David J. Linden - US grants

DESCRIPTION (provided by applicant): A guiding assumption in neurobiology has been that storage of information in the brain involves persistent, use-dependent alterations in neuronal electrical function. Determining the molecular basis of these forms of neuronal plasticity will provide the fundamental understanding necessary to provide therapies for diseases of memory as well as related persistent changes in electrical signaling that accompany addiction and epilepsy. Glutamate exerts its effects on the postsynaptic membrane through activation of two classes of receptor. Ionotropic glutamate receptors rapidly open integral cation channels while metabotropic receptors activate or inhibit G-protein coupled enzymes. Persistent modulation of fast neurotransmission, through both pre- and postsynaptic actions, comprises conventional LTP and LTD. Glutamate also acts upon postsynaptic mGluRs, particularly mGluR1 and mGluR5. Activation of mGluR1/5 stimulates phospholipase C2 and triggers a slow EPSC mediated by TrpC cation channels. Recently, we have reported that the mGluR1-mediated slow EPSC evoked by parallel fiber-Purkinje cell bursts shows strong LTD following either repeated climbing fiber-Purkinje cell synaptic activation or direct postsynaptic depolarization. Here, we propose to extend the characterization of this novel and unique form of synaptic plasticity. Aim 1. What molecular mechanisms underlie the induction of LTD(mGluR1) in Purkinje cells? We know that LTD(mGluR1) requires Ca influx for induction but we have a poor understanding of the parameters of the Ca signal. We shall use specific toxins of voltage- sensitive Ca channels to block LTD(mGluR1). This shall be done together with 2-photon Ca imaging of dendritic spines. We shall also attempt to induce LTD(mGluR1) by photolysis of postsynaptic caged Ca. We shall perform a screen to identify Ca-sensitive enzymatic processes (kinases, phosphatases, proteases, phospholipases) that may couple Ca influx to expression of LTD(mGluR1). Aim 2. What molecular mechanisms underlie the expression of LTD(mGluR1)? Is LTD(mGluR1) expression mediated by internalization of surface mGluR1? We shall test this by a) postsynaptically applying drugs and peptides to block mGluR1 endocytosis and b) transfecting Purkinje cells with our super-ecliptic-pHlorin-tagged mGluR1 to measure internalization with two-photon microscopy. The activity of mGluR1 is under dual regulation by Homer proteins and the prolyl isomerase PIN1. To test the hypothesis that this signaling cascade is involved in expression of LTD(mGluR1) we shall use a combination of drugs, peptides and knockout and mice. Aim 3. What are the consequences of LTD(mGluR1) for cerebellar circuit function &memory storage? We have shown that LTD(mGluR1) blocks the induction of conventional AMPA-R LTD at parallel fiber synapses. To extend analysis of meta-plastic effects, we shall assess the interaction of LTD(mGluR1) with short-term retrograde modulation of glutamate release that uses mGluR1-driven endocannabinoid signaling. PUBLIC HEALTH RELEVANCE: A guiding assumption in neurobiology has been that storage of information in the brain involves persistent, use-dependent alterations in neuronal electrical function. Determining the molecular basis of these forms of neuronal plasticity will provide the fundamental understanding necessary to provide therapies for diseases of memory as well as related persistent changes in electrical signaling that accompany addiction and epilepsy. Here, we propose to study the molecular mechanisms and functional role of a novel form of neuronal plasticity that we discovered recently, called "long-term synaptic depression of the neurotransmitter receptor mGluR1," abbreviated as LTD(mGluR1).

It has been appreciated for some time that glial cells in the brain express neurotransmitter receptors and can respond appropriately to application of exogenous neurotransmitter. These findings are exciting in that they suggest that glial cells might sense endogenous, synaptically released neurotransmitter and thereby have th capacity to respond to neuronal activity on a millisecond time scale. Recently, it has been shown using hippocampal microisland cultures, that stimulation of a single excitatory neuron can give rise to a rapid inward current in an adjacent glial cell (Mennerick and Zorumski, Nature 368:69, 1994). However, this current was almost exclusively mediated by the activation of electrogenic glutamate reuptake, not glutamate receptors. Reports of robust receptor mediated synaptic current recorded in glia are few, and are limited to specialized synaptoid contacts in the pituitary. in the cerebellar cortex, glial cells ensheathe the parallel fiber-Purkinje neuron synmapse unusually tightly and express a high density of inotropic glutamate receptors on the plasma membrane adjacent to these synapse. This suggests that cerebellar glial cells might be unusually well positioned to detect synaptically released glutamate. Using cell pair recording in cerebellar cultures from embryonic mouse, we have found that activation of a cerebellar granule neuron can give rise to a rapid inward current in an adjacent GFAP-positive glial cell, and that this current is mediated by activation of Ca-permeable AMPA/kainate receptors and is independent of glutamate reuptake or gap junctional coupling. We propose to extend these initial findings in several ways. First, simultaneous recordings will be made from a Purkinje neuron and glial cell innervated by a single granule cell in order to compare the neuronal and glial EPSCs in terms of receptor pharmacology, kinetics, quantal parameters, and short term presynaptic plasticity (facilitation). Second, as preliminary work has shown that these glial cells respond to exogenous GABA, stimulation of inhibitory interneurons will be undertaken to determine if the glial cells can detect synaptically released GABA as well. Third, all phases of this work will be extended to the slice preparation to determine if these novel properties of cerebellar glia are present in a preparation that maintains a greater degree of morphological fidelity to the intact cerebellum. These investigations of neuron/glia signaling hold promise for deepening our basic understanding of glutamatergic and GABAergic transmission, pathologies of which include epilepsy, hypoxic/ischemic damage and diseases of memory.

DESCRIPTION (Adapted from applicant's abstract): A guiding assumption in neurobiology has been that storage of information in the brain involves persistent, use dependent alterations in synaptic strength. One useful model system for this endeavor has been cerebellar long-term depression (LTD), in which co-activation of climbing fiber and parallel fiber inputs to a Purkinje neuron (PN), induces a persistent input specific depression of the parallel fiber PN synapse. This phenomenon has been suggested to be necessary for certain forms of motor learning including associative eye blink conditioning and adaptation of the vestibulo ocular reflex. Recently, the converse phenomenon, cerebellar long-term potentiation (LTP) has also been described, in which the parallel fiber PN synapse is strengthened by repetitive parallel fiber stimulation at intermediate frequencies, thus endowing this synapse with the capacity for use dependent bidirectional modification, a computationally important property. In recent years, this laboratory has focused upon defining the requirements for LTD induction using a cell culture model system in which parallel fiber stimulation is replaced by glutamate pulses and climbing fiber stimulation is replaced by direct depolarization of the PN. Most recently, we have developed several new protocols which have expanded the types of questions we may address: sustained recordings from single PNs in culture to investigate the late phase of LTD; recordings from two ultra reduced PN preparations that display LTD in the absence of dendritic spine compartments (acutely dissociated PNs and PN dendritic macropatches); and investigations of cerebellar LTP and LTD in granule cell PN pairs in culture. In addition, we now undertake "conventional" LTD experiments using a brain slice preparation. We propose to use these techniques to address the following questions. First, is cerebellar LTD, which we know to be expressed postsynaptically, mediated by an alteration in AMPA receptor kinetics? Second, which intracellular signaling pathways are engaged by the late phase of cerebellar LTD? Third, what are the requirements for the induction of LTP at the granule cell Purkinje cell synapse and where is its locus of expression? Finally, revisiting an ongoing controversy, what is the role of nitric oxide/cGMP signaling in cerebellar LTD induction? At the level of basic science, these investigations are central to an understanding of the cellular substrates of information storage in a brain area where the behavioral relevance of the inputs and outputs is unusually well defined. In addition, these investigations have potential clinical relevance not only for cerebellar motor disorders, but also for disorders of learning and memory generally.

DESCRIPTION: (Adapted from the Investigator's Abstract) A central hypothesis of modern neurobiology is that memory is stored through use-dependent changes in synaptic strength. Most work in this area has focused upon long-term potentiation and depression (LTP & LTD) of excitatory, glutamatergic synapses. One limitation of this approach is that the brain regions where LTP/LTD are most often studied, such as the hippocampus, receive information that is so complex that its content cannot be easily characterized. In contrast, in the cerebellum it has been possible to propose a "circuit diagram" for some simple forms of learning such as adaptation of the vestibulo-ocular reflex, and associative eyeblink conditioning. For example, it is possible to assign the conditioned (CS) and unconditioned stimuli (US) in associative eyeblink conditioning to specific pathways (mossy and climbing fibers, respectively). Over the last 20 years, a series of experiments that have used behavioral tasks together with extracellular recording, lesion and reversible inactivation have produced a strong case that the cerebellum is critical for these forms of motor learning. However, the precise synaptic location of the cerebellar engram has been elusive, with some studies favoring the synapses received by the Purkinje cell while others have implicated those received by the deep cerebellar nuclei (DCN). While the cellular electrophysiology of the Purkinje cell has been widely investigated, there are few studies which have examined the DCN. Recently, this laboratory has performed intracellular recordings from neurons of the DCN using a brain slice preparation. These have shown that activation of GABAergic Purkinje cell-to-DCN synapses (with a burst and pause stimulus that mimics natural firing patterns) results in a prominent rebound depolarization and associated spike burst which are evoked upon release from hyperpolarization, providing a mechanism by which inhibitory inputs can drive postsynaptic excitation. In these cells, LTP can be elicited by short, high-frequency.trains of IPSPs that reliably evoke a rebound depolarization in the DCN neurons. LTD is induced if the same protocol is applied while the amount of postsynaptic excitation is reduced (by postsynaptic hyperpolarization or an internal Na channel blocker). The polarity of the change in synaptic strength is correlated with the amount of rebound depolarization-evoked spike firing and the amplitude of the resulting postsynaptic Ca transient. In addition, we have preliminary data demonstrating LTD of the glutamatergic mossy fiber-DCN synapse. The present proposal seeks to build upon these initial results by addressing the following questions: What specific conductances contribute to the intrinsic excitability of DCN cells and how are they altered by neuromodulators such as acetylcholine and serotonin? What are the basic computational properties of LTP and LTD at the Purkinje cell-DCN synapse (optimal induction, saturability, reversibility, input specificity)? What Ca signals and second messenger systems are required for induction of LTP and LTD at the Purkinje cell-DCN synapse? What are the requirements for the induction of LTP and LTD at the mossy fiber-DCN synapse? At the level of basic science, these investigations are central to understanding the cellular substrates of information storage. In addition, they have potential clinical relevance for both cerebellar motor disorders and disorders of learning and memory generally.

It is remarkable that experience can modify neuronal function in a manner that is rapid and which can last for an entire lifetime as long-term memory. It is now well established that particular patterns of synaptic activity can give rise to alterations in synaptic strength, and these alterations (called LTP and LTD) are believed to underlie both memory storage and the activity-dependent fine-tuning of brain development. Both LTP and LTD have been shown to have late phases that require the synthesis of new proteins. Thus, synaptically-driven gene transcription is likely to be a key event in laying down long-term memories. There is general agreement that this process requires postsynaptic Ca influx. However, the details of how Ca influx is coupled to transcriptional events remain poorly understood. What are the spatial and temporal requirements for Ca signals to trigger transcription in neurons? Attempts to address this question have almost exclusively involved bath application of glutamate or high K to dissociated neuronal cultures. Given the nonphysiological nature of these stimuli, it is not surprising that conflicting results have emerged, with some investigators claiming a requirement for a Ca transient in the nucleus while others have reported that a Ca transient restricted to dendrites is sufficient. We will address this issue using a preparation that more closely resembles the intact brain. Here, we propose to stimulate glutamatergic synapses impinging upon dendrites of neurons in brain slices while measuring Ca concentration throughout the neuron at high resolution using multiphoton microscopy and simultaneously measuring both somatic membrane potential and transcription factor activity. The latter will involve both dynamic measurements using a CREB/CBP FRET system and posthoc analyses using high resolution in situ hybridization (CATFISH) and immunohistochemistry with phosphorylation-state specific antibodies. This analysis will be performed in two types of CNS neurons with different dendritic morphologies and firing properties. What are the critical nuclear targets for Ca triggered transcriptional events in neurons? Much attention has been directed towards the transcription factor CREB, to the exclusion of other potentially important targets. Cell culture experiments have indicated that the transcription factor SRF (Serum Response Factor) is robustly activated by Ca signaling in neurons. Moreover, many if not all activity-dependent IEGs contain binding sites for SRF and its associated factors. Thus we hypothesize that SRF is a key target of synaptically-driven Ca signals that supports initiation of new gene transcription. We propose to use novel spatial and temporal gene ablation techniques to address the necessity of SRF for the activation of a panel of neuronal immediate-early genes and in several forms of activity-dependent plasticity known to have transcription-dependent late phases (including hippocampal LTP and cerebellar LTD). Moreover, a novel small molecule SRF activator will be used to acutely induce SRF-dependent transcription in the absence of synaptic stimulation. The latter approach will be used to address the sufficiency of SRF-dependent transcription for long-term changes in synaptic efficacy and IEG transcription.

[unreadable] DESCRIPTION (provided by applicant): The Neuroscience Graduate Program, which was begun in 1983, has its headquarters in the Department of Neuroscience at The Johns Hopkins University School of Medicine. Consisting of 85 faculty drawn from various departments across the University, it serves as the hub of a broad spectrum of efforts for the training of graduate students, encompassing molecular, cellular, developmental, systems, cognitive and computational neuroscience as well as neurobiology of disease. Each year, from a pool of ~200 applicants, we typically matriculate 10-12 Ph.D. candidates as well as one to four candidates for combined M.D./Ph.D. degrees (who are admitted through a separate process). Students enter the program with diverse undergraduate backgrounds ranging from computer science to biochemistry. In the first year they are required to take a year-long integrative lecture course with lab entitled "Neuroscience and Cognition" as well as a seminar on "Science, Ethics and Society". Research opportunities are presented to students through a Departmental Retreat, Lab Lunches (which feature work-in-progress) and a Mini-symposium series by Program Faculty specifically designed to help first-year students choose their research rotations. This information is used to help pick three 12-week lab rotations which are typically completed by the end of the first academic year, following which, a thesis lab is selected. By the end of the second year, students complete six additional elective courses, many of which are chosen from a list of 12 small seminar-style courses in Neuroscience specialties. Following completion of a Comprehensive Exam at the end of Year Two, students write and defend a Thesis Proposal which is written in the form of a Predoctoral NRSA. Each student is advised by two Prethesis Advisors in Years One and Two (at three month intervals) and an individualized Thesis Advisory Committee thereafter (at six month intervals). Thesis Advisory Committees make reports to the Graduate Program Steering Committee which carefully tracks the progress of each student in the program as well as setting overall program policy. At present, 70 students are enrolled in the Neuroscience Graduate program. The average time to complete the Ph.D. has been 5.1 years. Of the students who have graduated from our program 92% have remained in biomedical research and 85% have remained in academic biomedical research. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Modern theories of memory storage have largely focused on persistent, experience-dependent changes in synaptic function such as long-term synaptic potentiation and depression (LTP & LTD). These phenomena are appealing in models of memory, in part because they typically display some degree of synapse-specificity, allowing for a very large number of independently modifiable units and, consequently, a very large storage capacity. In addition to these synaptic changes, evidence has now emerged for persistent changes in intrinsic neuronal excitability, what we call "intrinsic plasticity", produced by certain forms of training in behaving animals and artificial patterns of activation in brain slices and neuronal cultures. These intrinsic changes may function as a portion of the engram itself, or as a related phenomenon such as a trigger for the consolidation or adaptive generalization of memories, particularly non-declarative memories. Several years ago, we published the first report of persistent synaptically driven changes in intrinsic excitability in the brain. This, in the deep cerebellar nuclei (DON), a region which is central to memory storage for certain tasks such as associative eyelid conditioning. We have since performed an extensive parametric description of the induction requirements and the expression of this phenomenon. Here, we propose to extend these initial observations by investigating the cellular and molecular basis of intrinsic plasticity in the DCN. First, we wish to characterize the receptors involved in intrinsic plasticity with particular emphasis on receptors for glutamate, serotonin and norepinephrine. Second, we will address the role of second messenger cascades including protein kinases, phosphatases, lipases and Ca stores. Third, we shall seek to identify the particular ion channel(s) involved in the expression of intrinsic plasticity through recording and occlusion experiments. Fourth, we shall use confocal imaging and uncaging to determine the spatial extent of intrinsic plasticity. This is basic research to address the molecular mechanisms that underlie memory storage, using an unusually well-defined model system. It is hoped that this work will be useful for creating therapies and diagnostics for diseases of memory. Because these cellular and molecular processes are not only involved in memory storage, this work has implications for other brain diseases as well, including epilepsy and addiction.

DESCRIPTION (provided by applicant): Following local trauma, damaged axons in the adult mammalian central nervous system (CNS) regress, and subsequent regeneration of these damaged axons is very limited. This is thought to strongly constrain recovery of CNS function and contributes to paralysis, sensory dysfunction and cognitive impairment. To date, the study of brain axon regeneration has almost exclusively relied upon postmortem analysis of fixed tissue from intact preparations, which yields static images. These snapshots are suboptimal for evaluating therapeutic interventions: They often fail to distinguish regenerating axons from sprouting of undamaged fibers or spared axons at the lesion site. We have developed a model system in which long-distance regeneration of axons can be studied with time-lapse imaging in the intact adult mouse brain. Systemic treatment of adult rats with p-chloro-amphetamine (PCA) causes rapid regression of dorsal raphe serotonin axons, followed by a slow return of serotonergic innervation over many weeks. We have adapted this PCA protocol to adult BAC transgenic mice in which the complete extent of serotonin neurons is labeled with EGFP. Using a two-photon microscope and a cranial window, we can repeatedly image the same volume of neocortex and thereby track serotonergic axons before and e 13 weeks after lesion with PCA to provide time-lapse measurements of identified surviving, regressing and regenerating fibers. Here, we propose to develop and extend this model system. Aim 1. Do rapid dynamic events in regenerating serotonin axons following lesion with PCA predict features of stable regenerated axon morphology? To date, we have performed a low temporal resolution survey with weekly measurements. This provides an overview but has not allowed for examination of axons on a minutes-to-days timescale. We propose to augment our dataset with measurements at 10 min and daily intervals during two crucial periods: immediately following PCA to capture regression and ~ 7-8 weeks following PCA, when many pioneering fibers are entering the field of view. Aim 2. What are the key short and long-term structural dynamics of regenerating 5HT axons following a thermal lesion of the neocortex? We propose to repeat in vivo time-lapse imaging of serotonin axons, replacing PCA treatment with focal thermal lesions. Our goal is to have two well-defined model systems for axonal damage and regeneration, one conventional, pan-cellular and glial-scar-forming and the other cell-type-specific and non- scar forming in order to compare molecular interventions and candidate therapies for functional recovery. Aim 3. Is expression of integrin ¿1 in serotonin neurons required for all phases of axonal regeneration in the neocortex following thermal or PCA lesion? Serotonin neurons express high levels of integrin ¿1, a protein that forms part of the receptor for the permissive growth substrate laminin. We shall cross mouse strains: floxed integrin ¿1 with serotonin transporter-driven Cre, to selectively delete integrin ¿1 in serotonin neurons coupled with in vivo time-lapse imaging of serotonin axons in response to PCA and thermal lesions.

In the mammalian central nervous system (CNS), axonal arbors are most often large and highly branched. What is the computational function of the axonal arbor and its array of presynaptic terminals and how does its structure and function change as a result of experience? We have developed a preparation that allows us to image cytosolic Ca transients, a consequence of spike firing, in an axonal arbor in the intact brain of the unanesthetized adult rat. We propose to use this preparation to address three central questions in neuronal circuit function. Aim 1: When action potentials invade the cerebellar mossy fiber axonal arbor, how are they spatially distributed to evoke Ca transients in the array of presynaptic terminals? Preliminary data from spontaneous firing suggests that mossy fiber axonal signal distribution is neither entirely reliable nor accounted for by simple branch-point failure rules. Rather, one terminal can have spike-driven Ca signals while another, a few microns away, on the same axonal segment does not. To determine whether the spatial pattern of presynaptic Ca failure is stereotyped or stochastic, we shall implant electrodes in the middle cerebellar peduncle to artificially activate ascending mossy fibers with various patterns of stimuli and compare with those driven by tactile stimuli, spontaneous movement or proprioceptive signals. Aim 2: Is the spatio-temporal pattern of Ca signal distribution in the mossy fiber axon sculpted by GABAergic inhibition from Golgi cells? Golgi cell axons form boutons which release GABA onto mossy fiber terminals, thereby activating GABA-B receptors. We shall record both spontaneous Ca transients and those evoked by a range of stimuli (tactile, proprioceptive, peduncle electrode) together with GABA-B receptor agonists, antagonists and allosteric modulators applied to a port in the cranial window. Aim 3: In trace eyelid conditioning, a simple form of associative learning, is the spatial representation of the CS across an array of ponto- cerebellar mossy fiber terminals modified by experience? Specifically, is it altered by habituation, associative learning, pseudoconditioning or extinction? We shall record CS-evoked and electrode-evoked mossy-fiber Ca transients in 3-D mapped axonal arbors, during and following behavioral training.

? DESCRIPTION (provided by applicant): It is widely hypothesized that memory is stored in the brain as enduring changes in the wiring diagram and strength of synaptic connections between neurons. The receptive structure for most of these synapses is the dendritic spine, a micron-scale protrusion emitted from the neuron's dendrite which senses neurotransmitter released by the closely-opposed presynaptic terminal. Several lines of evidence suggest that dendritic spine density and turnover in the neocortex is modified by learning and is positively correlated with learning rate. Since the early 1990s, it has been established that dendritic spine density varies by about 35% over the course of the ovarian cycle in female rats, mice and nonhuman primates, being higher in proestrus and diestrus and lower in estrus. Similarly, surgical ovariectomy leads to an ~40% loss of spine density which can be rapidly reversed by 17-beta-estradiol (E2) treatment. These findings have suggested an important question: If memory is largely encoded in spiny synapses and if spine density fluctuates by ~40% over the ovarian cycle, then how does long-term memory persist in female mammals in the face of this fluctuation? Why doesn't memory degrade with each ovarian cycle? To date, measurements of spines in relationship to the ovarian cycle have relied upon traditional anatomical methods applied to postmortem tissue. These methods preclude within-animal and within-dendrite comparisons across time. Here, we propose to use in vivo two-photon microscopy together with existing transgenic mouse lines (including Thy1M-EGFP for sparse labeling of layer 5 pyramidal cells) to produce time-lapse images of identified spiny dendrites in the neocortex and thereby address two crucial questions. Aim 1: When spines are lost following ovariectomy, do new spine regrow in those same dendritic locations when estrogen levels rise following E2 treatment? Pilot studies will use traditional histological techniques to determine the regions of the neocortex in which the largest and most reliable loss/recovery of spines can be seen in layers I - III. This information will then be used to guide the placement of cranial windows for time-lapse imaging spanning overiectomy or sham surgery and subsequent E2 or vehicle treatment. Aim 2: When spines are lost during estrus do they tend to regrow in that same dendritic location when estrogen levels rise again in proestrus/diestrus? As in Aim 1, initial histological experiments will guide the placement of cranial windows. Then, daily monitoring will commence together with E2 ELISA and vaginal swab assays of ovarian cycle status. Time-lapse imaging of dendritic spines together with manipulation of estrogen levels either though exogenous manipulation (Aim 1) or natural cycles (Aim 2) will allow us to determine whether dendritic spines lost during low-estrogen conditions are regrown at the same dendritic location when high estrogen levels return. Imaging over several cycles (Aim 2) will allow us to determine if spine loss in estrus tends to occur in a particular subset of spine locations or spine types that define a high turnover pool. Finally, we will use Thy1M-GFP mice injected with virus encoding red fluorophore in posteromedial thalamus as a first attempt to determine whether regrown spines are re-contacting their original thalamo-cortical axons in neocortical layer 1. This work shall have important clinical implications for hormone replacement therapy after ovariectomy or menopause.

? DESCRIPTION (provided by applicant): The induction of gene expression by synaptic activity is essential for the later phases of long-term synaptic potentiation and depression (LTP & LTD). While activity-dependent transcriptional events in neurons appear to be necessary for memory consolidation, there is poor understanding of the specific target genes and the molecular interactions between transcription factors and the promoter regions of these target genes that are required for the later phases of synaptic plasticity. The field has been hampered by a series of technical limitations in this endeavor. These include an inability to record the gene expression-dependent late phase of synaptic plasticity in single mammalian neurons and to study late phase synaptic plasticity using precise activation of individual, defined and imaged synapses. Furthermore, it has not been straightforward to mutate and engineer defined promoter regions of candidate plasticity genes and to image the expression of target genes in individual neurons undergoing synaptic LTP/LTD. We have solved these technical limitations and so can perform long-term patch-clamp recording of the transcription-dependent late phase of cerebellar LTD in single Purkinje cells in cultures and brain slices, evoked by uncaging of glutamate at defined dendritic spines, while imaging spine and nuclear Ca and fluorescent markers of gene expression. We perform these experiments in neurons from null mice and transfected with engineered bacterial artificial chromosomes (BACs) to easily manipulate the promoters of relevant genes. This work has revealed that the late phase of LTD requires transcription of the Arc gene. We hypothesize that climbing fiber activation drives nuclear Ca and CREB binding to Arc CRE6.9 and that climbing fiber + parallel fiber co-activation is necessary to initiate phosphatase and MAL signaling cascades that originate in activated parallel fiber spines and propagate to the nucleus to result in MEF2D and SRF binding to MRE6.9 and SRE6.9 in the Arc promoter respectively. In this way, all three binding events in the SARE region of the Arc promoter are necessary to trigger Arc transcription and the late phase of LTD. If our hypothesis is correct, this would be a double-AND logical operator encoded in the structure of the Arc promoter. Associativity is the hallmark of many memory traces, yet remains almost entirely unexplored for late phases of LTP/LTD in any brain region or model organism.

Project Summary It is widely believed that damaged axons in the adult mammalian brain have little capacity to regrow, thereby impeding functional recovery after injury. Serotonin neurons appear to be a notable exception. We have produced time-lapse images of serotonin axons in the neocortex of the adult mouse. Serotonin axons undergo massive retrograde degeneration following amphetamine treatment and show subsequent slow, long-distance regrowth. A stab injury that transects serotonin axons running in the neocortex is followed by local regression of cut serotonin axons and by regrowth from cut ends that projects across the stab rift zone. By contrast, dopamine fibers originating in neurons of the ventral tegmental area (VTA) are also small-diameter unmyelinated axons, yet they are not damaged by amphetamine treatment and they fail to regrow after stab transection. What are the molecular specializations that allow serotonin axons to regrow while other injured axons in the brain fail to do so? To begin to address this question we have harvested pools of serotonin neurons from the dorsal raphe of adult mice in order to measure gene expression using genome-wide RNA sequencing (RNA-Seq). This was done 1 week after the end of either amphetamine or saline treatment. Our preliminary data have revealed a set of genes that are significantly upregulated in serotonin neurons during amphetamine-evoked axon regrowth. Here, we propose to complete and extend this screen as a first step to define the molecular requirements for serotonin axon regrowth. Aim 1. What genes are differentially expressed in serotonin neurons during amphetamine-evoked regrowth of their axons in the neocortex? Our preliminary data reveal ~50 upregulated and ~60 downregulated genes measured ~1 week after the end of amphetamine treatment as compared with saline-treated-controls (using a false-discovery rate of p<.05). To complete this data set, we shall extend these RNA-Seq measurements using a time course of amphetamine and saline-treated mice at t= 1 day, 1 week and 18 weeks. There will be 8 animals per group (4 male, 4 female), and we shall validate genes of interest with in situ hybridization and immunohistochemistry. We shall also compare the gene expression changes in response to amphetamine challenge in serotonin neurons with those changes produced in dopamine neurons of the VTA. Aim 2. What genes are differentially expressed in serotonin neurons during regrowth of serotonin axons following a neocortical stab lesion? In contrast to amphetamine lesions, stab lesions produce only local axonal regression of serotonin and dopamine axons. Stab lesions form a glial scar, as occurs in stroke and other forms of acute brain injury. We shall compare dorsal raphe serotonin neuron and VTA dopamine neuron expression profiles in response to stab and amphetamine injury using the same post-injury time points. The results of these two complementary injuries in regenerating and non-regenerating neurons will inform later manipulative experiments to modulate serotonin axon regrowth and, ultimately, confer the capacity to regrow on non-serotonergic neurons in the brain.

In recent years, mild therapeutic cooling, to about 32°C for 8-48 hours, has been shown to attenuate hypoxic brain injury that occurs following cardiac arrest or ischemic stroke. In addition, induced cardiac arrest necessitated by aortic arch surgery is now often accompanied by strong therapeutic cooling, to about 16°C for~60 min. Despite a wealth of preclinical and clinical data on the neuroprotective benefits of therapeutic brain cooling, little is known about how this process affects the structure and function of neurons and hence the extent and quality of recovery. Here we seek to answer the question: What are the transient and persistent changes in neuronal structure and function that accompany mouse models of therapeutic cooling? We find that intraperitoneal AMP injection coupled with an ambient temperature of 15°C or 30°C results in reduced heart rate and a stable core temperature of ~16°C or 32°C, respectively, in adult mice. These mild or strong torpid states were maintained for 1 or 8 hours and were rapidly reversed upon rewarming. In Thy1-M mice, which express EGFP sparsely in layer 5 pyramidal neurons, we have implanted cranial windows overlying the somatosensory cortex for time-lapse in vivo imaging to measure spine density and dynamics before, during and after induced hypothermia. Our pilot data with both mild and strong cooling show a selective loss of long thin spines that lasts for 1-2 days. In addition, we observe damage to apical dendrites with blebbing and distal retraction, which resolved within 7 days. Here, we propose to complete and extend these studies as a first step towards determining whether mild or strong therapeutic cooling alters neocortical function. Aim 1. What changes in the fine structure of neocortical neurons are evoked by mild or strong therapeutic cooling and do these changes persist? We shall use two-photon microscopy together with EGFP-expressing transgenic mice to produce time-lapse images of layer 5 pyramidal neurons, somatostatin interneurons and parvalbumin interneurons. This technique allows us to measure the dynamic aspects of neuronal structure that are inaccessible to fixed tissue studies such as spine turnover and dendritic retraction, blebbing and regrowth. Aim 2. Does mild or strong therapeutic cooling produce lasting changes in the electrophysiological function of identified neocortical neurons? We shall prepare ex vivo slices of the neocortex 1 and 7 days after mild or strong cooling to perform whole-cell recordings from identified layer 5 pyramidal cells as well as somatostatin and parvalbumin interneurons to assess basal firing rate, intrinsic excitability and the kinetic properties of miniature and evoked IPSCs and EPSCs. Together, these measurements will allow us to detect cooling-evoked transient or persistent changes to neocortical circuits.