Packaged neatly in the wrapping of a journal article or maybe a news story, there remain few hints of what they really took to produce – the struggle and surprises, the ingenuity and serendipity, the toil and triumph. Perhaps it’s no wonder that many members of the public (7 in 10, by one National Science Foundation measure) feel at least somewhat unclear about how scientists know what they know or do what they do.

To deconstruct and perhaps demystify discovery, let’s unwrap the inside story of a paper published by the lab of Troy Littleton, Menicon Professor of Neuroscience in the Departments of Biology and Brain and Cognitive Sciences. The study reported important findings both about a possible mechanism of seizures in epilepsy, which affects 60 million people worldwide, and also the underappreciated relationship between neurons and the brain cells called glia that help them function. Through four years of work led by former postdoc Shirley Weiss, Littleton’s team thoroughly unraveled the complex breakdown that makes fruit flies with a genetic mutation prone to seizures and showed multiple ways to intervene, including with human medicines. Published in April 2019 in eLife, a far-reaching “open access” journal that is free for all to read, the paper has since been viewed thousands of times.

Figuring out the puzzle of exactly how the mutation made glial cells fail to prevent seizures was a source of particular excitement for Weiss. Littleton adds that the discovery could open up new strategies for developing drugs to address epilepsy in humans, which the flies model well.

“One of the long-term motivations for the field in general, not just our lab, is there might be pharmacological access to glial cells that might have less side effects than would happen if you target neurons directly,” he said.

Too much excitement

First, a little biological background. Neurons are electrical. Their participation in the brain circuits that guide behavior, emotion, reasoning and memory depends on how they build up or dissipate electrical charge by taking in or ejecting ions of calcium, potassium and sodium. If they remain too electrically charged up because of an imbalance of these ions, they can become hyperactive and, in groups, produce seizures. In this study, it turned out that a certain kind of glial cells were responsible for regulating the balance of potassium ions around their neuronal neighbors to help govern their electrical charge and activity. The mutation, Weiss discovered, caused the glia to leave too much potassium outside the neurons, making it harder for the neurons to get rid of the potassium they had built up inside when they were electrically active. Without the ability to get their potassium out, the neurons stayed too excited, producing seizures.

The lab first discovered the mutation, which the Louisianan Littleton named “zydeco,” in 2005 when a team led by Zhuo  Guan used genetic screening techniques Littleton learned when he was a postdoc in the 1990s at the University of Wisconsin. The team’s broader goal was to learn about how neurons communicate with each other, so they looked for flies with mutations that either shut the process down, leading to a readily observable symptom of paralysis, or amped it way up, leading to seizures. Zydeco fell into that second category, making fruit flies seize dramatically when stressed by heat or by getting jostled around.

“It was so striking, it was hard to ignore,” Littleton said. “Whatever this gene was, it was doing something very important in the brain.”

When Weiss joined the Littleton lab after earning her PhD at Hebrew University in Jerusalem in 2013, the lab had just  published a new paper about zydeco. It was a long-awaited follow-up. Zydeco disrupted a gene on the fly’s X chromosome which at the time was poorly understood. Led by former graduate student Jan Melom, the lab finally was able to clone the gene that was mutated in Zydeco and showed that it specifically affected “cortex glial” cells and that it caused them to retain too much calcium. But what remained completely unclear was how this made the neurons that those glia contact so susceptible to seizures.

Though Weiss had a research specialty in studying calcium in brain cells, at first she worked on a few other ideas. But because Melom had left the lab, Weiss soon picked up the zydeco baton. In so doing, she was taking on what would become an especially extensive effort involving scores of experiments and a vast array of techniques, some of which she would have to learn along the way.

No hypothesis needed or heeded

Based on the 2013 findings, Littleton had formulated a working hypothesis about what might be going on to cause the seizures. He figured the excess of calcium in the cortex glia probably caused them to emit too much of some kind of signal to the neurons, in turn causing them to remain too active.

That turned out to be wrong, Littleton acknowledges with a smile.

“What Shirley did was to disprove my very strong impression of what was actually happening,” Littleton said. “Sometimes that’s very difficult to do. Once you have an idea of how you think the biology is working, that can reinforce the sorts of experiments you do and affects how you think about the project. It was very exciting in the end that Shirley was able to get past my pre-conceived notions and figure out what was really happening.”

The team didn’t fall into that trap because the experimental approach they chose didn’t depend on what they thought. Weiss’s key initial inquiries were based on a wide-open, free-ranging manipulation of the zydeco flies’ genes. Her strategy was to “knock down” or interfere with the cell’s ability to make use of 847 different genes covering a wide variety of potentially relevant glial cell functions. If knocking down any particular gene stopped the seizures, that would give them a huge clue about how the seizures happen. And whatever worked, if anything, would work regardless of anyone’s guesses up front.

“The great thing about using forward genetics is you don’t have to have a very strong hypothesis,” Weiss said. “You can let the genetics lead the way. I tried to be hypothesis free and to be as unbiased as I could be.”

The knockdown screening yielded about 50 genes where interference totally or partially alleviated the seizures. One in particular squared well with what Melom had observed about a specific cellular process (scientists call it a “pathway”)that related to handling calcium.

Around that time, though, life outside the lab intervened. In February 2015 Weiss and her husband Kfir Sharabi, also a postdoc, and their then four-and-a-half year old daughter, Amit, welcomed their second daughter, Ma’ayan, to the world. With two young kids and the rest of her family in Israel, Weiss came back from maternity leave and got back to investigating the most promising hits of the knockdown screen.

A calcium conundrum

The particular hit related to calcium that caught Weiss’s eye was a gene was called CanB2. Zydeco flies with that gene knocked down experienced no more seizure troubles at all. Moreover, she found that it was specifically helpful to knock it down in cortex glia and that knocking it down in healthy flies didn’t do any discernable harm.

So what does CanB2 do? In general the gene, along with two others, make a protein called calcineurin. No one had ever characterized what calcineurin does in glia. If Weiss could become the first to figure that out, she could whatever problems the zydeco mutation causes.

By manipulating all three calcineurin genes, Weiss was able to confirm that calcineurin activity was indeed crucial for zydeco seizures. She engineered cortex glia so that a glowing green protein would start to be expressed when calcineurin was active. She could see the protein light up under the microscope. This told her there was a lot more calcineurin activity in zydeco mutant brains than in normal fly brains. Apparently, the excess calcium in the cortex glia correlated with increased calcineurin activity.

There are human medicines that ratchet back calcineurin activity. They are typically used to suppress the immune system after a transplant. Weiss wanted to see whether they could reduce seizures in the zydeco flies. When she fed them the drugs the seizures did subside, providing a clear demonstration that intervening in this glial pathway could hold promise for drug development.

Any such effort, to be truly well targeted, would require more than just an association between calcineurin and seizures. Weiss was determined to discover the mechanism that linked the two.

Figuring out what that mechanism was and how it led to seizures, would turn out to be the heart of the discovery and the most challenging phase of her four year endeavor.

A potassium epiphany

Weiss needed to find out what process this excess calcineurin activity might be putting into overdrive. She went back to genetics. She performed a screen to knock down direct targets of calcineurin. It didn’t appear to yield anything helpful. She did another screen of pathways where calcineurin was implicated. In that case, a process called “endocytosis” came up and sure enough, Weiss found that by inhibiting the process in the cortex glia she could again stop seizures in the zydeco flies. Endocytosis is how cells ingest material from their surroundings, including regulating the content of their cell surface membrane proteins. The process can therefore affect the proteins they employ on the membrane to interact with the environment outside the cell. Excess endocytosis could mean that the way by which cortex glia interact with their environment is altered in zydeco, perhaps affecting the neurons they support. But how might that matter in this case?

Weiss struggled with this question for months. She received feedback and advice in meetings with Littleton and in lab meetings where the members discuss, challenge and refine ideas. The discussions were helpful, but it turns out that the key breakthrough came from Weiss attending a conference in Cold Spring Harbor, N.Y. in July, 2018.

Among Weiss’s genetic screen results of calcineurin targets was a gene called “SAND” that makes a protein in flies called “sandman” (the human version is called TRESK). Sandman, when deployed to the cell membrane, forms a channel (picture a portal though the cell’s surface) that allows a cell to bring in potassium ions from outside. At first this result didn’t strike Weiss as all that notable, but at the conference, potassium channels kept coming up as a topic in talks. An idea started to percolate as she took notes. Then at the conference posters she started talking with a scientist who said that problems with potassium channels in glial cells have been linked to epilepsy. Potassium channels apparently merited another look.

“I already had the result,” Weiss said. “I just didn’t connect the two dots.”

One of her screens indeed showed that knocking down SAND in healthy flies caused seizures just like the ones seen in zydeco mutants. Further genetic manipulations confirmed that SAND knockdown and zydeco affected the same pathway in cortex glia cells.

By September 2018, a new hypothesis was emerging: Elevated calcium in cortex glia triggered excess calcineurin activity, which spurred increased endocytosis that hindered sandman’s intake of potassium. This came at a good time as Weiss was nearing the point where she had to start thinking of wrapping up her postdoctoral appointment at MIT. The hypothesis, and the evidence she’d built up, seemed enough to submit a paper to a journal.

Always mindful that a paper was the goal, Weiss had been writing as she went and developing the key figures. When she had a draft done, Littleton then set to polishing it and giving her feedback. eLife wasn’t the first journal they submitted the paper to, but the editors received it enthusiastically. All three of the scientists who reviewed the manuscript for eLife, however, said the same thing: If endocytosis was pulling sandman back from the membrane of the cortex glia, thereby disrupting its ability to take in potassium ions, they wanted to see it happening. Weiss and Littleton not only agreed with that critique, they had even anticipated it.

“You sort of know your own holes in the story,” Littleton said. “This is what we were planning to do next anyway.”

Since sending in the paper, Weiss had already produced those smoking gun images, showing that in zydeco cortex glia, sandman was much less abundant on the membranes than in the non-mutant flies. This cemented the argument, neatly wrapped up in the paper, that neurons become more susceptible to seizing when zydeco cortex glia, saddled with too much calcium and resulting calcineurin activity, overdo the endocytosis of sandman potassium channels, leaving too much potassium outside of neurons, causing increased excitability and the onset of seizures.

Since publication, the paper has garnered some mentions in the scientific press. It has also earned a new National Institutes of Health grant for Littleton’s lab, where they are following in Melom’s and Weiss’ footsteps to study how calcium levels in glia affect the flux of membrane proteins, not just in disease, but as a matter of course in healthy cells. And for Weiss, the paper impressed funders in Israel, providing her with the money to support her new position where she continues studying glia, calcium and seizures.

It was a hard-earned success. Though their end product is knowledge, scientists spend the vast majority of their time with the unknown. Between the lines of most every paper are years of effort in which scientists persistently asked open questions with open minds so that the evidence could lead them to a discovery they could share with the world.

Like a treasure map, brain region emphasizes reward location
Picower Institute
June 1, 2020

We are free to wander but usually when we go somewhere it’s for a reason. In a new study, researchers at The Picower Institute for Learning and Memory show that as we pursue life’s prizes a region of the brain tracks our location with an especially strong predilection for the location of the reward. This pragmatic bias of the lateral septum suggests it’s a linchpin in formulating goal-directed behavior.

“It appears that the lateral septum is, in a sense, ‘prioritizing’ reward-related spatial information,” said Hannah Wirtshafter, lead author of the study in eLife and a former graduate student in the MIT lab of senior author Matthew Wilson, Sherman Fairchild Professor of Neurobiology. Wirtshafter is now a postdoc at Northwestern University.

Last year, Wirtshafter and Wilson, a professor of biology and of brain and cognitive sciences, analyzed measurements of the electrical activity of hundreds of neurons in the LS and the hippocampus, a region known for encoding many forms of memory including spatial maps, as rats navigated a maze toward a reward. In Current Biology they reported that the LS directly encodes information about the speed and acceleration of the rats as they navigated through the environment.

The new study continued this analysis, finding that while the LS dedicates a much smaller proportion of its cells to encoding location than does the hippocampus, a much larger proportion of those cells respond when the rat is proximate to where the reward lies. Moreover, as rats scurried toward the reward point and back again within the H-shaped maze, the pace of their neural activity peaked closest to those reward locations, skewing the curve of their activity in association with where they could find a chocolate treat. Finally, they found that neural activity between the hippocampus and the LS was most highly correlated among cells that represented reward locations.

“Understanding how reward information is linked to memory and space through the hippocampus is crucial for our understanding of how we learn from experience, and this finding points to the role the lateral septum may play in that process,” Wilson said.

Specifically, Wilson and Wirtshafter interpret the results of the two studies to suggest that the LS plays a key role in helping to filter and convert raw information about location, speed and acceleration coming in from regions such as the hippocampus, into more reward-specific output for regions known to guide goal-directed behavior, such as the ventral tegmental area. In the paper they discuss ways in which the hippocampus and the LS might be wired together to do so. They theorize that the LS may dedicate neurons to receiving reward-related location information from the hippocampus and may blend non-reward location information within neurons also tasked for processing other information such as motion.

“This is supported by our previous work that shows somewhat overlapping populations of place-encoding and movement-encoding LS cells,” Wirtshafter said.

Though it’s easy for most of us to take the brain’s ability to facilitate navigation for granted, scientists study it for several reasons, Wirtshafter said.

“Elucidating brain mechanisms and circuits involved in navigation, memory and planning may identify processes underlying impaired cognitive function in motor and memory diseases,” she said. “Additionally, knowledge of the principles of goal directed behavior can also be used to model context-dependent brain behavior in machine models to further contribute to artificial intelligence development.”

The National Defense Science & Engineering Graduate Fellowship Program and the JPB Foundation provided funding for the study.

Study finds ‘volume dial’ for turning neural communication up or down
Picower Institute
May 6, 2020

Neuroscientists at MIT’s Picower Institute for Learning and Memory have found that a protein acts like a volume dial for the release of neurotransmitters, the chemicals that neurons release across connections called synapses to stimulate muscles or communicate with other neurons in brain circuits. The findings help explain how synapses work and could better inform understanding of some neurological disorders.

Working in the model of fruit flies, the team determined that the protein Synaptotagmin 7 (SYT7), which is also found in humans and other mammals, constrains the number and availability of neurotransmitter-containing blobs, called vesicles, for release at the synapse. Neurons deploy vesicles to sites called “active zones” to release them across synapses, a process called “vesicle fusion.”  When the scientists reduced SYT7, they saw much more neurotransmitter release at synapses. When they increased the protein, neurotransmitter release dropped significantly.

“You can think of this as almost like a radio’s volume dial,” said senior author Troy Littleton, Menicon Professor of Neuroscience in MIT’s Departments of Biology and Brain and Cognitive Sciences. “If a neuron wants to send more signal out all it has to do is basically reduce the levels of SYT7 protein that it is making. It’s a very elegant way for neurons to turn up or down the amount of output that they are giving.”

The study’s co-lead authors are Zhuo Guan, a research scientist, and Mónica C. Quiñones-Frías, who successfully defended her doctoral thesis on the work May 4. She noted that by acting as that volume dial, the protein could change the nature of a synapse’s activity in a circuit, a property called “synaptic plasticity.”

“Syt7 regulates neurotransmission in a dose-dependent manner and can act as a switch for short term synaptic plasticity,” Quiñones-Frías said.

Research scientist Yulia Akbergenova is also a co-author of the study published in eLife.

Synaptic surprise

Important as they are, the study’s findings are not ones the team was originally looking for.

For decades, neuroscientists have known that the synaptotagmin protein family plays key roles in synaptic function. In fact, Littleton’s 1993 doctoral dissertation showed that SYT1 promoted a quick release of neurotransmitters when triggered by an influx of calcium ions. But even with SYT1 disabled, synapses could still release neurotransmitters on a slower timeframe. No one has found what promotes that subsequent slower release, but many scientists had pinned their hopes on it being SYT7.

“That’s been something that the whole field, including my lab, has really been searching for,” Littleton said. “So it was a real surprise when we knocked it out and saw just the opposite of what we expected.”

Mutants and microscopes

To study SYT7 the team focused its experiments on synapses in a well characterized locale: the junction between a fly neuron and muscle. The team not only wanted to see what differences changing the protein’s levels would make in synaptic activity there, but also track how it made those differences.

They changed the amount of SYT7 the neuron could produce by mutating and breeding flies in which the gene was completely eliminated, only one copy could be expressed, or in which the gene was overexpressed, producing more SYT7 than normal. For each of these fly lines they measured the surprising inverse relationship between SYT7 and synaptic transmission.

Also, using a technique the lab invented to visually flag neurotransmitter release every time it happens, they mapped how active individual synapses at the neuron-muscle junction were over time. In flies engineered to produce less SYT7 they saw many more synapses with a high propensity for release than they did in normal flies.

Once they confirmed SYT7’s restrictive role, the natural question was how does SYT7 constrain neurotransmitter release. Synapses are very complex, after all, and crucial aspects of SYT7’s role within that machinery had yet to be characterized.

When they compared synapses in normal flies and those missing SYT7 they didn’t see major differences in anatomy or calcium influx that could explain how SYT7 works to limit release.

They then turned their attention to the cycle in which vesicles release their neurotransmitter cargo and are then sent back into the cell to refill with neurotransmitter before rejoining a pool of vesicles ready for redeployment. Their experiments showed that neurons lacking SYT7 didn’t recycle the vesicles differently but they nevertheless had more vesicles in the readily releasable pool (RRP). Moreover, mutants in which SYT7 was overexpressed substantially limited the vesicles in that pool.

“SYT7 limits release in a dosage-sensitive manner by negatively regulating the number of synaptic vesicles available for fusion and slowing recovery of the RRP following stimulation,” they determined.

The final step was to track down where SYT7 resides in the synaptic machinery. Under the microscope they were able to pin it down in a network of tubes surrounding, but not within the active zones. The vantage point is right where other proteins regulating vesicle trafficking also reside, giving SYT7 a clear opportunity to interact with those proteins to regulate the return of vesicles to the active zones.

Implications for disease and plasticity

Understanding more about SYT7’s role at the synapse in mammals could matter in several ways, Littleton said. Two years ago, researchers showed that the protein is reduced in mice harboring a genetic cause of Alzheimer’s disease. And in February another paper showed that patients with bipolar disorder exhibited lower levels of the protein than people who do not have the disorder. Mice with SYT7 knocked out showed some manic and depressive behaviors.

More fundamentally, Littleton and Quiñones-Frías said, is the flexibility or plasticity it can afford. Because SYT7 regulates neurotransmitter release by slowing down the resupply of releasable vesicles, an increase in its levels can transform a synapse from being the kind that sends out large bursts of signal (and therefore transmits more information) early on and then peters out into one that builds up its signal over time. Such distinctions in release timeframe can make important differences in circuit information processing in the brain.

Although the team was able to identify SYT7’s effect at synapses and show key aspects of how it functions, they still hope to determine the exact mechanism that allows the protein to gate vesicle fusion. That work is ongoing.

The National Institutes of Health and the JPB Foundation provided support for the research.

Neuroscientists find memory cells that help us interpret new situations

Neurons that store abstract representations of past experiences are activated when a new, similar event takes place.

Anne Trafton | MIT News Office
April 6, 2020

Imagine you are meeting a friend for dinner at a new restaurant. You may try dishes you haven’t had before, and your surroundings will be completely new to you. However, your brain knows that you have had similar experiences — perusing a menu, ordering appetizers, and splurging on dessert are all things that you have probably done when dining out.

MIT neuroscientists have now identified populations of cells that encode each of these distinctive segments of an overall experience. These chunks of memory, stored in the hippocampus, are activated whenever a similar type of experience takes place, and are distinct from the neural code that stores detailed memories of a specific location.

The researchers believe that this kind of “event code,” which they discovered in a study of mice, may help the brain interpret novel situations and learn new information by using the same cells to represent similar experiences.

“When you encounter something new, there are some really new and notable stimuli, but you already know quite a bit about that particular experience, because it’s a similar kind of experience to what you have already had before,” says Susumu Tonegawa, a professor of biology and neuroscience at the RIKEN-MIT Laboratory of Neural Circuit Genetics at MIT’s Picower Institute for Learning and Memory.

Tonegawa is the senior author of the study, which appears today in Nature Neuroscience. Chen Sun, an MIT graduate student, is the lead author of the paper. New York University graduate student Wannan Yang and Picower Institute technical associate Jared Martin are also authors of the paper.

Encoding abstraction

It is well-established that certain cells in the brain’s hippocampus are specialized to store memories of specific locations. Research in mice has shown that within the hippocampus, neurons called place cells fire when the animals are in a specific location, or even if they are dreaming about that location.

In the new study, the MIT team wanted to investigate whether the hippocampus also stores representations of more abstract elements of a memory. That is, instead of firing whenever you enter a particular restaurant, such cells might encode “dessert,” no matter where you’re eating it.

To test this hypothesis, the researchers measured activity in neurons of the CA1 region of the mouse hippocampus as the mice repeatedly ran a four-lap maze. At the end of every fourth lap, the mice were given a reward. As expected, the researchers found place cells that lit up when the mice reached certain points along the track. However, the researchers also found sets of cells that were active during one of the four laps, but not the others. About 30 percent of the neurons in CA1 appeared to be involved in creating this “event code.”

“This gave us the initial inkling that besides a code for space, cells in the hippocampus also care about this discrete chunk of experience called lap 1, or this discrete chunk of experience called lap 2, or lap 3, or lap 4,” Sun says.

To further explore this idea, the researchers trained mice to run a square maze on day 1 and then a circular maze on day 2, in which they also received a reward after every fourth lap. They found that the place cells changed their activity, reflecting the new environment. However, the same sets of lap-specific cells were activated during each of the four laps, regardless of the shape of the track. The lap-encoding cells’ activity also remained consistent when laps were randomly shortened or lengthened.

“Even in the new spatial locations, cells still maintain their coding for the lap number, suggesting that cells that were coding for a square lap 1 have now been transferred to code for a circular lap 1,” Sun says.

The researchers also showed that if they used optogenetics to inhibit sensory input from a part of the brain called the medial entorhinal cortex (MEC), lap-encoding did not occur. They are now investigating what kind of input the MEC region provides to help the hippocampus create memories consisting of chunks of an experience.

Two distinct codes

These findings suggest that, indeed, every time you eat dinner, similar memory cells are activated, no matter where or what you’re eating. The researchers theorize that the hippocampus contains “two mutually and independently manipulatable codes,” Sun says. One encodes continuous changes in location, time, and sensory input, while the other organizes an overall experience into smaller chunks that fit into known categories such as appetizer and dessert.

“We believe that both types of hippocampal codes are useful, and both are important,” Tonegawa says. “If we want to remember all the details of what happened in a specific experience, moment-to-moment changes that occurred, then the continuous monitoring is effective. But on the other hand, when we have a longer experience, if you put it into chunks, and remember the abstract order of the abstract chunks, that’s more effective than monitoring this long process of continuous changes.”

The new MIT results “significantly advance our knowledge about the function of the hippocampus,” says Gyorgy Buzsaki, a professor of neuroscience at New York University School of Medicine, who was not part of the research team.

“These findings are significant because they are telling us that the hippocampus does a lot more than just ‘representing’ space or integrating paths into a continuous long journey,” Buzsaki says. “From these remarkable results Tonegawa and colleagues conclude that they discovered an ‘event code,’ dedicated to organizing experience by events, and that this code is independent of spatial and time representations, that is, jobs also attributed to the hippocampus.”

Tonegawa and Sun believe that networks of cells that encode chunks of experiences may also be useful for a type of learning called transfer learning, which allows you to apply knowledge you already have to help you interpret new experiences or learn new things. Tonegawa’s lab is now working on trying to find cell populations that might encode these specific pieces of knowledge.

The research was funded by the RIKEN Center for Brain Science, the Howard Hughes Medical Institute, and the JPB Foundation.

Chimeras offer a new way to study childhood cancer in mice
Eva Frederick | Whitehead Institute
March 5, 2020

In a new paper published March 5 in the journal Cell Stem Cell, researchers in Whitehead Institute Member Rudolf Jaenisch’s lab introduce a new way to model human neuroblastoma tumors in mice using chimeras — in this case, mice that have been modified to have human cells in parts of their nervous systems. “This may serve as a unique model that you can use to study the dynamic of immune cells within human tumors,” says Malkiel Cohen, a postdoc in Jaenisch’s lab and the first author of the paper.

Neuroblastoma is a rare and unpredictable form of childhood cancer that affects around 800 young children in the US each year. Neuroblastoma tumors often occur in parts of the sympathetic nervous system, which includes the nerves that run parallel to the spinal cord and the adrenal medulla, part of the glands that produce hormones such as adrenaline. Neuroblastoma is notoriously hard to study primarily because of its disparate behavior: the tumors often shrink spontaneously in infants, while in toddlers they are highly aggressive and often fatal. “The seeds for the cancer are sown during fetal life,” says Rani George, MD, PhD, an associate professor of pediatrics at Harvard Medical School and a neuroblastoma researcher and physician at Dana-Farber Cancer Institute and Boston Children’s Hospital, and a co-senior author on the paper. “For obvious reasons, you can’t really study the development of these tumors in humans.”

Until now, researchers didn’t have many realistic ways to study these tumors in animal models, either. They could create transgenic mice with cancer-causing genes, but the resulting tumors were mouse tumors, not human ones, and had some key differences. Another method involved taking human tumor cells and implanting them in a mouse — a process called xenotransplantation — but that only worked in mice with compromised immune systems, and didn’t allow researchers to study how the tumors formed in the first place or how they interacted with a fully functioning immune system. “This is where we think the new model is a perfect fit,” said Stefani Spranger, PhD, an assistant professor of Biology at the Massachusetts Institute of Technology (MIT) and the Koch Institute for Integrative Cancer Research at MIT and a co-senior author on the paper.

Human-mouse chimeras have been used in the past to study Alzheimer’s disease and brain development. Jaenisch, who is also a professor of biology at MIT, and his lab had been working for years to create chimeric mice with human cells in the neural crest — the group of developing cells that go on to form parts of the sympathetic nervous system — and published their findings in 2016. “In this study, we hoped to use these mice with human neural crest cells to study how neuroblastoma tumors form and respond to immune system attacks,” Jaenisch says.

To create these chimeric mice, Cohen and coauthors at MIT’s Koch Institute and the Dana-Farber Cancer Institute first engineered human pluripotent stem cells to express two genes known to be abnormal in neuroblastoma, MYCN and mutated ALK, and modified them so they became neural crest cells, from which human neuroblastomas are derived. The genes could be turned on and off with the addition of doxycycline, an antibiotic. They also inserted the gene for eGFP, a brightly glowing fluorescent protein originally isolated from jellyfish. This would allow the team to tell whether the cells were spreading correctly through the bodies of the mice, and would cause any tumors originating from these human cells to be luminous under fluorescent light.

The researchers injected mouse embryos with these cells, and watched over the course of embryonic development as the cells proliferated and human tissues crept into the developing peripheral nervous systems of the tiny mice. To activate the two cancer-causing genes, researchers spiked the pregnant mice’ water with doxycycline, and over the next few days in utero — and in the weeks and months after the pups were born — the researchers inspected the chimeras to see whether tumors would appear.

Over the course of the next 15 months, 14% of the mice developed tumors — 29 mice out of 198 total. The tumors mostly appeared in the space behind the abdominal cavity close to the nerves along the spinal cord, although one mouse developed a tumor in its adrenal gland. Both locations are common places for human children to develop neuroblastoma. The researchers took samples of the tumors and found that they contained the glowing protein eGFP, which confirmed that they were of human origin.

When the team examined the growth patterns of the cancerous cells, they found that the tumors were remarkably similar to human neuroblastomas: they contained cell markers typical of human tumors, and some grew in characteristic rosette shapes — features that did not often appear in tumors implanted in immunocompromised mice through xenotransplantation.

Having successfully induced neuroblastoma tumors in the chimeric mice, the researchers took the opportunity to examine the communication between immune cells and tumors — and specifically, how the tumors evaded destruction by anti-cancer immune cells called T cells. One factor that makes human neuroblastomas and many other cancers dangerous is their sophisticated strategy for avoiding being destroyed by T cells. “The cancer tricks the immune system,” Cohen says.  By activating chemical signals that exhaust the T cells, the tumors effectively weaken their attack. The tumors in the chimeric mice, Cohen found, use a similar method to human neuroblastomas to evade immune responses.

Cohen and others plan to test the new system’s potential for modeling other cancers such as melanoma, and to use it to investigate potential treatments for neuroblastoma patients. “The obvious next step is to study how treatment of these tumors will allow these chimeric mice to be cured,” he says. “This is a model that will allow us to approach not only how to get rid of the tumor, but also to fix the immune system and recover those exhausted T cells, allowing them to fight back and deplete the tumor.”

This research was funded by the National Institutes of Health, as well as grants from the Emerald Foundation, the LEO Foundation, the Melanoma Research Foundation, and the St. Baldrick’s Foundation.

Citation: Cohen, M., et al. Formation of Human Neuroblastoma in Mouse-Human Neural Crest Chimeras. Cell Stem Cell. March 5, 2020. DOI: https://doi.org/10.1016/j.stem.2020.02.001

***

Written by Eva Frederick

***

With these neurons, extinguishing fear is its own reward
Picower Institute
January 16, 2020

When you expect a really bad experience to happen and then it doesn’t, it’s a distinctly positive feeling. A new study of fear extinction training in mice may suggest why: The findings not only identify the exact population of brain cells that are key for learning not to feel afraid anymore, but also show these neurons are the same ones that help encode feelings of reward.

The study, published Jan. 14 in Neuron by scientists at MIT’s Picower Institute for Learning and Memory, specifically shows that fear extinction memories and feelings of reward alike are stored by neurons that express the gene Ppp1r1b in the posterior of the basolateral amygdala (pBLA), a region known to assign associations of aversive or rewarding feelings, or “valence,” with memories. The study was conducted by Xiangyu Zhang, a graduate student, Joshua Kim, a former graduate student, and Susumu Tonegawa, Professor of Biology and Neuroscience at RIKEN-MIT Laboratory of Neural Circuit Genetics at the Picower Institute for Learning and Memory at MIT and Howard Hughes Medical Institute.

“We constantly live at the balance of positive and negative emotion,” Tonegawa said. “We need to have very strong memories of dangerous circumstances in order to avoid similar circumstances to recur. But if we are constantly feeling threatened we can become depressed. You need a way to bring your emotional state back to something more positive.”

Overriding fear with reward

In a prior study, Kim showed that Ppp1r1b-expressing neurons encode rewarding valence and compete with distinct Rspo2-expressing neurons in the BLA that encode negative valence. In the new study, Zhang, Kim and Tonegawa set out to determine whether this competitive balance also underlies fear and its extinction.

In fear extinction, an original fearful memory is thought to be essentially overwritten by a new memory that is not fearful. In the study, for instance, mice were exposed to little shocks in a chamber, making them freeze due to the formation of fearful memory. But the next day, when the mice were returned to the same chamber for a longer period of time without any further little shocks, freezing gradually dissipated and hence this treatment is called fear extinction training. The fundamental question then is whether the fearful memory is lost or just suppressed by the formation of a new memory during the fear extinction training.

While the mice underwent fear extinction training the scientists watched the activity of the different neural populations in the BLA. They saw that Ppp1r1b cells were more active and Rspo2 cells were less active in mice that experienced fear extinction. They also saw that while Rspo2 cells were mostly activated by the shocks and were inhibited during fear extinction, Ppp1r1b cells were mostly active during extinction memory training and retrieval, but were inhibited during the shocks.

These and other experiments suggested to the authors that the hypothetical fear extinction memory may be formed in the Ppp1r1b neuronal population and the team went on to demonstrate this vigorously. For this, they employed the technique previously pioneered in their lab for the identification and manipulation of the neuronal population that holds specific memory information, memory “engram” cells.  Zhang labeled Ppp1r1b neurons that were activated during retrieval of fear extinction memory with the light-sensitive protein channelrhodopsin. When these neurons were activated by blue laser light during a second round of fear extinction training it enhanced and accelerated the extinction. Moreover, when the engram cells were inhibited by another optogenetic technique, fear extinction was impaired because the Ppp1r1b engram neurons could no longer suppress the Rspo2 fear neurons. That allowed the fear memory to regain primacy.

These data met the fundamental criteria for the existence of engram cells for fear extinction memory within the pBLA Ppp1r1b cell population: activation and reactivation by recall and enduring and off-line maintenance of the acquired extinction memory.

Because Kim had previously shown Ppp1r1b neurons are activated by rewards and drive appetitive behavior and memory, the team sequentially tracked Ppp1r1b cell activity in mice that eagerly received water reward followed by food reward followed by fear extinction training and fear extinction memory retrieval. The overlap of Ppp1r1b neurons activated by fear extinction vs. water reward was as high as the overlap of neurons activated by water vs. food reward. And finally, artificial optogenetic activation of Ppp1r1b extinction memory engram cells was as effective as optogenetic activation of Ppp1r1b water reward-activated neurons in driving appetitive behaviors. Reciprocally, artificial optogenetic activation of water-responding Ppp1r1b neurons enhanced fear extinction training as efficiently as optogenetic activation of fear extinction memory engram cells. These results demonstrate that fear extinction is equivalent to bona fide rewards and therefore provide the neuroscientific basis for the widely held experience in daily life: omission of expected punishment is a reward.

What next?

By establishing this intimate connection between fear extinction and reward and by identifying a genetically defined neuronal population (Ppp1r1b) that plays a crucial role in fear extinction this study provides potential therapeutic targets for treating fear disorders like PTSD and anxiety, Zhang said.

From the basic scientific point of view, Tonegawa said, how fear extinction training specifically activates Ppp1r1b neurons would be an important question to address. More imaginatively, results showing how Ppp1r1b neurons override Rspo2 neurons in fear extinction raises an intriguing question about whether a reciprocal dynamic might also occur in the brain and behavior. Investigating “joy extinction” via these mechanisms might be an interesting research topic.

The research was supported by the RIKEN Brain Science Institute, the Howard Hughes Medical Institute and the JPB Foundation funded the research.

Whitehead Institute team develops new method to study human brain cells
Nicole Davis | Whitehead
November 25, 2019

A groundswell of evidence connects defects in the function of microglia, the brain’s resident immune cells, to neurodegenerative diseases, yet the tools for studying these cells in the laboratory have been limited. Now, a team of Whitehead Institute scientists has developed a new experimental platform for generating microglia from human stem cells that includes transplantation into newborn mice. As described online November 26 in the Proceedings of the National Academy of Sciences (PNAS), this new method yields microglial cells that resemble those in the human brain more closely than previous approaches, which could help enable future studies aimed at unravelling the role of microglia in neurodegeneration and other brain disorders.

“The dysfunction of microglia is implicated in a wide variety of brain conditions, and yet our knowledge of them, especially in humans, is really quite limited,” says senior author Rudolph Jaenisch, a Founding Member of the Whitehead Institute and professor of biology at the Massachusetts Institute of Technology. “This new approach will help us lift the hood on these important yet enigmatic brain cells.”

Microglia are increasingly recognized as key players in brain health and disease, but the majority of what is known about them comes from studies of mice, not humans. Yet human and mouse microglia are quite distinct — in humans, the cells are much larger, and have a more branched appearance, suggesting significant differences in their biology.

To address this gap in knowledge, multiple research teams have recently devised methods to generate microglia using human stem cells and grow them under laboratory conditions that mimic their natural environment. However, this approach has a fundamental drawback: the cultured cells do not look like microglia nor do they behave much like them, even though they display the appropriate molecular hallmarks.

“That really suggests to us that this is not the optimal approach to study how microglia are behaving in healthy and diseased brains,” says first author Devon Svoboda, a postdoctoral fellow in the Jaenisch lab. “We set out to create a new method in which the stem-cell derived microglial cells can reside in the brains of mice — one of the best models of the human brain that we have.”

Transplanting human cells into mice — creating “chimeras” — is a well-established technique. However, Svoboda and her colleagues discovered they needed to use special strains of mice that carry human genes for certain growth factors, called cytokines, which are required for microglial development and survival. The researchers utilized mice that carry human genes for four crucial cytokines: CSF1, IL3, SCF, and GM-CSF.

“What is special about these chimeras is really the mice we are using,” says Svoboda. “They express the human alleles of these cytokines which is key because the mouse versions are not able to communicate with receptors on human microglia, so the cells die.”

After transplanting the stem-cell derived microglia into these mice, the research team examined the cells’ morphology and their molecular characteristics. They found that the transplanted cells closely resembled those found in the human brain.

Further analyses revealed some striking differences between the team’s “chimera-grown” microglia and those grown in the laboratory using conventional cell culture methods. Surprisingly, Svoboda and her colleagues found that the cultured microglia showed strong similarities to the diseased microglia from patients with multiple sclerosis, another brain condition in which the cells are implicated.

“If you want to learn more about the role of microglia in disease, then studying them in culture is probably not the best way,” says Svoboda. “The chimeras and the in vitro methods really complement each other, and we think there is a place for both systems in microglia research going forward.”

The Whitehead-led team plans to extend their initial studies in several ways. One is to identify which cytokines and other growth factors are most crucial to microglial development. That knowledge could help improve existing cell culture methods and enable them to more closely mirror the cells’ natural environment. Another key direction is to use the new chimera-based system to create models of neurodegenerative diseases, including Alzheimer’s and Parkinson’s disease, to understand how microglia respond to diseased neurons and, in turn, how diseased microglia can impair neuron function.

“Our chimera-based method will give us a good handle to begin to stringently test the role of microglia in brain health and disease,” says Jaenisch. “This is an important step forward for the field.”

Support for this work was provided by the Cure Alzheimer’s Foundation, MassCATS, and NIH Grants R01 AG058002-01, R01 MH104610, R37 CA084198, and U19 AI131135 (to R.J.). L.D.S. is supported by NIH Grants R24 OD26440, AI32963, and CA034196. J.S. is supported by the National Institute of Child Health and Human Development (K99HD096049).

Written by Nicole Davis

***

Rudolf Jaenisch’s primary affiliation is with Whitehead Institute for Biomedical Research, where his laboratory is located and all his research is conducted. He is also a professor of biology at Massachusetts Institute of Technology.

***

Paper cited:

Human iPSC-derived microglia assumer a primary microglia-like state after transplantation into the neonatal mouse brain.

PNAS, online November 26, 2019. DOI: 10.1073/pnas.1913541116

Devon S. Svoboda (1), M. Inmaculada Barrasa (1), Jian Shu (1,3), Rosalie Rietjens (1), Shupei Zhang (1), Maya Mitalipova (1), Peter Berube (3), Dongdong Fu (1), Leonard D. Shultz (4), George W. Bell (1), and Rudolf Jaenisch (1,2)

 

1. Whitehead Institute, Cambridge, MA 02142, USA

2. Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

3. Broad Institute of MIT and Harvard, Cambridge, MA 02142

4. The Jackson Laboratory Cancer Center, The Jackson Laboratory, Bar Harbor, ME 04609