Neuron, Vol. 12, 443-455, February, 1994, Copyright 0 1994 by Cell Press Synaptic Regulation of Glial Expression In Vivo John Georgiou,* Richard Robitaille,*+ William S. Trimble,’ and Milton P. Charlton* *Network of Centres of Excellence on Neural Regeneration and Functional Recovery and Medical Research Council of Canada Group in Nerve Cells and Synapses Department of Physiology Medical Sciences Building 3232 ‘Centre for Research in Neurodegenerative Diseases and Department of Physiology Tanz Neuroscience Building University of Toronto Toronto, Ontario Canada, M5S IA8 Summary We investigated signaling between individual nerve ter- minals and perisynaptic Schwann cells, the teloglial cells that cover neuromuscular junctions. When deprived of neuronal activity in vivo, either by motor nerve transec- tion or tetrodotoxin injection, perisynaptic Schwann cells rapidly upregulated glial fibrillary acidic protein. Addition of transcription or translation inhibitors to ex- cised muscles prevented this increase. Stimulation of cut nerves prevented glial fibrillary acidic protein in- creases even when postsynaptic nicotinic receptors were blocked, but not when neurotransmitter release was blocked with oconotoxin GVIA. We conclude that there is a nerve terminal to glial signal, requiring presyn- aptic neurotransmitter release, which regulates perisy- naptic Schwann cell genes. This may be a general princi- ple since many types of glia are sensitive to transmitters applied in vitro or released in situ. Introduction There is a growing body of evidence that neurotrans- mitters can regulate genes of target cells. Although this is established for neuron-neuron contacts (for review see Armstrong and Montminy, 1993) and for neuron-muscle contacts (Witzemann et al., 1991), the principle has not been demonstrated for contacts be- tween neurons and glia in vivo. Many aspects of glial cells make them likely targets for gene regulation by neurotransmitters. For in- stance, glial cells are closely associated with synapses (Kuffler and Nicholls, 1966; SpaEek, 1971; Orkand, 1982; Pomeroy and Purves, 1988) and have receptors that respond to transmitters (MacVicar et al., 1989; Cornell-Bell et al., 1990; Barres, 1991; Cornell-Bell and tpresent address: Departement de Physiologie, Universitk de Montreal, P. 0. Box 6128, Station A, Montreal, Qukbec, Canada, H3C 3J7. Protein Finkbeiner, 1991). Moreover, nerve stimulation has been shown to trigger Caz+ waves in astrocytes of cul- tured rat hippocampal slices (Dani et al., 1992), and 2+ such changes in Ca have been associated with al- tered gene activity in culture (for review see Fink- beiner, 1993). The ability to alter gene activity may be significant because the physiology and phenotype of glia change with injury. Various insults and diseases induce glial reactions (for review see Eng and DeArmond, 1982; Malhotra et al., 1990) including increased immunore- activity for a cytoskeletal protein, gliall fibrillary acidic protein (GFAP; Eng et al., 1971). For instance, stab wounds and chemical injury induce astrocytes to pro- duce GFAP (Bignami and Dahl, 1976; Brock and O’Cal- laghan, 1987), and ischemic injury increases levels of mRNA encoding GFAP (Kindy et al., 1992). Within 3 hr after blockade of auditory nerve function by tetro- dotoxin (T-TX) injection, there is an increase in GFAP- immunoreactive astrocytic processes in the ipsilateral cochlear nucleus (Canady and Rubel, 1992); 1 week later, after recovery of nerve activity, GFAP immunore- activity is not different from the contralateral controls. Some glial cells in the periphery share characteris- tics with those in the CNS. Perisynaptic Schwann cells (PSCs) are non-myelin-forming cells that cover the en- tire neuromuscular junction (NMJ), sending out fine processes around the nerve terminal to within 50 nm of transmitter release sites (Dreyer and Peper, 1974; Heuser and Reese, 1973, 1977). These teloglial cells respond to nerve-evoked synaptic activity with an in- crease in intracellular Cap in situ (Jahromi et al., 1992; Reist and Smith, 1992). In these cells, nerve stimula- tion can be mimicked by local application of the neurotransmitter acetylcholine (ACh) or ATP, which induce increases in Ca2+ due to release from intracel- lular stores (Jahromi et al., 1992). As with CNS glia, the physiology and morphology of PSCs also change dramatically with denervation. Eighteen hours after nerve transection, PSCs become immunoreactive for CAP-43 (Woolf et al., 1992), and within days of denervation, PSCs begin to synthesize and release ACh (Bevan et al., 1973). After denervation, PSCs elaborate processes that escape the confines of the former NMJ, but these processes retract upon reinnervation (Reynolds and Woolf, 1992). Thus, PSCs share many attributes with CNS astro- cytes in regard to specialization at synapses and re- actions following injury. Considering that glia are sensitive to neurotransmitter release and react to injurywith alterations in protein expression, we inves- tigated the possibility that synaptic activity may regulate the physiological status of gllial cells. This hypothesis was tested at PSCs because they offer dis- tinct advantages for this type of study; simplicity of the connections and morphology allows examination