Richard Bertram - US grants

Bertram
9981822
The investigator uses mathematical models to study dynamical
and biophysical mechanisms for the multiple modes of bursting
observed in pancreatic Beta-cells. In pancreatic islets,
Beta-cells burst with a period of 10-30 seconds, while isolated
Beta-cells burst with periods of either 2-5 seconds or 1-2
minutes. The primary goal of this project is to answer the
following questions: (1) How does a single oscillator (the
Beta-cell) produce such a wide range of bursting behaviors? (2)
What biophysical mechanisms drive the different modes of
bursting? (3) Why does the bursting behavior differ between
isolated cells and intact islets? To address these questions the
investigator develops a family of mathematical models that
employ several slow variables. He studies how well these models
reproduce key electrophysiological and calcium imaging data from
isolated Beta-cells, cell clusters, and intact islets.
Pancreatic Beta-cells are the only cells in the human body
that secrete insulin, a hormone required by other cells to
process sugars in the blood. If Beta-cells function improperly
then late-onset diabetes occurs, the most common form of
diabetes. Insulin is secreted from Beta-cells when the cells
generate electrical impulses, so understanding the electrical
activity of Beta-cells is crucial to understanding insulin
secretion. For many years mathematical models have provided
insight into the complex electrical behavior of Beta-cells. In
the current project, the investigator develops improved
mathematical models to understand how Beta-cells produce patterns
of electrical impulses that are key to normal insulin secretion.
Only when these normal patterns are understood can abnormal
electrical activity, and abnormal insulin secretion resulting in
diabetes, be understood.

Penn State Erie's REU Program in Mathematical Biology will involve undergraduates in one of the richest and fastest growing fields in mathematics. It gives the six participants research training in mathematical biology and introduces them to several areas of current interest: cancer modeling, pattern formation in active media, and the modeling of electrically excitable cells. The only prerequisites are the successful completion of a course in ordinary differential equations and an interest in biology. The goal of the program is two-fold. First the students receive an introduction to the theory and methods of nonlinear dynamical systems, including phase plane analysis and bifurcation theory. These techniques are then used to answer questions about mathematical models of biological phenomena. This is of great value since dynamical systems arise in many areas in addition to biology, such as physics, engineering, and chemistry. The experience gained in studying models in detail will serve them well in future courses regardless of the field they choose to pursue. Second, the students are exposed to current topics of research in mathematical biology. Many of the topics discussed during the program are at early stages of development, and offer many opportunities for future research. Students will have access to Pentium computers running the Linux operating system and will learn the mathematical software package XPP, a powerful tool for phase plane and bifurcation analysis. Seminars will be given throught the program by both faculty and student participants along with discussion groups at the end of each day.

Bertram
The investigator and his colleagues study the behavior of
insulin-secreting pancreatic beta-cells. These cells display a
wide range of complex dynamics that are best understood with the
aid of mathematical modeling and analysis. The investigator has
recently developed a new mathematical model for beta-cells that
postulates an intracellular calcium subspace compartment located
between the endoplasmic reticulum and the cell membrane. This
model is based on recent data from a collaborating lab, and is
being further developed as new experiments are performed. One
direction of development involves the investigation of complex
electrical bursting patterns that are often observed in
pancreatic islets and isolated beta-cells. The investigator
determines, through mathematical modeling, whether these complex
rhythms are due to oscillations in glycolysis -- that is, whether
the coexistence of a membrane-driven bursting oscillator and an
endogenous glycolytic oscillator is consistent with the published
data, and with data from collaborating labs. One key question
that is addressed is whether these oscillators involving
glycolytic rhythms can be synchronized by electrical coupling.
This is a necessary condition since beta-cells are electrically
coupled in islets. Finally, the effects of stochasticity through
channel noise are investigated in both single-cell and islet
models. Channel noise is evident in the electrical patterns of
single beta-cells, and suppression of this noise through
elecrical coupling may play a key role in the more regular and
slower bursting patterns observed in islets.
The study of pancreatic islets, composed of beta-cells, is
important for a number of reasons. From a clinical viewpoint,
they are key players in normal glucose homeostasis. Because
beta-cells are the only cells in the body that make and secrete
insulin, their malfunction leads to type II or late-onset
diabetes, the most prevalent form of the disease. To understand
what makes the cells malfunction, one must first understand their
normal functioning, which is the primary aim of this work. From a
biological viewpoint, beta-cells are of great interest as the
focal point of a number of biochemical and cellular processes.
They are similar in many ways to neurons, but have more complex
biochemical regulatory mechanisms. Thus, information gained from
the study of beta-cells largely transfers over to neurons, as
well as other secretory cells. Finally, from a mathematical
viewpoint, beta-cells display rich dynamics that can only truly
be understood with the aid of mathematics. Modeling and anlysis
of beta-cells has led to new mathematics, most of which can and
is being applied to the study of other bursting nerve and
endocrine cells. The investigator involves graduate and
undergraduate students in the project, providing them with
research opportunities at the important interface between
biological and mathematical sciences.

[unreadable] DESCRIPTION (provided by applicant): The overall aim of this project is to combine experimental studies with mathematical modeling to understand the interactions between the hypothalamus and the pituitary gland that lead to rhythmic secretion of the hormone prolactin. This work will be done in close collaboration between an experimental lab and a mathematical modeling lab; both labs reside at the same university, facilitating daily interactions. The primary goal of this research is to understand how neurosecretory cells within the hypothalamus interact with the pituitary gland to produce daily rhythms of prolactin secretion in the rat during pregnancy. Prolactin is one of the most versatile hormones of mammalian organisms, with over 300 separate biological activities. The prolactin secreted following the mating stimulus has many targets, including other endocrine glands, and is important for maintaining a normal pregnancy in the rat. Prolactin is secreted by pituitary lactotrophs. Secretion from these cells is tightly regulated by the hypothalamus, a region of the brain that transmits time-of-day information to the rest of the body. The interaction between hypothalamic neurons and lactotrophs is complex; the neurons influence each other as well as the lactotrophs, and prolactin from lactotrophs feeds back onto and influences the hypothalamic neurons. Such a complex system is ideal for mathematical modeling, which can provide insight into the influence of the various network interactions, and can be used as a tool for integrating information. In this project, mathematical modeling is combined with experimental studies. The model will be calibrated by experimental data, and will make predictions that will be tested in the laboratory. This joint experimental-computational approach is well suited for understanding the complex hypothalamus-pituitary network. Student training is an important element of this project. It is anticipated that graduate students and postdoctoral fellows will play very active roles in the research described herein. This participation will provide multi-disciplinary training that will be invaluable for an increasingly multi-disciplinary workplace. There are four specific aims in this proposal. First, a mathematical model will be developed for pituitary lactotrophs. This model, based largely on experimental data on cultured lactotrophs from our lab, will provide a mechanistic understanding of the activity patterns of these cells. It will also be used to understand how the activity is modified by hormones such as dopamine and oxytocin. Second, mathematical models will be developed of hypothalamic dopamine- and oxytocin-secreting neurons, using hypothalamus slice data from our lab. These neurosecretory cells regulate prolactin secretion from lactotrophs, and are themselves under the influence of neurons within the suprachiasmatic nucleus (SCN). The third specific aim is to develop a mathematical model of the network interactions among the various hypothalamic neurons and pituitary lactotrophs. This model will be minimal, focusing on the network interactions between cells rather than the detailed biophysical processes that take place within cells (the goal of the first two aims). Fourth, the role of rhythmic clock gene expression in dopamine- and oxytocin-secreting neurons will be investigated. If the expression patterns are shown to be rhythmic, then this suggests that these cells provide circadian input to the pituitary that is separate from, but may be entrained by, neurons within the SCN. These studies will support the mission of NIDA by establishing the way the normal brain functions in the absence of drugs of abuse to support the pituitary gland. It will then lead to studies of effects of drugs of abuse on brain-pituitary function. [unreadable] [unreadable]

The long-term goal of this research is to understand the mechanism of insulin secretion from beta-cells, which are clustered into islets of Langerhans within the pancreas. Beta-cells secrete insulin in a pulsatile fashion, with pulse period of approximately five minutes. Disruption of this oscillatory pattern is observed in diabetics and their near relatives. The two primary aims of this project are (1) to better understand the mechanism for pulsatile insulin secretion, and (2) to investigate potential mechanisms for the synchronization of the population of islet oscillators. We focus on oscillations in glycolysis, coupled to the electrical activity of the cell, as a mechanism for pulsatile insulin secretion. This is based on data in the literature and from a collaborating lab showing oscillations in mitochondrial variables. From a mathematical viewpoint, the model consists of two mutually coupled oscillators, which we call a dual-oscillator model. Glycolysis is the slowest of the two oscillators, while electrical bursting in the cell is the faster oscillator. Much of our analysis focuses on the dynamics of this dual-oscillator system. The pancreas contains a large number of islets, and their activity must be synchronized for the insulin release from the islet population to be oscillatory. In this project, two mechanisms for synchronization are investigated. One is the entrainment of islets by peripheral nerves in intrapancreatic ganglia. This will be investigated by applying periodic pulses to the dual-oscillator model and identifying conditions for entrainment and entrainment windows. The other synchronization mechanism is the feedback of insulin onto insulin receptors on the beta-cells. This provides a relatively weak coupling effect on the model islets, but it may be sufficiently strong to achieve synchronization.

Failure of beta-cells to secrete the proper amount of insulin in response to changes in the glucose level in the blood is a major factor for type II diabetes. For this reason, it is important to understand the biological mechanism for insulin secretion, and the coordination of the insulin-secreting cells. The beta-cells are very complex, and our approach to understanding their behavior is to combine mathematical modeling and computer simulations with experimental studies, performed at a collaborating lab. Both undergraduate and graduate students are involved in the mathematical modeling and computer simulations, and will meet with experimental collaborators to discuss data and future experiments motivated by the modeling. Our goal is to understand how proper insulin secretion is achieved at the cellular level, and then extend this to understand how beta-cell disfunctions can lead to type II diabetes.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

The long-term goal of this research project is to understand how the many subsystems of the pancreatic beta-cell interact to produce insulin oscillations in mice (rats and humans also exhibit oscillations). These oscillations are crucial for the normal regulation of blood glucose levels, and loss of oscillations is linked to type II diabetes. Rhythmic insulin secretion from pancreatic islets of Langerhans is due to rhythmic bursting electrical activity in the beta-cells, and a consequent rhythm in the intracellular calcium concentration. Both the intracellular calcium and adenosine triphosphate (ATP) feed back onto the cell's electrical subsystem, opening or closing ion channels and thus affecting the cell's electrical activity. This project focuses on the metabolic subsystem that produces ATP, and on a mathematical analysis of previously-developed models of metabolismdriven bursting and the fast electrical bursting that occurs in single beta-cells that are isolated from the islet. The central hypothesis is that the slow electrical bursting oscillations and episodic bursting that are often exhibited by pancreatic islets, and that have the same period as insulin oscillations observed in vivo, are driven by oscillations in metabolism. One mechanism for these oscillations is in glycolysis, the first stage of glucose metabolism. However, it is possible that oscillations inherent in one of the other two stages of metabolism, the citric acid cycle and oxidative phosphorylation, could be the slow process that drives slow bursting activity and that clusters faster bursts together into periodic episodes. This possibility will be investigated using mathematical modeling and analysis, as will what measurements of periodicity in citric acid cycle intermediates indicate about the mechanism of the oscillations. Returning to the glycolytic component of metabolism, modeling studies will be conducted in parallel with experimental studies in a collaborating laboratory to determine how the enzyme phosphosphofructokinase-2 (PFK-2) may modulate glycolytic oscillations.

Insulin secretion in mammals, including rats, humans, dogs, and humans, is pulsatile, with a period of about five minutes. These insulin oscillations, which can be measured in the blood, are important for normal glucose homeostasis, since disruption of the oscillations is linked to type II diabetes. Insulin is secreted from micro-organs in the pancreas called Islets of Langerhans, composed largely of insulin-secreting beta-cells. For more than a decade now, the principal investigator has been investigating the biophysical mechanism for the oscillations in insulin secretion. This research involves mathematical modeling and analysis, and parallel experimental studies in a collaborating laboratory. It is thus a truly multidisciplinary project. The current project uses a current mathematical model of pancreatic beta-cells to understand how oscillations in the metabolism of glucose by the beta-cells can lead to oscillations in the electrical activity and insulin secretion from the beta-cells. In addition, bifurcation analysis and recent mathematical methods in the area of Mixed Mode Oscillations will be used to understand the oscillatory electrical activity of beta-cells that have been isolated from an islet. The intention is to determine, using this mathematical analysis, how single-cell behavior is converted to the very different behavior of beta-cells in an intact islet. The long-term goal of this research is to better understand the normal functioning of islets, which will ultimately provide insights into the dysfunction of islets that occurs in type II diabetes.

Like humans, juvenile songbirds learn to imitate the vocal patterns of an adult during a sensitive period of development. Also like humans, songbird vocal patterns are controlled by a forebrain network that includes pre-motor, striatal, and auditory pathways for processing vocal sounds and gestures. This proposal seeks to understand how these pathways interact to produce stable vocal patterns. Our working hypothesis states that learned song results from an integration of pre-motor and striatal pathways encoding stereotyped and variable vocal patterns, respectively. This simple hypothesis makes clear predictions about the vocal effects of altering the relative strength of pre-motor and striatal pathways. For example, we have shown that weakening the premotor pathway in adult birds (via partial ablation) produces highly variable singing reminiscent of a learning juvenile. However, adult vocal patterns show surprising resilience as birds subsequently recover stable song within 1 week. This recovery depends on auditory feedback, indicating that adult vocal recovery involves instructive mechanisms similar to those that guide juvenile learning. Recently, we designed an experiment to identify the neural locus of these instructive mechanisms in adult birds and demonstrated that they cannot lie within their long-presumed location - the striatal pathway. Here, we propose experiments to 1.) test our model of pre-motor and striatal pathway function in juvenile birds and 2.) identify the neural loci and mechanisms by which auditory feedback promotes vocal learning (in juveniles) and maintenance of stable song (in adults). These experiments will provide new information about the functional architecture of the songbird vocal control network - this knowledge will offer new insight into the functional architecture of the analogous neural regions that control human vocal learning. 1

DESCRIPTION (provided by applicant): The Central Hypothesis of this renewal application is that under physiological conditions, prolactin (PRL) secretion is inhibited by dopamine (DA) and stimulated by oxytocin (OT) and that the activities of these neurons are regulated by higher hypothalamic input and PRL itself. We will test this hypothesis with a novel interdisciplinary experimental and mathematical approach to the following specific aims: (1) To determine the basis for a physiological interaction between DA, OT and PRL secretion, (2) To determine the physiological basis for the estrous cycle stage differences in the response of lactotropes to the PRL-releasing properties of OT, (3) To characterize the mechanism of action of OT on pituitary lactotropes and (4) To characterize the relationship of DA and other modulators of PRL release on OT-induced responses in lactotropes. These aims are significant because they will challenge accepted dogma that PRL secretion in physiological circumstances is only under inhibitory control of hypothalamic DA. PUBLIC HEALTH RELEVANCE: Current clinical approaches to manipulating prolactin secretion involve manipulation of dopamine. For example, the benign hypersecreting pituitary tumor known as a prolactinoma is treated with the dopamine agonists bromocryptine or cabergoline. Each of these drugs has side effects which have serious implications for patients'welfare. Indeed, as our preliminary data suggest, if prolactin secretion is also under the control of peptidergic releasing factors such as oxytocin, then this may be another approach to treating disorders of prolactin secretion. Moreover, prolactin has been implicated in disorders related to more than 300 of its other biological actions. Results from these studies may serve as an approach to treating these disorders.

This project is a hybrid mathematical/experimental approach to the analysis of the behavior of hormone-secreting pituitary cells. We record a cell's electrical activity and use features of this to calibrate a mathematical model. This calibration is done with the aid of a graphics processing unit (GPU) for fast optimization of a fitness function that compares features of the model voltage trace to those of the experimental recording. Once calibrated, the model is used to make predictions about the effects of changes in biological parameters, such as the conductances of various types of ionic currents. These predictions are then tested on the same cell that was used to calibrate the model, using the dynamic clamp technique to inject a model-based ionic current into a real cell. To help with our understanding of the model, and thus to make more useful predictions, we use geometric singular perturbation methods to understand the basis for spiking and bursting patterns of electrical activity, and the parameter ranges where different types of activity occur.

The ultimate goal of mathematical models for biological systems is to generate hypotheses that can be tested in the lab. If the model is well calibrated, then testing a prediction made by the model is the best way to determine if our understanding of the biology that is reflected in the model is correct. In particular, if a test of a model prediction fails, then it means that something is wrong with our understanding of the biology that was the basis for the model. However, this is only true if the model is well calibrated, since a model that is correct in its formulation, but incorrect in its parameterization, can lead to incorrect predictions. This caveat is particularly important given the great degree of cell-to-cell heterogeneity that exists in many biological systems. For example, a cell may exhibit one type of pattern of activity, while a neighboring cell of the same type may exhibit a very different pattern of activity, reflecting differences in biological parameter values. In this project, we combine fast model calibration with the dynamic clamp technique to calibrate a model based on a single cell's activity, and then test predictions made by the model on the same cell, overcoming problems associated with cellular heterogeneity.

How does the brain create orderly sequences of behavior? The problem of serial order in behavior permeates human language, the ability to play a musical instrument, or any number of other activities where human or animal intent can only be effectively communicated if several behavioral gestures are chained together in a specific sequence. How the brain accomplishes this task is currently unknown. One approach is to use an animal model (a songbird, the zebra finch) in which the brain region responsible for the serial ordering of song syllables has been identified (called "HVC", the acronym is the proper name). This project will investigate the hypothesis that the serial ordering of song syllables is mapped across several spatially-arranged chains of HVC neurons. This hypothesis is based on preliminary data indicating that HVC neural activity is largely confined to a single spatial axis during singing. Experiments will delineate the spatial organization of connections and electrophysiological properties of HVC neurons. Computational models based on the properties of HVC neurons will then be used to discover network configurations that produce orderly, sequential patterns of neural activity. Models will be validated with circuit-breaking experiments in behaving birds. Results will provide a first look at a network architecture used by an animal brain to create order and sequence in behavior, which in turn will provide a computational platform to understand how the process of learning new behavioral sequences utilizes or shapes such architectures. The research plan coordinates the activity of a faculty research team from three different academic departments (Psychology, Mathematics and Statistics), providing graduate and undergraduate students with access to the expertise of faculty researchers outside their home departments. Computational software tools as well as data from this project will be made available to the public at http://www.math.fsu.edu/~bertram/software/birdsong/ and at http://www.songbirdscience.com.

The purpose of this project is to elucidate how serial order is coded by the brain. If you were to read this sentence aloud, you would utter a precise, ordered sequence of over 50 distinct vocal sounds in less than 10 seconds. How does the brain store the individual sounds of speech and then coordinate the production of these sounds in a meaningful order so quickly? A similar question could be asked about the pianist performing a Mozart sonata, the songs of birds, or the mating dances of insects. How the brain stores these elaborate sequences of behavior remains unknown. Using the songbird zebra finch, a model organism that learns meaningful sequences of vocal sounds like humans do, the interdisciplinary research team will test the hypothesis that the brain encodes and stores sequences of behavior through two separate mechanisms that operate in parallel: one coding mechanism for the sequence of vocal sounds and one for the vocal sounds themselves. Given the diversity of animal species that display elaborate, meaningful sequences of behavior, the findings will influence understanding across a broad array of organisms, including humans. The research plan coordinates the activity of a faculty research team from four different disciplines (Neuroanatomy, Neurophysiology, Mathematics, and Statistics) and will provide students with a unique interdisciplinary training opportunity and environment.

Observed in nearly all animal forms (and exemplified by human speech) serial order in behavior consists of learning to organize a set of elemental gestural units into a purposeful sequence of action. Adult male zebra finches (Taeniopygia guttata) produce a highly quantifiable example of serial order in behavior (birdsong). Moreover, a premotor cortical region (HVC, proper name) is known to encode a consolidated premotor trace of song. Although consisting of similar cell types, the medial and lateral portions of HVC are hypothesized to encode the sequence (medial HVC) and syllables (lateral HVC) of song in parallel. The research team will test whether these two dimensions of song are encoded by physiological differences in 1) afferent input to medial and lateral HVC, or 2) the intrinsic network properties of medial and lateral HVC (or a combination of 1 and 2). However, parallel encoding of serial order in behavior should be hierarchical, with traces for sequence in a supervisory position over traces for elemental gestural units. The team will also test whether efferent axons emanating from medial and lateral HVC interact in a hierarchical fashion within vocal-motor cortex. Results will elucidate a network architecture for serial order in behavior and provide a computational platform to understand how learning new sequences shapes such memory architectures.

Type II diabetes makes up about 90% of all cases of diabetes world wide, and 368 million people were diagnosed with type II diabetes in 2013. Although it is a complex disease with many interacting factors, there is agreement that a major component is dysfunction of the pancreatic beta cells that release the hormone insulin. In normal individuals, these cells release insulin in pulses, with the pulse amplitude reflecting the glucose level. There is now conclusive evidence that pulsatile insulin is more effective than constant insulin at stimulating glucose-lowering actions of the liver. In addition, it has been shown that type II diabetics, and their nearest relatives, show disorganized insulin levels, rather than the more regular insulin pulses of non-diabetics. This project seeks to better understand the intracellular pathways that result in pulsatile insulin secretion from beta cells. It also seeks to understand how mice, which also exhibit pulsatile insulin secretion, compensate for genetic knockouts of key proteins in such a way that rhythmic cellular activity and insulin secretion is maintained. Overall, the project will yield insights into the complex biology of pulsatile insulin secretion, and will push forward mathematical analysis techniques that are useful in complex intracellular signaling systems such as those in pancreatic beta cells. Graduate students involved in this research project will receive broad interdisciplinary training in answering mathematical questions driven by experimental data. Publicly available software will also be produced.

The Dual Oscillator Model has been under continuous development by Bertram and associates since it was first published in 2004. This model describes how three signaling pathways, metabolism, intracellular calcium dynamics, and electrical activity combine to produce pulsatile insulin secretion that is modulated by the extracellular glucose level. The complexity of the model makes it difficult to understand the dynamics of the system, but application of the fast/slow analysis technique allows one to take advantage of the separation of time scales of the model variables and formally decompose the system into subsystems that vary at different rates. This powerful multi-scale analysis technique will be applied to improve our understanding of oscillations in activity of the model beta cell that are induced by glucose and other substrates, as well as oscillations that are present in cells from mice in which different key proteins are genetically knocked out. Because there are several slowly changing variables, there are a number of different ways in which the fast/slow analysis technique can be used. These different approaches are explored, and will yield complementary insights. While the focus of the project is on oscillations in beta cell activity, the analysis approach used is applicable to a wide range of dynamic models, so the mathematical developments made in this project can be adapted to other complex biological systems.

Many neuroscientists explain learning as a simple change in the number and/or strength of specialized connections between nerve cells called synapses. This research project tests a new idea, that learning also importantly involves molecular changes in other parts of nerve cells which control the production of electrical activity required for these cells to communicate with each other. Learning is studied in the context of how young songbirds learn to sing, which happens in a similar way to how humans learn to speak, play a musical instrument, or produce any complicated sequence of behavior. Using a combination of behavioral observation and recording together with anatomical, physiological, molecular, computational and statistical techniques, the research team will test the hypothesis that developmental singing changes are accompanied by changes in cell electrical activity that are not located at synapses. They will also identify the mechanistic basis for these electrical changes. This research will reveal important new details about how the brain can store learned information for an entire lifetime. It will also provide student research assistants with broad multidisciplinary training, as well as developing new analytical software and computational models that will benefit the broader neuroscience community (these will be made freely available through a publically-accessible web site). Neuroscience and mathematics videos tagged to specific State-mandated high school learning objectives will also be produced, which will be made available to teachers via a cataloged State portal. Finally, the research team conducts a wide variety of community education programs specifically related to neuroscience, birdsong, and learning. These activities include programs for K-12 schools as well as a scholarly course for senior citizens.

Much is known about the brain areas and circuits that underlie birdsong learning. Consequently, scientists know where to look for learning-induced changes (an area named HVC), but they do not know what neural changes encode the auditory memories of song. The proposed research tests the novel hypothesis that changes in specific, non-synaptic ionic currents in HVC neurons contribute to the encoding of auditory memories. Songbird auditory learning can be readily manipulated by controlling exposure to a tutor song. Preliminary data show that the intrinsic cell body/axonal channel properties of HVC neurons change in an experience-dependent manner during song development. The research team's recent characterization and modeling of the ionic currents that determine the physiology of HVC neurons in the adult finch now allow for the proposed developmental studies. Proposed studies will test hypotheses about the developmental emergence of specific ionic currents as they relate to sensory learning. This will lead to testable hypotheses about the underlying molecular mechanisms responsible for this learning. As part of this project, the team will also correct mistaken views about HVC of female zebra finches. Long thought to lack male-typical connectivity, new data show that female and male HVC have the same cell types and patterns of extrinsic connectivity. Since females show auditory learning of tutor song, but do not sing, analysis of female HVC will provide a critical test of the proposed hypotheses. Together, the proposed experiments will provide clues to fundamental questions about how early sensory information is stored as a stable memory trace that lasts a lifetime.

The precise mechanisms through which astrocytes modulate synaptic transmission are not clear. It is also not clear how dysfunction in these cells, including those caused by drugs of abuse, can lead to brain disorders. The objective of this proposal is to develop a method that will allow the culture of astrocytes within a microfluidic platform to enable real-time measurements of intracellular Ca2+ concentration ([Ca2+]i) via fluorescent probes simultaneously with secretion of neuroactive substances in response to various stimuli. Guided by our strong preliminary data, we will obtain the objective of this proposal by pursuing the following specific aims: 1) monitoring D-serine and glutamate secretion simultaneous with [Ca2+]i imaging within astrocyte cultures, and 2) quantitation of ATP release from astrocyte cultures. In the first aim, we will develop a microfluidic system to monitor, with high time resolution, the release of primary amines from astrocytes simultaneous with intracellular [Ca2+]i imaging. In the second aim, we will integrate an enzymatic assay for adenosine triphosphate to allow monitoring of all the major gliotransmitters from astrocytes. The research proposed in this application is innovative because it will utilize a broad approach to the measurement of multiple factors released from astrocytes combined with simultaneous [Ca2+]i imaging, providing a highly dynamic view of gliotransmission. The results of this system will be significant because it will provide the first tool capable of investigating release of all major gliotransmitters simultaneously in an automated fashion in response to a number of stimulants. Ultimately, this device will produce a novel tool for future studies regarding astrocyte biology and the roles these cells have in neuroprotective behavior, tripartite synaptic transmission, and the effects of drugs of abuse.