Yale Psychiatry Grand Rounds: "Neuroimmune Mechanisms of Brain Development and Vulnerability"

April 14, 2023

Information

ID: 9830
To Cite: DCA Citation Guide

Download Transcript

00:00Wait,
00:04okay, do I need to?
00:05Yes, I need to click that.
00:07But can you help because
00:11the mouse is over here.
00:13So I will first thank Alicia and thank
00:15Marina for that wonderful introduction
00:16and to all of you for being here.
00:18And it's really an honor to be here
00:20for the special lecture and just
00:22a pleasure to be here in general
00:23because so many of my colleagues and.
00:25Friends are here,
00:26and I've learned a ton already about
00:28all the research that's going on.
00:30Like just makes me want to spend another
00:32week or more hanging out more because
00:34this clearly wasn't enough time.
00:35But I am.
00:36I'm really excited to tell you
00:38about some of our work today.
00:39And as you just heard a little bit from
00:41the background in the introduction,
00:43you know,
00:44I think is increasing evidence that
00:46really suggests that brain development
00:48is impaired in a lot of mental illnesses,
00:50including schizophrenia and bipolar.
00:52But you know,
00:53we know this is potentially happening
00:55long before the onset of symptoms,
00:57but there are many questions
00:59and challenges for the field in
01:01particular when do things begin?
01:03This is a major question as a
01:05developmental neurobiologist and glial
01:06biologist that I've been long interested in.
01:07It's just going back in development
01:09and trying to understand
01:10when and also which circuits.
01:12And ultimately the biggest question of
01:14course is what are the mechanisms and with
01:17emerging genetics of the last decade or more,
01:20that's really.
01:20Exploded in the field and it's really
01:23illuminating a lot of potential
01:25pathways that we never had before.
01:27It turns out genetics are pointing to a lot
01:29of a variance in genes that are implicating,
01:32not surprisingly,
01:32the synops right the point of communication
01:35between neurons where of course if
01:37synopses aren't working properly,
01:38that's going to have all kinds of impact on
01:40brain development and circuits and behavior.
01:43But the other thing I want
01:44to highlight today is,
01:45you know with all of the,
01:46the,
01:46the define mapping and GWA studies in,
01:49in the context of schizophrenia
01:50in particular,
01:51where I'll focus a bit more today,
01:53there's both common variants and rare
01:55variants that are implicating not just
01:58generally synapses but specific genes,
01:59not just variants but specific
02:02genes and which also,
02:04you know,
02:05implicate another pathway or part of the
02:07the body which is the immune system,
02:09which.
02:10At the beginning was a bit mysterious
02:12why these immune molecules
02:13keep coming up in the genetics,
02:15but I hope to convince you today
02:17that some same genes that relate
02:18to the immune system,
02:20like complement C4 and another
02:22gene called CSM D1.
02:23Ironically enough,
02:24my lab had been studying these
02:26genes independent of knowing the
02:28genetics was going to point to
02:30them in the context of risk,
02:31and so it's been a really great example
02:34where genetics and biology can converge.
02:36But even still to try to understand
02:38the mechanism requires a lot of
02:39work to try to develop new tools
02:41and importantly new model systems because
02:43even with the genetic leads and even with
02:45some biological insight as Marina just said.
02:47How do we translate that into new biomarkers,
02:50new mechanistic biomarkers that will
02:52enable us to stratify patients and to
02:55be able to identify those that are at
02:57more risk really in development ideally,
02:59but also new targets of
03:01intervention like that is,
03:02is you know I think it's a challenge but.
03:04I hope to convince you today through
03:07collaborative efforts by many of me and my
03:09colleagues at the broad and the Stanley
03:10Center and and many collaborators outside.
03:12We are trying to work together to try
03:14to to develop a pipeline to be able
03:17to take gene variant all the way to
03:19pathway to function so that we can try
03:21to develop new mechanistic understanding.
03:23So as a developmental neurobiologist
03:25and I think back to when I was a
03:27graduate student start first became
03:29interested in neuroscience.
03:31It's this type of question that.
03:33Brought me to neuroscience,
03:34which this idea of how it is that the
03:38environment can sculpt developing circuits,
03:40right.
03:40So we know early in development with
03:42during this process of brain wiring,
03:44we start off with, you know,
03:46kind of a sparse immature
03:48synaptic connections.
03:49And then over a course of
03:51different critical periods,
03:52right,
03:52we start to develop this process by
03:54which some of those connections form
03:56and get strengthened and maintained.
03:58And other of these connections
04:00get permanently removed.
04:00This is a process called synaptic
04:03refinement or synaptic pruning,
04:04and we know what happens all across the
04:06brain in different critical periods now.
04:08It's been best studied in sensory
04:10systems like the visual system and for
04:12that matter many other sensory systems
04:13where these critical periods are
04:15often happening early in development,
04:17but other parts of your brain like the
04:19prefrontal cortex and cortical circuits.
04:21These are the sort of last areas
04:23to refine and mature and and
04:25incredibly important.
04:26You know,
04:27there's much less known about the
04:29mechanisms and and timelines that regulate
04:32those processes of synaptic refinement.
04:34So what we are interested in is
04:36trying to better define not only
04:38the timing of different regions
04:40and circuits and their maturation,
04:41but we want to try to better
04:44understand how perturbations of the
04:45environment and particular genetic
04:47pathways influence and change
04:49circuits and ultimately behavior.
04:51Now it's long known,
04:52you know,
04:52I should just mention another thing,
04:54that plasticity is usually thought
04:55to be a good thing.
04:57It's why my daughter can learn French
04:59seamlessly and I am terrible at French,
05:01right?
05:01My critical period for learning French is
05:03long closed and no matter how much I try,
05:05it is very challenging.
05:06But my both my daughters are in
05:08a French immersion program and
05:10started learning that when they
05:11were kindergarten first grade.
05:12And it's amazing, you know.
05:14So that's just one of one of many
05:16examples of why sort of this
05:17idea of lose it or use it.
05:19Lose it,
05:19use it or lose it and the idea
05:21that you know plasticity while
05:23enabling learning and and and
05:25adaptation to the environment,
05:26it can also lead to vulnerability.
05:28So plasticity opens the brain up for
05:30potential vulnerability because it's so
05:32dynamic and plastic and I think it's
05:34understanding these critical periods
05:36are going to be important to thinking
05:38about why there is particular windows
05:40of vulnerability and diseases and
05:42disorders like schizophrenia for example.
05:45Now schizophrenia, you know,
05:46we know is, you know,
05:48there's evidence from both imaging
05:50studies and some anatomical studies that
05:52there is a loss of great Gray matter,
05:55thinning of Gray matter and and
05:57even some evidence of loss of
05:59spines and synopses that have come
06:01from postmortem studies.
06:02And, and I think more and more
06:04evidence is suggesting that at least
06:06in some individuals of schizophrenia
06:07there's evidence of synaptic defects.
06:09And this is not true just of schizophrenia.
06:11This is true of other
06:13neurodevelopmental disorders as well.
06:14But I think the problem with
06:16these types of examples,
06:18even the imaging,
06:18it doesn't give you the resolution to look
06:21at synapses and the postmortem analysis
06:23where you can quantify synapses in patients.
06:25By the time you get those brains, you know
06:27there's many things that have happened.
06:28So you don't know whether that's
06:30cause or consequence, right.
06:31There are many things that could
06:32have led to the loss of synapses.
06:34So a lot of this idea of of
06:37synaptic pruning defects,
06:38those are all been, it's a hypothesis.
06:40We don't know for sure that
06:42synapse pruning or synapse loss.
06:44Is contributing to,
06:45to,
06:45to to some of these disorders and
06:47I think that's real obviously
06:49very challenging to study in,
06:50in people and for that matter
06:52even in animals.
06:53And so we also know that synapse loss
06:56and dysfunction is a is a hallmark
06:58of many other disorders as well.
06:59And my lab over the last many years has
07:02been studying normally what regulates
07:04synaptic pruning and synapse elimination,
07:06hoping that by understanding basic
07:07mechanisms in the normal healthy brain,
07:09in animal models and ideally
07:11in patient samples.
07:12That that can then based on
07:14that mechanistic understanding,
07:15we could then apply some of that
07:17to understand whether those same
07:19mechanisms become aberrantly activated
07:20to lead to pathological synapse
07:22loss in the context of disease.
07:24And we've also been studying this in
07:26other diseases including Alzheimer's
07:27and age-related neurodegenerative
07:28diseases and in fact in normal aging.
07:30And even though all of these disorders
07:32are remarkably different in many ways,
07:34as you all know,
07:35I believe there's still evidence of
07:37a convergence that I'm particularly
07:38interested in because different
07:40things can initiate the process.
07:42Genetics and environment,
07:43different different pathways,
07:45maybe even different circuits,
07:47but ultimately at least the data that
07:49we have suggests the possibility
07:50that there could be some converging
07:52mechanisms that then ultimately
07:53leads to the synaptic vulnerability
07:55and synapse loss.
07:56And that's what I want to talk
07:57to you about today.
07:58So why has progress been so
07:59slow on the biology
08:00side even though the genetics and
08:02the genomic studies have exploded?
08:03And that's wonderful because
08:05unbiased data is giving us, I know,
08:07new leads, new candidates and new
08:09ways of thinking about mechanism.
08:11But obviously this complex circuitry
08:13inaccessibility of the human brain,
08:14as I mentioned and as I'll highlight
08:16and I think many of you appreciate,
08:18there's been a lack of credible
08:20disease models.
08:20I mean, I don't think anyone model a mouse
08:23or monkey or other animal models can really,
08:26truly recapitulate the complexity
08:27of the human brain and cognition,
08:29but with the thoughtful way of thinking
08:31about trying to model disease mechanisms,
08:34not model the disease.
08:35That's giving us new,
08:36a new foothold into trying to understand
08:38the types of questions I raised before.
08:40And also we don't have biomarkers,
08:42right.
08:42We don't have really good mechanistic and
08:45predictive biomarkers given the complexity,
08:47heterogeneity and polygenicity
08:48of these disorders,
08:50how do you even know who to track and and
08:52follow and stratify and we don't have,
08:54we don't have that for for these disorders.
08:57So what I want to do today
08:58is kind of take you through.
09:00The thinking and the the way we've
09:01been thinking about how to tackle
09:03this and we meaning there's a lot
09:04of collaborators and colleagues
09:05at the Stanley Center at the broad
09:07and and within my lab.
09:08But just the idea that you know
09:10that there are many open questions.
09:11We still don't know even with all
09:13of the genetics that are pointing to
09:15variance that we don't many cases
09:17know what the genes are right they're
09:18they're close to in a particular
09:20chromosome and a particular loci but
09:22which genes in some cases we know
09:24the genes and in many cases we don't.
09:26We don't know the mechanism by which
09:27these genes actually alter cellular function.
09:29We don't know which pathways are
09:31relevant because these genes don't
09:33necessarily tell us what pathway.
09:35We don't know in many cases which
09:37cell types are are are most affected.
09:39There are loss of heterogeneity of
09:41cells in the brain, as you know,
09:43even with an excitatory neuron populations,
09:45inhibitory neuron populations,
09:46and glial cells remarkable heterogeneity.
09:49So even if you found a gene in a pathway,
09:51which cell type do you want to be
09:53studying right and what's which
09:54ones to focus on?
09:56And of course which synapse,
09:57right, it's not all synapses,
09:58maybe the vulnerable synapse,
09:59which circuit in the brain,
10:01I mean obviously this is going to sound
10:02like the most daunting thing ever,
10:03how we ever going to figure this out,
10:05but ultimately all of these questions
10:08are are are open in many ways and so.
10:10What we want to try to be able to
10:12do is think about ways to to step
10:14back and systematically go through.
10:16And in some cases where the leads
10:18are more reasonable,
10:19we know something about the biology.
10:20We're going deeper to try to
10:22understand mechanism.
10:22And in other cases we're
10:24letting the unbiased data come
10:25in to nominate pathways and then
10:27we're developing new models.
10:29We wish to test that.
10:30So I'm going to give you an example
10:32in the first part of my talk that's
10:34going to focus on one gene and
10:36pathway and mechanism that we have
10:37worked on for the last decade.
10:39And it involves the complement cascade as
10:41you heard a little bit in the introduction.
10:44But I just want to tell you a little
10:46bit about how this all started
10:47because we knew from the genetics.
10:49That by far the largest and most
10:51mysterious genetic result has always been
10:53this huge man in the Manhattan plot.
10:56There's clear evidence of the MHC locust,
10:58right?
10:59This is the tallest peak in the
11:00Manhattan plot, but yet,
11:01and there's hundreds of genes
11:03within that locust.
11:04But it's been a mystery in terms of what gene
11:06or genes underlie this incredible signal.
11:08And even now,
11:09with more and more genetics coming in,
11:10that peak is still the highest peak.
11:12And, and at the time,
11:14I know geneticists were working on this.
11:16And of course, more and more
11:17data means more and more insight.
11:19But it turns out that you know,
11:22one of my very outstanding geneticists and
11:26neurobiologists colleague Steve Mccarroll,
11:28his lab was studying this locust more deeply,
11:30was fine mapping and starting to
11:32look into the locust to see what,
11:34what's what could explain this
11:35huge genetic signal.
11:36And for many reasons,
11:37and of course this work is published,
11:38but I just want to take you through the main
11:40findings that sets up the rest of the stock,
11:42what he found and what Ashwin
11:44Sekhart found in his lab.
11:46Was that in within that locus
11:48are is a gene complement C4.
11:51Now C4 in humans there are
11:54two isoforms of C422 genes,
11:56C4 big A and C4 big B.
11:59And it turns out when you looked at the
12:02haplotype in terms of different individuals,
12:05you can have different
12:06combinations of C4A versus C4B.
12:08So you can have multiple copies of a.
12:10Multiple copies of B both and vice versa.
12:13And what Steve and Ashwin managed to
12:14do and they discovered that it's not
12:16so much whether you have the gene or
12:18not have the gene or whether there's
12:20a loss of function of that gene.
12:21How many copies of the particular
12:23structural form of C4 is what,
12:26what, what linked to risk.
12:27And there's a really all of the data
12:29was published number of years ago,
12:31but I just want to highlight the
12:33take home of this important study.
12:34One is that they went to correlate
12:36and they've mapped us back on
12:38to the genetics and found.
12:40Indeed it was the copy number
12:41of C4A the conferred risk.
12:43So the more copies of you of C4A1
12:46individual had it significantly
12:48increased schizophrenia risk.
12:50But they also measured the
12:51expression of C4 on the brain.
12:53This was the other really important
12:55finding C4 in complement is
12:56expressed in the whole body.
12:57But it was the brain expression
12:59that correlated with all
13:00this and when they measured
13:01the RNA in the brain.
13:03It was the higher expression
13:05of C4 was also related to the
13:07more copies of C4A you had.
13:09So you're making more of the C4A
13:11gene now around the time they
13:13Steve and Ashwin and his lab
13:15started to uncover this really
13:17exciting and important finding.
13:18Steve and I actually didn't realize that we,
13:20our labs are right across
13:21the street from each other.
13:22And and so we started having
13:24coffee and I started Ed Skulnik
13:26and Steve ***** started inviting
13:27me over to the Stanley Center.
13:29I'm like, why they want to hang out with me?
13:30Well, it turns out they wanted to
13:32hang out with me because we were
13:34studying this very same pathway since
13:36I was a postdoc in Ben Barris's lab,
13:37not thinking anything about
13:39the genetics of schizophrenia.
13:40But we were really going deep into
13:42the biology of this complement
13:43pathway in the context of synaptic
13:45elimination and synaptic development.
13:47So it's a great.
13:48Adipitous example of how a biology
13:50and genetics can come together
13:52and lead to collaborative science
13:54that can then enable one to try to
13:56understand the biology underlying
13:57those emerging genetic findings.
13:59And I'm just going to highlight
14:00some ongoing work and in no way have
14:02we figured this out by the way,
14:03but I'm going to tell you what we've learned.
14:05I just want to tell you if those
14:06that are already thinking about
14:08when they learn about compliment
14:09back in the day in Med school or
14:11in grad school and your head is
14:13starting to spin thinking about that,
14:14don't worry,
14:15I'm not going to make you relearn compliment.
14:17The take home here is it's very complex
14:19and even the immunologist yesterday I
14:21was over in the Immunobiology department,
14:23they said,
14:23you know,
14:24nobody really wants to talk about the
14:26compliment cascade because it's so complex.
14:28And and and in fact even like these
14:31incredible immunologists try to avoid it.
14:33So I I didn't feel so bad when
14:34I when I didn't know much about
14:36adaptive immunity or B cells.
14:38But I do know something about
14:40complement and that is these are
14:42a group of complement secreted
14:43proteins that exist throughout your
14:45body and basically what complement
14:47does is it helps it's our first
14:49line of defense against a pathogen.
14:51So think about a a pathogen,
14:53a bacterial infection.
14:54Before your slower adaptive
14:55immune system kicks in,
14:57the complement system kicks in really
14:59fast and these start secreted and they,
15:02if they're all there,
15:03present,
15:03whether that be in the periphery
15:04and the blood or in the brain,
15:05as I'll tell you.
15:06That can then lead to a kind
15:08of a domino effect,
15:09a cascade that ultimately leads
15:10to the tagging of some of those
15:13complement components like C4 and
15:14C3 on the surface of that pathogen.
15:17And that brings in macrophages the
15:19Pacman to then recognize and remove it.
15:21And then that's a very
15:23helpful process to get rid
15:24of infection or dead cells
15:26or debris very rapidly.
15:27Well, it turns out when I was a postdoc
15:29in Ben Barris's lab through an unbiased
15:31gene chip experiment way back in the day,
15:33we also uncovered an unexpected
15:35role for C1Q during this process of
15:38development and synapse pruning that
15:40we also were very surprised about.
15:42And what it turns out,
15:43I'm going to tell you about is the work we
15:45uncovered when I was still in Ben's lab.
15:47And this is a figure from
15:48that very first paper.
15:49What we realized is.
15:51And hypothesize,
15:52and this is really Ben and I
15:53brainstorming about what complement
15:54might be doing in the brain,
15:55the healthy brain, no infection,
15:57no disease, no challenge.
15:58But it turns out,
15:59just like the immune system,
16:01glial cells, neurons,
16:02they all express these complement
16:04components all the time.
16:05And they're present in the brain now.
16:06They're not always there at the same levels.
16:08In fact, during development,
16:10during critical periods of development,
16:12there's a lot more of it,
16:13not just present, but tagging synapses.
16:15So instead of tagging a bacterial cell,
16:17what we discovered was that these
16:19complement components can similarly tag.
16:21Leg of Synapse and that microglia,
16:23not macrophages,
16:24have expressed the receptors
16:26that recognize complement.
16:27And through that,
16:28that's one way that synapses,
16:30exuberant connections,
16:31axons and synapses then can get
16:33removed and engulfed by microglia.
16:35Now that was a hypothesis.
16:36Then over the last decade or more,
16:38my lab has gone on to test,
16:40but I I want to tell you 2 important points.
16:42One is that complement is there,
16:45you know, naturally in the brain,
16:47right under normal conditions.
16:48It tags subsets of immature synapses
16:50during critical periods and we
16:52studied this in the visual system.
16:54We also went on to show if
16:56you genetically knock out
16:59C1QC3CR3 on the microglia through
17:00a number of studies over the
17:02years by my lab and others,
17:03that does lead to defects in synapse
17:06number and synaptic connectivity.
17:08We really focused on the visual
17:10system and we've gone deep into the
17:12visual system because it's such an
17:13elegant model to study the process of.
17:15Synaptic elimination and I'm going
17:16to take you through some of those
17:18mechanisms now and I'm going to come
17:19back out again and tell you how that
17:21might then lead to us thinking about
17:23synapse and vulnerability in the context
17:25of of disorders like schizophrenia.
17:27So the question we had in early on
17:30was could microglia be similarly
17:31playing this role in the brain?
17:34And as you just heard,
17:35you know, back then again,
17:37while a long time ago,
17:3815 years ago,
17:39microglia were really studied in the context
17:41of injury and disease or infection and,
17:43you know, other challenges in the brain.
17:44So we knew microglia are really important.
17:46But this observation made initially
17:48by my colleague at Stanford and
17:51others using two photon imaging,
17:52this is a movie by my own
17:54graduate student Janelle Wallace,
17:56just shows this.
17:57Really important observation that
17:58if you were to put a microscope
18:00in your head right now and watch
18:02microglia as you listen to my talk,
18:03they might be doing something like this.
18:05They're always moving their processes,
18:07they're moving around,
18:08they're touching synopses and a
18:09lot of other cells in the brain.
18:11This is just a sparse example
18:12where you can see the interactions
18:14between synopses and microglia.
18:16But this led to all kinds of just
18:18this observation alone led me to
18:19wonder all kinds of questions.
18:21What are they sensing when
18:22they're touching a synapse?
18:23What are the molecular cues
18:24that might be recruiting them?
18:25Find me signals to bring them
18:27in when they get there,
18:28some of them land and hang out
18:30with synapses longer than others,
18:31and others pull back.
18:32What are those molecules and what
18:34are the the sort of bidirectional
18:36signaling mechanisms between microgly
18:38and neurons that could be, you know,
18:40leading to this sort of interaction?
18:41But most importantly.
18:42You know, what's the functional consequence?
18:44What might they be doing?
18:45And as a resident phagocyte,
18:47and based on the mechanisms I
18:48told you about with complement,
18:49we hypothesize that maybe subsets of
18:51those synapses during development
18:53might actually be engulfed or phagocyte
18:55toast or pruned away by microglia.
18:57The weaker synapse is what we
18:59hypothesize and indeed in the
19:01visual system and the retina
19:02geniculate system that we study.
19:04We have a way thanks to the work
19:06of Carla Schatz and Shinfei Chen
19:07and and many others that have used
19:09so Huebel and weasel the system to
19:11study activity dependent refinement.
19:13You can basically in a mouse or a
19:16cat or a monkey, but we do mice.
19:18You can put tracers in the eyes of both mice,
19:21both eyes of of a mouse in red and and blue,
19:24and then that it leads to the
19:25tracing of the
19:26projection into the visual thalamus.
19:28And a cartoon of the relay neuron
19:30on the thalamus might look something
19:32like this early development.
19:33There's a neurons innervated by both
19:35eyes and they're pretty weak inputs.
19:37But during the critical period
19:39in Post Natal development,
19:40some of those inputs,
19:41the weak ones get pruned away and weakened
19:43and others get strengthened and maintained.
19:45This is the idea of use it or lose
19:47it and it's activity dependent.
19:48So based on all that work,
19:50we wondered a a micro engulfing
19:52or pruning synapses and if so,
19:55are they doing it in a selective way,
19:56is it activity dependent?
19:58The alternative less interesting hypothesis
20:00is they're just cleaning up the debris.
20:02And a lot of data supports that's that.
20:03It's the first that they're not just
20:05cleaning up and picking up the debris,
20:06but they're having a more active
20:08role in this process.
20:09Because work by Dorothy Schaefer in
20:10my lab showed that when you manipulate
20:13activity in the two eyes drive competition,
20:15there was a selective elimination
20:17or removal or phagocytosis of
20:19those presynaptic axons in,
20:21in a way that they were selectively
20:23engulfing the less active input.
20:24So this is an important sort of early
20:26finding because it did tell us.
20:28That it's, you know,
20:29instructive cues must somehow be
20:31regulating this process so that the
20:32less active inputs perhaps may have
20:34molecular cues that are different
20:35than the than the red inputs and
20:37then that might lead to microglia to
20:39then come and engulf those synapses.
20:41And we have a quite a bit of data
20:43to support that hypothesis,
20:44at least in the mouse visual system.
20:46So how does this relate to the complement
20:48C4 story I told you in genetics?
20:50Well,
20:50we started collaborating with Steve
20:51Mccarroll and we collaborated
20:52with also with an immunologist,
20:54a wonderful immunologist at Harvard,
20:56Michael Carroll,
20:57who actually cloned C4 and have
20:59been studying C4 in the context of
21:01lupus and in many other contexts
21:03in the immune system and using
21:05a combination of of mouse models
21:07that Mike's lab generated,
21:09C4 knockouts and other model I'll
21:11tell you about in a moment and a
21:13human IPS neurons and human tissue
21:15we showed that much like C1Q and
21:17C3 we could also see C4 decorating.
21:20Operating synopses in in neurons.
21:23And we also using the same kind
21:25of model system that I told you
21:27about earlier when we looked at
21:28pruning and refinement.
21:30C4 knockout mice Pheno copy the
21:33C1QC3CR3 mice again suggesting that the
21:35C4 similarly at least in a mouse was
21:38was was mediating the synaptic pruning.
21:40That's a loss of function but that's
21:42not what the genetics is telling us.
21:43Remember it's telling us too
21:45much C4 too much complement.
21:48And so too much C4 could mean
21:49more tagging of synapses,
21:50but it could also mean more
21:52activation of the classical comp
21:54and cascade because C4 is necessary
21:55for the whole pathway to happen.
21:57So the question was?
21:59Increase C4 levels,
22:01it lead to excessive burning.
22:02So this is more recent work.
22:04And so this necessitated the need to develop
22:06new models and in particular humanized
22:09mouse models that we didn't have before.
22:11And so Mike Carroll's lab developed a
22:14C4A humanized mouse model when he could
22:17introduce human C4A alleles into the
22:19mouse genome using back DNA transgenesis.
22:21So you can basically introduce
22:23human C4A or B in this case.
22:25I'm just telling you about a.
22:26Based on the you put it in AC4 knockout
22:29background and based on the the crosses
22:31that you do hets versus homozygous,
22:34you can also control copy number.
22:35So it's pretty cool.
22:36So as a proof of concept we wanted to know
22:39by making a mouse that has a lot of the
22:42genetic risk variant C4A over express.
22:45Versus C4B,
22:46which I'm not showing you.
22:47Does that in fact lead to using
22:49similar assays over pruning more
22:51microgly engulfment less synapses?
22:53And in the first study that was
22:55led by Mike Carroll's lab in
22:57collaboration with all of us showed
22:58using many of the same mechanisms,
23:00I just told you in the assays that
23:02indeed in the visual system and in
23:04the in the frontal cortex and in other
23:07regions they showed enhanced engulfment.
23:09And also,
23:10they also went on to show a spine loss,
23:13which,
23:13you know,
23:14this was really interesting because
23:15a lot of the work we had done
23:16in the visual system was a bit
23:18more on the presynaptic side.
23:19So this is now looking in the cortex,
23:20getting closer to where
23:21I was wanting to head,
23:23but we weren't quite brave
23:24enough to go there yet.
23:25So what is all this telling us?
23:27And there's a lot more I'm not telling you.
23:28This paper was just published in 2020 if
23:31you want to learn more about this study.
23:34But the working model,
23:36and this is building on a lot
23:38of ongoing work in the lab.
23:39We're really interested in understanding
23:41the specificity of this process.
23:42So I remember giving this talk
23:44early on compliment work both to
23:46neuroscientists and to immunologists.
23:48And the question I always got was,
23:49well, these are secreted.
23:51How do you get any specificity
23:52in this process if they just are
23:54around and binding things in
23:56this sort of nonspecific way?
23:57I said, well,
23:58I think they're probably
23:59not binding things in it.
24:00They can be binding in a nonspecific way,
24:02but I think there could
24:03be a specificity in this,
24:04in this process and indeed.
24:06The working mechanistic model,
24:08especially bringing an activity
24:10is perhaps it's it's the case
24:12where there's some receptor,
24:13some molecular difference and
24:17the surface of synapses that's
24:18recruiting or or enabling complement
24:20to bind to certain synapses,
24:22let's say the blue ones but not the red ones.
24:24And there may also be molecules that are
24:27protecting the synapses you want to keep.
24:29Now this was a hypothesis that came out of,
24:31you know,
24:32actually I remember giving
24:33this talk to immunologists
24:34at Harvard. Early, early on,
24:35where they said, well, could there be
24:36don't eat me signals in the brain.
24:38And I didn't even know what those
24:39were because I was, you know,
24:41neuroscientist trying to be an immunologist.
24:43But it turned out I had a fantastic
24:45graduate student who wanted to learn
24:46about these don't eat me signals.
24:47And it turns out there's a whole bunch
24:49of inhibitors and don't eat me signals
24:51that when we looked in the brain,
24:53immunologists didn't pay attention
24:54to the brain.
24:55But we had antibodies and we
24:57looked by expression profiling.
24:58Well, there's a whole slew of them on the
24:59brain and we better have those there.
25:00Otherwise the compliment system is going to.
25:03Kind of will be on OverDrive and they
25:04needs to be this tightly regulated system.
25:06So just as important as
25:08having compliment there,
25:09there needs to be the regulators,
25:10the brakes that keep it from activating.
25:13And so we started looking and I'm not
25:14going to tell you about all of these,
25:16but intriguingly in addition to a bunch of
25:18don't eat me signals that we've identified,
25:21I want to show you the working
25:23model we identified for example.
25:25CD47 which is also there's been some
25:27genetics that are not as powerful but
25:29there's some some hints that CD47 serve.
25:31Alpha could be playing a role in
25:33some of these in some of these
25:35GY studies still early days.
25:37But regardless of that we showed in
25:40the brain that CD47 is a classic don't
25:42eat me signal in the immune system and
25:44it actually prevents a macrophage from
25:47eating a self cell or healthy cell.
25:49In fact healthy cells and self
25:50cells have a whole shield of don't
25:53eat me signals that say.
25:54You know what do not even come
25:56near me because I need to stay.
25:57And the idea is when they're there
25:59there's apartosis or when it's
26:00some kind of a damage signal,
26:02those get down regulated and relocalized,
26:04and it exposes parts of cells that can
26:06then lead to the removal by a macrophage.
26:07And what my graduate student went
26:09on to show is CD47 is one of those
26:12signals highly expressed in neurons.
26:13Microglia have SERP Alpha,
26:14which is the receptor that recognizes that
26:17and essentially that says don't eat me.
26:19And one model is that.
26:20The stronger synapses have a lot of
26:23those protective molecules so that even
26:25if microglia are tagged by complement,
26:27synapses have complement on them.
26:29They won't get removed.
26:30And that's again a working model.
26:31But we do have data to support
26:33that idea because when we knock
26:34out the don't eat me signals,
26:35there's not only do we lose that
26:38activity dependent specificity,
26:39but you get an over pruning as well.
26:41So that's one example.
26:42We have others that I'm not going
26:43to tell you about today because
26:45that's what I talked about yesterday.
26:46But I want to tell you about
26:48another one and some,
26:49if you guys are paying
26:50attention to my second slide,
26:51you might have remembered
26:53another molecule called CSM D1,
26:54which is a gene that came up really
26:57early in GWAS schizophrenia.
26:58That's been largely nothing
27:00known about CSM D1 in the brain.
27:02And so there was a I went to a meeting
27:04in this is like a small world that I
27:06went to a compliment meeting in in
27:08Greece would not a bad place for a meeting.
27:10And I met this woman who scientists
27:12who came up to me and said I've been
27:14really wanting to meet you because
27:16we've been studying the CSM D1 and
27:18we have evidence that inhibits C4 in
27:20not in the brain but in a in a test
27:22tube in the in the in the laboratory
27:24and and it's a putative compliment
27:26inhibitor I'm like.
27:27Well, this is amazing.
27:28So we started talking more
27:29and and for many reasons,
27:31we started thinking that this
27:32would be an interesting molecule,
27:33not just because of the genetics,
27:35but if it is involved in the
27:36classical complement cascade,
27:37we want to know more about what
27:38it might be doing.
27:39And this is ongoing work.
27:40The paper's on bio archive,
27:41but we're just revising it now.
27:42So there's more to come,
27:43but just want to highlight
27:44this work and again how, how?
27:46Again,
27:47two now common variants are
27:48coming together that suggests some
27:50role in the complement system.
27:52So CSM D1 is a well localized
27:54US signal as I mentioned with
27:56schizophrenia risk on chromosome 8.
27:58And although it's enriched in the brain,
27:59in fact if you look expression
28:01by protein and gene and and RNA,
28:03there's unlike C4,
28:04it is in nowhere else except the
28:06testes like it is blazing in the brain.
28:08So therefore, wow,
28:09how can't we know what this thing does?
28:11Okay.
28:11And so we don't know about lost,
28:14we don't know directionality or anything.
28:16There's a lot more genetics to
28:17be done in terms of the mapping,
28:19but we wanted to look more closely
28:21at this and so.
28:22A postdoc, Matt Johnson,
28:23now a group leader at the Stanley
28:25Center who's been Co leading this
28:26with me and collaboration with Steve
28:28Mccarroll and our graduate student
28:30Matt Baum is an MDPHD student who finished,
28:33I'll tell you about what he
28:34what he uncovered.
28:35So it turns out,
28:36see as I mentioned,
28:37it's very highly expressed both at the
28:39RNA and protein level in the brain.
28:41So that was, you know,
28:42obviously something worth looking at no
28:44matter what but when we started looking.
28:47Thanks to an amazing antibody
28:49that we had access to,
28:51we started staining and obtained
28:53ACSM D1 knockout mouse.
28:54Just to ask just like we did with complement,
28:56you know what,
28:57what's the phenotypes and where
28:59is the protein.
29:00So that was also a great tool
29:01for our antibody because the
29:03antibody is really specific,
29:04not only does it light up the
29:05brain in interesting regions of
29:07the brain especially you could
29:08see hippocampus and certain areas
29:09of the brain at lomag.
29:10But if you really is a man, I hard to see.
29:12We saw a lot of punctate staining
29:14and it colocalizes with subsets.
29:16Of inhibitory and excitatory synopses.
29:19So it's at synopsis not only at synapses,
29:21but we have some data that's enriched
29:23in synopsis because we've done
29:24some synapticome preps as well,
29:25which I'm not going to tell you about,
29:26but it's in the paper actually.
29:28This might be it.
29:29Yes, it is. OK.
29:29So there is a little bit of data
29:30there to show that it's an,
29:31it's an enriched in the synapticomes.
29:33So again at the right
29:34time and the right place,
29:35nothing known really about its biology,
29:37but that one interaction at the compliment
29:39meeting in Greece made me start to think
29:41about could the two be related in some way.
29:44And so the idea would be, okay,
29:45we already know that C4 is high and that
29:48leads to activation of the cobman cascade.
29:51At least in animal models that could
29:54lead to over pruning could C4C,
29:56SM D1.
29:57Which the other reason why this
29:58is really exciting is it is
30:01a huge extracellular domain,
30:02it's giant and it has a the reason
30:05why it's called CSM D1 is it's
30:08it's got cub and sushi domains.
30:10And these are the domains
30:11that are expressed in many,
30:12many complement inhibitors
30:13in the immune system.
30:14And that's why my my colleague
30:16was interested in it.
30:18And so that led us to wonder,
30:20could could it be that what csmd one
30:23is normally doing is putting the
30:25brakes on C4 and keeping it in check?
30:28Therefore,
30:29if there was a genetic mutation or loss
30:32of function or an ability to disrupt csmd,
30:35one's function could then that
30:37therefore that would lead to
30:38again an over activation of C4.
30:40Again,
30:40this is all hypothesis and I'll just
30:42tell you the way we've been testing
30:44it and it's still ongoing work.
30:45One way we can also look at this
30:47to see whether there'd be more
30:48complement activation or more tagging
30:50of complement was this experiment.
30:52We not only do we have the knockout mice,
30:54so we can look at whether there's
30:56too much complement tagging
30:57and removal of synapses.
30:58We also took advantage of IPS
31:00stem cell models where we could
31:02use isogenic controls in a CSMD,
31:04knockout neuronal differentiation
31:06these N G2 neuron protocols.
31:09And basically did a very classic
31:11immunology experiment where we could
31:13sensitize with an anti surface antibody
31:14which would then bind to the surface.
31:17And then we wanted to know
31:18if we then add tag C3,
31:20do we get more tagging or the
31:23localization of synopses with C3
31:24in mice that don't have CSM D1,
31:26not sorry mice,
31:27but yes mice but in this case cells.
31:29OK.
31:29So this is just an example of
31:31the type of assay and indeed.
31:33Much more data on this than I'm showing you,
31:35but just to illustrate what what
31:36Matt and others found was there was
31:39a sort of a selective and enhanced
31:41tagging of C3 in the neurites even in a
31:43mixed culture that didn't have CSCSM D1.
31:46So that was proof, not proof,
31:48but certainly evidence that suggests
31:50that it's somehow regulating complement
31:52deposition and we think activation to
31:54other experiments we're working on.
31:56And then in the in the mouse model,
31:58we I'm not going to show all the
32:00data in the interest of time.
32:01But we went on and did the same
32:03kind of experiments by looking at
32:05complement and localization at synapses.
32:07And we wanted to know do we see
32:09more complement tagging of synapses
32:11in the in the CSM D1 knockout mice
32:13and in the visual system at least
32:15has increased complement tagging.
32:17We see an enhanced refinement by a retina,
32:19geniculate experiments,
32:20decrease in synapses and some
32:22electrophysiology work that's in progress.
32:24So together suggesting that some
32:26interesting phenotypes and that may be 1
32:29mechanisms to the complement inhibition.
32:31And so that and we also have some in
32:33vitro and in vivo work ongoing that also
32:36show that microglia can overprune or
32:38engulf much like the C4A mice we're doing.
32:40So that's sort of the working model and
32:43in no way does this explain that this
32:45is the mechanism that that the genetics,
32:47this is the only way this might be working.
32:49But I think it is together suggest a
32:52mechanistic model that we're going to
32:55continue to test in in various ways.
32:57So this is a lot of the basic
32:59science that I wanted to start with,
33:01but now I want to zoom out and I
33:03want to talk more about the timing
33:04part because that's the part that
33:05I'm particularly interested in,
33:07is understanding why,
33:09for example,
33:11adolescence is a is a time where we know is
33:15an onset for a lot of these mental illnesses.
33:18In particular,
33:19schizophrenia,
33:20even even though potentially
33:22there could be issues earlier,
33:24but the emergence tends to
33:25happen in laid adolescence,
33:26early adulthood and this is
33:27true of other things as well.
33:29So that raised all kind of questions
33:30as a developmental neurobiologist was,
33:32could there have always been issues
33:34but then something kind of opens
33:35up and and it emerges at this
33:37time in adolescence or is there
33:39actually something happening during
33:41adolescence that leads to that?
33:42And so this is now in the
33:43last part of my talk,
33:44I want to switch down to this part,
33:45the harder part,
33:47understanding which circuits are relevant,
33:49which brain regions and the timing.
33:51And this is really trying to get
33:53us to now get into the identifying
33:55the circus in the timing.
33:57So then we can start doing our
33:59perturbations of our mechanisms and
34:00then start to know what see what things
34:02to read out in terms of phenotypes.
34:04So it's around that time.
34:06When we realized we can't just
34:07study the visual system forever,
34:09even though I love it,
34:10I started collaborating with
34:12auto sabertini's lab and we
34:14recruited a terrific pH,
34:15now postdoc,
34:16who was a postdoc called Kevin
34:18Mastro to join our labs to work
34:21together on this project.
34:22And in particular we wanted to try to
34:24better understand and define the normal
34:27developmental refinement and and changes
34:28that are going on during adolescence
34:30in particular in the frontal cortex,
34:32the prefrontal cortex advice and
34:34ultimately in human and and non human
34:37primates which which is not an easy thing.
34:40And and I talked to Amy and
34:42many others about this,
34:43but I'm going to tell you about some
34:44of the the ways we're going about it
34:46today and this is all unpublished and
34:48ongoing work and I'm happy to to get
34:50feedback from this group but I think.
34:52You know, we've done a lot of our
34:54work over here in early development
34:55and of course that's super important.
34:57But we wanted to better understand sort
34:59of this next stage and we wanted to again,
35:01after the sensory systems develop and refine,
35:03now we want to get to the prefrontal,
35:05which began as the last area of the
35:07brain to myelinate and mature and all
35:09those connections are still being built.
35:10So we wanted to zone in on adolescence,
35:12but to do that we need to define
35:14what do we mean by adolescence and
35:15start to do what we've done in the
35:17visual system in the prefrontal and
35:19and so then we can start to ask, OK,
35:21how does environmental stressors,
35:22how do genetic stressors at different
35:25times impact circuit maturation
35:27and ultimately behavior.
35:29So this is before we could do any
35:30of that we needed to establish
35:32the neurotypical development and
35:34understand some of those milestones
35:35and we're doing it not only in mice.
35:37But through collaborations and
35:39work at the Stanley Center,
35:41we have a marmoset colony and
35:42we've been trying to have,
35:43we have a smaller colony devoted
35:45to these developmental studies.
35:47Go Ping Feng and many others at the road
35:48are leading the charge with the marmosets.
35:50But we've been very fortunate to be
35:51able to start to do some of this work.
35:53And this is all work led by Kevin
35:54Mastro that I'm telling you about.
35:55And there's Kevin who's amazing and he's
35:57on the job market and you guys should
35:59recruit him and Bernardo Sabatini,
36:00who's amazing collaborator,
36:02electrophysiologist.
36:02And So what Kevin really wanted to know is,
36:05you know, let's try to better
36:07understand again the match.
36:08But also we want to link this to
36:11behavioral readouts of decision making,
36:13cognitive flexibility,
36:14risk taking,
36:15things that we know are relevant
36:17to this adolescent window.
36:18And so that's going to require
36:20developing and identifying behavioral
36:21readouts that we can use in these
36:23different time points and and models.
36:26So Kevin wanted to know you know
36:28what circuit mechanisms support
36:29these changes and with a sort of
36:32focus on cognitive flexibility,
36:33readouts and decision making and so.
36:37The questions he's asking over what
36:39time skills do does behavior change
36:41and what circuits are changing?
36:42And can we link changes in circuits
36:44to these changes in behavior?
36:45Right.
36:45Many people are trying to do this,
36:47but we really want to do this
36:48over this developmental window.
36:50And then if we identify circuit changes,
36:52do they drive the behavioral changes?
36:54So then you want to go in and start
36:55to manipulate aspects of the circuit.
36:57And so we started thinking a lot
36:58about what time points are we
37:00do we talking about here and we
37:01of course did our due diligence,
37:03looked into the literature of course lots
37:04of work's been done in the prefrontal.
37:06So. So that's that's amazing
37:07because we could build on that.
37:09But what we were a little bit surprised
37:11to find is most times when we read these
37:13papers but the sort of end point of
37:15development of the prefrontal was like
37:17P60 or something and we're like okay,
37:19is that really the end because it
37:21seems pretty early to me and Kevin
37:23and so we started thinking about.
37:26Expanding this for this window,
37:27so not just stopping at PP60 you know,
37:30but to to broaden this out into
37:32early adulthood and to do it sort of
37:34go after the second phase of this
37:35using a combination of approaches
37:38from electrophysiology,
37:39slice Physiology and vivo
37:40Physiology and behavior.
37:42And as you'll see in a bit overlaying
37:44things like single cell multi omic
37:46sort of characterization on top of of
37:49of the characterization by Physiology.
37:51We were naive enough to think,
37:52oh,
37:52we'll just the first year we'll
37:54do neurotypical development and
37:55then we'll just start doing all
37:56the cool stuff three years later.
37:58Four years later,
37:58we're just wrapping up the
38:00first neurotypical paper.
38:01But it was required because if you
38:03don't have robust readouts and
38:04really understand the development,
38:06then you don't know what you know,
38:07what's your readout.
38:08And I think that was the goal.
38:09So is the brain done developing at P60?
38:12How many people think? The answer is yes.
38:14Oh, good, cause it's not.
38:17OK,
38:17So what kinds of now this is
38:18all new to me because I'm not a
38:20behavioral neuroscientist.
38:21So thank God for Kevin and Bernardo's
38:23lab because they've developed a
38:24lot of really great tools and and
38:26behavioral tasks and obviously all
38:29the electrophysiological readouts.
38:30So what Kevin decided to do first is
38:32start with a simple reversal task.
38:34And so mice were placed in a
38:35box with three nose port,
38:37they had to learn to initiate by poking in
38:39the center and then deciding right or left.
38:42And over the course of these trial errors,
38:43they would learn that one side
38:45is rewarded and the other's not.
38:46So after 20 tile,
38:4820 trials,
38:48then everything's reversed.
38:49And So what you're going to
38:51be measuring is switching.
38:52And we want to know how switching happens.
38:54That's plasticity and flexibility,
38:55how that changes with age.
38:58And what Kevin found is first of
39:00all just to say that the it's not
39:03really due to the number of trials.
39:05So he did a lot of controls on this.
39:06But what's really cool is in
39:09assessing the differences between
39:10the start and the end of the of the plateau.
39:13What he's finding, oops,
39:14I got, I clicked too fast,
39:15is that the young animals were much better
39:18at adapting their behavior than when the
39:20rules shifted than the older animals.
39:22And when I say young and old,
39:23we're talking about between this P60 up to P,
39:26right? Like. Okay.
39:27So, so there's a big,
39:29there's a lot happening up to P120 here,
39:32right, and beyond just the simple
39:34sort of deterministic reversal
39:35task where this is now showing
39:37the two ages in blue and red.
39:39And that's really, you know,
39:40the first important point I just showed
39:42you just graphed slightly differently.
39:43What's more interesting is
39:45when he started bringing in.
39:46Probabilistic 2 armed bandit task
39:48which is now going to start to
39:50get at behavioral strategies the
39:51animals use during this process.
39:53So the two armed bandit task is a task
39:54used by many groups and Bernardo's lab
39:56has done a lot on this but never really
39:58looked during development and aging.
40:00So that's what Kevin wanted to do.
40:01And essentially it's it's a different
40:03task because you can adjust the
40:05probability that each port is rewarded.
40:07You can kind of switch between
40:0990 and 10 probabilities,
40:10high versus low reward and then you can
40:12start to introduce sort of this this.
40:14You know,
40:15you know this sort of variation
40:16in the tasks and the animals
40:18have to adjust their strategy.
40:19It's harder, right?
40:20So what does the results show on the right?
40:22You can see again,
40:24the younger animals are much
40:25better at dealing with this
40:27and switching versus the older.
40:29So they they switch and they
40:31they're much more flexible and they
40:33switch to the reward report faster.
40:35So then the question is what's
40:36going on in the brain during this?
40:38And so Kevin started to record the
40:40activity using fiber photometry using
40:43the calcium indicator G Camp 6 and he
40:45can record activity during the task.
40:48And what's really cool about these
40:50experiments is what he found is that nice
40:52essentially during the two on bandit task,
40:55it turns out,
40:55as you can see here that you can actually,
40:58you know you can align to the
41:00initiation choice and out the outcome.
41:02You can see there's a huge difference
41:03in the calcium activity during the
41:05rewarded but not the unrewarded trials.
41:07And interestingly there's
41:08major age-related differences.
41:10It's actually showing that there's
41:12essentially a decrease, right?
41:13You can see there's a decrease
41:15in activity during the switching
41:16which was a little bit unexpected,
41:18but actually makes sense with other
41:19data that you'll see in a second.
41:21So there's age-related differences
41:23in terms of.
41:24Of of of how the activity patterns
41:26are aligning with these behavioral
41:28switching behavior.
41:29So.
41:30So now the question is what's
41:31happening at the circuit level and
41:32this is where slice Physiology comes
41:34in and he's been doing a lot on this.
41:36I'm only going to highlight the
41:37key results here.
41:38He's been looking at the input changes,
41:39sort of looking at refinement and
41:42connectivity, also local local circuit
41:43changes in the in the different ages
41:46and what he found by doing slice
41:49electrophysiology in the prefrontal cortex.
41:51Is in in the punchline of this given the
41:53interest of time is it's a systematic
41:56shift from excitation to inhibition,
41:58they become more inhibition dominant during
42:00this window of P60P90 to 120 exactly when
42:04we're seeing the behavioral changes, right.
42:05So you can see in the in on the
42:07bottom the way the data is graphed,
42:09he plots E / e E plus I.
42:12It's really enabling us to look at the,
42:14the, the inputs that are more excitation
42:16dominant versus inhibition dominant.
42:17Hopefully we could see as this big shift.
42:20Right.
42:20And so this is a lot of slice Physiology
42:23is down over many different animals
42:25and then he can go on and and do
42:26more because he can then start to
42:28look at the PV inner neurons and
42:30using viral strategies where you
42:32can target and label the PV cells,
42:34he could then also record from the
42:36PV neurons during the same task.
42:38And what's cool about that?
42:39This is just some of the the viral tools
42:41we can use in mice and marmosets that
42:43were developed by by Gord for Shell's
42:45group and others and Ben Deverman's lab.
42:47We can essentially he could find the opposite
42:50in terms of activity changes in the PV cells.
42:53So in the excitatory neurons when he recorded
42:55from those right it's a shift from E to I,
42:58but in the PV cells they shift
42:59up right so is the opposite.
43:01So it's really cool and so it now
43:03it all makes sense but at the time
43:04we didn't we didn't know this.
43:06So now this is really suggesting
43:07that the shift from excitation
43:09inhibition might be underlying the
43:10behavioral shift that he saw with
43:12the two armed bandit and in very,
43:14very like very new data now.
43:16He's starting to test that
43:18hypothesis more directly.
43:19So the real killer experiment then
43:20is to block inhibition in the older
43:22mice and see if you can shift
43:24them back to becoming a younger,
43:25more plastic, more flexible mouse.
43:28That could be interesting.
43:29So we did that experiment
43:30and this is very new data.
43:31So basically he used the dreads,
43:33the chemogenetic way to do with this.
43:35So if you inhibit the PV
43:37activity using dreads.
43:38The older animals P120 were
43:40trained on the two armed bandit.
43:42All the animals received the drug treatment,
43:44half had M cherry and orange
43:45and the other had the inhibitory
43:47dread and the PV neurons in blue.
43:49And he found that upon acquiring the
43:51two armed bandit task the adult mice.
43:53With less inhibition performed
43:55better than the control as shown
43:57by their increased reward.
43:58So they essentially the animals are
44:00better at tracking reward over time.
44:01And then, you know,
44:02the same animals were then placed on
44:04saline and SEM CNO and this is cool,
44:06the effect went away so you can reverse it.
44:08So there's a lot more to do here.
44:09But the data suggesting the
44:11inhibition of PV network improves
44:13performance in the older mice.
44:15Now, there's a lot more to do here
44:17in terms of asking how how long,
44:18how old you are to switch him back.
44:20But I think this is pretty cool because
44:22he's doing it in a very select way.
44:24And now of course,
44:25we're broadening out even more to say, well,
44:26what else is happening mechanistically here,
44:28not just in mice but also in marmosets?
44:31Because we're basically Kevin's training
44:33the marmosets to do the same task.
44:35So we can track marmosets
44:36from six months to two years,
44:38and he's already taught them
44:39how to do the two armed bandit.
44:40They learn really well.
44:41And we're also starting to take
44:43the brains of mice and marmoset.
44:45And start to look in a more unbiased
44:47way using on the genetics using
44:49cellular and single cell and Moxy
44:51omics to look at which synapses
44:53and cells are changing over time.
44:54And again I just showed you some of
44:56the behavior and and the Physiology
44:58to try to zip all this together.
44:59But the hope is at the end of
45:01the day probably many years from
45:02now when all this data comes in,
45:04it might give us more of a mechanistic
45:06handle on which cells are changing,
45:08which circuits are changing and then how
45:11those mechanisms relate to the behavior.
45:13And just to say real quick you know
45:14the single cell is starting to come
45:16out super interesting already.
45:17It's starting to highlight cell
45:18types like astrocytes and all the
45:20dendrocytes which is pretty cool.
45:21So it's pretty early days I was
45:23mentioning these marmosets if
45:24you've never seen a marmoset they're
45:26they're really pretty cool they
45:28can actually learn to do this task.
45:29They can touch the touchscreen
45:31instead of licking like a mouse they
45:33can Kevin's trained them how to how
45:35to do this sort of sort of reversal
45:37learning and and to to on bandit task.
45:40They do quite well.
45:41They,
45:42they they like to do the task and so
45:43stay tuned for some of this data.
45:45But already he's starting to show
45:47in parallel some of these animals
45:49were starting to be able to then do
45:51some of the work knowing what we
45:53know in mouse and applying this to
45:55the to the marmosets and and and
45:56ultimately this is sort of a summary
45:58of where we're going.
45:59We can start to also look at anatomical
46:01tracing experiments.
46:02And we can also start to then manipulate C4C,
46:06SM D1,
46:07some of the schema mutants like Rin
46:082A and then we can then pair that
46:10with environmental challenges.
46:11So first we look at genetics and
46:13see what these genetic leads,
46:15knockouts,
46:15wild types are doing to circuits
46:18and behavior.
46:18And then we can,
46:19based on the adolescent period do
46:21a second hit and see how environmental
46:23challenges like social isolation
46:25or social stress.
46:26Pairs with and combines with
46:28these genetic challenges and how
46:29ultimately that leads to changes in
46:31these behaviors as a starting point.
46:33And so we hope ultimately that at
46:35the end you might learn more and in
46:38more systematic way as we get more
46:40on the genetics and the biology,
46:41we we we can then apply what we
46:43learn to these other sort of circuit
46:45and cognitive level readouts.
46:47And then ultimately the goal is,
46:48Marina said in her introduction is
46:50to try to translate some of this
46:52work to the clinic to the patients.
46:54And and one example,
46:55C4 happens to be secreted and it has
46:58come out in CSF and we can measure
46:59it and Steve Mccarroll has done that
47:01and some of that works published.
47:02C4 can be not only read out
47:04and relates to copy numbers.
47:06So that's a good proof of concept,
47:08but we can read out other molecules
47:10that we're studying in the context
47:11of other disorders like Alzheimer's,
47:13a lot of neuroimmune molecules and
47:15synaptic markers we can read out in CSF.
47:17And thanks to a really amazing
47:19collaborative effort by many here that
47:21are also involved in the schizophrenia
47:23Spectrum biomarker consortium.
47:24The idea is what if we could
47:26start to bank and collect CSF
47:27not just from the later stage,
47:29but from this early stage of individuals
47:30that are at risk for developing
47:32schizophrenia or schizophrenia,
47:33bipolar and control and start to
47:35measure some of the leads that
47:37are coming out of the genetics.
47:38And then that would hopefully enable
47:40us to start to stratify patients.
47:42And as we get more information,
47:43we'll have better ways of reading
47:45out these different markers
47:46in the samples of patients.
47:47And relate this to the biology
47:49and ultimately to to cognition.
47:51So longterm goal that I think C4 is
47:54a good example of a genetic lead.
47:56We understand a little bit about the
47:58biology and we can in fact read out it
48:00on the genetic level and in the CSF
48:02and and I think this is the first start,
48:04but I think we have a long ways to go.
48:06But it really is going to require a village,
48:08a lot of collaboration and a lot of
48:10feedback from folks like you to think
48:12about how we could then you know
48:13expand some of this work into other
48:15models or into other disease areas.
48:17I focused on schizophrenia,
48:18but I think this is highly relevant
48:20to other disorders as well.
48:21So I'll end by thanking an amazing
48:23group of collaborators and support
48:24the human microglia in my lab.
48:26They're pretty cool.
48:27I'm the nucleus there,
48:29and this is the rest of them.
48:30Be being goofy at one of our retreats.
48:32So thanks very much.