Elzbieta Jankowska

A long‐lasting GABA‐dependent increase in the excitability of afferent fibres, and thus modulation of the sensory input to the spinal cord, may be evoked by epidural polarization. However, the direct effects of fibre polarization are short‐lasting and the sustained increase in their excitability appears to be secondary to the release of GABA from nearby astrocytes. We have now investigated whether the modulation of spinal sensory input by stimulation of a peripheral nerve, previously attributed to synaptically evoked intraspinal field potentials, is evoked in a similar way. However, as neither its dependence on GABA nor its relays have been investigated, we addressed the question of whether the increase in the excitability of epidurally stimulated afferent fibres following a peripheral nerve stimulation does or does not depend on GABA and whether it might be mediated by astrocytes. The effects of conditioning stimulation of the tibial nerve were evaluated from changes in the excitability of both Group I and II muscle afferents, estimated from action potentials recorded in peripheral nerves and in field potentials recorded in the dorsal horn respectively in acute experiments on deeply anaesthetized rats. The excitability of the afferents was increased by stimulation of Group II and/or cutaneous but not Group I muscle afferents. The effects were significantly weakened by blocking GABA channels by gabazine and by astrocyte toxin L‐alpha‐aminoadipic acid (L‐AAA), indicating that the excitability of both Group I and II afferent fibres may be modulated by GABAergic astrocytes, the new role played by astrocytes.

A long-lasting GABA-dependent increase in the excitability of afferent fibres, and thus modulation of the sensory input to the spinal cord, may be evoked by epidural polarization. However, the direct effects of fibre polarization are short-lasting and the sustained increase in their excitability appears to be secondary to the release of GABA from nearby astrocytes. We have now investigated whether the modulation of spinal sensory input by stimulation of a peripheral nerve, previously attributed to synaptically evoked intraspinal field potentials, is evoked in a similar way. However, as neither its dependence on GABA nor its relays have been investigated, we addressed the question of whether the increase in the excitability of epidurally stimulated afferent fibres following a peripheral nerve stimulation does or does not depend on GABA and whether it might be mediated by astrocytes. The effects of conditioning stimulation of the tibial nerve were evaluated from changes in the excitability of both group I and group II muscle afferents, estimated from action potentials recorded in peripheral nerves and in field potentials recorded in the dorsal horn respectively in acute experiments on deeply anaesthetized rats. The excitability of the afferents was increased by stimulation of group II and/or cutaneous but not group I muscle afferents. The effects were significantly weakened by blocking GABA channels by gabazine, and by astrocyte toxin L-alpha-aminoadipic acid (L-AAA), indicating that the excitability of both group I and group II afferent fibres may be modulated by GABAergic astrocytes, the new role played by astrocytes.

Modulation of input from primary afferent fibres has long been examined at the level of the first relays of these fibres. However, recent studies reveal that input to the spinal cord may also be modulated at the level of the very entry of afferent fibres to the spinal grey matter before action potentials in intraspinal collaterals of afferent fibres reach their target neurons. Such modulation greatly depends on the actions of GABA via extrasynaptic membrane receptors. In the reported study we hypothesized that the increase in excitability of afferent fibres following epidural polarization close to the site where collaterals of afferent fibres leave the dorsal columns is due to the release of GABA from two sources: not only GABAergic interneurons but also glial cells. We present evidence, primo, that GABA released from both these sources contributes to a long‐lasting increase in the excitability and a shortening of the refractory period of epidurally stimulated afferent fibres and, secondo, that effects of epidural polarization on the release of GABA are more critical for these changes than direct effects of DC on the stimulated fibres. The experiments were carried out in deeply anaesthetized rats in which changes in compound action potentials evoked in hindlimb peripheral nerves by dorsal column stimulation were used as a measure of the excitability of afferent fibres. The study throws new light on the modulation of input to spinal networks but also on mechanisms underlying the restoration of spinal functions.

Modulation of input from primary afferent fibres has long been examined at the level of the first relay neurons of these fibres. However, recent studies reveal that input to the spinal cord may also be modulated before action potentials in intraspinal collaterals of afferent fibres reach their target neurons, even at the level of the very entry of afferent fibres to the spinal grey matter. Such modulation greatly depends on the actions of GABA via extrasynaptic membrane receptors. In the reported study we hypothesized that the increase in excitability of afferent fibres following epidural polarization close to the site where collaterals of afferent fibres leave the dorsal columns is due to the release of GABA from two sources: not only terminals of GABAergic interneurons but also glial cells. We present evidence, primo, that GABA from both these sources contributes to a long-lasting increase in the excitability and a shortening of the refractory period of epidurally stimulated afferents fibres and, secondo, that effects of epidural polarization on GABA-containing terminals of GABAergic interneurons and on glial cells are more critical for these changes than direct effects on the stimulated fibres. The experiments were carried out in deeply anaesthetized rats in which changes in compound action potentials evoked in hindlimb peripheral nerves by dorsal column stimulation were used as a measure of the excitability of afferent fibres. The study throws new light on the modulation of input to spinal networks but also on mechanisms underlying the restoration of spinal functions.

Few attempts have so far been made to define the mechanisms underlying the hour‐long effects of trans‐spinal stimulation combined with epidural polarization. In the present study, we investigated the potential involvement of non‐inactivating sodium channels in afferent fibres. To this end, riluzole, a blocker of these channels, was administered locally to the dorsal columns close to the site of the excitation of afferent nerve fibres by epidural stimulation in deeply anaesthetized rats in vivo. Riluzole did not prevent the induction of the polarization‐evoked sustained increase in the excitability of dorsal column fibres but tended to weaken it. It likewise weakened but did not abolish the sustained polarization‐evoked shortening of the refractory period of these fibres. These results lead to the conclusion that the persistent sodium current may contribute to the sustained post‐polarization‐evoked effects but is only partly involved in both the induction and the expression of these effects.

Integrative functions of spinal interneurons are well recognized but the relative role of different interneuronal populations in this process continues to be investigated. It therefore appeared useful to review the principles of integration of afferent information by the interneurons analyzed so far as these principles should apply also to those remaining to be analyzed. Considering the results of both functional and morphological studies of spinal interneurons and of the morphology and immunochemistry of afferent fibers that provide input to them, the following five basic principles of processing of afferent information by them will be outlined; 1) afferent information of any origin is forwarded to several neuronal populations, 2) information from any sources of input is distributed unevenly, 3) input from several sources is integrated by individual neurons as well as by their populations, 4) specific combinations of input are integrated by different neuronal populations, and 5) afferent input to spinal interneurons is only one of the features distinguishing their functional populations. As the spinal neuronal organization and properties of neurons and afferent fibers in the so far investigated species (cat, rodents, and primates) have been found to resemble, future studies using molecular techniques in the mouse should allow the new data to integrate with those of the preceding studies and the principles outlined earlier as well as any new ones should apply also in humans.

The main question addressed in this study was whether the refractoriness of nerve fibres can be modulated by their depolarisation and, if so, whether depolarisation of nerve fibres evokes a long‐term decrease in the duration of the refractory period as well as the previously demonstrated increase in their excitability. This was investigated on nerve fibres within the dorsal columns, dorsal roots and peripheral nerves in deeply anaesthetised rats in vivo. The results revealed major differences depending on the sites of fibre stimulation and polarisation. Firstly, the relative refractory period was found to be shorter in epidurally stimulated dorsal column fibres than in fibres stimulated at other sites. Secondly, the minimal effective interstimulus intervals reflecting the absolute refractory period were likewise shorter for nerve fibres within the dorsal columns even though action potentials evoked by the second of a pair of stimuli were similarly delayed with respect to the preceding action potentials at all the stimulation sites. Thirdly, the minimal interstimulus intervals were reduced by epidurally applied cathodal direct current polarisation but not at other stimulation sites. Consequently, higher proportions of dorsal column fibres could be excited at higher frequencies, especially following their depolarisation, at interstimulus intervals as short as 0.5–0.7 ms. The results demonstrate that epidural depolarisation results in long‐lasting effects not only on the excitability but also on the refractoriness of dorsal column fibres. They also provide further evidence for specific features of afferent fibres traversing the dorsal columns previously linked to properties of their branching regions.

The review surveys various aspects of the plasticity of nerve fibers, in particular the prolonged increase in their excitability evoked by polarization, focusing on a long-lasting increase in the excitability of myelinated afferent fibers traversing the dorsal columns of the spinal cord. We review the evidence that increased axonal excitability 1) follows epidurally applied direct current (DC) as well as relatively short (5 or 10 ms) current pulses and synaptically evoked intrinsic field potentials; 2) critically depends on the polarization of branching regions of afferent fibers at the sites where they bifurcate and give off axon collaterals entering the spinal gray matter in conjunction with actions of extrasynaptic GABA_A membrane receptors; and 3) shares the feature of being activity-independent with the short-lasting effects of polarization of peripheral nerve fibers. A comparison between the polarization evoked sustained increase in the excitability of dorsal column fibers and spinal motoneurons (plateau potentials) indicates the possibility that they are mediated by partly similar membrane channels (including noninactivating type L Cav⁺⁺ 1.3 but not Na⁺ channels) and partly different mechanisms. We finally consider under which conditions transspinally applied DC (tsDCS) might reproduce the effects of epidural polarization on dorsal column fibers and the possible advantages of increased excitability of afferent fibers for the rehabilitation of motor and sensory functions after spinal cord injuries. NEW & NOTEWORTHY This review supplements previous reviews of properties of nerve fibers by surveying recent experimental evidence for their long-term plasticity. It also extends recent descriptions of spinal effects of DC by reviewing effects of polarization of afferent nerve fibers within the dorsal columns, the mechanisms most likely underlying the long-lasting increase in their excitability and possible clinical implications.