Primary afferent depolarization produced in Aδ and C fibres by glutamate spillover? New ways to look at old things

In 1957, Frank & Fuortes described a new mechanism for the regulation of the synaptic effectiveness of sensory fibres in the vertebrate spinal cord, namely presynaptic inhibition. However, it was not until the early sixties that Eccles and his collaborators related this inhibition to primary afferent depolarization (PAD; for review see Rudomin & Schmidt, 1999).

It is now fairly well established, from both electrophysiological and morphological investigations, that the terminal arborizations of large cutaneous and muscle afferents in the vertebrate spinal cord are the targets of specific sets of GABAergic interneurones. Activation of GABA_A receptors in the subsynaptic regions of the afferent terminals increases their permeability to chloride ions and produces PAD. Presynaptic inhibition would result from the depolarization produced by the outward chloride currents, as well as from increased membrane conductance, which may reduce, or prevent, conduction of action potentials at branch points within the intraspinal arborizations. There is in addition activation of GABA_B receptors, which also reduces synaptic effectiveness without producing PAD (for review see Willis, 1998; Rudomin & Schmidt, 1999).

In the mammalian spinal cord, PAD elicited in intraspinal branches from muscle spindles and tendon organs has a local character; namely, it may remain confined to some of the intraspinal collaterals of a single afferent fibre, without spreading to nearby or distant collaterals of the same fibre (Lomelí et al. 1998). This allows, at least in principle, a functional decoupling of neuronal activity which may otherwise be correlated with common sensory inputs, as well as independent activation of specific sets of spinal neurones (see Rudomin & Schmidt, 1999).

The observations of Russo et al. presented in this issue of The Journal of Physiology, show that in transverse slices of the isolated turtle spinal cord, the dorsal root potentials (DRPs) produced by supramaximal stimulation of dorsal roots are depressed, but not eliminated, after blockade of GABA_A receptors; however, they are eliminated by additional blockade of NMDA and AMPA receptors. This suggests that glutamate may also be involved in the generation of DRPs. Other observations indicate that GABA_A, NMDA and AMPA receptors may also contribute to the generation of a TTX-insensitive DRP.

To the authors the most plausible explanation for these observations is that glutamate is released by action potentials conducted in Aδ and C afferent fibres possessing TTX-insensitive sodium channels, and spills over onto the same and/or nearby afferent terminals, where it produces PAD. To them, this non-spiking circuitry represents an effective way for the local regulation of transmitter release around active primary afferents in the same manner as has been found at hippocampal mossy fibre synapses (see Min et al. 1998).

Clearly this is an attractive proposal, particularly because GABAergic axo-axonic synapses at central terminals of fine afferent fibres are rare, at least in the spinal cord of higher vertebrates. Yet, C fibres in the dorsal horn are depolarized by activation of skin afferents, suggesting volume-mediated autocrine and paracrine interactions (see Willis, 1998; Rudomin & Schmidt, 1999).

At this stage, it seems important to determine whether the TTX-insensitive PAD is indeed due to the glutamate that is released by the stimulated fibres and spills over onto surrounding fibres, or whether it is due to accumulation of potassium ions in the extracellular space, released by the activated fibres, which may in turn induce glutamate release from neighbouring neuronal elements and glia. To the extent that the TTX-insensitive PAD is due to glutamate spillover, it should also be sensitive to changes affecting glutamate uptake and diffusion in the extracellular space (Min et al. 1998). This could be a relevant factor determining whether the spatial spread of glutamate remains confined within the environment of the activated Aδ and C fibres, or whether it also spreads to nearby inactive fibres, from muscle as well as from cutaneous afferents. Another question that needs to be addressed, in order to pursue the proposal that glutamate spillover plays a significant role in the self-regulation of the synaptic effectiveness of the activated afferents, is the extent to which the TTX-resistant PAD reduces transmitter release in fine afferents.

The observations of Russo et al. (2000) are consistent with the view that the afferent terminals are not passive elements for information transmission, but rather are potential routes where information flow can be modulated by a variety of mechanisms. Those mediated by GABA_A receptors appear to be highly selective, and could be largely involved in the execution of specific motor tasks and during sensory processing, while others, such as glutamate or potassium spillover, could form part of a more diffuse control system of synaptic efficacy in the same manner as other autocrine and paracrine interactions (see Rudomin & Schmidt, 1999, for review). It may be anticipated that the observations of Russo et al. (2000) will promote further investigation into self-regulatory mechanisms of synaptic efficacy in the vertebrate spinal cord, which although well documented for Aδ and C fibres in the mammalian spinal cord, could also be in operation in large muscle and cutaneous afferents (Zytnicki & Jami, 1998).