| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Faculty of Pharmaceutical Sciences, Division of Pharmacology and Toxicology, The University of British Columbia, Vancouver V6T 1Z3, Canada
Submitted 17 November 2003; accepted in final form 15 April 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
, 2001a; Taepavarapruk et al. 2002). In particular, the spontaneous and glutamate-evoked responses of DSCT neurons are markedly reduced during active sleep and are mediated by the "fast" inhibitory amino acid neurotransmitters GABA and glycine. Evoked potential studies conducted in the 1960s revealed that peripheral nerve-evoked responses recorded in the cerebellum were indeed suppressed only during intense eye movements of active sleep (Carli et al. 1966, 1967a,b). These findings suggest that the magnitude of low-threshold, peripherally evoked responses of DSCT neurons may also depend on behavioral state.
In this study, we investigated this issue for the first time by applying low-intensity stimuli to the sciatic nerve in the intact chronic cat preparation and quantifying the evoked responses of individual DSCT neurons as a function of sleep and wakefulness. We were also interested in determining if modulation of peripheral nerve-evoked responses was specifically related to the presence of eye movement events that hallmark the state of active sleep, in view of earlier studies that employed simple field potential recordings and indicated sensory transmission was suppressed in association with eye movements (Carli et al. 1966, 1967a,b).
Finally, if peripheral nerve-evoked responses of DSCT neurons were indeed suppressed during active sleep, we wished to determine if GABA, a classical inhibitory neurotransmitter in the spinal cord, underlies this phenomenon, and if so, whether it does specifically during non–eye movement periods and/or eye movement events of active sleep.
The rationale employed here was that direct activation of the axons of primary afferent neurons would bypass their normal transduction mechanisms and impulse transmission from peripheral receptors. If the response magnitude of DSCT neurons evoked by stimuli applied directly to the axons of primary afferent neurons decreases during active sleep, this would indicate that state-specific inhibition was occurring at the level of the spinal cord at synapses located between primary afferents and the recorded cells. Moreover, juxtacellular administration of appropriate inhibitory neurotransmitter antagonists near the recorded DSCT neurons should block this state-dependent inhibition. Such findings would support the theory that a neural pathway from higher brain centers is engaged to dampen ascending sensory inflow as opposed to a withdrawal of afferent input or disfacilitation of DSCT neurons arising as a consequence of motor atonia that is present during active sleep.
Here we report that synchronous synaptic activation of DSCT neurons by primary afferents is subjected to GABAergic inhibitory influences during tonic non–eye movement periods and phasic eye movement events, the latter of which hallmark the behavioral state of active sleep.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
All experimental procedures reported herein were performed on a total of three adult intact, unanesthetized cats and complied with (inter)national and institutional (University of British Columbia Animal Care Committee) statutes. Under deep gaseous anesthesia (45–60% N2O in 1.5–2.5% halothane/oxygen mixture), the animals were implanted with head- and lumbar-restraining devices. Electrodes were implanted into the frontal sinus [electroencephalogram (EEG)], lateral geniculate nucleus of the thalamus [ponto-geniculo-occipital (PGO) wave activity], the orbital plate [electro-oculogram (EOG)], and neck muscles (EMG). Through the use of these electrodes, each animal's behavioral states of wakefulness and sleep could be determined. A complete recovery over a 3-mo period was required before any recordings from DSCT neurons were performed. Procedures for surgical implantation, the gradual adaptation to painless head and lumbar restraint during the latter half of the 3-mo recovery period, identification of sleep/wake states, and antidromic identification of spinal projection neurons in the chronic intact cat have been previously described in detail (Soja et al. 1995).
Sciatic nerve stimulation
A cuff electrode was implanted on the sciatic nerve during a separate surgical procedure performed under gaseous anesthesia 
3 wk before recording sessions commenced (Soja et al. 1995). Electrode leads were tunneled subcutaneously and externalized to a 10-pin connector plug affixed to the lumbar restraint (see Fig. 1 in Soja et al. 1995). During neuronal recording sessions, single shock stimuli (0.05 ms duration, 1.0 Hz interval) were applied at current intensities sufficient (1.0 T) to synaptically activate DSCT neurons as evidenced by recording short or longer latency action potential(s).
|
Each preparation readily cycled between sleep and wakefulness during each experimental recording session that typically lasted 5–6 h. Recording sessions were performed over 4 consecutive days (Soja et al. 1996). DSCT neuron activity was recorded over multiple sleep cycles. Each sleep cycle consisted of the following states: wakefulness, quiet sleep, active sleep, and awakening from active sleep. Electrophysiological criteria for determining states of wakefulness and sleep were identical to those reported in detail previously (Soja et al. 1995, 1996, 2001a,b; Taepavarapruk et al. 2002).
Antidromic identification and recording procedures
Low-intensity search stimuli (0.05 ms, 0.5 Hz, 100–300 µA) were applied to a removable stimulating microelectrode that was stereotaxically positioned within the anterior cerebellar lobule during each recording session to "backfire" DSCT neurons using conventional criteria, i.e., constant latency, collision between spontaneous and antidromically propagated spikes, and high-frequency following (Soja et al. 1995, 1996). All DSCT neurons reported herein satisfied these criteria for antidromicity. Antidromic search stimuli did not cause any indications of arousal, EEG desynchronization, or interfere with the animal's normal cycling between sleep and wakefulness.
All behavioral state data (EEG, EOG, PGO, and EMG activities) were recorded continuously and digitized "on-line" (sampling rate: 0.2 kHz) into separate waveform channels using a Pentium 4 computer equipped with commercially available software (Spike 2, version 4.01, Cambridge Electronic Design, Cambridge, UK) and hardware (1401-plus, Cambridge Electronic Design). DSCT neuron activity was recorded extracellularly (1,000x, band-pass: 0.3 kHz, –10 kHz) using an AC-coupled amplifer (model 1800, AM-Systems, Carlsborg, WA).
DSCT neuron spike acquisition procedures
The signal-to-noise ratio of DSCT neurons exceeded 4:1 (Soja et al. 1995). The action potential height and shape of each DSCT neuron was constantly monitored on a storage oscilloscope, and the stability of the response was verified throughout the duration of each cell's entire recording time. Substantial ROM memory allocation was required to capture all behavioral state signals and DSCT action potential activity. Hence, spike sorting routines available in Spike 2 were not utilized or required. Instead, a spike processor (model D130, Digitimer, Hertfordshire, UK) was used to discriminate "on-line" continuous spike activity recorded from DSCT neurons across sleep and waking states (Fig. 1). DSCT neuronal spikes crossing a voltage level discriminator were converted to transistor transistor logic (TTL) pulses, time-stamped, and saved in a separate Spike 2 channel (Figs. 1B and 2B). Sciatic nerve stimulus presentations were also captured as time-stamped TTL pulses in a separate Spike 2 channel and were used in conjunction with the DSCT neuron spike TTL channel to construct peristimulus time histograms (PSTHs; see PSTH analyses).
|
Computer-generated PSTHs were used to quantitatively assess the responsiveness of DSCT neurons to low-intensity stimulation of the sciatic nerve. Primary and secondary PSTH analysis methods were employed in this study. The primary PSTH analysis method was used to determine if DSCT neuron responsiveness was dependent on behavioral state. Here, PSTHs were constructed using 75–200 consecutive sciatic nerve-evoked responses obtained during wakefulness, quiet sleep, active sleep, and awakening from active sleep. Responses during active sleep included those occurring during non–eye movement periods as well as eye movement events. TTL pulses corresponding to DSCT neuron action potentials were cumulatively summed and plotted in 0.5-ms binwidths against the stimulus TTL pulse at time 0. A Spike 2 script was used to determine the average latency-to-onset (ms) and response magnitude (spikes/stimulus) from user-defined epochs corresponding to early (2–4.5 ms) and/or late (6–200 ms) synaptic responses (see RESULTS).
A secondary PSTH analysis routine and Spike 2 script was used to quantify sciatic nerve-evoked activity for each DSCT neuron specifically around eye movement events of active sleep. Eye movements were identified and gated as an event (Fig. 2A). Stimulus and response TTL markers "during" eye movement events were tagged and matched with markers "before" and "after" each eye movement event (Fig. 2A). A trio of PSTHs, (0.5 ms bindwidth) corresponding to before, during, and after eye movement events were constructed from all corresponding "before," "during," and "after" TTL markers (Fig. 2D). Each of the three PSTHs during active sleep comprised the same number of nonconsecutive trials across the entire state of active sleep. For comparison, equivalent trial PSTHs were also constructed during wakefulness, quiet sleep, and awakening from active sleep. The average latency-to-onset (ms) and response magnitude (spikes/stimulus) were calculated over user-defined epochs corresponding to early and late responses.
Eye movement events
Analyses were performed off-line with data acquisition software (Spike 2, Cambridge Electronic Design) and analysis subroutine scripts that were specially developed in-house for Spike 2 software. These scripts are available from P. J. Soja on request. The behavioral state of active sleep is comprised of hallmark phasic periods where eye movement events of variable length and intensity occur and tonic periods where no eye movement events are present (referred to as the non–eye movement periods of active sleep). Briefly, off-line analyses employed a Spike 2 script that demarcated eye movement episodes from the EOG waveform activity. Marker pulses depicting peaks and troughs of EOG waveform activity crossing voltage window levels that were preset to correspond to baseline EOG (Fig. 2A, EOG trace) were saved in new channels in the Spike 2 file for each cell. These markers, in turn, were used to create a gate corresponding to the onset and offset of active sleep-related eye movement events or eye movement events during wakefulness. A minimum duration of 250 ms was used as a default for these eye movement events. Eye movement event durations were summed and the eye movement density (i.e., summed eye movement event time/total active sleep time) determined across all recording experiments.
Microiontophoresis procedures
In several experiments, the single barrel recording microelectrode in the spinal cord was removed and replaced with a commercial multibarrel glass micropipette (Carbostar-4, Kation Scientific, Minneapolis, MN) for extracellular unit recording and simultaneous juxtacellular drug microiontophoresis (see Fig. 1A, magnified box enclosure in Taepavarapruk et al. 2002). The main carbon fiber-containing recording barrel was surrounded by barrels filled with the inhibitory amino acid agonist GABA (0.5 M, pH 3.5), the inhibitory amino acid antagonist bicuculline methiodide (Curtis et al. 1971; Krnjevic et al. 1977) (BIC: 0.02 M, pH 4.0), or sodium chloride (NaCl: 4 M). All drugs used for microiontophoresis were dissolved in distilled water and were obtained from Sigma-Aldrich. Drugs were ejected using positive currents. Negative currents (–10 nA) were continuously applied to each drug barrel to minimize spontaneous leakage. Automatic current neutralization procedures were performed using a drug barrel containing 4 M NaCl during drug microiontophoresis (Taepavarapruk et al. 2002).
In these experiments, a two sleep cycle paradigm was employed whereby the DSCT neuronal response magnitude to sciatic nerve stimulation was first determined during active sleep (non–eye movement and eye movement portions) and compared with preceding episodes of wakefulness, quiet sleep, and awakening from active sleep. The response magnitude was tested in the second sleep cycle while the GABAA receptor antagonist, bicuculline (Curtis et al. 1971; Krnjevic et al. 1977), was continuously ejected. Bicuculline was microiontophoretically released in quantities sufficient to block GABA-evoked inhibition of sciatic nerve-evoked responses during the state of wakefulness or quiet sleep prior to active sleep in this second sleep cycle. Once this pharmacological control was achieved, the bicuculline ejection was maintained throughout the subsequent episode of active sleep during the second (test-bicuculine) sleep cycle. We have previously utilized multiple control and test sleep cycle microiontophoretic paradigms for determining the neurotransmitter basis for the inhibition of spontaneous spike activity of DSCT neurons during active sleep (Taepavarapruk et al. 2002).
Statistical analyses
Values are reported as means ± SE in the text and figures. Each recorded DSCT neuron served as its own control. The magnitude and latency-to-onset of DSCT neuron responses to sciatic nerve stimulation determined from each PSTH were assessed as a function of behavioral state. Most of the population data reported herein were not normally distributed. Therefore, parameters of response magnitude and response latency were universally compared during active sleep with eye movement events, active sleep with non–eye movement periods, wakefulness, quiet sleep, and awakening from active sleep using a nonparametric Friedman repeated-measures ANOVA on ranks to determine if a statistical difference was present. Then, to isolate the group or groups that differed from the others, a multiple comparison procedure was performed (Student–Newman–Keuls method). Relative percent changes in response magnitude during non–eye movement and eye movement periods of active sleep were based on the control state of wakefulness. A Spearman rank-order correlation test was used to investigate the relationship between the change in sciatic nerve-evoked activity during non–eye movement active sleep relative to wakefulness and the change in sciatic nerve-evoked activity during eye movement events relative to non–eye movement active sleep. In all tests, the level of significance was set at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Experiments were performed on a total of 60 antidromically identified DSCT neurons obtained from three cats, all of which displayed spontaneous spike activity during the state of wakefulness. Fifty of these neurons were examined using a single barrel micropipette, and the remaining 10 neurons were recorded using a multibarrel micropipette. The antidromic latency for all 60 cells measured 3.28 ± 0.10 (SE) ms (range, 2.5–6.5 ms). The estimated average axonal conduction velocity measured 79.5 ± 1.72 m/s (range, 38.5–101.2 m/s). These observations are consistent with those of previous studies of DSCT neurons recorded across sleep and wakefulness (Soja et al. 1996, 2001a; Taepavarapruk et al. 2002, 2003b).
DSCT neuron responses to sciatic nerve stimulation during wakefulness
During the control state of wakefulness, threshold single shock stimuli applied to the sciatic nerve (0.05 ms, 0.2–0.5 mA, 1.0 Hz) produced a single action potential or a train of action potentials in 50 DSCT neurons tested. The nature of the response varied among DSCT neurons. In 18 of these neurons, a single spike response was recorded with a group mean latency of 3.27 ± 0.15 ms. In 40 other DSCT neurons, a burst of two to six action potentials occurred at a mean latency of 23.7 ± 2.23 ms. In our population of 50 DSCT neurons, 8 DSCT neurons displayed both short latency and longer latency spike responses.
Modulation of sciatic nerve responses during wakefulness and sleep
The primary PSTH analysis method was used to determine if DSCT neuronal responsiveness to sciatic nerve stimulation depended on behavioral state. PSTHs were constructed using an average of 125 ± 10 consecutive sciatic nerve stimuli presented during wakefulness, quiet sleep, active sleep, and awakening from active sleep. The response magnitude and onset latency were analyzed and compared as a function of state from these PSTHs. The "early" responses in 18 of 50 DSCT neurons did not differ in their group mean (±SE) magnitude between the states of wakefulness (1.07 ± 0.07 spikes/trial), quiet sleep (0.98 ± 0.06 spikes/trial), or awakening from active sleep (1.01 ± 0.08 spikes/trial; P > 0.05, repeated-measures ANOVA). Overall, for these same 18 neurons, the short latency response during active sleep (0.79 ± 0.09 spikes/trial) decreased by 26% compared with the control state of wakefulness (P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). An example of a DSCT neuron responding to consecutive sciatic nerve stimuli during sleep and wakefulness is presented in Fig. 1. In contrast to response magnitude, the latency of the early response did not significantly differ between wakefulness (3.27 ± 0.15 ms), quiet sleep (3.37 ± 0.23 ms), active sleep (3.43 ± 0.24 ms), or awakening from active sleep (3.48 ± 0.23 ms; P > 0.05, repeated-measures ANOVA).
The "late" responses were also analyzed during sleep and waking states. The group mean magnitude of the late response in 40 of 50 DSCT neurons did not differ between the states of wakefulness (3.01 ± 0.32 spikes/trial), quiet sleep (2.91 ± 0.29 spikes/trial), or awakening from active sleep (2.92 ± 0.32 spikes/trial; P > 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). Overall, the long latency response during active sleep (2.59 ± 0.31 spikes/trial) decreased 14% compared with the control state of wakefulness (P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). The onset of the "late" response did not differ significantly (P > 0.05) between wakefulness (23.7 ± 2.23 ms), quiet sleep (24.3 ± 2.17 ms), active sleep (23.7 ± 1.93 ms), or awakening from active sleep (23.8 ± 3.31 ms).
Hence, experimentally evoked monosynaptic and polysynaptic transmission between primary afferent terminals and identified DSCT neurons is significantly decreased specifically during the state of active sleep.
General characteristics of eye movement events during active sleep
Eye movement events hallmark the state of active sleep. The mean number of eye movement events, individual event duration, and cumulative eye movement event time across a total of 65 active sleep episodes measured 15.3 ± 1.3 events, 2.3 ± 0.1 s, and 40.4 ± 7.4 s, respectively. The mean total active sleep duration measured 449.2 ± 63.8 s. Finally, the density of eye movement events detected during the state of active sleep in this study measured 9.6% of this behavioral state.
Modulation of sciatic nerve-evoked responses of DSCT neurons during eye movement events of active sleep
Since the density of eye movement events during active sleep amounted to 10% of this entire state, any suppression of synaptic transmission between primary afferents and the recorded DSCT neurons reported above using primary PSTH analyses may have been masked by the proportionally larger number of responses averaged during non–eye movement portions of the active sleep state. To investigate this possibility, we analyzed the response magnitude and latency-to-onset data of early and late responses derived from PSTHs of the same sciatic nerve-evoked DSCT neurons before, during, and after individual eye movement events of active sleep using secondary subroutines developed in-house. An average number of 41 ± 6 stimulus trials was used to construct these cumulative PSTHs across each state and around eye movement events. PSTHs were compared on an equal trial basis with PSTHs constructed from data in the preceding states of wakefulness, quiet sleep, or awakening from active sleep.
The group mean (±SE) magnitude of the early sciatic responses did not differ between the states of wakefulness, quiet sleep, or awakening from active sleep (Fig. 3A; P > 0.05, repeated-measures ANOVA). However, during the non–eye movement events of active sleep, the group mean response magnitude decreased by 25.8% before and 24.3% after eye movement events (Fig. 3A; P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). During eye movement events of active sleep, the early sciatic nerve-evoked response magnitude decreased further by 14.4% (Fig. 3A; P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method), which corresponds to an overall 36.4% decrease in the response magnitude compared with response magnitude during the preceding state of wakefulness (Fig. 3A; P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method).
|
|
Figure 4 indicates that the late response in 15 of 40 (37.5%) DSCT neurons underwent >25% suppression in response magnitude, while 2 of 40 (5%) cells underwent >25% facilitation before eye movement events of active sleep. During the eye movement events of active sleep, response magnitude decreased >25% in 23 of 40 (57.5%) of the DSCT neurons. Only one neuron (2.5%) underwent >25% facilitation during eye movement events (Fig. 4). The dashed curve in Fig. 4 indicates that the suppression of the late response increased during eye movement periods of active sleep for the entire population of 40 neurons. The latency-to-onset of the late response did not differ across the sleep wake cycle or around eye movement events of active sleep (Fig. 3B; P > 0.05, repeated-measures ANOVA).
Overall, the results indicate that peripheral nerve-evoked activity of identified DSCT neurons is significantly attenuated during eye movement events of active sleep compared with non–eye movement periods of active sleep. To investigate if the extent of eye movement–related suppression was related to the amount of suppression of the sciatic nerve-evoked activity that occurs during non–eye movement active sleep relative to wakefulness, these two ratios were plotted against each other (Fig. 5). There was no significant relationship (correlation coefficient: –0.23; P > 0.05, Spearman rank-order correlation) identified between these two parameters. Thus, in this study, it is unlikely that ongoing inhibition during non–eye movement portions of active sleep masks the neuronal suppression that DSCT neurons undergo during eye movement events of active sleep.
|
The GABAA receptor antagonist bicuculline was assessed for its ability to block the suppression of DSCT neuron activity around eye movements during active sleep. This was accomplished in 10 DSCT neurons that responded to sciatic nerve stimulation; 4 of these responded with a short (early) latency response, and the remaining 6 responded with a longer (late) latency response. The experimental paradigm taken in these particular experiments also required monitoring each cell's activity over a minimum of at least two consecutive sleep cycles. The first sleep cycle served as a control to establish baseline levels of active sleep-specific neuronal suppression. The second sleep cycle served as a test to assess active sleep-related suppression of DSCT neuron activity in the presence of the sustained release of bicuculline. During the second sleep cycle, bicuculline ejection currents (67.8 ± 7.7 nA, 15.4 ± 2.7 min, n = 10 neurons) were first adjusted for each neuron by confirming the drug's ability to block the suppressive actions of GABA pulsed at regular 10 s intervals (46.2 ± 9.2 nA, n = 10 neurons) on sciatic nerve-evoked responses. Once this was confirmed (Fig. 6), the ejection current applied to the bicuculline barrel for each cell was maintained into and throughout the state of active sleep and subsequent awakening from active sleep.
|
|
Similarly, the magnitude of the late response recorded from six other DSCT neurons decreased in the first (control) sleep cycle by 27.8, 41.5, and 23.7% before, during, and after eye movement events, respectively, compared with the preceding episode of wakefulness (Fig. 7A; P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). The additional 13.7% suppression of the late response observed during the eye movement events of active sleep was also statistically significant from that observed before eye movement events of active sleep (Fig. 7A; P < 0.05, repeated-measures ANOVA, Student-Newman-Keuls method). During the second sleep cycle, in the presence of bicuculline (69.3 ± 12.2 nA, 15.3 ± 4.0 min), suppression of the late response in the same six DSCT neurons was also abolished (Fig. 7A; P > 0.05, repeated-measures ANOVA).
The latency of the late response did not differ between behavioral state in the first (control) and second (test-bicuculline) sleep cycles (Fig. 7B; P > 0.05, repeated-measures ANOVA).
These data suggest that the inhibitory neurotransmitter GABA participates in the active sleep-specific suppression of early and late synaptic responses of DSCT neurons evoked by low-intensity sciatic nerve stimulation. GABA mediates the tonic suppression during non–eye movement events as well as the increased phasic suppression during eye movement events of active sleep.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Nature of sciatic nerve-evoked responses of DSCT neurons in the chronic preparation
In this study, sciatic nerve stimulation resulted in the synaptic activation of DSCT neurons that was characterized by short latency (early) and long latency (late) spike discharges. The early spike response can be attributed to a monosynaptic linkage of large diameter group I primary afferents with recorded DSCT neurons (Eccles et al. 1961). Indeed, intracellular records published from earlier studies indicate that the latencies of peripheral nerve-evoked DSCT neuronal action potentials crowning monosynaptic excitatory postsynaptic potentials (EPSPs) are similar to the latencies of our spikes recorded extracellularly in response to sciatic nerve stimuli (Curtis et al. 1986; Eccles et al. 1961; Eide et al. 1969; Xi et al. 1997). The late response is likely due to polysynaptic inputs from a variety of receptors via cutaneous and muscular afferent fibers and corresponds closely in terms of onset latency to the type III responses of earlier investigations performed in anesthetized acute animal preparations (Eccles et al. 1961; Eide et al. 1969; Mann 1973; Osborn and Poppele 1983, 1988; Walmsley 1989).
The silent period that often precedes polysynaptic responses is reminiscent of studies performed by earlier investigators where a similar pause occurs in spike activity of DSCT neurons following peripheral nerve stimulation. This pause has been attributed to inhibitory postsynaptic potentials (IPSPs) set up by local circuit or extrasegmental interneurons (Knox et al. 1977).
Despite the similarities in peripheral nerve-evoked response patterns of DSCT neurons between earlier studies using acute anesthetized animal prepartions and ours, important differences also exist. Most importantly, the neural distortion caused by depressant actions of anesthetics is absent in this study. Indeed, barbiturate anesthetics significantly reduce response magnitude and increase the latency-to-onset of early responses in DSCT neurons by 23 and 28%, respectively (Soja et al. 2002). Moreover, as shown in this report, the response magnitude is clearly dependent on behavioral state.
State-dependent suppression of sciatic nerve-evoked responses
In the mid-1960s, Carli et al. (1966, 1967a,b) first suggested that sensory transmission was reduced during active sleep. They reported that hindlimb nerve-evoked responses recorded as a population or as field potentials from ipsilateral T12 ventrolateral quadrants or the cerebellum were occasionally reduced in magnitude during phasic rapid eye movement (REM) episodes of active sleep as opposed to other states such as wakefulness, quiet sleep, or tonic non-REM periods of active sleep (Carli et al. 1966, 1967a,b). These investigators were unable to unequivocally determine if the modulation of evoked responses during active sleep was due to changes occurring specifically in the input from dorsal spinocerebellar tract neurons, per se, and/or other ascending pathways. The data reported herein, using chronic single unit recording methodologies, clearly indicate that identified upper lumbar DSCT neurons constitute one ascending second order sensory pathway that contributes to the population response observed in earlier studies. The suppression of DSCT neuron responses to sciatic nerve stimulation we observed is different from earlier evoked potential studies in that evoked DSCT neuron activity undergoes phasic inhibition not only during eye movement episodes but also sustained tonic suppression throughout the non–eye movement portions of active sleep.
Any distinct change in the transfer of sensory input from primary afferents to the recorded neurons is masked if data during phasic eye movement periods are analyzed together with data from non–eye movement periods. The results presented clearly emphasize this point. For example, when the primary PSTH analysis method was used, which by default, included non–eye movement and eye movement periods, there was an overall 26% suppression of the early response. Similarly, when the secondary methods of PSTH analysis were used that selectively parceled out responses during non–eye movement periods of active sleep, nearly the exact same amount (25.8%) of suppression occurred. However, when response magnitude was analyzed around, i.e., before, during, and after eye movements in a paired fashion, further significant suppression (14.4%) of synaptic transmission occurred specifically during eye movement events of active sleep. This finding is characteristic for the vast majority of DSCT neurons recorded in this study and does not seem to be related to the degree of ongoing tonic suppression that is also present during non–eye movement portions of active sleep.
The source and neural pathway(s) conveying descending influences inhibiting DSCT neurons during non–eye movement and eye movement events of active sleep is not completely understood. The spinal inhibitory neurons recruited by sciatic nerve stimulation that have been suspected of mediating the pause that often precedes the late response of DSCT neurons recorded in the acute preparation (Knox et al. 1977) are unlikely to play a significant causative role since the latency of the late response did not vary as a function of state or after bicuculline administration.
Lumbar motoneurons are known to be tonically hyperpolarized during non–eye movement periods of active sleep by a process of glycine-mediated postsynaptic inhibition (Chase et al. 1989; Soja et al. 1991). The amplitude of monosynaptic EPSPs intracellularly recorded from motoneurons decreases during non–eye movement periods of active sleep compared with quiet sleep. During eye movement events, the EPSP amplitude is further decreased (Chandler et al. 1980; Morales and Chase 1981). During eye movement events of active sleep, the membrane potential activity of motoneurons is also characterized by the presence of complex phasic hyperpolarizations which have yet to be characterized pharmacologically (Chase and Morales 1982; Morales and Chase 1981).
It is possible that certain brain stem nuclei may relay descending influences to spinal neurons during eye movement events. In this regard, reticulospinal neurons backfired from the L3 spinal cord segment display firing rates reciprocal to DSCT neurons during sleep and wakefulness, being 10-fold higher during non–eye movement events of active sleep compared with preceding episodes of wakefulness or quiet sleep. During eye movements of active sleep, reticulospinal neuron discharge is increased another twofold (Soja et al. 2003; Taepavarapruk et al. 2003a; see also Wyzinski et al. 1978).
Irrespective of exactly how reticulospinal neurons are coupled to recorded cells (i.e., direct inhibitory connections and/or excitatory connections with inhibitory interneurons that, in turn, impinge on target neurons or afferent terminals), sciatic nerve-evoked monosynaptic responses such as the early responses recorded from DSCT neurons in this study or monosynaptic EPSPs recorded from motoneurons in the previous studies (Chandler et al. 1980; Morales and Chase 1981), undergo increased phasic suppression during eye movement events. Perhaps parallel interneuronal circuits are recruited by reticulospinal neurons to suppress sensory inflow via the DSCT as well as motor outflow. Such a divergence of reticulospinal inhibitory pathways would potentially obtund ascending sensory inputs and movements that might compromise the continuance of the active sleep state. However, further studies are warranted to investigate this possibility.
Another key issue our data addresses is whether active sleep-specific suppression of the sciatic nerve-evoked responses in DSCT neurons can be adequately explained by withdrawal of afferent input, i.e., disfacilitation, that is secondary to motor atonia and the hyperpolarization of motoneurons (Chase et al. 1989; Morales and Chase 1981; Soja et al. 1991). Disfacilitation is unlikely to account for the reduced synaptic responses in this study because sciatic nerve stimulation bypasses peripheral sensory receptors. Moreover, the magnitude of the monitored incoming volley in several experiments where the sciatic nerve-evoked field potential was recorded in Clarke's column (unpublished data) did not change during the sleep/wake cycle. Hence, active sleep-related inhibition directed toward the early monosynaptic response likely occurs in the spinal cord between the central terminals of large diameter primary afferents and the recorded DSCT neurons. The site of active sleep-specific suppression for the late responses may occur at presynaptic afferent terminals, and/or postsynaptically on last-order interneurons conveying the late responses to DSCT neurons, and even the recorded cells themselves.
State-dependent inhibition of DSCT neurons by GABA
Recently, we reported that the GABAA receptor antagonist bicuculline and the glycine antagonist strychnine blocked the active sleep-specific suppression of spontaneous spike activity, as well as the excitatory amino acid-driven responses of DSCT neurons, suggesting that GABA and glycine pre- and/or postsynaptically inhibit sensory inflow via the DSCT (Taepavarapruk et al. 2002). We hypothesized that supraspinal systems are activated during active sleep and inhibit, via GABAergic or glycinergic neurons, second-order DSCT neurons. Microdialysis experiments have also shown an increased efflux of these inhibitory neurotransmitters in Clarke's column during active sleep (Taepavarapruk et al. 2003b). Accordingly, appropriate antagonists delivered juxtacellularly to the recorded cell should block the reduction in response magnitude during active sleep. Indeed, using a multiple sleep cycle recording paradigm, we clearly showed that each cell was suppressed during active sleep in a control sleep cycle. During the second test sleep cycle, where the same neuron was exposed to the microiontophoretically administered GABAA receptor antagonist bicuculline, active sleep-related tonic suppression during non–eye movement periods and further phasic suppression during eye movement events did not occur. Therefore, on the basis that bicuculline is considered to be a competitive antagonist at GABAA receptors (Curtis et al. 1971; Krnjevic et al. 1977), we propose that GABA mediates the tonic suppression of sciatic nerve-evoked responses of DSCT neurons occurring during non–eye movement portions of active sleep as well as the phasic increased inhibition that occurs during the eye movement events characteristic of this state.
The release of GABA from, as yet unidentified, last-order neurons is not known and could occur indirectly on the afferent pathways. Anatomical substrates for presynaptic inhibition in Clarke's column have been identified, and GABA has been implicated in the presynaptic inhibition of group I afferents in the spinal cord (Curtis et al. 1986; Walmsley et al. 1987). The further reduction in the amplitude of monosynaptic EPSPs recorded from motoneurons during eye movement events of active sleep is thought to be due to presynaptic inhibitory processes since the EPSP decay phase did not differ before or during eye movement events of active sleep (Chandler et al. 1980).
In the mid-1960s, Morrison and Pompeiano employed Wall's technique (Wall 1958), an index of presynaptic inhibition, and showed that the terminal arbors of large diameter afferents depolarize [primary afferent depolarization (PAD)] during active sleep (Morrison and Pompeiano 1966; Pompeiano et al. 1967). Whether PAD of group I afferent fibers that terminate in Clarke's column occurs during active sleep is not known. Recent dual PAD experiments in acute animal preparations performed by Rudomin et al. have shown that segmental and ascending collaterals of the same muscle spindle afferent that, respectively, project to lower lumbar motoneuron pools and L3 Clarke's column, are independently and selectively controlled by impulse traffic in spinal or supraspinal PAD pathways (Lomeli et al. 1998; Rudomin 1999, 2002). Future work is needed to address whether selective PAD-like processes influence segmental and ascending collateral branches of primary afferent fibers during sleep versus wakefulness.
Functional significance of state-dependent inhibition of DSCT neurons
Malfunctioning GABA control mechanisms may adversely influence DSCT neurons in certain sleep disorders. For example, benzodiazepines effectively minimize the sensory dysesthesias associated with restless legs syndrome (RLS) and sensorimotor effects accompanying REM behavior disorder (RBD). Patients with RLS experience an imperative desire to move their lower legs because of paraesthesias or dysesthesias that occur at rest and this often delays sleep onset. Periodic limb movements often accompany RLS symptoms and worsen during active sleep in RLS patients (Patel et al. 1997; Trenkwalder 1998). Ineffective muscular atonia mechanisms and/or excess facilitatory drives onto motoneurons are thought to underlie the varying degrees of sensorimotor activation in RBD (Chase and Morales 1990; Schenck and Mahowald 2002). Taken together, impaired glycine and GABA-mediated inhibition that normally occurs in motor (Chase et al. 1989; Soja et al. 1991) and ascending sensory pathways such as Clarke's column DSCT neurons (Soja et al. 1993, 1996, 2001a,b; Taepavarapruk et al. 2002) may contribute to these clinical syndromes. Future electrophysiological studies are warranted to investigate the state-dependent regulation of GABAergic (and glycinergic) pathways inhibiting sensory inflow to higher brain centers under natural conditions and conditions emulating RLS and RBD.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: P. J. Soja, Faculty of Pharmaceutical Sciences, The Univ. of British Columbia, 2146 East Mall, Vancouver, BC V6T 1Z3, Canada (E-mail: psoja{at}interchange.ubc.ca).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Carli G, Diete-Spiff K, and Pompeiano O. Cerebellar responses evoked by somatic afferent volleys during sleep and waking. Arch Ital Biol 105: 499–528, 1967a.[Medline]
Carli G, Kawamura H, and Pompeiano O. Transmission of somatic sensory volleys through ascending spinal hindlimb pathways during sleep and wakefulness. Pfluegers Gesamte Physiol Menschen Tiere 298: 163–169, 1967b.
Chandler SH, Chase MH, and Nakamura Y. Intracellular analysis of synaptic mechanisms controlling trigeminal motoneuron activity during sleep and wakefulness. J Neurophysiol 44: 359–371, 1980.
Chase MH and Morales FR. The atonia and myoclonia of active (REM) sleep. Annu Rev Psychol 41: 557–584, 1990.[CrossRef][ISI][Medline]
Chase MH and Morales FR. Phasic changes in motoneuron membrane potential during REM periods of active sleep. Neurosci Lett 34: 177–182, 1982.[CrossRef][ISI][Medline]
Chase MH, Soja PJ, and Morales FR. Evidence that glycine mediates the postsynaptic potentials that inhibit lumbar motoneurons during the atonia of active sleep. J Neurosci 9: 743–751, 1989.[Abstract]
Curtis DR, Duggan AW, Felix D, and Johnston GA. Bicuculline, an antagonist of GABA and synaptic inhibition in the spinal cord of the cat. Brain Res 32: 69–96, 1971.[CrossRef][ISI][Medline]
Curtis DR, Gynther BD, and Malik R. A pharmacological study of group I muscle afferent terminals and synaptic excitation in the intermediate nucleus and Clarke's column of the cat spinal cord. Exp Brain Res 64: 105–113, 1986.[ISI][Medline]
Eccles JC, Oscarsson O, and Willis WD. Synaptic action of group I and II afferent fibres of muscle on the cells of the dorsal spinocerebellar tract. J Physiol 158: 517–543, 1961.
Eide E, Fedina L, Jansen J, Lundberg A, and Vyklicky L. Properties of Clarke's column neurons. Acta Physiol Scand 77: 125–144, 1969.[Medline]
Knox CK, Kubota S, and Poppele RE. A determination of excitability changes in dorsal spinocerebellar tract neurons from spike-train analysis. J Neurophysiol 40: 626–646, 1977.
Krnjevic K, Puil E, and Werman R. Bicuculline, benzyl penicillin, and inhibitory amino acids in the spinal cord of the cat. Can J Physiol Pharmacol 55: 670–680, 1977.[Medline]
Lomeli J, Quevedo J, Linares P, and Rudomin P. Local control of information flow in segmental and ascending collaterals of single afferents. Nature 395: 600–604, 1998.[CrossRef][Medline]
Mann MD. Clarke's column and the dorsal spinocerebellar tract: a review. Brain Behav Evol 7: 34–83, 1973.[ISI][Medline]
Morales F and Chase MH. Postsynaptic control of lumbar motoneuron excitability during active sleep in the chronic cat. Brain Res 225: 279–295, 1981.[CrossRef][ISI][Medline]
Morrison AR and Pompeiano O. Depolarization of central terminals of group Ia muscle afferent fibres during desynchronized sleep. Nature 210: 201–202, 1966.[Medline]
Osborn CE and Poppele RE. Cross-correlation analysis of the response of units in the dorsal spinocerebellar tract (DSCT) to muscle stretch and contraction. Brain Res 280: 339–342, 1983.[CrossRef][ISI][Medline]
Osborn CE and Poppele RE. The extent of polysynaptic responses in the dorsal spinocerebellar tract to stimulation of group I afferent fibers in gastrocnemius-soleus. J Neurosci 8: 316–319, 1988.[Abstract]
Patel R, Kerr R, and Maxwell DJ. Absence of co-localized glutamic acid decarboxylase and neuropeptides in noradrenergic axons of the rat spinal cord. Brain Res 749: 164–169, 1997.[CrossRef][ISI][Medline]
Pompeiano O, Carli G, and Kawamura H. Transmission of sensory information throug ascending spinal hindlimb pathways during sleep and wakefulness. Arch Ital Biol 105: 529–572, 1967.[Medline]
Rudomin P. Presynaptic selection of afferent inflow in the spinal cord. J Physiol 93: 329–347, 1999.
Rudomin P. Selectivity of the central control of sensory information in the mammalian spinal cord. Adv Exp Med Biol 508: 157–170, 2002.[ISI][Medline]
Schenck CH and Mahowald MW. REM sleep behavior disorder: clinical, developmental, and neuroscience perspectives 16 years after its formal identification in Sleep. Sleep 25: 120–138, 2002.[ISI][Medline]
Soja PJ, Fragoso MC, Cairns BE, and Jia WG. Dorsal spinocerebellar tract neurons in the chronic intact cat during wakefulness and sleep: analysis of spontaneous spike activity. J Neurosci 16: 1260–1272, 1996.
Soja PJ, Fragoso MC, Cairns BE, and Oka JI. Dorsal spinocerebellar tract neuronal activity in the intact chronic cat. J Neurosci Methods 60: 227–239, 1995.[CrossRef][ISI][Medline]
Soja PJ, Lopez-Rodriguez F, Morales FR, and Chase MH. The postsynaptic inhibitory control of lumbar motoneurons during the atonia of active sleep: effect of strychnine on motoneuron properties. J Neurosci 11: 2804–2811, 1991.[Abstract]
Soja PJ, Oka JI, and Fragoso M. Synaptic transmission through cat lumbar ascending sensory pathways is suppressed during active sleep. J Neurophysiol 70: 1708–1712, 1993.
Soja PJ, Pang W, Taepavarapruk N, Cairns BE, and McErlane SA. On the reduction of spontaneous and glutamate-driven spinocerebellar and spinoreticular tract neuronal activity during active sleep. Neuroscience 104: 199–206, 2001a.[Medline]
Soja PJ, Pang W, Taepavarapruk N, and McErlane SA. Spontaneous spike activity of spinoreticular tract neurons during sleep and wakefulness. Sleep 24: 18–25, 2001b.[Medline]
Soja PJ, Taepavarapruk N, and McErlane SA. Reticulospinal (RS) neuron activity around rapid eye movement (REM) events of active sleep. Soc Neurosci Abstr 2003. Online.
Soja PJ, Taepavarapruk N, Pang W, Cairns BE, McErlane SA, and Fragoso MC. Transmission through the dorsal spinocerebellar and spinoreticular tracts: wakefulness versus thiopental anesthesia. Anesthesiology 97: 1178–1188, 2002.[CrossRef][ISI][Medline]
Taepavarapruk N, McErlane SA, and Soja PJ. State-related inhibition by GABA and glycine of transmission in Clarke's column. J Neurosci 22: 5777–5788, 2002.
Taepavarapruk N, McErlane SA, and Soja PJ. Reciprocal state-dependent changes in reticulospinal (RS) and dorsal spinocerebellar tract (DSCT) neuron activity. Soc Neurosci Abstr 2003a. Online.
Taepavarapruk N, Taepavarapruk P, John J, Lai YY, Siegel JM, McErlane SA, and Soja PJ. State-dependent release of glycine and GABA in Clarke's column of the upper lumbar spinal cord. Sleep 26: A11, 2003b.
Trenkwalder C. Sleep dysfunction in Parkinson's disease. Clin Neurosci 5: 107–114, 1998.
, early response; 
, late response) of DSCT neurons during non–eye movement active sleep relative to wakefulness vs. change in response magnitude during eye movement events relative to non–eye movement active sleep. No significant relationship between the 2 parameters was identified (P > 0.05, Spearman rank-order correlation), indicating that ongoing neuronal suppression during the non–eye movement parts of active sleep did not mask the effect of eye movement events on sciatic nerve-evoked activity.
