Synaptic plasticity

On-demand webinar

webinar-image

Summary:

Understanding the processes of brain rewiring is a vital aspect in the field of neuroscience. Get insights on the latest developments in this live interactive digital session.

Synaptic plasty

Speakers:

Dr. Jennifer Y Sun, UCL Institute of Ophthalmology, United Kingdom
Experience-Dependent Circuit Reorganization in the Visual Cortex
Time stamp: 00:06:51

Dr. Dion Dickman, University of Southern California, United States
A Glutamate Receptor C-tail recruits CaMKII to suppress retrograde homeostatic signaling
Time stamp: 00:46:07

Panel Discussion led by Dr. Tom Otis
Time stamp: 01:32:20

Moderator:

Dr. Tom Otis, Sainsbury Wellcome Centre, United Kingdom

Video Transcript

00:00 - 00:15: Hi and thank you for joining the last and final session of Spotlight on Neuroscience,
00:16 - 00:20: a month-long virtual conference. We have two wonderful talks for you today,
00:20 - 00:27: followed by a riveting panel discussion being led by Dr. Tom Otis. My name is Sian Constantine
00:27 - 00:32: and I am the Strategic Marketing Manager for Neuroscience at Abcam. Before we begin,
00:32 - 00:37: I'd like to run through a couple of housekeeping notes. All attendees are automatically muted,
00:37 - 00:42: however please feel free to submit your questions in the question and answer box at the bottom of
00:42 - 00:46: your screen and these will be addressed during the dedicated Q&A session after each talk.
00:47 - 00:51: If we don't get to your question, we will get a response to you shortly after the event,
00:51 - 00:57: so please don't worry. I'm excited to inform you that Abcam is approved as a provider of
00:57 - 01:04: continuing education programs in the clinical laboratory sciences by the ASCLS PACE program.
01:05 - 01:09: PACE credits are available for this event and a link to the requested credit can be sent in the
01:09 - 01:16: chat box at the end of the event. So, I'd now like to talk to you about how Abcam is working
01:16 - 01:22: to advance the needle in science within the neuroscience space. So, we have three strategic
01:22 - 01:27: areas that we're working on to make sure there are research tools available for your research.
01:28 - 01:33: One area is new development, the second is neurobiological processes, which is where
01:34 - 01:41: the synaptic plasticity sits, and then we also have neurological diseases. We work
01:42 - 01:49: well with collaborators to make sure that these research tools that we release are appropriate
01:49 - 01:55: to the targets that you're working on, researching and working with, and are really keen to hear from
01:55 - 02:00: you about what those targets in the future need to be and what needs to be in our catalog,
02:00 - 02:04: but also how you find our products. So, please do feel free to reach out to us
02:05 - 02:09: using the email neuroscience@abcam.com.
02:14 - 02:21: So, we have multiple different product types that span in the neuroscience portfolio.
02:22 - 02:25: What's not a surprise to most people is the antibodies that we sell,
02:26 - 02:31: and also the ELISA kits and proteins and peptides. So, that's the cornerstone of our
02:31 - 02:40: business, but we also have products to support your research. For example, multiplex microRNA
02:40 - 02:46: immunoassays for cellular biochemical assays. We also have cell lines and lysates. So,
02:47 - 02:53: for example, we have the glutamatergic neuron IPSC line that's particularly helpful for
02:54 - 03:02: neuroscience research. At the moment, if we don't have a product that you need for your research,
03:02 - 03:08: you can also use our customization services for that product to be created. We're really keen
03:08 - 03:12: to make sure that these products are accessible and ready to be ordered straight away. So,
03:12 - 03:19: do get in touch with us with that neuroscience@abcam.com email if you think there's a
03:19 - 03:23: particular target or product that we should be working on to bring to you in our catalog.
03:26 - 03:31: So, this slide just gives you a flavor of the different types of products we have in each
03:31 - 03:37: product range, specific to synaptic plasticity. So, on the bottom right-hand side there,
03:37 - 03:43: you can see the glutamatergic neuron that I mentioned earlier. And then across the slide,
03:43 - 03:48: you can see multiple different product types, sorry, different product targets that sit in
03:48 - 03:56: under the different product types. Now, I'd like to take a moment to introduce the moderator for
03:56 - 04:03: today's event, Dr. Tom Otis. Tom is the Chief Scientific Officer at the Sainsbury Wellcome
04:03 - 04:07: Centre for Neural Circuits and Behavior and holds a professorship in neuroscience at the
04:07 - 04:14: University College London. He received his BSc and MSc degrees in biological sciences in 1988
04:14 - 04:20: and his PhD degree in neuroscience in 1993, all from Stanford University.
04:21 - 04:27: Prior to SWC and UCL, he led a team of 45 scientists conducting early-stage drug
04:27 - 04:33: development in neurodevelopmental disorders and psychiatry at the Hoffman-LaRoche Pharmaceutical
04:33 - 04:39: Company. Before Roche, he served as the Edith Agnes Plum Chair of the Department of Neurobiology at
04:39 - 04:45: the University of California in Los Angeles. His research is focused on the cellular and circuit
04:45 - 04:51: function of the cerebellum and hippocampus, motor systems function and motor learning,
04:51 - 04:57: and preclinical models of neurological disorders, spinocerebellar ataxia and
04:57 - 05:03: amyotrophic lateral sclerosis. Thank you. I will now hand the mic over to Tom.
05:04 - 05:13: Thanks very much, Sian. And I want to start by thanking Abcam for organizing this webinar.
05:14 - 05:21: And today we have, I think, a very exciting program. We have two leading scientists in the
05:21 - 05:29: field of synaptic plasticity. And I'll introduce Jennifer Sun, our first speaker, in a moment,
05:29 - 05:37: and then later in the program, we'll have a talk from Dion Dickman. And I only want to say before
05:37 - 05:45: they start that surely everyone on this webinar appreciates that synapses are really where a lot
05:45 - 05:54: of the action is in the brain. We'll hear today about how synapses help to determine how neurons
05:55 - 06:01: respond to the stimuli that they respond to. And synapses are the place that a lot of the changes
06:02 - 06:10: that occur in the brain during learning happen. They also happen to be a focus for many disease
06:10 - 06:16: mechanisms. We've learned a tremendous amount over the past decade or two in terms of genetics
06:17 - 06:24: and mechanisms in human disease. And synapses are one of the places where genetics,
06:25 - 06:34: at least to the extent that certain human disorders have a cause that can be traced to
06:34 - 06:41: genetics, synapses are where many of these genes cluster. So for good reason, a lot of research
06:41 - 06:49: focuses on synapses, on how synaptic connections change and on how they're maintained during
06:49 - 06:57: normal health and disease. So with that, let me now turn to introducing Dr. Jennifer Sun.
07:00 - 07:06: Jennifer has recently started her lab at the University College London, where I am. And
07:06 - 07:13: indeed, she's an affiliate member of the Sainsbury Wellcome Centre, also where I am. So we're really
07:13 - 07:19: delighted about that. She leads the Visual Plasticity Lab based in the Institute of
07:19 - 07:30: Ophthalmology. She obtained her PhD from USC, where she worked with Zhang Li and did some beautiful
07:30 - 07:39: work on how auditory responses and neurons in primary auditory cortex develop their tuning
07:39 - 07:46: properties involving changes in synaptic inputs to those neurons. She went on and did a postdoc
07:46 - 07:53: with Roger Nicoll and Mike Stryker at UCSF, where she turned to the visual system and again
07:53 - 07:59: looked at mechanisms underlying tuning properties, in this case, in primary visual
07:59 - 08:06: cortical neurons. And now she will focus her lab on the cellular and circuit mechanisms of
08:06 - 08:14: neuroplasticity and developing an adult brain, focusing on vision using state-of-the-art imaging
08:15 - 08:21: together with molecular, physiological, and computational approaches to try to understand
08:21 - 08:27: vision. So with that, I'm going to turn the program over to Jennifer to hear about her
08:27 - 08:36: research. Welcome, Jennifer. Thank you, Tom, for the warm welcome and for the kind introduction.
08:37 - 08:44: And also, I would like to take this opportunity to thank Abcam for the invitation to this webinar
08:44 - 08:53: to allow me to introduce my past research and also my research vision at my current lab at UCL
08:53 - 09:03: to study the visual plasticity. And as Tom talked about, I just started my lab recently, so today
09:03 - 09:11: I'm going to focus mainly on the past work that led me to my current research. And then at the
09:11 - 09:19: end of the talk, I'll talk a little bit more about the outlook of my research and the approaches
09:19 - 09:28: my lab is proposing to do. So today, the topic of my talk is experience-dependent circuit
09:28 - 09:35: development and plasticity in the visual cortex. And when we talk about the circuit...
09:35 - 09:38: Jennifer, I don't believe you're sharing your screen just yet.
09:39 - 09:44: Oh, my apologies. No worries. Just didn't want you to get too far. So down at the bottom,
09:44 - 09:46: there's the shared screen icon. Yep.
09:53 - 10:00: This looks perfect. Thank you very much. Thank you. So as we were talking about
10:01 - 10:05: this circuit in the visual cortex, I would like to
10:09 - 10:15: show you what most people probably would think of when we talk about a circuit, where there's
10:15 - 10:23: lots of components and wires that would allow us to execute certain functions. And in a sense,
10:23 - 10:32: that's true. Our brain is like one of those circuits where it has billions of neurons and
10:32 - 10:39: interconnected through trillions of synapses. And it is still one of the most complicated and
10:39 - 10:47: powerful circuits to allow us to execute multiple functions that currently cannot be achieved by the
10:47 - 10:57: electric circuits yet. And if we look at that, indeed, those numbers and the complexities of the
10:57 - 11:06: level of the circuit probably are the basis that allow us to execute those complicated functions
11:06 - 11:17: in our brain only by us, not by a computer at the moment. And to understand how our system,
11:17 - 11:25: how our biological circuits would enable us to conduct those functions, we actually can break
11:25 - 11:32: this down into multiple scales. Even just a simple behavior like reading a book actually
11:32 - 11:40: contains multiple pathways. The ascending pathway would send those visual information from the eye
11:40 - 11:48: up to the cortical area and then feed to the prefrontal cortex for learning, perception,
11:48 - 11:57: understanding, as well as a descending pathway to move our eye and allow us to keep going on.
11:57 - 12:05: And within this area, we need to execute our function to allow us to remember the content
12:05 - 12:12: while we move on to that, which involves short-term plasticity as well. And if we look at each
12:12 - 12:24: single, each single pathway, there's a small neuron circuit that actually comprises of multiple
12:24 - 12:33: neuron types, as well as different neurotransmitters that would allow them to communicate
12:33 - 12:41: with each other in a very precise and effective way. And of course, when we are talking about
12:41 - 12:52: their communication side, that brings us to this synapse, the really one of the most top loci
12:52 - 13:01: that all the learning plasticity can happen at this tiny junction there. And with that,
13:01 - 13:08: we're getting to know more detail about their molecular and genetic basis for these approaches.
13:08 - 13:17: And adding to that, as I said, we begin to develop more approaches to look at the longitudinal
13:17 - 13:26: changes or the dynamics of each of those scales to understand how the changes and development
13:26 - 13:36: adaptation can allow us to properly and reliably process the information that we need for us to
13:38 - 13:46: even accomplish this simple task of reading a book. And in the past, as Tom mentioned,
13:46 - 13:54: my research is mainly concentrated on the sensory processing. The reason is kind of straightforward
13:54 - 14:04: that we need our sensory system. We rely on our vision to track a flying rugby, and we need
14:04 - 14:12: our hearing to listen to and communicate with each other effectively. And we need those sensory
14:12 - 14:21: information to allow us to execute more complicated functions like emotion, learning, and even memory.
14:22 - 14:30: And we know that in the sensory process can be impaired and affect our daily life in multiple
14:30 - 14:38: diseases, especially neurodevelopmental disorders and a lot of aging-related diseases.
14:39 - 14:48: And what I've been looking at is to see how the sensory processing is achieved in the brain
14:48 - 14:57: circuit. And to begin with, let's first take a quick look at the circuit. So our sensory stimulus
14:57 - 15:05: are converted to electrical signals at the peripheral receptors that would then directly
15:05 - 15:12: or indirectly sent to the thalamus, which then relates to the cortical area for sensory
15:12 - 15:21: perception. And at the cortical level, there is a very stereotypical structure that the layer four
15:21 - 15:28: recipient layer would receive a lot of thalamic connections, and that would relate to the layer
15:28 - 15:36: two, three, which then project onto layer five, six. And that signal would then be able to send
15:36 - 15:42: to other cortical areas or the subcortical projections for further information processing.
15:43 - 15:52: And in part of the research, I've been focusing on those sensory circuits, but also the development
15:52 - 16:00: and the plasticity of them to see how the circuit has ensured to achieve a proper sensory
16:00 - 16:09: representation of that information. And the two key questions that I have been asking are, one:
16:09 - 16:15: what is the developmental process that can give rise to this highly specialized function in the
16:15 - 16:23: circuit? And what are their underlying synaptic basis? And two: how the sensory circuit
16:23 - 16:30: could rewire to adapt to the altered environment, both in the developing and adult brain?
16:31 - 16:37: What strategies are they exploring at the circuit, cellular, and molecular level
16:37 - 16:41: to ensure that this plasticity is precisely regulated?
16:44 - 16:51: And to begin with, I want to talk about my past research at San Francisco with Dr. Michael
16:51 - 16:59: Stryker to study this ocular dominance plasticity in the developing visual cortex. As I mentioned that
17:01 - 17:09: at the thalamus, this visual information is passed through the thalamus and then up to the visual
17:09 - 17:18: cortex. And back in the 1960s, when Hubel and Wiesel first established this method to study ocular
17:18 - 17:25: dominance plasticity using monocular deprivation, which is the suturing of one eye, they
17:27 - 17:35: established this method in such a beautiful way that quickly has been developed and acknowledged
17:35 - 17:43: as one of the best models to study synaptic plasticity and as the reason that this is the very
17:43 - 17:49: first stage, the cortical neurons can receive the converged information from both eyes. So,
17:49 - 17:57: that would give us a very unique opportunity to perturb and look at the neuronal mechanisms
17:57 - 18:05: and the synaptic basis of the plasticity and the development in the cortical area. Moreover,
18:05 - 18:18: this has been later shown to also persist in other animal species from ferrets, monkeys, and mice,
18:18 - 18:28: which all nicely mimic the condition in the human clinical condition called amblyopia or the lazy eye,
18:28 - 18:37: which is that if during the early developmental stage, you close one of the eyes, the brain circuit
18:38 - 18:47: from eye to the cortex would dramatically shift to adapt to these changes. As before, the majority
18:47 - 18:54: of neurons would listen to the contralateral eye input versus with just monocular deprivation
18:54 - 19:01: during development, you would significantly see that the majority of the neurons would shift their
19:01 - 19:11: preference to the remaining open eye. And indeed, further anatomical studies show that there's a drastic
19:11 - 19:18: decrease of the thalamocortical connection there that accounts for the major loss of this
19:18 - 19:28: contribution from the closed eye. And later studies that focus on the higher cortical area
19:28 - 19:37: where the upper layers of the cortex that project to the higher cortical areas actually found
19:38 - 19:46: that ocular dominance plasticity is even faster. That only takes about one day in cats
19:46 - 19:53: and up to three days in mice to manifest this drastic change from the contralateral eye
19:53 - 19:59: to the ipsilateral eye if you do the ocular dominance.
20:00 - 20:17: And this drastic change is ahead of the layer four neurons, which we can see here, and also is ahead of the anatomical changes that we observed at the thalamocortical level.
20:17 - 20:27: So, what could the anatomical basis or the physical loci for these changes to happen?
20:28 - 20:35: And to answer this question. Again, in short, it comes down to the synapses.
20:35 - 20:54: There are already multiple studies in the adult animal showing that the excitatory or the inhibitory synapses in the adult brain can quickly change in response to altered sensory input.
20:54 - 21:01: And these changes can be fully...
21:01 - 21:29: These changes actually underlie the functional plasticity that is observed at the cortical level. So, what we want to study here is to look whether those synapses are the basis in the developing brain that can actually accomplish and serve as the basis for those functional
21:30 - 21:32: changes that we observed.
21:32 - 21:55: And to do this, we actually take advantage of the transgenic animal, where we apply the PSD95, which is the most salient postsynaptic marker at these excitatory synapses and
21:56 - 22:09: inject this construct into the embryonic animals, where in that way we can label all the excitatory postsynaptic structures in the layers two and three of the neurons.
22:10 - 22:29: And you can see that there is one neuron that's being sparsely labeled, and then all its synapses can be fully presented by the PSD95 driven by the GFP.
22:30 - 22:52: And what we can do then is to track those single neurons over days during the development using two-photon imaging, so that here we show the same neuron over the course of eight days, and in some of the animals, we can do the monocular deprivation,
22:52 - 23:01: and then see whether the neurons would be affected by this manipulation, and how their synapses would change over time.
23:04 - 23:27: What I did then is to develop this pipeline, where we can do the automatic image stitching registration over different days, and also the automatic detection and linking over different days to compare the synapses that might disappear or be newly added in those cases.
23:27 - 23:41: And as you can see here, in the control animals, we can see relatively stable synaptic structures over the time course of the eight days.
23:41 - 24:02: However, in the animals that experienced the monocular deprivation, we can see that there is a drastic reduction of the overall PSD95 labeled puncta structure.
24:03 - 24:17: And this actually came back to the baseline level after six days of monocular deprivation and persists relatively there, comparable to the control animals.
24:18 - 24:33: And indeed, to match these changes with the functions, I also conducted intrinsic imaging to directly measure the population responses of those brain areas.
24:34 - 24:49: And we calculate this ocular dominance index by taking the difference normalized by the overall responses of the cortical neurons contributed by the input from each eye.
24:50 - 25:07: And you can see here that, as expected, there is a drastic decrease of the ocular dominance. That means the cortical responses are indeed shifting from the contralateral eye to the open eye.
25:07 - 25:36: So, what we have shown thus far is to demonstrate that these postsynaptic marker changes can nicely correlate with the population changes that have been shown in other studies, where after three days of monocular deprivation, there is a decrease of the postsynaptic response.
25:37 - 25:53: And this is shown by the following spine numbers, which aligns with other observations where the responses from the open eye stay more or less the same, but there's a drastic decrease of the responses to the closed eye.
25:54 - 26:13: On the other hand, at six days of MD, there are some homeostatic processes happening, which drive the increased responses to the visual input from the open eye, so that the overall cortical responses stay more or less the same.
26:14 - 26:20: And that, again, is consistent with what we observed in the spine number changes.
26:20 - 26:31: And one step further, we look at how the spine would, at the single neuron level, correlate with their functional changes.
26:32 - 26:52: To do this, we did a similar approach by looking at individual neurons after their structural imaging, and by loading the COSM indicator and applying two-photon imaging, by measuring their functional activity at each individual neuron.
26:52 - 27:19: And you can see here, one of the neurons that have more preference to the contralateral eye actually suffers a loss of the postsynaptic structures, in this case, versus when we see other neurons that have more preference to the open eye actually have relatively stable or even an increase of their spine numbers.
27:19 - 27:33: So in this way, we demonstrated that the postsynaptic structure likely serves as an anatomical basis for this rapid ocular dominance plasticity observed in layers 2-3.
27:34 - 27:57: And one step further, we also look at the presynaptic structure. And to notice the difference that by looking at the presynaptic structure, we would have the opportunity to explore both the connection from excitatory neurons to other excitatory neurons, as well as to the inhibitory neurons.
27:58 - 28:15: So this is one of the markers showing that the synaptophysin is being labeled by the GFP, and that co-localizes with the nicely tdTomato labeled neuron.
28:16 - 28:33: And what we can see here is that during development, there is a gradual increase of the total presynaptic structures in the normal animals.
28:34 - 28:54: However, for the monocular deprived animals, we see that their presynaptic structure stays more or less the same. That means the de novo formation of the presynaptic structure was disrupted by monocular deprivation.
28:55 - 29:15: Therefore, by looking at both the pre- and postsynaptic structures, our study demonstrated that it's very likely it is the excitatory postsynaptic structure that serves as the anatomical basis for the rapid ocular dominance plasticity in layers 2-3.
29:16 - 29:31: And also, during development, there is a de novo formation of excitatory to inhibitory synapses, which is dependent on visual experience.
29:32 - 29:45: The monocular deprivation disrupts this process, which then stops or largely eliminates the observed formation of these synapses.
29:46 - 30:00: So then I'm going to quickly switch gears to the second part of my talk, which is how the sensory circuits can also change throughout life.
30:01 - 30:19: As we know that the adult brain is not that plastic, and in fact, in old days, even the famous Ramón y Cajal had these comments that the adult center are something fixed and immutable.
30:19 - 30:23: Everything may die, nothing may be regenerated.
30:23 - 30:30: So later on, we found that it's not completely right in that sense.
30:30 - 30:43: However, indeed, the changes that can happen in the immature brain, it's very, very slow, and their magnitude is much, much smaller in comparison to the developing brain.
30:44 - 30:56: Therefore, that accounts for very difficult recovery in adult patients with developmental disorders or severe brain injury.
30:57 - 31:20: And that's why the science is really geared to develop ways to rejuvenate the brain through those pharmacological approaches, stem cell transplantation approach, as well as electro-stimulation to see whether we can boost up the brain plasticity in the adult cortex.
31:21 - 31:31: And more recently, there is some hope to show that running can actually enhance the plasticity in the adult brain.
31:32 - 31:52: And one of the studies has been shown that simply by exposing the animal to the visual stimulus, and then giving it the opportunity to run, there is a significant increase of the visual representation of the visual stimulus being presented.
31:53 - 32:19: And this probably requires some temporal correlation of the two conditions, the running and the visual input, because if we put the animal on a running wheel for a couple of hours, and then put it back to a home cage where it can only have minor exercise, and then present the visual stimulus,
32:19 - 32:25: you can see that the effect is very minimal and comparable to the control state.
32:26 - 32:47: And what I did furthermore then is to design this closed loop visual experience, where the animal will have the opportunity to see one type of visual stimulus while it's running on a ball, and then during still another type of visual stimulus can be presented to it.
32:48 - 33:09: In this way, I can then look at after five days of this state-dependent visual experience, how the population of visual cortical neurons change in response to different visual stimuli.
33:10 - 33:36: And what we found is that there are neurons that would specifically upregulate their visual responses to the visual stimulus that only presented during running, which can be seen in the red box, versus the responses to the presentation that's equally presented,
33:36 - 33:56: but only during the still state, it stays more or less the same. And indeed, among nine animals, hundreds of neurons, their population response showed a significant increase only to the visual stimulus that was presented and correlated with the running behavior.
33:57 - 34:11: This is not too surprising because even in another study over time, it was indicated that there are certain neuron computation that are being modulated only during running.
34:12 - 34:24: And this instantaneous modulation is likely administered by some neuromodulator, for example, noradrenaline.
34:26 - 34:46: And indeed, other studies have shown that during the running state, multiple neuromodulatory circuits can be recruited that then exert their regulatory effect onto multiple circuits that regulate the overall visual output.
34:47 - 35:07: So, my lab would be very interested to see how those neuromodulatory functions would affect the long-term visual plasticity. And to assess that, we will employ this viral tracing technique in combination with the freely moving recording as well as the two-photon imaging
35:08 - 35:21: to see if the precise manipulation of specific neuromodulatory pathways can directly perturb and regulate the long-term plasticity in the visual cortex.
35:22 - 35:42: And also, we would like to see if at the circuit and synaptic level, those changes can be manifested by using the patch-clamping recording that we can see if these changes are also observed and manifested.
35:43 - 35:58: And in addition to that, I would like to look at the microscale where the synapses are the loci where those plasticity manifested.
35:59 - 36:24: And the past years, people have started to have more interest and also appropriate techniques to closely look at those synaptic structures, which is just more than just the excitatory and inhibitory synapses, but also surrounded by the supporting glial cells.
36:25 - 36:52: And indeed, there are multiple research studies showing that there is a close interaction between the excitatory neurons, inhibitory neurons, glial cells, as well as the blood vessels, that they closely modulate those changes and then form a very tightly regulated environment for the plasticity.
36:53 - 37:21: Therefore, my lab would be very interested to use the transgenic approaches to look at those different cell populations, including excitatory, inhibitory, as well as those glial cells, and also develop appropriate neurophysiological assessments to see how those different parties can all contribute
37:21 - 37:26: to work together to regulate and enable the plasticity.
37:27 - 37:44: So, indeed, my new lab at the UCL Institute of Ophthalmology would be very interested to know how the circuit from the eye to brain enables visual perception.
37:44 - 37:57: And also, we would love to explore all the circuits, neuromodulatory pathways, as well as the local environment that can all contribute to this process.
37:58 - 38:11: And finally, I would like to express my acknowledgment to my past training, as well as to my current funding agent, and I will be happy to take any questions. Thank you.
38:15 - 38:32: Thank you, Jennifer. That was a great talk. Really fascinating work. And we have some questions here in the chat. So let me jump into those.
38:33 - 38:47: You talked a bit about the issue of critical periods and adult, you know, the relative reduction in plasticity mechanisms in adult animals.
38:48 - 38:57: There's a question about whether these findings have been extended in some specific way to humans. Is this also seen in humans?
38:58 - 39:08: Yeah, indeed. There are some very quick follow-up studies in human patients that have amblyopia.
39:09 - 39:25: But sadly, it's kind of controversial where some of the patients were set on static cycling and then given high contrast visual stimuli, and then they're being assessed.
39:26 - 39:49: But I think because those studies used a little bit different approaches and assessment, there's at least so far, there's no great conclusion about whether it's effective or not, because they are still under assessment by multiple groups. Yeah.
39:49 - 39:58: Yeah. And I would also add that, and I know that this certainly isn't at the level of specific neuronal circuits.
40:00 - 40:10: For procedures, very invasive or serious procedures for childhood epilepsies, it's well known that
40:11 - 40:16: those surgeries can be performed in large sections of cortex, for example,
40:16 - 40:22: can be resected early on in life for young children. But in older patients, this would
40:22 - 40:30: never be done. There's a much more long-lasting deficit. And so in young children,
40:30 - 40:34: there can be hemispherectomies in the case of very severe epilepsies, for example.
40:34 - 40:41: Okay. There's a question about,
40:47 - 40:55: so after visual recovery, after the monocular deprivation, are there lasting? I think the
40:55 - 41:01: question is getting at whether there are lasting changes. And I think your microphone is muted,
41:01 - 41:09: Jennifer. Yeah, I see. So, in short, it's very likely that there's some after effect.
41:09 - 41:17: For what we see that even at the functional level, there can be a quick reverse to the baseline.
41:17 - 41:25: But based on previous studies showing that animals that have been previously monocularly
41:25 - 41:33: deprived have a much higher chance to suffer those plasticity or the changes in their later
41:33 - 41:40: adulthood. And there are some studies showing that the prolonged plasticity can actually involve
41:40 - 41:48: changes much more than just the cortical changes, so that which might not be reversible afterwards.
41:49 - 41:56: I see. Yeah. And if I recall, isn't it the case that if one re-sutures the same eye,
41:56 - 41:58: that some of the plastic changes come on more quickly?
41:59 - 42:02: Exactly. Yes. Yeah. That's the classical.
42:02 - 42:06: So, yeah, exactly. So there's something that's not, it's not a complete reversal,
42:06 - 42:14: essentially, of everything. Yeah. A question about the work implicating neuromodulatory pathways
42:15 - 42:25: is, I think the question is getting at whether, what the mechanism of that
42:26 - 42:34: modulatory effect is. And if I can jump in and ask, we know that neuromodulators of this sort
42:34 - 42:40: can affect the excitability of the neurons. And we know just the baseline excitability, of course,
42:40 - 42:47: can affect plasticity mechanisms. Do you think that these neuromodulatory mechanisms
42:49 - 42:55: are relying on more than just effects on the excitability of neurons? Is there something
42:55 - 43:02: else there? Is there any evidence for that? That's a really good question. I think,
43:03 - 43:09: in short, as you said, the excitability is one of the fundamental things that the
43:09 - 43:21: neuromodulators insert. And also, I would expect that there are also some changes in terms of the
43:21 - 43:28: firing pattern. For example, they can turn the phasic activity, the transient activity,
43:29 - 43:38: into a more long-lasting effect. Because we know, for example, the reward that recruits the dopamine
43:38 - 43:47: can actually prolong the activity of specific neurons. And also, for acetylcholine, they might
43:47 - 43:56: actually sharpen the responses of the neurons, which might lead to or eliminate the plasticity.
43:56 - 44:04: And those changes need to be observed at the intracortical level. And probably the most
44:04 - 44:12: suitable approach is to look at the slice recording and directly measure their electrophysiological
44:12 - 44:19: properties. Yeah. And so I suppose there'd be implications with regard to the associativity.
44:20 - 44:26: You know, you might broaden the associativity if the neuron's excitability was extended over
44:26 - 44:30: a longer time window. And so you could bind more things together versus when it's sharpened.
44:32 - 44:40: Yes. Yeah. That's definitely a possibility to increase the synchrony of multiple brain areas.
44:40 - 44:46: That might be a really great way to open the plasticity window. Yeah.
44:46 - 45:01: Yeah. Okay. I think we could turn to the next speaker. Let me see. Yeah. So I think we've got
45:02 - 45:10: all of the questions in the chat. I will say, and I think this will come up in the discussion,
45:10 - 45:17: I'm really impressed, and I expect Dion's going to extend this theme with the way that you've
45:17 - 45:22: placed questions about synaptic plasticity, questions at the level of individual synapses
45:23 - 45:29: into a systems context. I think that's one of the things that is clearly a challenge going forward
45:29 - 45:38: in neuroscience in general, to try to explain how phenomena that we might see and can detail
45:38 - 45:48: at a molecular level that involve identified genes that can be recorded electrophysiologically,
45:48 - 45:54: how those fit into the bigger question of what's it doing for the neuron, for this type of neuron,
45:54 - 46:02: given what it's signaling. So we'll pick up on that theme in the roundtable at the end, I think.
46:03 - 46:13: Okay. So I will now introduce Dr. Dion Dickman, who will give the next talk. Dion is an associate
46:13 - 46:21: professor of neurobiology at USC, at the University of Southern California. I learned this, even though
46:21 - 46:29: I've known Dion for years, I learned he was born in Hawaii. He did his undergraduate degree at
46:29 - 46:36: Washington University in St. Louis, studying, working with Josh Sainz, who is now at Harvard.
46:37 - 46:42: And he earned his PhD in neuroscience at Harvard University, not working with Josh Sainz,
46:42 - 46:53: but working with Tom Schwartz, where he must have first been exposed to the Drosophila model.
46:54 - 47:02: And he went on to do postdoctoral work with Graham Davis at UCSF. And it was there that he
47:03 - 47:12: undertook a forward genetic screen and identified genes, really revealed genes for the first time
47:12 - 47:17: that were involved in a form of plasticity that Jennifer mentioned called homeostatic
47:17 - 47:23: plasticity that I'm sure Dion is going to tell us a lot about, which is critical for
47:24 - 47:30: maintaining balance at all synapses and in all nervous systems so that the pre and the
47:30 - 47:37: postsynaptic elements can be within the same range and function properly. So we'll hear,
47:38 - 47:46: I expect, a lot about that, and not only what these genes are, but also, of course, with knowledge of
47:46 - 47:54: the genes, Dion's gone on to explore how those genes exert their effects mechanistically in the
47:54 - 48:01: Drosophila system. So his lab is interested in general and molecular mechanisms that achieve
48:01 - 48:06: and maintain stable synaptic function. And he uses, within the Drosophila system, a combination
48:06 - 48:13: of genetics, electrophysiology, pharmacology, and imaging approaches. With that, I will turn
48:13 - 48:22: the program over to Dion. Great, thank you, Tom and to Abcam for organizing this great
48:22 - 48:28: webinar. I've participated in a few and watched a few of the previous ones. I think this is a
48:28 - 48:32: great program. So today I'm going to talk about, we're going to switch gears a little bit, and I'm
48:32 - 48:39: going to talk about the molecular and cellular signaling mechanisms that stabilize homeostatic
48:39 - 48:46: plasticity using the Drosophila neuromuscular junction as our model system. So as we all know,
48:46 - 48:55: you know, the nervous system is remarkably complex with this amazing ability to learn and remember
48:55 - 49:01: and adapt throughout life. But this poses a major challenge to stability in the brain. You can
49:01 - 49:08: imagine, you know, if the flexibility in our nervous systems are a little too open, you'll
49:08 - 49:13: get uncontrolled excitation and seizures and things like that. But what's, you know, what's
49:13 - 49:19: really amazing is that from early development to growth and into old age, nervous system function
49:19 - 49:26: remains remarkably stable. So this has given rise to this idea that the brain, that the nervous
49:26 - 49:36: system, is endowed with robust homeostatic mechanisms that maintain stability and
49:36 - 49:45: stabilize activity. So homeostasis is a concept in physiology. It's been studied for over 200 years.
49:45 - 49:50: I think many of you are probably familiar with this concept. It's a fundamental form of
49:50 - 49:57: physiological regulation that maintains internal stability. It's been studied for over 200 years
49:57 - 50:03: in the context of processes like temperature, osmotic balance, hunger, and thirst. And very
50:03 - 50:09: interesting to me, sleep is under homeostatic control. But only over the last 20 years or so
50:09 - 50:15: has this concept of homeostasis been brought to electrical activity in the nervous system.
50:17 - 50:23: So if the nervous system is under homeostatic control, then every feature of a homeostat should
50:23 - 50:29: be embedded in this system. Temperature is one of the best understood homeostats, and so we'll
50:29 - 50:35: kind of go over what we understand about that. Homeostasis implies a set point in the context
50:35 - 50:41: of temperature regulation, its body temperature. There must be sensors of the system that can
50:41 - 50:48: detect the state of the system itself. In the case of temperature, there are ion channels that can
50:48 - 50:54: open and close in response to differences in temperature and send signals to an integrator.
50:55 - 51:01: In the context of temperature and actually many homeostatic processes, the hypothalamus in the
51:01 - 51:07: brain is thought to be the integrator that can detect the state of the system and compare it
51:07 - 51:13: to a set point. If there are deviations in the system, then there are effector mechanisms that
51:13 - 51:19: kick in. If you're too cold, you shiver to generate heat. If you're too hot, you sweat.
51:20 - 51:27: All of it being used to re-target back to stabilize the set point of the system.
51:28 - 51:34: So if the nervous system in general and electrical activity in the nervous system
51:34 - 51:38: is under homeostatic control, then every feature of this system must be maintained.
51:39 - 51:45: Now, stress is this word and concept that's kind of thrown around and used very loosely,
51:45 - 51:51: but I like to use the definition of stress to be any perturbation that forces the system under
51:51 - 51:57: homeostatic control to deviate from set point values. So in the case of body temperature,
51:57 - 52:03: a stress could be exercise, it could be jumping into a cold lake, but whatever it is, it challenges
52:03 - 52:11: the set point of the system and triggers an adaptation to get you back to that set point.
52:11 - 52:20: So sweat or shivering could be adaptations. So the fundamental hypothesis that undergirds much
52:20 - 52:26: of the research in my lab and indeed our field is that latent forms of homeostatic plasticity
52:26 - 52:33: can be revealed at synapses that are experiencing stress. But I want to emphasize that under normal
52:33 - 52:39: basal conditions, you're not going to see any of these mechanisms. You must trigger them by
52:39 - 52:48: challenging the synapse or the nervous system with a stress that perturbs a physiological process
52:48 - 52:56: under homeostatic control. So the homeostat that we're going to talk about today is a homeostatic
52:56 - 53:02: signaling system that stabilizes synaptic strength itself. And this is kind of diagrammed here,
53:02 - 53:08: and it's the amplitude of synaptic strength that we think is under homeostatic control.
53:08 - 53:13: Now in our system, we're studying the Drosophila neuromuscular junction. Neuromuscular junctions,
53:14 - 53:22: a similar process of homeostatic control of synaptic strength is observed at neuromuscular
53:22 - 53:28: junctions in mammals and indeed humans. And parallel processes like this function in the
53:28 - 53:32: central nervous system of mammals. So it's a fundamental form of plasticity.
53:33 - 53:38: So before we get to the evidence of homeostatic control, I'll tell you about our system. Most
53:38 - 53:47: of our experiments are done in this Drosophila third instar maggot. We dissect it open to reveal
53:47 - 53:52: the repeated musculature shown here, the central nervous system, the brain, and then the motor
53:52 - 53:58: nerves that innervate these repeated muscles. Now if we zoom in on one of these
53:58 - 54:04: neuromuscular junctions, it's a glutamatergic synapse, which is different from mammalian
54:04 - 54:10: neuromuscular junctions, which are cholinergic, but we think parallel many aspects of glutamatergic
54:10 - 54:17: synapses in the mammalian brain. We can place a recording electrode in the postsynaptic muscle
54:17 - 54:23: and detect the spontaneous release of individual synaptic vesicles, so-called minis or miniature
54:23 - 54:31: excitatory postsynaptic potentials. These are small depolarizations in voltage that we use as
54:31 - 54:36: a measure of the postsynaptic sensitivity to neurotransmitters, really the state of the
54:36 - 54:43: postsynaptic glutamate receptors. We can also place a stimulating pipette on the motor nerve,
54:43 - 54:48: inject current to induce an action potential, which invades the presynaptic terminal and leads
54:48 - 54:54: to the synchronous release of many synaptic vesicles. This triggers a large depolarization
54:54 - 55:02: in the postsynaptic muscle called an excitatory postsynaptic potential, and it's the amplitude
55:02 - 55:11: of this EPSP that we think is under homeostatic control. Okay, so what's the evidence that this
55:11 - 55:16: synapse is under homeostatic control? So under basal conditions, you get relatively
55:17 - 55:23: stable levels of miniature activity and evoked amplitude shown here.
55:24 - 55:30: And when you challenge this system, for example, by genetically ablating a subset of the
55:30 - 55:37: postsynaptic glutamate receptors, you would expect the minis to be reduced and the evoked amplitude
55:37 - 55:42: to be similarly reduced because there are fewer postsynaptic glutamate receptors. However,
55:42 - 55:48: what's observed is that the minis are indeed reduced because there are fewer
55:49 - 55:56: glutamate receptors. However, synaptic strength or the evoked EPSP amplitude
55:57 - 56:04: is similar to baseline levels. And so this is a really remarkable signaling system that implies
56:04 - 56:11: a dialogue between the postsynaptic muscle and the presynaptic neuron. And somehow the presynaptic
56:11 - 56:18: neuron detects this decrease in excitability or in the state of the neurotransmitter receptors,
56:19 - 56:26: induces a signaling system that drives a retrograde signal back to the presynaptic
56:26 - 56:32: motor neuron, and precisely transmutes this signal into an increase in neurotransmitter
56:32 - 56:37: release and the amount of glutamate release to compensate for fewer postsynaptic receptors.
56:37 - 56:43: So this process is called presynaptic homeostatic potentiation because it's an
56:43 - 56:51: increase in neurotransmitter release or quantum content that is the mechanism of adaptation of
56:51 - 56:58: this form of plasticity. And again, this same process happens in rodent and human neuromuscular
56:58 - 57:07: junctions as well as in the central nervous system of mammals. So a lot is known, a lot has been
57:07 - 57:13: uncovered over the past 10 to 15 years from forward genetic screens and others that have
57:13 - 57:19: really filled in a lot of the mystery about what happens in the presynaptic motor neuron
57:19 - 57:25: that leads to this adaptive increase in neurotransmitter release. To date, about 25
57:25 - 57:30: genes have been found that function in the presynaptic compartment, and they all sort of
57:30 - 57:36: converge to two mechanisms that lead to potentiation of neurotransmitter release.
57:36 - 57:42: Number one is there is an enhancement in calcium influx at the active zone, and number two,
57:42 - 57:47: there's an increase in the number of synaptic vesicles available for release. And these two
57:47 - 57:54: mechanisms, as well as likely others, converge to tune presynaptic neurotransmitter release
57:55 - 58:01: to compensate for reduction in the sensitivity of the postsynaptic cell to neurotransmitter.
58:02 - 58:09: But a big question in our field right now is how is this entire process of PHP induced to begin
58:09 - 58:16: with? How does the postsynaptic compartment detect this change in excitability or sensitivity to
58:16 - 58:24: glutamate release and induce retrograde signaling in response? And so this is the major question
58:24 - 58:29: that my lab's very interested in, and today I'm going to tell you about some insights we've
58:30 - 58:33: uncovered into how this process is induced to begin with.
58:35 - 58:42: So, since the original discovery of presynaptic homeostatic potentiation that was found in Corey
58:42 - 58:50: Goodman's lab over 20 years ago, the major hypothesis in the field was that a decrease
58:50 - 58:58: in postsynaptic calcium was induced, is needed, is the fundamental signal that drives PHP signaling.
58:58 - 59:03: So, the idea is that you remove or perturb some of the postsynaptic glutamate receptors.
59:03 - 59:09: There should be a decrease in calcium. This decrease in calcium might impact calcium
59:11 - 59:18: response elements, such as CAM kinase 2, and this is the fundamental mechanism that senses
59:18 - 59:23: the perturbation and induces retrograde signaling. And I'm going to tell you that
59:23 - 59:29: this hypothesis is actually wrong. So, there are two fundamental subtypes of
59:33 - 59:37: postsynaptic glutamate receptors at the Drosophila neuromuscular junction
59:37 - 59:44: called GluRA receptors or GluRB. They're thought to be heterotetramers that contain the common
59:44 - 59:52: subunits, GluR2C, D, and E for both, and they have either the A or the B subunit.
59:52 - 59:58: So, what's important to know is that there are two subtypes and that GluRA receptors
59:58 - 59:59: drive the major postsynaptic perturbation.
60:00 - 60:09: Currents shown here, GluRB receptors desensitize very rapidly and transmit very little current.
60:09 - 60:21: So, here's staining at the Drosophila neuromuscular junction for GluR2A-containing subunits, GluR2A-containing receptors, B, and then the common subunit D shown here.
60:21 - 60:28: The other background I need to tell you is a little bit about how CAM kinase 2 is regulated.
60:28 - 60:34: So, as many of you know, CAM kinase 2 is one of the most abundant proteins in our brains.
60:34 - 60:39: About 5% of all protein in our brain is CAM kinase 2.
60:39 - 60:43: It forms a 12-mer holoenzyme.
60:43 - 60:47: In mammals, there's a threonine residue at 286.
60:47 - 60:50: In Drosophila, it's at 287.
60:50 - 61:03: And phosphorylation at this threonine residue is a key change that is thought to open up CAM kinase 2 into an active conformation.
61:03 - 61:14: So, generally, there's an increase in calcium influx that helps drive phosphorylation of this threonine residue, and that activates CAM kinase 2.
61:14 - 61:24: And the whole point of this is that CAM kinase 2 can detect a transient change in calcium and can transform that into a long-lasting change in activity.
61:24 - 61:29: And it's really a fundamental node of plasticity.
61:29 - 61:35: So, this is the evidence that calcium and CAM kinase 2 might be involved in PHP.
61:35 - 61:49: So, many years ago, Corey Goodman's lab and more recently my lab has validated this, that when you take a GluR2A mutant, you have only GluR2B receptors shown here.
61:49 - 61:59: So, the amplitude is small, but the evoked amplitude is similar to wild-type because of this adaptive increase in neurotransmitter release shown here.
61:59 - 62:17: So, when you express in the postsynaptic compartment a constitutively active form of CAM kinase 2, where there's a negatively charged amino acid change at 287, there's no major change in basal transmission.
62:17 - 62:32: But when you overexpress constitutively active CAM kinase 2 in a GluR2A mutant, you block retrograde homeostatic signaling, and that's shown here where you get no change in presynaptic neurotransmitter release.
62:32 - 62:45: So, the amplitude is still small, but the evoked amplitude is reduced. So, that's quantified here where I'm showing you the miniature amplitude in either genotype in a GluR2A mutant as a percentage of its baseline.
62:45 - 63:04: So, you reduce minis by about 50%, but you increase quantum content, which is the number of vesicles released per action potential, by about twofold. And when you overexpress constitutively active CAM kinase 2, there's no increase in presynaptic release to compensate.
63:04 - 63:08: And so this is evidence that retrograde signaling is blocked.
63:08 - 63:29: It's consistent with this idea that a reduction in postsynaptic calcium might lead to a reduction in CAM kinase 2 activity, and then this drives retrograde signaling, and if you perturb this by driving a constitutively active CAM kinase 2, you block retrograde signaling.
63:30 - 63:39: The major problem in the field was there were no good ways to really monitor CAM kinase 2 activity because of a lack of reagents.
63:39 - 63:56: Most of the antibody staining was done using mammalian CAM kinase antibodies, and it really hasn't been, this model has not really been tested or revisited since the early studies over 10 years ago.
63:56 - 64:16: We generated our own reagents. Here I'm showing you a really nice and specific antibody we made against Drosophila total CAM kinase 2. When we stain NMJs with it, we see a nice co-localization with the postsynaptic density marker DLG, just a homolog of PSD95.
64:16 - 64:36: So, CAM kinase 2 is enriched at postsynaptic compartments. We also made a CAM kinase 2 antibody that's specific for the phosphorylated or active form of CAM kinase 2 using this phosphorylated peptide at this threonine 287 of the Drosophila peptide.
64:36 - 64:55: And very interestingly, there's a very different staining that you see at the postsynaptic compartment with pCAMK2, where it looks much more punctate and seems to co-localize, not with the postsynaptic density marker PSD95, but rather the glutamate receptor itself, GluR2A.
64:55 - 65:24: So, using these really nice reagents now, we looked at what happens to CAM kinase 2 activity in a GluR2A mutant where PHP signaling is induced. And very interestingly, we see a complete absence of the active form of CAM kinase 2 when we look at a GluR2A mutant shown here, no change in GluR2D receptors, and no major change in overall CAM kinase 2 activity.
65:25 - 65:44: When we do an RNAi, we see against GluR2A, we see a similar impact. And when we overexpress GluR2A, we see high levels of GluR2A as expected and high levels of active CAM kinase 2 without changing total CAM kinase 2 levels.
65:44 - 66:08: So, this suggests an intimate relationship between CAM kinase 2 activity and the abundance of the GluR2A subunits. But this doesn't really uncouple whether it's calcium or some other mechanism that might be driving CAM kinase 2 activity, because when you have high levels of GluR2A, you'll have high, presumably high calcium influx and current.
66:08 - 66:14: So, we really needed to differentiate this.
66:14 - 66:29: To accomplish this, we use CRISPR-Cas9 to generate a calcium impermeable GluR2A subunit. And here at the M2 pore filter, there's normally a Q residue which allows calcium to come through.
66:30 - 66:47: We switched this. We edited the genome to include a positively charged arginine. And in other types of receptors, like AMPA receptors, it's been shown that this arginine can kick out calcium and render these receptors calcium impermeable.
66:47 - 67:03: Now, the GluR2A, these glutamate receptor subunits in Drosophila are actually closer to kainate-type receptors than AMPA receptors. So, we weren't positive that this would work, but we did calcium imaging.
67:03 - 67:21: And so, to do this, we looked at GCaMP6F localized at postsynaptic compartments. We can see nice little individual miniature events at neuromuscular junctions. We can stimulate and see a large synchronous release of many synaptic vesicles here.
67:21 - 67:48: When we did this imaging in a GluR2A mutant, we saw about 50% of the amount of calcium that we normally see in wild-type, as expected because of this reduction in the large conductance GluRA receptors. And similarly, in our Q-to-R mutation of GluR2A, we saw about a 50% decrease in calcium influx, which is really the same as a GluR2A mutant.
67:48 - 68:04: So, in terms of calcium, this single amino acid change really rendered calcium to be the same as a GluR2A mutant. And again, in this condition, retrograde PHP signaling is induced. And here's the quantification shown here.
68:04 - 68:16: So, we've stained receptors. We saw normal levels of GluR2A, B, and D receptors in the Q-to-R allele, so it did not disrupt trafficking of the receptor or anything like that.
68:17 - 68:30: And when we looked at CAM kinase activity, we saw absolutely no change in active CAM kinase 2 in the Q-to-R allele, nor did we see a change in total CAM kinase 2.
68:30 - 68:47: So, this suggests that activity of CAM kinase 2 is insensitive to a reduction in postsynaptic calcium, which is very surprising and very different than how CAM kinase 2 is thought to work in mammalian nervous systems.
68:47 - 69:05: So, but there's still about 50% of calcium in the postsynaptic compartment, which could, in principle, allow some activation of CAM kinase 2. So, we engineered a botulinum neurotoxin and expressed it in neurons, Botox-C.
69:05 - 69:25: And expressed in presynaptic terminals of a neuron, it completely, it cleaves the SNARE complex needed for vesicle fusion, and you get absolutely no miniature or evoked transmission. So, all synaptic calcium influx should be disrupted in this manipulation.
69:25 - 69:37: And what we see when we do GCaMP imaging, we see absolutely no calcium signaling in the postsynaptic compartment. So, there is zero calcium in this condition.
69:37 - 69:53: And we looked at CAM kinase activity here. And here, very surprisingly, we saw absolutely no difference in active CAM kinase 2 compared to wild-type NMJs in Botox-poisoned synapse.
69:53 - 70:07: So, as we can tell, CAM kinase 2 in this system at the postsynaptic compartment of the fly neuromuscular junction operates completely independently of calcium. Very surprising.
70:08 - 70:22: And when we did physiology in the Q-to-R mutation, we saw no change in miniature transmission, which is expected because most of the current is driven by sodium. And we saw no change in the evoked amplitude.
70:23 - 70:37: So, increasing postsynaptic calcium was not sufficient to induce retrograde PHP signaling, in which case minis might be normal, but you would expect an increase in the evoked amplitude, which we did not see and which is quantified here.
70:37 - 70:51: So, we think all of this evidence really rules out this simplistic decrease in calcium model that's sufficient to trigger retrograde PHP signaling through CAM kinase 2.
70:51 - 71:08: So, we kind of racked our heads and wondered, well, what kind of a mechanism might be involved? And we were inspired by a really seminal study by Uli Baer and Howard Schulman, where they found that the C-terminal tail of a glutamate receptor in mammals, the NMDA
71:08 - 71:24: receptor C-tail, could bind directly to CAM kinase 2 and lock it in an active conformation. So, here this model would be a scaffold model where a biochemical interaction drives activation or locks activation of CAM kinase 2.
71:24 - 71:42: So, in mammals, it's a cooperative relationship between calcium signaling and biochemical interaction with the NMDA receptor C-tail. Here in Drosophila, we don't think calcium seems to play any role, but we wondered whether a similar mechanism of the GluR2A
71:42 - 72:04: C-tail might be involved. And indeed, when we looked at the structure of the GluR2A C-tail, out at the terminal 20 amino acids of the GluR2A C-tail, we found a motif that closely matches the motif known in NMDA receptors to interact with CAM kinase 2, shown here.
72:04 - 72:25: So, here's the GluR2A C-tail. There was an antibody already available that bound the extracellular domain of GluR2A. We generated a new GluR2A-specific antibody that specifically recognizes the C-terminal tail, shown here. We then used CRISPR mutagenesis to truncate the C-terminal tail of GluR2A.
72:26 - 72:49: So, here's the wild-type terminal 20 amino acids of GluR2A, and our CRISPR mutagenesis generated a truncation of the last 20 amino acids, as well as the last six amino acids. We also made a truncation in the Q-to-R mutation of the GluR2A C-tail, so we could both truncate this putative
72:49 - 73:12: interacting domain with CAM kinase 2, as well as block a calcium influx. So, here's a schematic of these alleles that we made. We then stained receptors. GluR2A is gone, but present in all of these alleles at normal levels, so we didn't disrupt trafficking.
73:13 - 73:28: But the C-terminal tail-specific antibody of GluR2A is completely absent in all of these, as expected, because we, you know, truncated this tail. Total receptor abundance has not changed in any of these alleles.
73:28 - 73:41: So, when we looked at CAM kinase 2 activity, we were surprised it was completely absent in any and all of these C-terminal tail truncations, while total levels of CAM kinase 2 were normal.
73:41 - 74:05: This suggests that this C-terminal tail motif of about 20 amino acids is necessary to stabilize CAM kinase 2 activity in a locked conformation of high activity, and actually suggests that calcium seems to play no role in it. It's really the presence of the C-terminal tail of the GluR2A receptor.
74:06 - 74:22: We then looked whether a loss of this C-terminal tail was sufficient to allow retrograde PHP signaling to happen, and we saw no change in minis and no change in baseline transmission in the C-terminal tail truncations.
74:22 - 74:45: And even when we combined the truncation with the Q-to-R mutation, we saw normal mini amplitudes and normal evoked. So, we don't think that this loss of the tail is sufficient to then drive retrograde PHP signaling. There's likely some other signaling system induced on top of it that allows this process to happen.
74:46 - 74:59: So, we wondered, so what is the point of having this CAM kinase 2 dependent mechanism, and does it play any role in retrograde PHP signaling? And to address this question, we generated chimeric receptors.
74:59 - 75:15: So, here's the GluR2A subunit with its C-terminal tail. Here's the GluR2B subunit. And we molecularly swapped the A tail with the B tail, generating a chimeric GluR2B receptor with the GluR2A tail.
75:16 - 75:38: And when we express the GluR2B, the wild-type GluR2B, in a null mutation that removes A and B receptors, we see that it basically looks like a GluR2A mutant. We see no GluR2A staining, but we see normal levels of GluR2B in this condition.
75:38 - 75:45: So, we can basically phenocopy a GluR2A mutant, which normally induces retrograde PHP signaling.
75:45 - 75:59: When we put in the chimeric receptor tail, we now see the C-terminal tail as expected. We do not see GluR2A staining when recognizing the non-C-terminal tail.
76:00 - 76:15: And when we look at pCAMK2 activity, surprisingly, we see completely normal levels of active CAM kinase 2 when we swap the C-terminal tail onto the B receptors shown here.
76:15 - 76:27: So, we think this now demonstrates that the C-terminal tail of GluR2A is both necessary and sufficient to recruit active CAM kinase 2 to postsynaptic compartments.
76:27 - 76:31: Now, what does this have to do with retrograde PHP signaling?
76:31 - 76:44: So, we did physiology in these conditions when we replaced postsynaptic glutamate receptors with just a wild-type GluR2B receptor, which is basically the same as a GluR2A mutant.
76:44 - 76:53: We see small minis and a normal evoked, which demonstrates that PHP signaling is induced and expressed normally.
76:53 - 77:02: However, when we put in this chimeric receptor, minis are similarly reduced because these B receptors transduce very little current.
77:02 - 77:13: But now there is no change in presynaptic neurotransmitter release, and the EPSP is reduced, which demonstrates that retrograde PHP signaling is blocked in this condition.
77:13 - 77:28: So, despite the highly reduced miniature activity, which is essentially the same as a GluR2A mutant, the physical presence of this GluR2A C-tail completely blocks retrograde signaling.
77:28 - 77:31: So, this is our current model for how this is involved.
77:31 - 77:43: Retrograde PHP signaling allows a disinhibition of retrograde PHP signaling and really is sensing the physical presence of the GluR2A C-tail.
77:43 - 77:48: We think calcium influx is not involved at all in this signaling process.
77:48 - 78:05: And so, one major step to allow retrograde homeostatic signaling to occur is that CAM kinase 2 is sensing the physical loss of the GluR2A tail of GluR2A receptors.
78:05 - 78:14: And so, when this physical loss is sensed, you lose active CAM kinase 2, and that now allows retrograde PHP signaling to happen.
78:17 - 78:36: So, what we think is actually very different in how Hebbian plasticity works and how GluR2A, how CAM kinase 2 activity works at postsynaptic compartments in the context of Hebbian plasticity and interacting with the NMDA receptor C-tail.
78:36 - 78:44: This is really a novel mechanism to regulate CAM kinase 2 and indeed retrograde homeostatic signaling.
78:44 - 78:59: This work was really pioneered by an excellent postdoc in the lab, Sarah Perry, and a great graduate student, Yi-Fu Han, together really made all these wonderful reagents and revealed this signaling system, as well as the rest of the members of my lab,
78:59 - 79:02: on other projects I don't have time to talk about.
79:02 - 79:12: So, I would be happy to take any further questions. Thanks very much.
79:12 - 79:24: Thanks very much, Dion. That was really interesting. So, I wonder, I'm going to ask, I can't help but ask a couple of questions myself.
79:25 - 79:47: So, the complete loss of phospho-CAM kinase 2 was kind of interesting because it suggests that not only is that interaction necessary with the glutamate receptor C-terminal tail, but also that the phosphorylation only occurs when that interaction occurs.
79:48 - 79:59: Am I correct in assuming that, right? So that there's no other way, which is really, really unusual because you might expect for CAM kinase that calcium would do this, but you don't see that. Is that right?
79:59 - 80:00: Right. Yeah.
80:00 - 80:06: Or do you think, like, that the antibody, that there's something about the shape of the protein
80:06 - 80:11: when it's bound to the C-terminal tail that could make the antibody, you know, that impacts whether
80:11 - 80:16: the antibody interacts? Yeah, that's a great question. And, you know, the antibodies are
80:16 - 80:22: crucial. The specificity of the antibodies is crucial to really confirm our model. So,
80:22 - 80:27: we spent a lot of time, we put on CAM kinase 2 inhibitors. We overexpress, you know,
80:27 - 80:34: constitutively active and inactive forms, various peptides that we know inhibit CAM kinase 2,
80:34 - 80:40: and all of them, the antibodies look the way you would expect them to look if they were specific.
80:40 - 80:46: So, when you get rid of the active, you know, the active form of CAM kinase 2, independently of
80:46 - 80:55: anything with the glutamate receptor or with that T287D, you still inhibit or block the active,
80:55 - 81:00: you know, the antibody from recognizing active CAM kinase 2. So, we're quite confident in the
81:00 - 81:06: specificity of the antibodies, which is important. But you're absolutely right. In mammals, almost
81:06 - 81:12: all the data that I know of, even the ones where there's an interaction with the glutamate receptor
81:12 - 81:19: tail, the NMDA receptor, calcium still plays a role to help change that activity. And here,
81:19 - 81:24: it's a big surprise. At least in the context of homeostatic plasticity, it appears to be
81:24 - 81:29: completely calcium independent. And it's also interacting with the kainate receptor, which are
81:29 - 81:36: enigmatic and kind of mysterious, as you know, in mammals in this kind of really interesting way
81:36 - 81:44: where it shares some mechanistic parallels with what happens with NMDA receptors, but it's also
81:44 - 81:51: quite novel. So, yeah, we think this is an ancient kind of evolutionary adaptation that's really just
81:51 - 81:57: sensing the physical presence of this receptor and allowing the system to either be constitutively
81:57 - 82:04: inhibited or allowed to be expressed. And the antibodies work on western blots, on denatured
82:04 - 82:12: protein as well. We've done a few of those, yes. They do seem to work, but you lose localization.
82:12 - 82:17: We actually have it be interesting to know if they would cross-react with mammalian systems. We
82:17 - 82:21: haven't actually tested that. So, if anyone wants to try some fly antibodies against mammals, that
82:21 - 82:27: would be great. Interesting. Yeah. So, there are, as you might imagine, there are a bunch, there are
82:28 - 82:35: questions in the chat about what, you know, I think just to rephrase the question, you know,
82:35 - 82:43: both what's upstream of CAM kinase 2. Is it really, so it seems to me that your results
82:43 - 82:49: and your interpretation was that the loss of the receptors is what is triggering this signal,
82:50 - 82:58: this PHP signal. And what's upstream of that, do you think? Yeah. So, it's a great question. We were
82:58 - 83:05: hoping that just the physical loss of the receptor tail would be sufficient to now induce retrograde
83:05 - 83:10: homeostatic signaling. And clearly, we did the experiment. We even combined that with loss
83:10 - 83:17: of calcium or reduction in calcium. And together, that was not sufficient to induce. So, this is,
83:17 - 83:23: we think of this as a constitutive break on the system that just, you know, that disrupts, that
83:23 - 83:29: can block that whole signal from happening. And you need to relieve that to allow the process to happen.
83:29 - 83:34: But there's something else involved. There are several candidate mechanisms. We know new gene
83:34 - 83:41: translation is involved. So, there must be some new genes, we think, that are being translated,
83:41 - 83:46: new protein synthesis. But the targets of that, I would love to know. We have some candidates who
83:46 - 83:52: are working on it, but we don't have that full process out. In terms of the downstream signals,
83:52 - 83:58: so there are a few candidate retrograde signals that, you know, we found and, you know, that was
83:58 - 84:05: found in Gray Davis's lab as well. One is multiplexin, which is an extracellular matrix
84:05 - 84:13: molecule, as well as plexin, you know, plexin signaling, you know, that's involved in axon
84:13 - 84:18: guidance actually might be involved in this process as well. There's still candidate retrograde
84:18 - 84:23: signals. So, what happens downstream is still unclear, I would say, but there are a few candidates
84:23 - 84:29
that we can, that have been put out there. Yeah, and there's a question about a voltage-dependent
84:29 - 84:35: calcium channel that just appeared in the chat. And I would imagine the changes in calcium influx,
84:35 - 84:39: I mean, they have to involve voltage-dependent calcium channels, a presynaptic terminal, yeah?
84:41 - 84:46: Right, right, right, right. So, that's involved in the expression mechanism of increasing
84:46 - 84:51: neurotransmitter release. And yes, the voltage-gated calcium channels, the CAV2 channels that are at
84:51 - 84:56: the active zone, those definitely seem to be modulated. Actually, in that case, there seems
84:56 - 85:02: to be an increase in the apparent abundance of these channels at active zones very rapidly in
85:02 - 85:09: the context of homeostatic clusters. Oh, interesting, okay. And are you, is, are the
85:09 - 85:15: working hypotheses in terms of the retrograde, the actual signal, that it could be kind of a
85:15 - 85:19: protein-protein interaction with large proteins spanning the synaptic cleft?
85:20 - 85:27: Right, well, so, yeah, it's quite an interesting idea. We have preliminary evidence, actually, that
85:27 - 85:34: that this retrograde signaling can happen in the absence of any synaptic activity,
85:34 - 85:38: which might suggest that it's actually conformational protein-protein interactions
85:38 - 85:44: that are trans-synaptically being, you know, transmitted between the postsynaptic compartment
85:44 - 85:51: to the pre. And so, yeah, extracellular matrix in the synaptic cleft is a very attractive idea.
85:51 - 85:56: There are large extracellular domains of the calcium channel and various active zone components
85:56 - 86:02: that jut out into the synaptic cleft, and the glutamate receptor itself is bound, we know,
86:02 - 86:07: to a lot of other postsynaptic scaffolds. So we think this whole, you know, signaling module
86:07 - 86:13: is in tight interactions, and conformational signals can directly be transmitted to actually
86:13 - 86:19: change synaptic structure and function in a very rapid and acute manner, yeah.
86:19 - 86:24: Yeah, and we've certainly got a list of more and more proteins that seem to link pre- and
86:24 - 86:28: post-synaptic, you know, I'm thinking of neurexins and all kinds of...
86:28 - 86:30: Neuroligins, we're looking at those as well, yeah.
86:30 - 86:31: Sure.
86:33 - 86:40: Yeah, that's cool. Let's see, I want to make sure. So there's a question in the chat about
86:40 - 86:46: parallels between these findings of Drosophila and in mammalian synapses in rodents.
86:47 - 86:53: Yeah, so there are a lot of interesting parallels, but also quite a lot of distinctions here. So
86:54 - 87:00: almost every form of Hebbian plasticity in mammals has been shown to involve CAM kinase 2.
87:00 - 87:05: Here we're looking at homeostatic plasticity, which is actually a counteracting force,
87:05 - 87:11: we think, that's happening. A few studies have linked CAM kinase 2 to homeostatic plasticity as
87:11 - 87:19: well in mammals, but it's less clear how direct that interaction is, I guess. Clearly, there's
87:19 - 87:25: a intimate interaction with the C-terminal tails of glutamate receptors in both systems.
87:25 - 87:33: NMDA, as everyone knows, is crucial for synaptic plasticity of glutamatergic synapses. In our case, we're looking at a kainate receptor.
87:38 - 87:44: I think the biggest difference and the most surprising is that in mammals, calcium, you know,
87:44 - 87:51: dynamic changes in calcium influx are intimately coupled with activation of CAM kinase 2,
87:51 - 87:57: as well as with other scaffolds. And here, you know, we really think we've ruled out any calcium
87:57 - 88:02: modulating CAM kinase 2 activity, at least in this postsynaptic compartment.
88:02 - 88:07: It really seems to be a sentinel for just detecting the presence of this receptor.
88:07 - 88:12: So, there's various evolutionary reasons we can get into about why that might be the case,
88:12 - 88:14: but that's probably the biggest difference.
88:14 - 88:22: Yeah. And so, is it also consistent that putative phosphorylation targets of CAM kinase 2,
88:23 - 88:27: they don't seem to be involved? I mean, maybe you already said this, but...
88:27 - 88:31: Oh, sorry. Yeah. So, there are actually a lot of common homologies.
88:31 - 88:39: We have not yet linked them to homeostatic plasticity, but PSD95, for example, we know
88:39 - 88:47: is phosphorylated in the postsynaptic compartment in flies, just as it is a target in mammals also.
88:48 - 88:52: There's a few other scaffolds that, you know, there's a potassium channel and a few other
88:52 - 88:58: things that might be attractive targets. And so, we're looking at DLG PSD95 right now
88:58 - 89:03: with homeostatic plasticity, and we actually have evidence it's involved, but we're looking at the
89:03 - 89:08
link with CAM kinase 2 right now. So, in terms of the substrates downstream, that might be a
89:08 - 89:15: point of convergence, actually. Yeah. Yeah. And I suppose it could be
89:15 - 89:20: that CAM kinase 2 is a linchpin in a large multi-protein complex, you know.
89:20 - 89:24: Exactly. Yeah. It needs both to be there
89:24 - 89:28: in terms of its protein structure, but also maybe in terms of its kinase activity.
89:28 - 89:32: Right. And its scaffolding as well, right? Yeah.
89:35 - 89:43: Let's see. Good. So, I wanted to ask you about phenotype. And, you know, one thinks about
89:43 - 89:49: homeostatic plasticity and how when you manipulate it, as you do so elegantly in these experiments,
89:49 - 89:56: that you could peg a system. What happens to the flies? You see that there's a consequence.
90:02 - 90:11: Yes, exactly. Or if you push it to one extreme,
90:02 - 90:05: you know. Yeah. So, it's a great question.
90:06 - 90:11: So, there are behavioral defects. So, you know, the neuromuscular junctions needed for movement.
90:11 - 90:18: And if you look at how the flies, you know, the larvae are moving in mutants that we know
90:18 - 90:23: disrupt homeostatic plasticity, the movement, you know, there's problems in the movement.
90:23 - 90:28: In the central nervous system, actually, that's where we're more interested in the links of
90:28 - 90:33: behavior. And so, in fly mutants that disrupt homeostatic plasticity, some of the mutants we
90:33 - 90:38: found that disrupted at the NMJ, they have behavioral defects in learning and memory,
90:38 - 90:43: and very interestingly in sleep as well, and fly courtship. Actually, all of these have been linked.
90:44 - 90:50: Some of the fly homologs have been linked to human psychiatric diseases, actually. So,
90:50 - 90:55: one of the genes that we found in flies that blocks homeostatic plasticity is a mutation in
90:55 - 91:01: the fly homolog of dysbindin. And in human, you know, GWAS studies and others, dysbindin is a
91:01 - 91:07: susceptibility factor for schizophrenia. So, there's a lot of interesting links with disrupted,
91:07 - 91:15: synapses and seemingly disparate, complex neuropsychiatric diseases that might share some common etiology in stabilizing synaptic stress.
91:15 - 91:20: Yeah, yeah. And I actually, I have just overheard, the rumor mill is spinning,
91:25 - 91:34: and there's a new publication from the Psychiatric Genome Consortium, where they're
91:34 - 91:41: going to reveal additional hits, you know, so it's building on the GWAS studies, and
91:42 - 91:47: my understanding is that there'll be a lot of details about higher risk-odd
91:48 - 91:52: genes. So, you know, more penetrant, if you will, and more rare
91:54 - 91:59: hits. And these are synaptic. Most of these are synaptic structural. So, there'll be some
91:59 - 92:03: interesting stuff. We'd love to get that list, and we'll screen those for…
92:04 - 92:10: You'll recognize some genes on the list, I have no doubt. And I think also copy number stuff
92:10 - 92:16: has been revealed. So, really interesting links to schizophrenia in this case.
92:17 - 92:22: Good. I think at this point, we should invite Jennifer back in, and we can start
92:22 - 92:30: the roundtable bit. Please, for those of you who've joined the webinar, please continue
92:30 - 92:36: posting in the chat, and we'll try to liven up the roundtable discussion.
92:38 - 92:49: So, I think, you know, moving kind of beyond both of your elegant research programs and
92:49 - 92:57: the findings that you're looking at, it'd be good to hear your thoughts about big questions
92:57 - 93:04: that are remaining. And I'd be interested also in hearing how you think about big questions that
93:04 - 93:13: also touch on the kind of systems elements from your points of view and your respective systems,
93:13 - 93:19: right? What don't we know where some form of synaptic plasticity is very likely involved?
93:19 - 93:24: Where is there a big gap, and what might it explain? So, why don't we start, Jennifer, with you?
93:24 - 93:34: Yeah. So, I guess, in terms of those big questions, for example, when we study this
93:35 - 93:42: system neuroscience, it's kind of, by default, we look at more the Hebbian rules and how it
93:42 - 93:50: drives the plasticity. But as D.L. mentioned, this homeostatic plasticity is so essential,
93:50 - 93:56: it's so essential, and that serves as kind of the fundamental that we can maintain this
93:57 - 94:05: stability and a healthy environment for us to carry on and be able to be Hebbian. So,
94:06 - 94:12: I think, but that's relatively less studied, I would say, in system neuroscience.
94:13 - 94:23: One of the issues is, like, it doesn't take effect within a short period of time in the
94:24 - 94:31: behaving system. Sometimes, for example, during sleep, I think there are some very elegant studies
94:31 - 94:38: that have been touching upon that phase, where how the general plasticity is being downregulated
94:38 - 94:47: or to maintain the homeostasis over time and things like that. But so far in the field,
94:48 - 94:57: we still lack proper tools or assessment to directly answer that question at this level,
94:57 - 95:03: whereas we know that they're being developed so elegantly at the synaptic level and the
95:03 - 95:14: molecular level to address that question. Yeah. Yeah, I mean, I think that those are great points.
95:14 - 95:19: And, you know, I think for me, I have three big things that I think are really, at least for me,
95:19 - 95:26: are very interesting. So, number one, you know, we tend to study a single form of plasticity,
95:26 - 95:31: in my case, homeostatic plasticity, a lot of labs. And as Jennifer talked about,
95:31 - 95:36: they're studying, you know, Hebbian plasticity rules. But really, in the dynamic nervous system,
95:36 - 95:43: these plasticities, these forms of plasticity are working in conjunction and need to be integrated
95:43 - 95:50: somehow. And so, how are homeostatic and various other forms of plasticity integrated? And to
95:50 - 95:55: maintain coherent information transfer, I still think it's quite remarkable that we don't walk
95:55 - 96:01: around as kind of raving lunatics half the time. You know, something is stabilizing our cognitive
96:01 - 96:07: function under normal times. And, you know, you've got to really have a firm grasp on maintaining
96:08 - 96:13: stability while allowing the flexibility that's necessary for learning and memory.
96:14 - 96:19: I also think, you know, how do these processes change throughout development in humans, you know,
96:19 - 96:27: from early youth, you know, young adults, you know, and maturation? And really, interestingly,
96:27 - 96:33: this plasticity changes in late stages of life. And why does it break down? Why doesn't it stay
96:33 - 96:40: as robust? Are there ways to change that? How does that happen in disease? And how are Hebbian
96:40 - 96:46: and homeostatic processes targeted in disease processes? And then personally, for me, I think
96:46 - 96:53: sleep is really fascinating. And what is happening during sleep? Almost every psychiatric disease
96:53 - 96:58: has a sleep disorder associated with it. What happens during sleep in terms of Hebbian and
96:58 - 97:03: homeostatic plasticity and renormalization of synaptic strength and things like that?
97:04 - 97:13: I think sleep is a really interesting next frontier. Yeah. Yeah. I was just going to say,
97:13 - 97:18: I mean, Jennifer, you as a computational neuroscientist, and certainly where I sit,
97:19 - 97:23: you can't get anywhere without some homeostatic mechanism, a basic homeostatic mechanism for
97:23 - 97:29: rescaling. And if one looks at what happens to any neuron in the nervous system,
97:30 - 97:33: certainly in a mammalian nervous system, they grow so much, there are all these additional
97:33 - 97:39: ion channels. It has to happen. There have to be robust and probably numerous mechanisms here.
97:40 - 97:48: So this is but a start, but I think we're lost without it. Anyway, you were going to add
97:48 - 97:56: something, Jennifer. Oh, yeah. I want to follow up on Dion's point that it is really amazing how the
97:56 - 98:05: brain should or is able to recruit so many different ways to maintain this fine balance between the
98:05 - 98:14: stability and flexibility. Because I think at a certain stage, we actually appreciate this
98:14 - 98:22: stability. We don't want to see something or hear something, get one set of responses today versus
98:22 - 98:32: tomorrow. And this stability actually sometimes can prevent us from being flexible or recover
98:32 - 98:41: from those damages or so. So if there are some general rules we can uncover to see how to
98:41 - 98:50: perturb this fine balance when we need it. And also when we, depending on different
98:50 - 99:00: developmental stages, then we can really utilize it to help design some relative translational
99:00 - 99:08: work to be able to manipulate that in human patients. And again, that brings to another
99:08 - 99:15: question is whether there are some general rules that can be transferable and to what degree
99:15 - 99:22: between different systems we use and also different animal species we're looking at.
99:22 - 99:28: And even there are so many ways to induce plasticity or drive it,
99:28 - 99:33: whether there is something that's converging for us to be able to extract.
99:37 - 99:42: Yeah. Do you, I know that this is probably technically difficult Dion, but are there any
99:42 - 99:48: indications that the mechanism, the homeostatic plasticity mechanisms have a kind of critical
99:48 - 99:54: period or that in adults they're less or more active? Yeah, it's a great question.
99:55 - 99:59: There's a lot of evidence that various
100:00 - 100:05: states of stress or I would say conditions have been shown to modulate the set point
100:05 - 100:10: of homeostatic plasticity and change it. So we've actually looked at what happens to synapses
100:10 - 100:16: during injury. So when you injure the motor neuron, there's a really interesting dialogue
100:16 - 100:22: that happens between the pre- and post-synaptic compartments. And the ultimate change is that
100:22 - 100:28: the homeostatic set point of synaptic strength is reduced to a new level, but it's homeostatically
100:28 - 100:33: maintained at that new level in the context of injury to the neuron. So if you injure the neuron,
100:33 - 100:39: this process happens. There's a lot of evidence that in aging, there's cognitive decline,
100:39 - 100:47: less robust Hebbian plasticity, sleep is fragmented, and homeostatic changes can kind
100:47 - 100:56: of be maladaptive in that context. Very interestingly, the highest susceptibility
100:56 - 101:01: of seizures in humans is between ages three and five during a condition of immense neuronal
101:01 - 101:07: growth. And almost certainly those are breakdowns in the homeostat that maintains stability there.
101:07 - 101:15: So I think there's lots of interesting areas in which other forms of stress, disease, injury,
101:15 - 101:22: growth, and development can challenge homeostatic stability in a seemingly normal situation.
101:22 - 101:29: And you can see the consequences of it, even when there are no forms of neural diseases like
101:29 - 101:35: seizures and things like that. Yeah. And as Jennifer touched upon,
101:36 - 101:43: there are good reasons why nervous systems might want to kind of turn a bit towards stability and
101:43 - 101:49: maybe constrain some of the mechanisms. But of course, the consequences in the face of disease
101:49 - 101:59: or insult can be very severe, but nature maybe has said, well, if we can get this far,
102:01 - 102:07: this is the way we'd like to do it in the interest of stability. But right, understanding those
102:07 - 102:14: things could give us real advances in terms of therapeutics for a number of diseases,
102:14 - 102:21: late-life diseases, that's for certain. I do want to ask both of you about,
102:21 - 102:27: neither of you really talk directly about inhibition, but I want to ask you, I've been
102:27 - 102:34: struck over the years and I've dabbled in this field a bit. We know relatively much less about
102:35 - 102:42: plasticity of inhibitory circuits, but I think we also know that this happens. This is a big part
102:42 - 102:47: of learning. This is a big part of circuit development. Why do you think this is?
102:51 - 102:56: Jennifer, I think this is right up your alley with inhibition, right?
102:56 - 103:03: Yeah. I felt like, first of all, we just get the chance to really look at those interneurons
103:03 - 103:09: at relatively much later stages because their population is so sparse in comparison to
103:09 - 103:16: excitatory neurons. And most of them are more local projections versus the excitatory neurons
103:16 - 103:22: are like automatic manifestations of the function that they give out to the majority of the
103:22 - 103:28: responses they project to other cortical areas, most of the time, although now we know some of
103:28 - 103:37: the inhibitory tone also play a part. But in terms of their electrical properties, just using
103:37 - 103:43: electrophysiology for recording, it's really hard to distinguish them. So
103:45 - 103:49: people just use a biased way to sample the excitatory neurons most of the time.
103:49 - 103:59: But I think now it's really a prime age for people to use all the advanced imaging tools,
103:59 - 104:06: transgenic tools to directly look at those different types of interneurons. I'm sure
104:07 - 104:14: everyone's aware of the Allen Institute and other places that have all this cell type
104:14 - 104:23: transcriptomics and those maps really showing that the interneurons are very diverse in comparison
104:23 - 104:29: to excitatory neurons. And then they give out all those different functions that wait for us to
104:30 - 104:38: really take a peek at what they are doing and what they can do. And just to talk a little bit about
104:38 - 104:46: the adult plasticity that I didn't have the chance to say much is that when people do stem cell
104:46 - 104:54: transplantation, if you do the excitatory neurons, that doesn't work versus if you do the interneurons,
104:54 - 105:03: sometimes several types of interneurons, they can actually effectively perturb the balance
105:03 - 105:10: of excitatory-inhibitory tone in that sense. And then they will be able to induce the new
105:11 - 105:18: time window for the plasticity to happen. And those parts of things sort of indicate
105:19 - 105:25: and along with other studies showing that those inhibitory neurons might be even more plastic
105:26 - 105:32: than the excitatory neurons. And then that could be a good opening venue for us to really
105:32 - 105:38: look at it and use it as a biomarker or therapeutic target. Yeah.
105:38 - 105:45: Yeah. So the inhibitory plasticity involving inhibitory circuits, you feel is more
105:46 - 105:50: active in later life. Is that another way to put it? Yeah.
105:50 - 105:59: Yeah. And I think D.L. mentioned this three to five-year-old epileptic symptoms.
105:59 - 106:07: I felt like, so at the early age, the inhibitory tone is kind of low, but also the excitatory
106:07 - 106:13: tone is that. And then when they are sort of picking up, but not yet that mature, that's where
106:14 - 106:22: infantile seizures and other things can actually result. So it is this fine balance they need to
106:24 - 106:28: achieve at that moment. Yeah. Yeah.
106:29 - 106:33: Yeah. I would just add, you know, gating, you know, that you've talked about critical
106:33 - 106:39: periods, Tom, and inhibitory circuits are really important, as I think Jennifer well knows,
106:40 - 106:44: in gating that critical period. And in homeostatic plasticity, I didn't directly
106:44 - 106:50: answer your question, but there's a lot of evidence that upstream circuits depend on
106:50 - 106:56: inhibition to stabilize excitatory-inhibitory balance and the overall nervous system
106:57 - 107:01: and within circuits. And so I think that's also a good point as well.
107:01 - 107:08: Yeah. And it seems very, very likely that retrograde mechanisms are going to be important
107:08 - 107:13: in many of these forms of plasticity, right? Because the, I mean, fundamentally, the interneurons
107:13 - 107:20: have to understand how excitable the principal neurons are, right? In order for this to work.
107:21 - 107:27: So you're going to have to have that, you know, principal excitatory neuron to
107:27 - 107:33: inhibitory neuron plasticity as well as a big part of the picture. And I think the point you
107:33 - 107:39: made, Jennifer, about the diversity of roles for interneurons, which are quite diverse,
107:40 - 107:46: we're understanding and the classes of these interneurons, it won't be a surprise if learning,
107:46 - 107:53: there's a relative abundance of learning rules for these kinds of plasticities, given the different
107:53 - 107:59: roles of the interneurons. I mean, it isn't to say that excitatory neurons don't do many things,
107:59 - 108:05: but I think we're learning about the specific roles that different classes of inhibitory neurons
108:05 - 108:11: have, and they're quite specific. Yeah. And adding to that at the local level,
108:11 - 108:18: at the synapse level, those interneurons, because they tend to receive much more
108:18 - 108:25: input from their nearby area. So those supporting cells tend to actually have a much greater effect
108:25 - 108:34: on them as well, like the glial cells. And even the, I think the general network,
108:35 - 108:44: when people are saying how to restrain the local environment, that's more on the interneuron to
108:44 - 108:47: make sure they are relatively stable there. Yeah.
108:51 - 108:58: And in terms of kind of disease biology and genetics and things like that, are there things
108:58 - 109:05: that each of you pay attention to? I mean, where do you see, where do you see sort of,
109:05 - 109:11: let's just take genetics, genetic research into psychiatric disease, just to sort of constrain
109:11 - 109:16: the question. Where are there advances that you're going to be paying attention to, each of you?
109:19 - 109:25: Dion, you can go first. Yeah. Well, so I think the incredible bioinformatics
109:25 - 109:31: and GWAS studies that I think you've mentioned, Tom, and there's been an explosion in the last
109:31 - 109:37: few years that are, I think, really converging on a lot of interesting and potentially novel
109:37 - 109:42: genes that have been implicated in psychiatric disease. It's really interesting, neurodegenerative
109:42 - 109:48: disease, I feel like it's very obvious what's wrong. We see cell deaths of certain neurons,
109:48 - 109:54: how to fix it is not clear, but it's not a major question what's happening. Whereas in psychiatric
109:54 - 110:01: diseases, there's something really subtle happening that's having really major cognitive
110:01 - 110:06: changes that you can't pin down and just look at a brain and say, okay, clearly this is how
110:06 - 110:12: it's happening. So to me, that says, at least in the case of neuropsychiatric diseases, there's some
110:12 - 110:18: problem in tuning of functional neural circuits that you really are going to need to understand
110:18 - 110:24: how our neurons are talking to each other and changing over time, which is perfect for this
110:24 - 110:30: idea of synaptic plasticity, Hebbian and homeostatic changes there. So I'm really looking, it's hard
110:30 - 110:37: to model psychiatric diseases, I think in rodents, but especially in flies. So I don't try to model
110:37 - 110:44: the diseases in flies. I look for wisdom in the human studies to find new genes and processes,
110:44 - 110:51: and then test and interrogate them in our experimental models. But I'm sure Jennifer
110:51 - 110:58: has some interesting ideas as well. Yeah, I think I haven't been having a lot of
110:59 - 111:06: chances to look at the psychiatric diseases, but for some of them, including even autism,
111:06 - 111:14: it seems like there's also some perturbed balance or proper function at the sensory levels.
111:14 - 111:23: So for psychiatric diseases and those autism, it seems like the balance between excitation and
111:23 - 111:30: inhibition at those places, at least at the cortical level, are definitely abnormal.
111:30 - 111:38: And it could be there's something downstream or neuromodulatory already sort of
111:38 - 111:44: playing a role because they are actually broadcasting all over to the cortical area.
111:44 - 111:54: And I think it's, as Dion mentioned, it's really hard to establish a psychiatric disease model for
111:54 - 112:02: people to approach it. And for now, what I've been following a little bit is about those,
112:03 - 112:11: the drugs that tend to mimic those psychiatric conditions, like LSD or so, they sort of recruit
112:11 - 112:19: some of the neuromodulators and then they tend to then exert or broadcast their
112:20 - 112:30: functional activities. And that might give us a way to look into those diseases or the condition,
112:30 - 112:35: temporarily altering the balance between the excitation and inhibition at those sensory
112:35 - 112:45: cortices. So I think maybe this might be a good thing to explore first at the slice level,
112:45 - 112:51: the in vitro level, for us to get a good understanding of what are the cellular and
112:51 - 113:01: molecular bases are there. And then maybe there's some insight for us to develop a proper
113:02 - 113:07: assessment at the behavior level to really look into it. Yeah.
113:07 - 113:17: Yeah. Well, I couldn't echo more your lament about modeling psychiatric disease in rodents. I mean,
113:17 - 113:22: you know, in my time at Roche, we of course, there were teams who thought very, very deeply
113:22 - 113:28: about this and, you know, all of pharma. And it is tough. There's not a lot of
113:28 - 113:35: predictive validity in many of the models and the behaviors, of course, that are key
113:36 - 113:44: symptoms of psychiatric diseases are difficult to find homologs for in rodents. What I will say,
113:45 - 113:53: which I think is, you know, cause for a lot of optimism, you can see in neurodevelopmental
113:53 - 113:59: disorders that there are, you know, there are about two dozen now well-understood
113:59 - 114:06: monogenic forms, syndromic forms that have comorbid autism and their mouse models can be
114:07 - 114:12: really helpful. It doesn't mean that the symptomology, symptom presentation of the
114:12 - 114:19: rodents is going to be necessarily the same. I think that needs, deserves deep scrutiny, but
114:19 - 114:28: the genes give rise to disorders and those disorders have some relationship to the human
114:28 - 114:40: disease. And those rare genetic alleles are highly penetrant, right? They're monogenic
114:41 - 114:47: forms that lead to these severe neurodevelopmental disorders. In schizophrenia, this has been less
114:47 - 114:55: the case, right? It's been, as Dion said, big GWAS studies, very small odds risk ratios. You
114:55 - 115:02: know, the biggest hits give you a change from 1% lifetime odds for being diagnosed with
115:02 - 115:10: schizophrenia to maybe 1.5% or 1.4%, right? So this is tough. And you've got a list of 200 genes
115:10 - 115:18: and counting, right? So as we, you know, if indeed we can see some rare single-gene forms
115:18 - 115:26: of psychiatric disorders that have similarity to schizophrenia or that are schizophrenia,
115:26 - 115:33: this will be, I think, you know, this will enable some research development because those can be
115:33 - 115:39: then modeled in various systems and we can see what happens. Humility is in order. You know,
115:39 - 115:44: again, you can't assume that that's going to be a schizophrenic mouse or a mouse that has
115:44 - 115:51: Angelman syndrome or, you know, severe, I mean, many of the very severe early neurodevelopmental
115:51 - 115:58: diseases are, you know, are quite similar, but I think a mouse is not a person, of course.
115:59 - 116:06: So, yeah, there is some, as we learn more in genetic advances, I think this is definitely a
116:06 - 116:17: space to watch. Good. I think, I don't know, do you want to raise any issues before we
116:18 - 116:20: wrap things up to either of you?
116:20 - 116:25: I can just start, you know, I think it's an exciting time for research and I think, you know,
116:25 - 116:32: with CRISPR-Cas9, all the genetic studies that Tom talked about, I mean, I think there's,
116:33 - 116:38: it's a really difficult question, lots of intractable problems that have been difficult
116:38 - 116:44: to really get our, you know, hands around, but I also think there's a lot of exciting
116:45 - 116:50: things happening now and even more so with the development of these new technologies
116:50 - 116:56: and approaches. And so, you know, my hope is that I want to end on a positive note about
116:56 - 117:03: the potential, and I think there's a lot of really great new insights and technologies
117:03 - 117:06: that are, you know, in the next five years we're really going to see come out. And so,
117:07 - 117:11: and I think that that will elucidate a lot. You know, the company, you know,
117:12 - 117:16: that that will elucidate a lot. You know, the combination of the genetic studies that Tom
117:16 - 117:22: talked about, but, you know, I think CRISPR-Cas9, it's revolutionized my lab in three years,
117:22 - 117:29: you know, for what we can do, and I'm sure all across research it'll have a big effect.
117:32 - 117:39: Yeah, I think really this is the age with advanced biotechnology and all the
117:39 - 117:43: interdisciplinary studies coming out, so that gives us a lot of hope. And
117:45 - 117:53: also this, recently I've been looking at those new studies from immunology
117:54 - 118:01: to look at the neurons, and apparently there's so much interaction at the neuroglia,
118:01 - 118:07: neuroimmune level that we weren't able to closely look at it. And now with all the tools here,
118:08 - 118:13: especially visualizing or perturbing them so easily with CRISPR and everything,
118:13 - 118:19: we really have the chance to answer those questions that can either be
118:21 - 118:29: done only electrophysiology or only at the cellular biology histology. Now everything can
118:29 - 118:37: be sort of combined together and survey a question at all different levels and scales,
118:37 - 118:44: and then give us an integrated view at the end to really maybe provide a much
118:45 - 118:54: stronger answer than before. Yeah. Yeah, that's a wonderful way to wrap up, and,
118:54 - 119:00: you know, I would add that we're all aware as scientists that, you know, you stand on the
119:00 - 119:06: shoulders of your predecessors, and I think the richness of not only understanding
119:06 - 119:14: the details, mechanistic details at the molecular level and the sort of the tools,
119:14 - 119:22: the incredibly, the essential tools that those give you to do more specific experiments,
119:22 - 119:31: those are on some of these shoulders, but also the ever more detailed understanding
119:31 - 119:37: of the systems that we work in, and the perspectives about why you might have
119:38 - 119:43: different types of synaptic plasticity, or why forms of synaptic plasticity might be
119:44 - 119:51: engaged at different stages of the animal's life, for example. Those kinds of things,
119:51 - 119:59: I think, bode well for the future, because that context will also drive the creativity about
120:00 - 120:17: how things might work, but also enable sophisticated assays and interpretations. We didn't even talk about population recording of how information's coded, but I think that bears on this as well, and obviously, there's been a great advance there.
120:17 - 120:30: Good, so I want to thank everyone who attended, and thank Jennifer and Dion, and I think at this point, I now will hand over to Ali from Abcam.
120:30 - 120:41: Thanks, Tom. What a great session. On behalf of Abcam, I would like to thank you all for attending today's event. We really appreciate your feedback, and we'll be conducting a short poll, which will appear on your screen.
120:42 - 120:53: Answering the poll is really helpful for us, because it gives us the opportunity to take your feedback on board and shape further sessions like the ones we've run over the last month.
120:53 - 121:03: We're going to be putting the information in the chat box as well about the PACE credits, should you wish to take the option of claiming those for today's session.
121:03 - 121:25: Finally, I want to thank our speakers, Jennifer and Dion, fantastic job, and a special thanks to Tom for moderating today's session. Bringing speakers together and getting a great dialogue and a conversation going is a challenge remotely, and I think we've done an excellent job on that today, so thank you to the three of you very much.
121:26 - 121:37: Now that the spotlight session on the neuroscience series has come to an end, our month's been incredibly busy with six sessions, but we're finally at an endpoint. I would like to thank you all for tuning in and participating throughout the month.
121:37 - 121:46: If you're unable to catch some of the live sessions, don't fret, because you're going to receive an email in a few weeks, which will give you all the on-demand links.
121:47 - 121:55: Any sessions that you haven't been able to attend, you can continue to take on board the insights and education that's available there.
121:55 - 122:05: Thank you very much for your time, and I do believe the weekend is almost upon us, certainly in Europe. Have a great weekend, guys, and thank you very much for your time.

You may be interested in...