5. Physiological Studies of Transmitter Function in the Thalamus

It has been known for many years that L-glutamate and L-aspartate can excite thalamic relay neurones (Curtis & Davis 1962; Haldeman & McLennan 1972). More recently, the excitatory action of L-homocysteate has been characterised in the LGN (Jones & Sillito, 1992) and VB thalamus (Eaton & Salt, 1992). In contrast, the dipeptide NAAG does not appear to have direct excitatory or facilitatory actions in these brain areas (Henderson & Salt, 1988; Jones & Sillito, 1992): this would appear to rule out a transmitter action of NAAG via excitatory amino acid receptors in the thalamus, although it is of course possible that this dipeptide exerts a physiological effect by some other means when it is released.

5.1 Input to Relay Neurones

The use of selective NMDA receptor and non-NMDA ionotropic receptor antagonists has allowed substantial advances in knowledge. In vivo work in the VB thalamus and LGN has shown that physiological activation of either low-threshold somatosensory (Salt, 1986, 1987; Salt & Eaton, 1989) or visual afferents (Funke et al., 1991; Jones & Sillito 1992; Sillito et al., 1990a,b), respectively, evokes responses which are sensitive to non-NMDA receptor antagonists (e.g. CNQX) and/or NMDA receptor antagonists (eg AP5 or CPP). Intracellular recording experiments, both in vivo and in vitro, together with selective antagonists, indicate that the postsynaptic response of thalamic relay neurones comprises an initial non-NMDA-receptor-mediated component followed by a NMDA-receptor mediated component (Crunelli et al., 1987; Esguerra et al., 1992; Hu et al., 1994; Leresche 1992; Salt & Eaton, 1991a; Scharfman et al., 1990; Turner et al., 1994). This compound EPSP is typically followed and curtailed by a GABAergic IPSP. It seems to be the case that the NMDA-receptor mediated component becomes more apparent upon repetitive activation of afferents (Salt, 1986; Turner et al., 1994). Several mechanisms could account for this: firstly, the voltage dependent responses of NMDA receptor channels would be suitable to allow the required temporal integration of synaptic inputs; secondly, it is possible that repetitive afferent input recruits activity of the cerebral cortex, which in turn projects back to the relay neurones, possibly via an NMDA receptor (see below); thirdly, it may be that the GABAergic IPSPs are depressed following repetitive activation (a well-known phenomenon), thus allowing the NMDA-receptor EPSP component to become more apparent.

The use of the quinoxalinedione antagonists CNQX, DNQX and NBQX provides powerful evidence for the involvement of AMPA/kainate receptors in thalamic neurotransmission. The data obtained with these compounds must be treated cautiously however, unless careful controls have been carried out. Such controls are especially important when using antagonists like CNQX and DNQX, since they can reduce NMDA receptor-mediated responses via an action the glycine site associated with the NMDA receptor (section 3.2) (Birch et al., 1988; Long et al., 1990). The specificity of the antagonist for non-NMDA receptors versus NMDA receptors will then depend on the concentration of glycine in the extracellular space and also on the NR2 subunit types which may be present. Nevertheless, the involvement of AMPA/kainate receptors in thalamic sensory input seems to be beyond doubt, but it has not been possible to discriminate between the various GluR and KA subtypes using the quinoxalinediones. An initial advance in this direction is the finding that Evans Blue can reduce miniature EPSCs (which are known to be part-mediated by AMPA/Kainate receptors (Pfrieger et al., 1992)) at low concentrations in cultures of rat thalamic neurones (Lessmann et al., 1992): this indicates that the response may be mediated via GluR1 or heteromeric combinations of GluR1 and/or GluR2 or GluR3, rather than GluR6 or GluR3 alone (Keller et al, 1993). Clearly, more work is required to determine which subunits are involved in the various glutamatergic synapses in the thalamus.

The transmitter function of the cortico-thalamic pathway has been less extensively studied than the direct sensory input to the thalamic relay neurones, even though this is a substantial projection on anatomical grounds. This is largely due to technical reasons. Firstly, it has been difficult to activate this pathway selectively in in vivo studies, as electrical stimulation of the cortex or fibre tracts is also likely to activate thalamo-cortical axons antidromically, and the projections of these axons and the cortico-thalamic axons to the nrt and onto intrinsic Golgi II inhibitory neurones typically results in a massive inhibitory input to relay neurones which may obscure any excitatory input. Secondly, in vitro studies have been hampered because of similar difficulties in selective stimulation, although the removal of certain parts of the circuitry during slice preparation may be an advantage in some cases. Thirdly, it is unclear as to what pattern of stimulation may be required in order to activate cortico-thalamic fibres in a physiologically-relevant manner, and it is almost certain that the approaches taken to date have not been optimum in this respect. Nevertheless, a body of evidence has been accumulated which indicates that stimulation of the cortico-thalamic pathway can activate excitatory amino acid receptors. In particular, NMDA antagonists applied by various means, both in vivo and in vitro, can reduce EPSPs in thalamic relay neurones evoked by stimulation of presumed cortico-thalamic afferents, thus providing evidence in favour of NMDA receptor participation (Deschênes & Hu, 1990; Eaton & Salt, 1995; Scharfman et al., 1990). Furthermore, repetitive stimulation of the cortical input to the cat VL thalamus produces a potentiating EPSP (Deschênes & Hu, 1990) which is reminiscent of an NMDA-receptor-mediated EPSP which has been described in the cortex (Thomson, 1986). It is noteworthy, however, that in such studies there are also residual EPSP components which are not reduced by NMDA antagonists and which may therefore be mediated by AMPA/kainate or other receptors.

In addition to ionotropic excitatory amino acid receptors, there is also evidence suggesting that postsynaptic mGluRs may participate in synaptic responses. The first indication that this may be the case came from the finding the selective mGluR agonist trans-ACPD can excite thalamic neurones (Hall et al., 1979; Salt & Eaton, 1991b), and that responses of rat thalamic neurones to nociceptive stimuli can be antagonised by a variety of novel mGluR antagonists which share the property of being Group I mGluR antagonists (Eaton et al., 1992, 1993; Salt & Eaton, 1994). It has also been shown that trans-ACPD can depolarise thalamic neurones in vitro via a reduction in a K+ conductance: this bears a resemblance to a slow EPSP which is thought to be of cortico-thalamic origin, thus suggesting that cortico-thalamic input may also activate mGluRs (McCormick & Von Krosigk, 1992). The postsynaptic location of mGluR1 and mGluR5 in the thalamus (Godwin et al., 1995; Martin et al., 1992; Romano et al., 1995) makes these receptors likely candidates for the mediation of these responses.

5.2 Input onto GABAergic Inhibitory Neurones

The Golgi type II GABAergic neurones which are intrinsic to thalamic relay nuclei (apart from rodent VB and VL nuclei) also receive their inputs from sensory afferents and the cerebral cortex, and thus it is likely that these inputs are glutamatergic. There is however no data available concerning which receptors may be involved in these pathways. It would be of great interest and importance to make recordings from identified Golgi type II neurones and to then characterise the receptors which might mediate their synaptic input, especially if the synaptic pharmacology were to be different from that which impinges upon relay neurones. It is however interesting that immunostaining for mGluR1a and mGluR5 has been seen on GABAergic dendrites within the LGN (Godwin et al., 1995).

The synaptic pharmacology of the GABAergic neurones of the nrt has been studied in vitro (De Curtis et al., 1989), and stimulation of the internal capsule evokes a dual-component EPSP which may be of cortico-thalamic origin, although it is not possible to exclude the possibility that at least some of this response may be due to collaterals of thalamo-cortical fibres. The late phase of this EPSP has the characteristics associated with NMDA-receptor mediated events (voltage dependence and sensitivity to Mg++ and NMDA antagonists), whereas the early phase of the EPSP does not share these properties and may thus be mediated by non-NMDA ionotropic receptors. The differences in AMPA/kainate receptor subunit distribution between some of the thalamic relay nuclei and the nrt are also worthy of comment. In particular, the greater presence of GluR6,7 and KA1,2 subunits in nrt suggests that functional kainate receptors may play a more prominent role in transmission in this part of the thalamic circuitry. The rapid desensitisation to kainate seen with some kainate receptor subunit compositions may thus be of functional relevance in the nrt. The use of subunit-selective antagonists (see section 3.1) in electrophysiological studies of nrt would provide important information concerning this matter. In addition, the prominence of mGluR3 in nrt (see section 4.1) indicates that this receptor may be important in the responses of these neurones to glutamatergic input from thalamic relay neurones and/or cortical afferents. Given that this receptor couples to an inhibitory cyclic AMP cascade in isolated systems (Tanabe et al., 1993), the functional consequences of activating this receptor on nrt neurones are difficult to predict. Clearly, physiological studies with selective agonists and antagonists for mGluR3 are required in this brain area.

5.3 Presynaptic Modulation of Transmission

There is extensive evidence, gathered from various parts of the CNS (e.g., hippocampus, striatum, cerebral cortex, spinal cord), that both inhibitory (GABAergic) and excitatory (glutamatergic) transmission can be modulated by presynaptic excitatory amino acid receptors (Baskys & Malenka, 1991; Burke & Hablitz, 1994; Calabresi et al., 1992; Forsythe & Clements, 1990; Hayashi et al., 1993; Ishida et al., 1993; Jane et al., 1994; Poncer et al., 1995; Sladeczek et al., 1993; Stefani et al., 1994; Takahashi et al., 1992). It is likely to be that these receptors are of the mGluR type(s). Little is however known concerning the possibility of such mechanisms in the thalamus. Recently, we (Salt & Eaton 1995a) have shown that agonists which are active at either Group II or Group III mGluRs are able to reduce IPSPs in VB, and that such effects are blocked in a selective manner by a variety of novel mGluR antagonists (Salt & Eaton 1995a). Given the lack of Golgi type II interneurones in the rat VB, these actions are likely to be mediated via presynaptic mGluRs (both Group II and Group III) on the terminals of nrt neurones which project back into VB. This hypothesis is supported by recent anatomical evidence (Van Horn et al., 1995). There is however no anatomical evidence for axo-axonic synapses onto nrt terminals, and thus the functional significance of these receptors remains to be elucidated.

It is intriguing that so far no physiological evidence has been found for a presynaptic modulation of the excitatory inputs (sensory or cortical) to the thalamus by mGluRs. Whilst this may be due to lack of data, it is noteworthy that effects on excitatory transmission by mGluR agonists such as L-AP4 or CCG-I have not been seen in situations where they might have been expected to be seen (e.g. Salt & Eaton 1991b, 1995a): this is in marked contrast to other brain areas, such as the hippocampus, where presynaptic modulation of excitatory transmission is readily visible. Whether this reflects a true lack of presynaptic receptors on excitatory terminals or a receptor which is unaffected by the agonists tested so far remains to be determined: it is possible that mGluR5 represents this receptor, or an entirely novel mGluR may have such a function.

Please note: this document is part of the HTML version of a paper originally published in print in Progress in Neurobiology.

Contents Previous Section Next Section Reference List

Return to Tom Salt's Home Page