Development and Function of Astrocytes
Shortly after they are born, mammalian astrocytes invade synapse-rich regions of the brain and become intimately associated with nearly all synapses, the fundamental unit of neuron-neuron signaling. Why are astrocytes so closely coupled with synapses? Recent work points to important roles for astrocytes in synapse formation, maturation, and efficacy, and it is likely that glia are also active participants in synaptic signaling and plasticity. We wish to understand the molecular bases of these interactions and the precise roles of glia in central nervous system (CNS) information processing.
Are there astrocytes in Drosophila? Using a variety of molecular-genetic approaches, we recently identified a cell type in the Drosophila brain that bears a striking resemblance to mammalian astrocytes at the morphological and molecular levels. For example, the fly astrocyte takes on a tortuous morphology, extending profuse membrane specializations throughout the brain that associate closely with CNS synapses (Figure 1). These cells, like mammalian astrocytes, express transporters essential for the clearance of major neurotransmitters from the synaptic cleft (e.g., glutamate and GABA), arguing for an important role for fly astrocytes in modulating neurotransmitter tone in the brain. Using our collection of newly developed tools we have embarked on the first molecular-genetic analysis of astrocyte development and function in an organism amenable to rapid genetic analysis. We are exploring a number of questions: How do astrocytes take on their tortuous morphologies? Do astrocytes occupy unique spatial domains in the CNS? Is astrocyte morphology or synapse association responsive to neural activity? Are astrocytes actively remodeled during development? How do astrocytes modulate synaptic signaling?
Neuron-Glia Signaling Sculpts Axonal and Synaptic Connectivity
It is widely believed that glial cells play an important role in wiring the nervous system, but compelling evidence supporting a requirement for glia in the establishment of neural connectivity has remained scarce. We are exploring the mechanisms by which glia sculpt neural connectivity during development in two contexts: the adult Drosophila antennal lobe, and the larval neuromuscular junction (NMJ). Through glial-specific knockdown of a number of signaling molecules we have found that developing adult brain glia are critical players in the wiring of the adult olfactory system (Figure 2). This observation implicates glia in establishment of the precise spatial map of olfactory receptor neuron (ORN) axon connectivity in the antennal lobe. We are now seeking to determine how neuron-glia interactions sculpt axonal connections in this tissue.
To explore how glia sculpt synaptic fields, we turned to the Drosophila larval NMJ (in collaboration with Vivian Budnik, University of Massachusetts Medical School). During larval life, the NMJ increases ~100-fold in size, and thereby acts as a developmental model for synaptic plasticity. We find that suppressing glial phagocytic activity leads to the accumulation of massive amounts of presynaptically derived debris at the NMJ (Figure 3) and a dramatic decrease in the formation of new boutons. We believe that normal synapse expansion entails large-scale shedding of presynaptic membranes, that glia transiently invade the NMJ to engulf this material, and that constitutive clearance presynaptic material is essential for normal synaptic expansion. We hope to use the Drosophila NMJ to probe additional fundamental questions in glial biology by (1) developing it as a model for a "tripartite synapse" where the effects of glial cells on synaptic physiology can be determined, (2) exploring in live preparations the effects of neural activity on glial motility and Ca2+ signaling, and (3) determining additional ways in which glia can modulate synaptic connectivity.
Molecular Pathways Mediating Glial Immune Function
Glia have the ability to sense any neural injury (e.g., spinal cord injury, ischemia, or neurodegenerative disease) and respond by undergoing reactive gliosis, a process whereby glia exhibit dramatic changes in morphology and gene expression patterns, migrate to the injury site, and manage brain responses to trauma. Reactive gliosis has been a major topic of study in the field of glial cell biology for more than a decade, but molecular pathways mediating neuron-glia signaling after injury have remained largely undefined.
We recently developed a simple assay for acute nerve injury in Drosophila and showed that severed Drosophila axons degenerate after a defined latent phase and elicit potent morphological and molecular responses FROM glia: within hours after injury, glia up-regulated the expression of Draper (the Drosophila ortholog of theCaenorhabditis elegans cell corpse engulfment receptor CED-1; Figure 3 ), extended membranes toward severed axons, and phagocytosed degenerating axonal debris. In draper mutants, glia failed to respond to axon injury or clear axonal debris from the CNS. These observations demonstrated that Drosophila glia can respond to brain trauma and phagocytose degenerating axonal debris, and identify Draper as a central mediator of these events. In addition, based on the requirement for the cell corpse engulfment receptor Draper for clearance, this work suggests for the first time that cell corpses and degenerating axons may present similar engulfment cues to local phagocytes. We are now defining the Draper signaling pathway and identifying additional signaling molecules that mediate the neuron→glia signaling events after axon injury.
Molecular Mechanisms of Axon Autodestruction
We are interested in understanding how axons undergo autodestruction and autonomously tag themselves for engulfment by glia. For more than a century, severed axons were thought to waste away passively due to a lack of nutrients from the cell body. However, the identification of the slow Wallerian degeneration (Wlds) molecule, which can suppress axon degeneration in mice for weeks after axotomy, forced us to reconsider this notion. It is now thought that Wallerian degeneration may be an active program of axon autodestruction, akin to apoptotic death, that is somehow suppressed by Wlds. We recently discovered that severed Drosophilaundergo Wallerian degeneration that can be strongly suppressed by mouse Wlds (Figure 4); thus the molecular mechanisms driving Wallerian degeneration and Wlds function are conserved in Drosophila and mammals.
We also found that severed Wlds-expressing axons do not elicit any response from glia, indicating that axonal production of the neuron→glia injury signal that activates glial phagocytosis is genetically downstream of Wlds. This work opens the door to a genetic analysis of axon autodestruction pathways. We are now immersed in trying to understand (1) the signaling pathways that actively drive axon autodestruction, (2) how Wldsprotects severed axons from autodestruction, and (3) the molecular identity of axonal injury signals.
