Spinal Structures

The spinal cord provides the primary input stage for sensory afferents and the final output stage for PLMS. At this stage it is possible, in principle, to reduce sensory complaints such as those in RLS by active movements. This is a necessary condition to explain the cessation of symptoms (i.e., that they cease with movement) and, vice versa, the dominance of symptoms in periods of rest.

Incoming sensory afferents and the motoneuron system are interconnected by the spinal interneuronal system, which is responsible for the generation of simple reflex movements as well as complex motor patterns ranging from tonic to phasic movements and from a co-contraction of antagonistic muscles to the coordinated patterning of locomotor activity (for a review, see Schomburg¹). The activity of these interneurons, like the activity of all spinal interneurons and the activity of spinal nociceptive neurons in the dorsal horn, is modulated by descending pathways with both inhibitory and excitatory components.⁴ During normal voluntary movements, reafferent sensory input continuously converges on the spinal circuits that are activated by these descending motor commands; this locus may serve as the anatomical correlate for the earliest suppression of reafferent sensory input⁵ (Fig. 14-1). Also, presynaptic inhibition of spinal afferents produced by descending commands may effectively reduce synaptic transmission at the initial synapse.⁶

FIGURE 14-1. On the function of FRA in movement control. The descending command for the performance of a movement is assumed to be mediated by spinal interneurons, which receive an excitatory or inhibitory multisensory convergent input from FRA activated by the movement (mechanosensitive and nociceptive cutaneous afferents, joint afferents, group II muscle afferents from secondary spindle endings, groups III and IV muscle afferents from free nerve endings). This general FRA feedback may be suppressed by monoaminergic or enkephalinergic pathways (experimentally by L-DOPA and opioids). Group Ib afferents, not belonging to FRA, are included in this schematic illustration, because they may partly converge onto common interneurons with FRA. Group Ib pathways are not blocked by L-DOPA and opioids.

Modified from Jankowska E, Lundberg A. Interneurones in the spinal cord. TINS 1984;4.

All these interactions have to be generated quickly, and they depend on immediate motor requirements, that is, on the phase and condition (including possible external disturbances) of a movement and the position of the trunk and limbs. This so-called state-dependent modulation of neurotransmission is an important tool to adapt the spinal interneuronal systems to the requirements of performing movement in response to the existing conditions. Under conditions of disturbed descending control, such as in spasticity after a stroke, identical—or subthreshold peripheral—input becomes itself capable of inducing muscle contractions. In contrast, such as under normal conditions, reflexogenic activation is largely downregulated to prevent inappropriate reflexogenic movements from the continuously activated peripheral receptors. In this downregulated situation, the activity of individual receptor systems is generally unable to evoke reflexogenic movement. If at all, a distinct multisensorial convergence of different receptor systems is required to elicit a movement.⁷ During sleep, in particular during rapid eye movement (REM) sleep, state-dependent modulation is further downregulated to preclude any unwanted movements during dreaming.

Of all possible spinal structures involved, the spinal flexor reflex afferent (FRA) system and a disturbance of its supraspinal control seem to play the crucial role in the generation of PLMS. The FRA system encompasses the following features:

Dopamine in the Spinal Cord

The spinal cord is virtually devoid of dopaminergic cell bodies. The major source of the spinal dopaminergic projection into the dorsal horns and intermediolateral tracts is area A11.¹⁰ The behavioral effects of a first animal model on A11 lesion have not been convincing, most likely because the A11 neurons are spread over a comparatively large area and an isolated lesion appears to be very difficult to achieve.¹¹ Nevertheless, an additional intriguing argument in favor of an A11 lesion as a source for RLS is its anatomical location close to the suprachiasmatic nucleus, which largely controls circadian rhythms and which could explain its predominance at night. The RLS-A11 hypothesis would require a rather selective neurodegeneration of A11 with increasing age and a concomitant reduction in the descending modulation of the dorsal horn or other related structures in the spinal cord.¹²

Because no RLS animal model is readily available, most data have been obtained in the course of pain studies. Dopamine may ameliorate chronic pain. In rats, intrathecal administration of the dopamine agonist apomorphine produced an analgesia that was antagonized by the prior intrathecal administration of a dopamine receptor antagonist but not by the administration of α₂, serotonin, or opiate antagonists.¹³ Further data beyond the scope of this review confirm the antinociceptive effects of dopamine D₂ receptors (for the most detailed review, see Millan¹⁰), probably via a direct inhibitory action at either the nociceptive-responsive primary afferent fiber terminals or projection neurons. An inhibitory influence of presynaptic dopamine D₂ receptors on Ca²⁺ currents provides a potential mechanism for a presynaptic reduction of release from primary afferent fiber terminals in the dorsal horn.¹⁴ Also, L-DOPA inhibits the early flexor reflex, regardless of whether it is evoked by nociceptive or nonnociceptive afferents.¹⁵

State dependence usually depends on dopaminergic stimulation, but there are contrary cases. For example, in acute phase–dependent (during locomotion)^16–19 or limb position–dependent²⁰ reflex transmission, dopamine is not primarily involved.

The influence of dopamine on the spinal cord is best investigated in the spinalized cat. The findings in spinalized cats are transferable to the intact animal in principle, but in the intact cat, they are under strong supraspinal control. Thus, in the spinalized animal, the functional state of a spinal cord free of predominantly inhibitory supraspinal control can be studied not only in isolation but also in a disinhibited, and thus enhanced, functional state.

Within a medium range of spinal dopamine concentration in the spinalized and paralyzed cat, locomotor activity (“fictive locomotion”; i.e., neuronal rhythmic activity resembling locomotor pattern without real movements in paralyzed animals) can be induced. During fictive locomotion, motoneurons can develop a bistable firing characteristic, caused by their membrane potentials switching between two levels. This state is induced by L-DOPA. Depolarization during active phases was characterized by stable fixed levels and could partly be initiated by excitatory inputs and terminated by inhibitory inputs.²¹ Similar bistable characteristics have been postulated for RLS.²² It appears, however, unlikely that an L-DOPA–induced bistability is related to the PLM-associated firing pattern in RLS, because L-DOPA suppresses PLM instead of inducing such bursting. However, this is not a direct motoneuronal effect but has to be regarded as a consequence of the L-DOPA–induced suppression of short-latency FRA pathways (see earlier). The intricate function of the dopaminergic system in spinal motor control becomes evident when L-DOPA is applied in different dosages. Without the application of L-DOPA in the spinal animal, repeated afferent FRA stimulation evokes the “normal” short-latency flexor reflexes. After L-DOPA in a low dosage that is subthreshold for induction of rhythmic locomotor activity, the same stimulation evokes longer-lasting tonic responses. If the dopamine concentration is raised above the level that causes rhythmic locomotor activity, a spontaneous irregular motor hyperactivity occurs.¹⁸ It can be assumed that the responses to afferent stimulation during the latter stages of hyperreactivity are due to an increased responsiveness of the late FRA system. At the spinal level, L-DOPA thus heavily depresses the short-latency FRA pathways (probably its therapeutical action in RLS), whereas the transmission in the long-latency pathways, and in parallel the generation of locomotion, are facilitated (see Table 14-1). Because the depression of polysynaptic reflexes seems to be mainly mediated by D₃-dopamine receptors and dopamine may cause an increase of these reflexes in the isolated spinal cord of D₃KO mice, it has been assumed that an impaired D₃ activity could contribute to the RLS.²

The importance of FRA for RLS has been delineated in two clinical spinal reflex studies.^23,24 Here, RLS patients had significantly increased spinal cord excitability with lower threshold and greater spatial spread of the flexor reflexes, which were comparatively more prominent during sleep. Multiple late responses were seen during sleep in all patients and in some control subjects at higher thresholds. The most prominent of these responses had a very long duration and a latency range of 250 to 800 milliseconds, and because of its close temporal relation to the FRA stimulus, the authors considered it as a late, high-threshold component of the FR (FR3). This finding cannot immediately be compared with PLMS, which, as described later, decrease with depth of sleep stages. These data can, however, be modeled on the basis of the wide convergence of nociceptive and non-nociceptive flexor reflex afferents onto common interneurons in spinal reflex pathways.^25,26

Relation to Sleep Stages

With increase in sleep stages, from light stage 1 sleep to deepest non-REM (NREM) sleep and even further in REM sleep, PLMS are continuously suppressed compared with a relaxed and awake state in bed.²⁷ Nicolas and colleagues²⁷ found that REM sleep was characterized by the shorter duration and the lowest frequency of PLMS, due most likely to the inhibition of spinal motoneurons that prevails in REM sleep. It is most likely that the system involved in REM sleep motor depression is different from the RLS generators because it is able to attenuate PLMS but does not eliminate them completely. Thus, two different systems can be assumed—one dependent on the dopamine level and responsible for PLMS, and the other responsible for suppression of movements in REM sleep and heavily dependent on GABA neurotransmission.^28,29 Nevertheless, Nicolas and colleagues²⁷ concluded from their result that during REM, the duration of sleep intervals returned to values obtained in stage 1 sleep and that these two stages shared similar patterns of electroencephalographic activity—and that a single state-dependent mechanism may be responsible for the periodicity of PLM noted both during sleep and wakefulness. It is, of course, potentially important that RLS symptoms peak at night when dopamine levels are lowest.

An altered circadian pattern in spinal dopamine synthesis was described in D₃KO animals.³ Low dopamine levels (1 μmol/L) decreased the monosynaptic “stretch” reflex (MSR) amplitude in wild-type (WT) animals but increased it in D₃KO animals. Higher dopamine concentrations (10 to 100 μmol/L) decreased MSR amplitudes in both groups, but always more strongly in WT mice. Like low dopamine, the D₃ receptor agonists pergolide and PD 128907 reduced MSR amplitude in WT but not D₃KO mice. Conversely, D₃ receptor antagonists (GR 103691 and nafadotride) increased the MSR in WT but not in D₃KO mice. In comparison, D₂-preferring agonists bromocriptine and quinpirole depressed the MSR in both groups. Low dopamine (1 to 5 μmol/L) also depressed longer-latency (presumably polysynaptic) reflexes in WT but facilitated responses in D₃KO mice. Additionally, in some experiments (e.g., during 10 μmol/L dopamine or pergolide in WT), polysynaptic reflexes were facilitated in parallel to MSR depression, demonstrating differential modulatory control of these reflex circuits. Thus, low dopamine activates D₃ receptors to limit reflex excitability. The authors conclude that in D₃ ligand–insensitive mice, excitatory actions are unmasked, functionally converting the modulatory action of dopamine from depression to facilitation.

Augmentation: A Paradoxical L-DOPA–Induced, Long-Latency FRA Effect

In some patients, after once-nightly treatments with carbidopa or levodopa³⁰ but also after treatment with other dopamine agonists,³¹ a delayed augmentation of the RLS symptoms was observed the following day. This augmentation is a potentially serious problem for all dopaminergic treatments in RLS patients. It has been speculated that augmentation may arise as a kind of rebound or a result of a suppression of the circadian rhythm in the striatal, hypothalamic, or spinal dopaminergic system by the higher L-DOPA dosages that are usually associated with augmentation.³² However, according to experimental neurophysiological results, it seems to be more probable that the augmentation of the RLS symptoms is due to a second, delayed action of L-DOPA, which consists of facilitation of long-latency FRA pathways and of an introduction of rhythmic, partly locomotor movements or a more-or-less arrhythmic hyperactivity (cf. Table 1; review by Schomburg¹). In the spinal animal, these symptoms may occur about 20 minutes to 1 hour after intravenous injection of L-DOPA (after premedication with a monaminoxidase-blocker) and could last for several hours (up to ≥8 hours, Kniffki and associates).¹⁸ More recently, we argue that in terms of receptor contribution, augmentation may be caused by a relative D₁ receptor overstimulation (Paulus and Trenkwalder, 2006).³³ Iron deficiency and 6-hydroxydopamine lesions in A11 nuclei differentially altered the D₁, D₂, and D₃ receptor expression and binding capacity in the lumbar spinal cord of RLS animal model, which were accompanied by changes in locomotor activities.³⁴

Most interestingly, A11 diencephalospinal dopaminergic neurons lack presynaptic synthesis modulating D₂ autoreceptors in mice. Thus, there is no counterregulation of elevated dopamine levels in augmentation.³⁵ The therapeutic benefit and the unwanted side effect of L-DOPA could thus be results of the two opposite effects of L-DOPA at the spinal level: a suppression of the early short-latency FRA pathways and a facilitation of the late long-latency FRA pathways. This assumption finds further support in the fact that similar augmentation of the RLS symptoms were not observed after RLS treatment with opioids and that the augmentation occurring after levodopa can be successfully treated by opioids, which are known to suppress the long-latency FRA pathways and their ability to generate rhythmic spinal motor activity³⁶ (cf. also Table 14-1).

Opioids in the Spinal Cord

The dramatic therapeutic response to opioids in RLS patients warrants the assumption of a central role in pathophysiology. The efficacy of opiates in the relief of the symptoms is possibly partly related to the dopaminergic system.³⁷ However, the sensitivity to opioids does not imply a crucial role of the nociceptive system in the development of RLS. Like L-DOPA, opioids do not need to have a selective antinociceptive effect at the spinal level because they suppress segmental pathways from all FRA in the same way, from group II spindle afferents, mechanosensitive skin or joint afferents, as well as from nociceptive skin and muscle afferents^1,25 (Figs. 14-2 and 14-3).

FIGURE 14-2. Neurotransmission of afferent information in the superficial dorsal horn of the spinal cord. Possible interactions between nociceptor afferent fibers, local interneurons, and descending fibers in the dorsal horn of the spinal cord. Nociceptive fibers terminate on second-order spinothalamic projection neurons. Local enkephalin-containing interneurons (ENK) exert both presynaptic and postsynaptic inhibitory actions at these synapses. Serotonergic and noradrenergic neurons in the brainstem activate the local opioid interneurons and also suppress the activity of spino-thalamic projection neurons.

FIGURE 14-3. Action of opiates in the superficial dorsal horn of the spinal cord: Activation of nociceptors leads to the release of glutamate and neuropeptides from sensory terminals in the superficial dorsal horn, thus depolarizing and activating projection neurons. Opiates decrease the duration of the action potential of nociceptors, probably by decreased Ca²⁺ influx, and thus decrease the release of transmitter from primary afferent terminals. In addition, opiates hyperpolarize the membrane of the dorsal horn neurons by activating a K⁺ conductance. Stimulation of the nociceptor normally produces a fast excitatory postsynaptic potential in the dorsal horn neuron; opiates decrease the amplitude of the postsynaptic potential.

Reprinted with permission from Basbaum AI, Jessell TM. The perception of pain. In Kandel ER, Schwartz JH, Jessell TM [eds]. Principles of Neural Science. New York, McGraw-Hill, 2000, pp 472-491.

Thus, the enkephalinergic system may cooperate with the descending monoaminergic system to provide a mechanism for suppression of the relatively unspecific FRA feedback during supraspinally induced movements, which require a strictly goal-directed performance without any disturbances from the periphery.³⁸ An intriguing interaction between both the monaminergic and the opiate system seems to be ahead in some genetic forms of RLS.³⁹ Supraspinal motor actions of opioids are more ambiguous because they may partly increase such locomotor activity in intact animals^40–43 as a result of supraspinal interactions, probably including actions on the striatum.^44,45 However, at the spinal level, opioids merely depress locomotor rhythm generation³⁶ (see also Table 14-1). Regarding the fact that locomotor rhythm generation is at least partly a function of the spinal FRA system^36,46 and the above-mentioned observation of the depressing action of opioids on FRA pathways, it can be assumed that therapeutical opioidergic influence on RLS largely results from a spinal action of opioids.

Summary

We sought to bring together two separated areas of research, investigation of opioid and dopaminergic mechanisms in the spinal cord of spinalized animals and clinical features of RLS (Paulus and Schomburg, 2006).⁴⁷ The most important correlation that can be drawn is a reduction of activity of the early flexor reflexes by both substances, whereas the late reflexes are facilitated by L-DOPA. The latter mechanism may contribute to the phenomenon of augmentation. The fact that the opioids reduce both may offer a rationale to the use of opioids in the treatment of augmentation.