A role of TARPs in the expression and plasticity of calcium-permeable AMPARs: evidence from cerebellar neurons and glia

Abstract

The inclusion of GluA2 subunits has a profound impact on the channel properties of AMPA receptors (AMPARs), in particular rendering them impermeable to calcium. While GluA2-containing AMPARs are the most abundant in the central nervous system, GluA2-lacking calcium-permeable AMPARs are also expressed in wide variety of neurons and glia. Accumulating evidence suggests that the dynamic control of the GluA2 content of AMPARs plays a critical role in development, synaptic plasticity, and diverse neurological conditions ranging from ischemia-induced brain damage to drug addiction. It is thus important to understand the molecular mechanisms involved in regulating the balance of AMPAR subtypes, particularly the role of their co-assembled auxiliary subunits. The discovery of transmembrane AMPAR regulatory proteins (TARPs), initially within the cerebellum, has transformed the field of AMPAR research. It is now clear that these auxiliary subunits play a key role in multiple aspects of AMPAR trafficking and function in the brain. Yet, their precise role in AMPAR subtype-specific regulation has only recently received particular attention. Here we review recent findings on the differential regulation of calcium-permeable (CP-) and -impermeable (CI-) AMPARs in cerebellar neurons and glial cells, and discuss the critical involvement of TARPs in this process. This article is part of the Special Issue entitled 'Glutamate Receptor-Dependent Synaptic Plasticity'.

Keywords: AMPA receptors; Calcium-permeable AMPA receptors; Cerebellum; Glutamate receptors; Plasticity; Synaptic transmission; TARPs.

Fig. 1

Main cerebellar neurons and glial cells: location, connectivity, glutamate receptor and TARP content. Subunits indicated in bold form a majority of the AMPARs, NMDARs, and KARs involved in fast synaptic transmission. Other subunits listed are also expressed, but evidence of their contribution to synaptic currents is unclear. Note that extrasynaptic receptors can be activated by glutamate spillover, and that mGluRs are usually located perisynaptically.

Fig. 2

Developmental and activity-dependent changes in EPSC rectification reflect increased relative expression of CI-AMPARs at parallel fibre-to-stellate cell synapses. (A) Schematic diagram depicting how replacement of GluA2-lacking CP-AMPARs by GluA2-containing CI-AMPARs results in a decreased proportion of AMPARs blocked by intracellular spermine at positive potentials (strong versus weak rectification). Note that the single-channel conductance of CI-AMPARs is lower than that of CP-AMPARs. (B) Decreased AMPAR-mediated EPSC rectification during development. Recording made in slices from the cerebellum of 8, 18 and 28 day-old rats. Top panels, EPSCs evoked in stellate cells at membrane potentials ranging from −60 mV to +40 mV. Bottom panels, corresponding peak I–V relationships (RI; Rectification Index, calculated as conductance ratio: +40/−40 mV). Note in P8 stellate cells the strong current block at positive potentials, characteristic of CP-AMPAR expression. (C) High-frequency stimulation of parallel fibres is followed by a decreased EPSC rectification. Top panels from left to right: responses to a train of 100 stimuli at 50 Hz, averaged parallel fibre-evoked EPSCs at −60 and +40 mV in the same cell before and after high frequency stimulation, and corresponding I–V relationships. In the left-hand panels, note the fast AMPAR-mediated synaptic currents (and associated stimulation artefacts) together with mGluR-dependent slow currents. The switch from CP- to CI-AMPARs causes a decrease in the amplitude of EPSCs at negative potential (reflecting reduced single-channel conductance) and a reduction in rectification (reflecting increased spermine block at positive potentials). Bottom panels, same as top panels but in the presence of two mGluR antagonists. Note that the antagonists prevent both the slow current and the change in RI that follows the stimulation. (B and C, modified from Soto et al. (2007) and Kelly et al. (2009), respectively).