Kainate receptor (GluR5)-mediated disinhibition of responses in rat ventrobasal thalamus allows a novel sensory processing mechanism

Abstract

Kainate receptors have been studied extensively in vitro, but how they might function physiologically remains unclear. We studied kainate receptor modulation of synaptic responses in the rat ventrobasal thalamus using the novel antagonist LY382884 and the agonist ATPA (selective for GluR5-containing kainate receptors) as tools. No evidence could be found for a direct contribution of kainate receptors to responses of thalamic relay cells to lemniscal (sensory) input in thalamic slices studied with the aid of intracellular and field potential recordings, using selective AMPA and NMDA receptor antagonists and LY382884. However, the GluR5 agonist ATPA reduced the IPSPs originating from the thalamic reticular nucleus. Extracellular single-neurone recordings in anaesthetised rats showed that excitatory responses evoked by physiological vibrissa afferent stimulation were reduced by LY382884 applied iontophoretically at the recording site. This action of the antagonist was occluded when GABA receptors were blocked, indicating that the reduction in excitatory sensory responses by LY382884 is due to an action on GABAergic inhibition arising from the thalamic reticular nucleus. Further experiments showed that these actions depended on whether inhibition was evoked during activation of the excitatory receptive field rather than when inhibition was evoked from a surround vibrissa. We suggest that GluR5 is located presynaptically on inhibitory GABAergic terminals of thalamic reticular nucleus neurones, and that it is normally activated by glutamate spillover from synapses between excitatory afferents and relay neurones during physiological stimulation. We propose that this GluR5-activated disinhibition has an important novel role in extracting sensory information from background noise.

Figure 5. The effect of LY382884 on responses to principal and secondary whisker stimulation

A1, the route of activation of inhibitory neurones during principal whisker stimulation, as indicated by the arrows from the afferent onto the relay cell (RC) and then onto the TRN cell. A2, the route during secondary whisker stimulation. B1, histograms from a VB neurone where stimulation of the principal (B4) vibrissa elicited an excitatory response. The cumulative response to 12 air jets directed at the B4 whisker is shown during the control period (left), after ejecting LY382884 for 10 min (centre), and 10 min after the cessation of antagonist ejection (right). LY382884 reduced the response to stimulation of whisker B4. B2, histograms from the same neurone illustrating the inhibition produced by stimulating the B2 whisker when the baseline activity of the neurone was elevated by ejection of carbachol. The cumulative activity during 10 stimulation cycles with an air jet directed at the B2 whisker is shown during the control period (left), after ejecting LY382884 for 10 min (centre), and 10 min after the cessation of antagonist ejection (right). LY382884 did not increase the degree of inhibition in these circumstances. The inset shows the relative positions of whiskers B4 and B2, stimulation of which produced excitation and inhibition, respectively.

Figure 1

A, responses of a thalamic relay neurone to stimulation of the medial lemniscus (100 μA). A1 shows five superimposed EPSPs (each evoking an action potential) in response to 0.1 Hz stimulation either under control conditions or following perfusion of AMPA and NMDA antagonists (GYKI52466 plus MK-801 plus AP5). The antagonists completely blocked the EPSP, as can also be seen in the overlay of averaged responses (n = 5). A2 shows the same neurone in the continued presence of the AMPA and NMDA antagonists, with 1 s trains of 50 Hz medial lemniscus stimulation (five superimposed responses) in an attempt to reveal additional synaptic response components. No further responses were revealed by this stimulation protocol. B, IPSPs recorded in response to stimulation of the TRN in the presence of GYKI52466, AP5 and MK-801. In the upper row, each trace is an average of responses to 10 repeated stimuli applied to the TRN (presented at the time indicated by the arrowhead). Application of ATPA to the bathing medium reduced the IPSP amplitude after 5 min, and this effect was reversed when LY382884 was then added for 10 min. The right panel shows superimposed traces from the panels on the left. The lower row of traces shows averaged responses to a −0.1 nA current pulse injected through the recording electrode.

Figure 2. Field EPSPs recorded in response to stimulation of the medial lemniscus (five at 20 Hz)

Each trace (A–C) is an average of 10 responses to stimulus bursts presented at a rate of 0.1 Hz. All data were obtained in the presence of GABA_A and GABA_B receptor antagonists (100 μm picrotoxin and 3 μm CGP55845). A, responses to the five stimuli under control conditions (with GABA antagonists), 15 min after addition of LY382884 (10 μm) to the bath, and after 15 min washout (Wash) of this antagonist. B, enlarged traces showing field EPSPs in response to the first (EPSP1) and second (EPSP2) stimuli of each train shown in A. C, overlay of EPSP1 and EPSP2 responses obtained under control conditions and during LY382884 application. D, averaged data (± s.e.m.) from five experiments. Field EPSPs have been normalised as a percentage of EPSP1 under control conditions. LY382884 had no significant effect on EPSPs or on the ratio of EPSP2 to EPSP1 or EPSP5 to EPSP1.

Figure 3. Effect of LY382884 on sensory and agonist responses of a VB neurone

PSTHs of action potential spikes counted into epochs of either 200 ms (A-C) or 1000 ms (D), recorded from a VB neurone in vivo. A–C, cumulative histograms of the responses to five air-jet stimuli to the principal vibrissa (B4) before, during and after the iontophoretic application of LY382884. Note the reduction of vibrissal responses by the antagonist. Arrows indicate the time points in D from where these histograms were computed. D, continuous record from the same neurone in response to regular sensory stimulation of the principal vibrissa with groups of five air jets and iontophoretic ejections of NMDA, ATPA and fluorowillardiine (Fluoro), as indicated by the marker bars beneath the record. Continuous ejection of LY382884 (marker bar) resulted in a reduction of the responses to vibrissal stimulation, with little effect on responses to iontophoretically applied agonists.

Figure 4. Reduction of LY382884 effect by a GABA antagonist

A series of PSTHs showing the cumulative response of a VB neurone when air jets were presented to its principal whisker (B3). Five stimuli were presented with a 10 s interstimulus interval. Action potential spikes were collected in 200 ms bins. The PSTHs show: a, the control response; b, the reduction of the response when LY382884 was ejected; c, the return of the response to control levels after cessation of antagonist ejection; d, a control response; e and f, the increase in the response when 20 nA SR95531 was applied; g and h, the effects of LY382884 during the ejection of SR95531; and i, the final response after cessation of LY382884 ejection but in the presence of SR95531. For this neurone, the responses to NMDA and ATPA were, respectively, 92 and 91 % of the control level during the ejection of LY382884. The records were taken at 5 min intervals.

Figure 6. Average data from eight neurones where the effects of LY382884 on responses to principal vibrissa stimulation and secondary vibrissa stimulation were investigated

A, effects of the antagonist on excitatory responses to principal vibrissa stimulation and to iontophoretically applied NMDA and carbachol. Values are expressed as percentages (± s.e.m.) of the value prior to antagonist ejection. B, effects of the antagonist on inhibitory responses to secondary vibrissa stimulation. Values are percentages of the carbachol response (± s.e.m.) before and during LY382884 ejection (see text for details).