Gliotransmission at central glutamatergic synapses: d-serine on stage

Abstract

Long ignored and only considered as housekeeping cells for neurons, astroglial cells in the last decade have gained increasing attention as key players of higher functions in healthy brain, but also in diseases. This revolution in our way to think the active brain culminates in the concept of a tripartite synapse, which considers glial cells and notably astrocytes as an integral dynamic partner of synapses. Glia not only listens but also talks to neurons through the release of neuroactive substances. Recently much attention has been paid to the role played by the atypical amino acid d-serine in this signalling pathway. This molecule synthesized through racemization of l-serine fulfils most criteria as a gliotransmitter and as the endogenous ligand for the strychnine-insensitive glycine binding site of the NMDA receptors. d-Serine is considered to be a permissive factor for long-term changes in synaptic plasticity and neuronal migration through activation of NMDA receptors. It is also known that disturbance of NMDA receptors activity can cause cell death. Not surprisingly, then, d-serine has also been found to promote neurons death in experimental models of β-amyloid peptide-induced neuroinflammation and of ischaemia by overactivating the NMDA receptors. Finally, in a more recent past, studies have pointed to the molecular mechanisms leading to d-serine release into and removal from the synaptic cleft.

Introduction

Glutamate, the main excitatory neurotransmitter in the mammalian CNS, mediates slow and fast neurotransmission through three classes of ionotropic receptors (NMDA, AMPA and kainate) (Dingledine et al., 1999). The N-methyl d-aspartate (NMDA) subtype of glutamate receptors is member of a class of ionotropic receptor channels, organized as heteromeric assemblies composed of a spliced variant of NR1 subunit, combined with at least one of four NR2 (A–D) subunits (Dingledine et al., 1999, Kemp and McKernan, 2002). A third subunit, NR3, which is less common, can co-assemble with NR1/NR2 complexes. Based on the subunit composition, the pharmacology and the distribution of NMDA receptors (NMDARs) vary. Activation of NMDAR is complex. The ion-channel integral to the NMDAR is voltage-dependently blocked by magnesium. Depolarization removes this block, allowing the permeation through NMDARs of Ca²⁺ ions inside the depolarized cells (Dingledine et al., 1999, Kemp and McKernan, 2002). Thus the receptor acts as a coincidence detector, linking glutamate activation with the electrical state of the neuron. Accordingly, the magnitude or duration of NMDA-receptor-mediated Ca²⁺ influx has been shown to dictate the type and sign of plasticity induced (Perez-Otano and Ehlers, 2005). A useful simplified framework has been that small amounts of NMDA-receptor-mediated Ca²⁺ influx produce Long-Term Depression (LTD) whereas strong activation of NMDARs leads to Long-Term Potentiation (LTP). Furthermore, recent studies have shown that NMDARs serve a homeostatic role ensuring that plasticity is kept within a working range (away from saturation) and control the threshold shift for LTP and LTD induction (Perez-Otano and Ehlers, 2005). Thus, the activity of NMDARs is critical for normal brain function and NMDARs knock-down mice display severe brain impairment with defect in neuronal migration and in synapses stability (Lee et al., 2005, Mohn et al., 1999). On contrast, chronical and excessive overstimulation of NMDARs causes excitotoxicity (Shleper et al., 2005, Takano et al., 2001) driving the search for NMDA antagonists as neuroprotective agents. Too much NMDAR activity is harmful as too little. Thus the elucidation of the factors maintaining a balanced activity for these receptors is important.

Furthermore, in addition to glutamate, activation of NMDAR requires the binding of a coagonist on the NR1 subunit. Since the late 1980s, it has been assumed that glycine is the coagonist (Johnson and Ascher, 1987). Nevertheless, a number of reports in the last decade have led to the demonstration that d-serine is the major neuromodulator for the NMDAR at the glycine site (Miller, 2004).