New roles for astrocytes: Regulation of CNS synaptogenesis Michal Slezak and Frank W. Pfrieger Max-Planck/CNRS Group, UPR 2356, Centre de Neurochimie, 5 rue Blaise Pascal, F-67084 Strasbourg, France The notion that astrocytes have a profound influence on the function of synapses between CNS neurons implies that the development of synaptic connections and their glial neighbors are controlled by reciprocally acting signals. Currently, however, synaptogenesis is considered a purely neuronal affair. This article sum- marizes recent experimental evidence suggesting that this may not be the case. Astrocytes may indeed regu- late the formation, maturation and maintenance of synapses. The recent advances caution that synapses cannot develop correctly without astrocytes. Further progress on this issue requires new experimental models to identify signaling pathways and to scrutinize the relevance of glia – synapse interactions in vivo. During the past decade, an ageing star has reappeared on the CNS stage, playing a new role in an old piece that before featured only two actors. That star is the astrocyte and the play is the function of synaptic connections between two partner neurons [1–5]. This new triad of actors has prompted the concept of the ‘tripartite synapse’, which states that ‘synaptically associated astrocytes should be viewed as integral modulatory elements of synapses’ [6]. However, recent reviews on the development of synapses [7–13] reveal that this fundamental process is still considered an entirely neuronal event. This article summarizes evidence that this view could be outdated and that astrocytes might take part in this process as well. As is true for all liaisons, the life of a synaptic connection between two partner neurons can be divided into three phases: (i) establishment of a first – physical – contact; (ii) a maturation process, during which each connection acquires its characteristic properties; and (iii) a stabiliz- ation or elimination phase, during which only robust matches are maintained. In this review, evidence for an astroglial contribution to each of these phases is evaluated (Figure 1). Owing to space limitations, the article focuses mainly on connections in the CNS. However, it should be noted that glial cells also appear to influence synaptogen- esis in the peripheral nervous system (PNS) [14–19]). Astrocytes control synapse number The idea that astrocytes play a role in the formation of synaptic contacts arises from a conspicuous temporal correlation between synaptogenesis and the differen- tiation of this glial cell type [20] (Figure 2). In rodents, for example, astrocytes are generated around birth, whereas massive synaptogenesis starts at the end of the first postnatal week and continues for two-to-three weeks [20,21]. The long duration of synaptogenesis contrasts with the notion that individual synaptic contacts are formed in less than an hour [22,23]. Evidently, it may simply take weeks to form the incredibly high number of Figure 1. Different stages of synapse development and possible contributions by astrocytes. (a) At the embryonic stage, before astrocytes (green) are generated, neuronal processes [presynaptic (yellow) and postsynaptic (white)] form few and immature synapses. (b) The massive increase in synapse number during the post- natal stage might be enabled by astrocyte-derived components (green text). (c) Excess synapses might be eliminated by astrocytic processes invading the synaptic cleft or by release of proteases that digest synapse-stabilizing com- ponents. Astrocyte-derived signals can stabilize surviving synapses and induce presynaptic and postsynaptic maturation processes, including accumulation of vesicles and modifications in the molecular complexes that mediate transmitter release (colored ovals) and reception (thick lines). Question marks indicate that as-yet unidentified factors might also be involved. Abbreviations: ADNF, activity- dependent neurotrophic factor; ECM, extracellular matrix; MARIA, muscarinic acetylcholine receptor-inducing activity; TNF-a, tumor necrosis factor a; VIP, vasoactive intestinal polypeptide. TRENDS in Neurosciences Synapse formation Cholesterol via lipoproteins Growth factors ECM components ? Presynaptic and postsynaptic maturation MARIA VIP-induced release of ADNF TNF-α ? (a) (b) (c) Stabilization or elimination Proteases Protease inhibitors ? Corresponding author: Frank W. Pfrieger (fw-pfrieger@gmx.de). Review TRENDS in Neurosciences Vol.26 No.10 October 2003 531 http://tins.trends.com 0166-2236/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.tins.2003.08.005