| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Zoology, University of Cambridge, Cambridge, United Kingdom
Submitted 13 July 2006; accepted in final form 30 September 2006
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
; Santhakumar and Soltesz 2004). For example, variability in the excitatory synaptic drive has been associated with irregular spiking in spinal cord, thalamic, and cortical neurons (Blitz and Regehr 2003; Calvin and Stevens 1968; Harsch and Robinson 2000; Zador 1998), and simulation and experimental studies have suggested that changes in the variance of inhibitory synaptic inputs (independently of changes in the mean amplitude) can significantly affect simulated cellular and network activity (Aradi and Soltesz 2002; Aradi et al. 2004). While variability may seem at odds with reliable function, some degree of variability may be essential in any flexible or adaptive systems. Thus healthy physiological systems are intrinsically variable, with highly regular activity being associated with pathological states (Buchman 2002; Sterling 2004). Variability can also be exploited by the nervous system during development or learning to allow a system to sample the effects of different responses before settling on an optimal output (Barnes et al. 2005; Brezina et al. 2006).
Variability is common in motor systems. This has been accepted for some time with respect to voluntary movements (Newell and Corcos 1993; Harris and Wolpert 1998) and is now becoming recognized in rhythmically active motor systems where regular activity had been assumed to be the normal or optimal condition (Horn et al. 2004; Parker et al. 1998; Pearlstein et al. 2005; Szucs et al. 2005; Zhang and Grillner 2000). While variability may be a necessary component of any flexible system, mechanisms must exist to ensure that it does not reduce reliability and accuracy once a particular movement sequence has been selected (Harris and Wolpert 1998; Stein et al. 2005). Variability could be regulated by feedback or feedforward inputs to locomotor networks or by the intrinsic cellular or synaptic properties of network neurons.
This study examined variability in the lamprey spinal cord. Glutamatergic agonists can activate the network in the isolated spinal cord to evoke alternating bursts of activity in ventral roots on the left and right sides of the body ("fictive locomotion"; Cohen and Wallen 1980). Fictive activity is usually only analyzed when a regular network output is generated (Wallèn and Williams 1984). However, as in other rhythmically active motor systems (Barthe and Clarac 1997; Horn et al. 2004; Pearlstein et al. 2005; Szucs et al. 2005), the activity can show species-dependent irregularity (Parker et al. 1998; Wallèn and Williams 1984; Zhang and Grillner 2000).
Electrophysiological and ultrastructural analyses have been used here to examine cellular and synaptic variability in the lamprey spinal cord and its modulation by the neuropeptide substance P (Aradi and Soltesz 2002). The results show that there is neuron and synapse-specific variability in the spinal cord and that substance P can modulate this variability to make the excitatory drive to motor neurons more regular.
|
|
METHODS |
|---|
|
Pieces of spinal cord 
15 segments long were isolated from the notochord and placed ventral side up in a Sylgard-lined chamber and superfused with Ringer containing (in mM) 138 NaCl, 2.1 KCl, 1.8 CaCl2, 1.2 MgCl2, 4 glucose, 2 HEPES, and 0.5 L-glutamine. The Ringer was bubbled with O2, and the pH was adjusted to 7.4 with 1 M NaOH. The Ringer in the experimental chamber was kept at a temperature of 10–12°C.
Paired intracellular recordings were made from identified interneurons and motor neurons in the cell body region of the spinal cord using thin walled micropipettes filled with 3 M potassium acetate and 0.1 M potassium chloride (resistances of 60–80 M
). Motor neurons were identified by recording orthodromic extracellular spikes in associated ventral roots after current injection into their somata. Excitatory interneurons (EINs; Buchanan et al. 1989) and ipsilateral inhibitory interneurons (SiINs; Buchanan and Grillner 1988) were identified by their ability to elicit monosynaptic excitatory postsynaptic potentials (EPSPs) or inhibitory postsynaptic potentials (IPSPs), respectively, in ipsilateral motor neurons. Crossing inhibitory and excitatory interneurons (IScINs and EScINs, respectively) were identified by their ability to elicit monosynaptic IPSPs or EPSPs in contralateral motor neurons, respectively. The SiINs and ScINs could be distinguished from the lateral and crossed caudal interneurons because they had higher input resistances (49 ± 19 vs. 12.7 ± 2.5 M; see also Buchanan 2001), which presumably reflected their smaller cell bodies (Buchanan 1993; Buchanan and Grillner 1988; Ohta et al. 1991). These neurons also have relatively short axonal projections (Ohta et al. 1991), and therefore, unlike the lateral and crossed caudal interneurons, they did not evoke extracellular spikes in suction electrodes that covered the whole of the ipsilateral or contralateral region of the spinal cord 10 segments caudal to their cell bodies (Parker 2003a). Monosynaptic PSPs were identified by their reliability and constant latency after presynaptic stimulation at 20 Hz. An Axoclamp 2A amplifier (Axon Instruments) was used for voltage recording and current injection. In all experiments, the membrane potential in the presence of drugs was kept at a control level of –70 mV by injecting depolarizing or hyperpolarizing current using single electrode discontinuous current clamp (DCC; the mean resting membrane potentials for the motor neurons and EINs in this study were 71.3 ± 2 and 67.7 ± 1 mV, respectively). This was necessary to examine the synaptic variance independently of any potential changes in the membrane potential. DCC was also used to inject positive or negative current pulses to examine input resistances and excitability. Data were acquired, stored, and analyzed on computer using an A/D interface (Digidata 1200, Axon Instruments) and Axon Instruments software (pClamp 9).
The effects of substance P on synaptic properties should ideally be examined during ongoing locomotor network activity. However, it is extremely difficult to examine individual synapses in detail when networks are active (Manor et al. 1999; Parker, unpublished data). Cellular and synaptic variability was instead examined by repetitively activating network interneurons and motor neurons. The presynaptic stimulation protocol consisted of bursts of five spikes at 20 Hz delivered every 500 ms (this is a physiologically relevant stimulation pattern; Buchanan and Cohen 1982; Buchanan and Kasicki 1995). Only 20-Hz stimulation was examined here because preliminary analyses showed that there was no significant effect on synaptic variability at lower stimulation frequencies (5 and 10 Hz; Parker, unpublished observations). This matched the metaplastic effects of substance P on activity-dependent synaptic plasticity, which also only occurred with 20-Hz stimulation (Parker and Grillner 1999). The stimulation episode consisted of 50 bursts, beause effects over this number of bursts match the effects seen when more bursts are given (Parker 2000a). Individual action potentials in each burst were evoked by injecting 1-ms depolarizing current pulses of 10–60 nA to ensure a constant 50-ms interval between presynaptic spikes. This removed any variability in presynaptic spiking and thus allowed effects on synaptic variability to be examined in isolation. PSP amplitudes during spike trains were measured as the peak amplitude above (EPSP) or below (IPSP) the baseline immediately preceding the PSP. The degree of PSP summation varied in different connections, presumably depending on the number of PSPs that were evoked and their time-course (Parker 2003b). Where PSPs summate, this can alter the PSP amplitude because of changes in driving force and input resistance. We have previously corrected for these effects using the formula of McLachlan and Martin (1981) (Parker 2003b). Because this correction did not result in any significant changes in the measured synaptic responses (corrected EPSPs changed by only 0.05–0.1 mV) or alter the properties and plasticity of inputs during spike trains, we did not apply it here.
Synaptic variability was examined by calculating the variance of PSPs within bursts (intraburst variability; Fig. 1) and across bursts (interburst variability). The intraburst variability will influence the summed synaptic input within each burst and could thus alter the duration and pattern of spiking: this will in turn influence the frequency and regularity of the network output. The interburst variability was calculated by measuring the variance of the initial PSP amplitude in each burst (Fig. 1, A and B). This will influence the variability of the synaptic input over successive bursts, which could affect the onset of the burst and thus the interburst variability.
|
90% of control and facilitation as an increase to >110% of the control PSP amplitude over the train of five spikes. Connections in which the PSP amplitude did not fall below 90% or above 110% of the initial PSP amplitude during the train were defined as unchanged. If the PSP amplitude fluctuated above 110% or below 90% of the initial amplitude but there was no sustained change during the spike train, the connection was also classified as unchanged.
The variability of spiking was either examined during ongoing N-methyl-D-aspartate (NMDA)-evoked network activity or in response to depolarizing current pulses (0.5-5nA) designed to mimic rhythmic membrane potential depolarizations during network activity (50 current pulses of 100- or 500-ms duration given every 200 or 1,000 ms, respectively; Fig. 1C). This pattern of stimulation was designed to simulate repetitive activation during network activity at frequencies of 1 or 5 Hz (Buchanan and Cohen 1982; Buchanan and Kasicki 1995; i.e., within the frequency range over which the effects of substance P were examined on network activity; Parker et al. 1998). Only cells with resting potentials of less than –50 mV and overshooting action potentials were examined in these experiments to avoid potential effects on spiking caused by damage to the cell body (10 µm; Buchanan et al. 1989). Because motor neurons and EINs were found most often, it was only in these cells that a sufficient sample size matching these criteria was obtained (synaptic inputs from excitatory and inhibitory interneurons were examined if the presynaptic action potential did not match these criteria, because action potentials that propagated to the terminal to evoke transmitter release should not have been affected by damage at the cell body). The lowest resting membrane potential of a cell used in this study was –57 mV. Cells current clamped to –70 mV from different membrane potentials did not show any differences in their spiking properties or the effects of substance P. Two measures of spike regularity were used (Harsch and Robinson 2000): the coefficient of variation of the interspike intervals (CV; SD of interspike interval/mean interspike interval), and the Fano factor. The Fano factor is the ratio of the variance of the spike interval to the mean spike count over fixed time windows (these time windows were 5 ms for spiking in response to depolarizing current pulses and 50 ms for spiking during network activity where the frequency of spiking was lower). The reliability of spiking was defined as the proportion of spikes that were repeatable within a window over repeated responses (i.e., a spike was very reliable if it occurred consistently within the same time window over repeated trials). The precision of spiking was determined by calculating the SD of the timing of reliable spikes (i.e., the fluctuation of the spike occurrence when spikes occurred repeatedly in the time window; Harsch and Robinson 2000). Values were obtained by measuring 50 bursts during network activity or in response to current injection.
Electron microscopy
Changes in synaptic ultrastructure were examined in 63 synapses in three animals. Two pieces of spinal cord were taken from each animal. One piece was treated by exposure to substance P (1 µM) for 10 min (Parker et al. 1998). Control and treated pieces of spinal cord taken 30 min after substance P application (this time was chosen to match the duration of the electrophysiological recordings) were transferred to 3% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) and fixed at 4°C overnight. The samples were washed in 0.1 M phosphate buffer (pH 7.4) and postfixed in buffered 1% OsO4 for 1 h at 4°C. After washing in 0.1 M phosphate buffer, the samples were dehydrated through an ethanol series, transferred to propylene oxide for two 15-min washes, and embedded in araldite CY212 (Agar Scientific). Ultrathin sections were cut in cross-section on a Reichert OmU2 microtome and mounted on uncoated copper mesh grids. The sections were counterstained with uranyl acetate and lead citrate and visualized using a Phillips EM 300 electron microscope. Each section contained the cell body region of the spinal cord. No attempt was made to label specific interneurons because this prevents the analysis of presynaptic properties (Buchanan et al. 1989). Micrograph negatives were scanned at 1,200 dpi using an Epson Perfection 2450 Photo scanner. The scanned images were imported into Adobe Photoshop for the analysis of the diameter of vesicle profiles in different synapses (Colliver et al. 2000). Round vesicles were assumed to be present at excitatory synapses and flattened/pleomorphic vesicles present at inhibitory synapses (Peters et al. 1991). The analysis examined docked vesicles (i.e., those in contact with the presynaptic active zone) and nondocked vesicles (i.e., those located in the vesicle cluster around the release site but which were not in contact with the presynaptic active zone). Random sample sections were analyzed blind a second time to ensure that there was no bias in the analysis of properties in control and treated sections.
Drug application
Locomotor network activity was evoked by continuous bath application of 50–200 µM NMDA (Parker et al. 1998). NMDA-evoked network activity can differ markedly in its frequency and regularity in different experiments (Parker et al. 1998; Zhang and Grillner 2000), and the pattern can change spontaneously, particularly within the first hour after NMDA application (Parker et al. 1998). Experiments only began when the pattern of activity had been stable for 1 h (Parker et al. 1998). Substance P was applied for 10 min at a concentration of 1 µM in all experiments (Parker et al. 1998). Because of its long-term effects, substance P was only applied once to each piece of spinal cord. Glutamate receptor antagonists and altered Ringer calcium solutions (Parker 2000a) were applied for 10 min before substance P application. Protein kinase inhibitors and the slow intracellular calcium chelator EGTA-AM were applied for 1 h before substance P application to ensure that they had entered the cell and were evoking their effects (Parker 2000a). It was not possible to record from the interneurons long enough to compare effects in single cells in control and 1 h after protein kinase inhibitors or EGTA-AM application. After 1 h, new connections were found in which to examine the effects of substance P in the presence of these drugs. Statistical significance was examined using paired or unpaired t-test, one-way ANOVA, or 
2. The text states the proportion of experiments in which individual effects occurred (all cells were included in the statistical analyses of a particular effect whether a change occurred or not). The bar graphs summarize the data obtained on all cells in a particular analysis, again whether an effect occurred or not. Differences in variance were examined using the F-test (Aradi and Soltesz 2002). All drugs were obtained from Sigma or Calbiochem.
|
|
RESULTS |
|---|
|
) and increased the frequency and amplitude of subthreshold locomotor-related depolarizations in motor neurons (Fig. 2, Ai, Aii, and B; P < 0.05, n = 14). In addition, it reduced the variance of the locomotor-related depolarization amplitude (P < 0.05, F-test; n = 14; Fig. 2B) and improved the regularity of spiking in motor neurons (Fig. 2, Ci–Ciii). The improved regularity of spiking was shown by a reduction in the interspike CV (Fig. 2Di) and Fano factor (Fig. 2Dii) and by the improved reliability and precision of individual spikes (Fig. 2, Diii and Div). These effects were not dependent on NMDA concentration or the initial frequency of network activity, because they occurred with 50–200 µM NMDA and at initial burst frequencies of 0.2–2.4 Hz (data not shown).
|
), but it failed to improve the regularity of spiking (Fig. 3, A and B).
|
). The firing properties of different EINs vary (Buchanan 1993): cells could adapt (i.e., firing was terminated before the end of the current pulse, n = 8 of 28), not adapt (firing was maintained at or close to the initial frequency, n = 16 of 28), or stutter (firing was interrupted during the current pulse, n = 4 of 28; data not shown). The EINs in which the effects of substance P were examined either adapted or were nonadapting. In both types of EIN, substance P usually increased the excitability (n = 7 of 9; Fig. 3, C and D), reduced the spike variability (n = 7 of 9; Fig. 3, C–E), and improved the reliability and precision of individual spikes (n = 7 of 9; Fig. 3F). The reduction of the interspike variability was greater with lower current pulses (Fig. 3G) and was more pronounced over the later part of the current pulse where variability was greatest in control (Fig. 3, C and D). The effects on EIN excitability and spike variability were unchanged for as long as recordings could be maintained (
40 min after washing off substance P; Fig. 3C). Because the variability of spiking was less over the initial part of the current pulse in control and in substance P (i.e., where excitability is higher; Fig. 3C), the reduced variability on the latter part of the current pulse after substance P application could simply have reflected the increase in excitability. That the changes in excitability alone could not account for the effects on the spike regularity is shown for the 2-nA current pulse in Fig. 3C. In this case, the excitability was not altered by substance P, but the regularity of spiking was improved.
Substance P increased the input resistance in a proportion of the EINs (from 15 ± 10 to 22 ± 12 M, n = 6 of 14; data not shown), but it did not usually affect the input resistance in motor neurons (n = 20 of 22; data not shown). An increase in input resistance could potentially have influenced the effects of substance P on the EIN excitability and spike variability. However, in the seven cells in which the excitability and interspike variability were modulated, the input resistance was only increased in three cells. It could thus not account for the effects on spiking in all cases. The increase in input resistance also meant that a 2-nA current pulse after substance P application should result in the same level of depolarization as a 3-nA current pulse in control (40–45 mV; Fig. 3G). However, despite the similar depolarization in both cases, the interspike variability was still lower for the 2-nA current pulse in substance P (Fig. 3G). This further suggested that the reduced interspike variability was not dependent on the change in input resistance.
Several cellular mechanisms could account for the effects of substance P on the increase in EIN excitability and the reduction in spike variability (see DISCUSSION). Only one of these mechanisms was examined here. The EINs have a slow afterhyperpolarization after the action potential that is mediated by a calcium-dependent potassium conductance (KCa; Buchanan 1993). This conductance was examined because it is suggested to affect locomotor network activity (Grillner et al. 1995). Substance P did not significantly affect the amplitude or duration of the slow afterhyperpolarisztion evoked in the EINs after single action potentials evoked at 0.1 Hz (n = 8, P > 0.05; Fig. 3H), suggesting that its cellular effects were not caused by the modulation of the KCa conductance.
Effects on synaptic properties
The effects of substance P on EIN spiking could have contributed to the reduction in the variability of the excitatory drive to motor neurons and their spiking during network activity. In addition, direct changes in the variability of synaptic inputs could also have influenced the variability of the synaptic drive to motor neurons. While inhibitory and excitatory synaptic inputs in the lamprey are reliable (i.e., they do not usually fail), they do fluctuate during spike trains (Fig. 4, A and B). This fluctuation could alter the integration of synaptic inputs and thus the net depolarization in motor neurons. Synaptic variability was examined here by measuring the variance of PSPs from inhibitory and excitatory interneurons in motor neurons across and within repeated spike trains (interburst and intraburst variance, respectively; see METHODS; Fig. 1, A and B).
|
The effects of substance P on the synaptic variance were associated with synapse-specific effects on the PSP amplitudes (Fig. 4E). Because the synaptic variance was positively related to the PSP amplitude at each connection (Fig. 5, Ai–Aiv and legend for EIN- and SiIN-evoked inputs), changes in the PSP amplitude could have influenced the effects of substance P on the synaptic variance. This effect thus had to be considered. The amplitude of EIN-evoked EPSPs was significantly increased by substance P (P < 0.05, n = 15 of 28; Fig. 4E). However, this could not account for the reduced EPSP variability because, if anything, an increase in the EPSP amplitude should have increased the variance (Fig. 5, Ai and Aii). Substance P did not significantly affect the amplitude of EScIN-evoked EPSPs (n = 8, P > 0.05; Fig. 4F) or SiIN-evoked IPSPs (P > 0.05, n = 16; Fig. 4F), and thus changes in PSP amplitude cannot account for the effects on the synaptic variance at these connections. However, the significant reduction of the IScIN-evoked IPSP amplitude by substance P (P < 0.05, n = 4 of 6; Fig. 4F) could have contributed to the marked reduction in variance at this connection.
|
Substance P thus had synapse-specific effects on the amplitude and variance of monosynaptic EPSPs and IPSPs. Because the EINs were found most often during experiments, this connection was used to examine the modulation of the synaptic variability in detail.
EIN-evoked EPSPs in motor neurons exhibit various forms of activity-dependent plasticity (Parker 2003b). Activity-dependent changes in the EPSP amplitude during the spike train will alter the PSP amplitude and could thus contribute to the intraburst synaptic variance. Substance P can alter the activity-dependent properties of EIN-evoked EPSPs to evoke metaplasticity (Parker and Grillner 1999). Because changes in the EPSP amplitude could alter the intraburst variability (Fig. 5, Ai and Aii), the influence of activity-dependent plasticity and metaplasticity on the modulation of the synaptic variance had to be examined.
The variance of EIN inputs to motor neurons (from Parker 2003b) was 0.22 ± 0.01 mV at depressing connections (n = 103), 0.22 ± 0.03 mV at facilitating connections (n = 61), and 0.16 ± 0.02 mV at unchanged connections (n = 26). The intraburst synaptic variance was significantly lower at unchanged connections (P < 0.05, F-test), but there was no significant difference at depressing and facilitating connections (Fig. 5B). As expected, the depression or facilitation of inputs during spike trains increased the synaptic variance. The main effect of substance P on activity-dependent plasticity was to convert unchanged or depressing connections into facilitating connections (U-F and D-F, respectively; n = 10 of 16, 62% of connections; Fig. 5C). From the relationship of synaptic variance to activity-dependent plasticity (Fig. 5B), this should either have increased (U-F connections) or had no significant effect on the synaptic variance (D-F connections). The metaplastic effects of substance P thus could not account for the reduced intraburst variability, suggesting that substance P was acting directly on mechanisms that influenced the synaptic variability.
Mechanisms underlying the reduced synaptic variance
Insight into the mechanisms underlying changes in cellular and synaptic properties can help to link these effects to changes in the network output (Bevan and Parker 2004). The mechanisms underlying the reduced variance of EIN-evoked EPSPs were thus examined.
The network effects of substance P on the burst frequency and burst regularity modulation are triggered by different intracellular pathways [protein kinase C (PKC) and protein kinase A (PKA), respectively] (Parker et al. 1998). The second messenger pathways underlying the effects of substance P on the synaptic variance were also examined. The PKA and PKG antagonist H8 (10 µM, n = 6; Parker et al. 1998) and the specific PKA antagonist RpcAMPs (10 µM, n = 5; data not shown) did not significantly affect the intraburst and interburst synaptic variance (P > 0.05; Fig. 6Ai). However, substance P did not usually reduce the synaptic variance in the presence of these antagonists (n = 5 of 6 in H8; Fig. 6, Ai and Aii; n = 5 of 5 in RpcAMPS, n = 5; data not shown). The number of connections in which the synaptic variance was reduced in PKA antagonists was significantly reduced compared with control (P < 0.05, 2 test), which suggested that the effect was PKA dependent.
|
) also failed to significantly affect the intraburst and interburst variance (mean intraburst synaptic variance 0.16 in control to 0.15 mV in chelerythrine, and the interburst synaptic variance from 0.17 in control to 0.19 mV in chelerythrine; P > 0.05, n = 12; data not shown). However, in chelerythrine, substance P reduced the synaptic variance in 6 of 12 experiments (from 0.15 and 0.19 mV to 0.11 and 0.14 mV; data not shown). The proportion of connections in which the synaptic variance was reduced in chelerythrine and substance P did not differ significantly to the proportion of experiments in which the variance was reduced by substance P alone (P > 0.05, 2; data not shown), which suggested that the effect was not dependent on PKC.
The network effects of substance P are calcium dependent (Parker et al. 1998). The effects of calcium on the synaptic variability were examined using the slow intracellular calcium chelator EGTA-AM (10 µM; Parker et al. 1998). EGTA-AM consistently and significantly reduced the amplitude of synaptic inputs during spike trains (because of a reduction in the calcium-dependent vesicle replenishment; Parker 2000a) and increased the intraburst and interburst synaptic variance (from a mean of 0.24 and 0.28 mV in control to 0.52 and 0.49 mV in EGTA-AM, respectively; P < 0.05, n = 7 of 7; Fig. 6Bi). Substance P did not significantly affect the intraburst and interburst synaptic variance in EGTA-AM (n = 6 of 7; Fig. 6, Bi and Bii), which suggested a potential calcium-dependent component to the modulation.
Substance P pre- and postsynaptically potentiates glutamatergic synaptic transmission. The postsynaptic effect is caused by the specific modulation of NMDA receptors (Parker and Grillner 1998). The role of glutamate receptors in the synaptic variance modulation was examined by applying substance P in the presence of glutamate receptor antagonists. Blocking non-NMDA glutamate receptors with DNQX (10 µM; Parker et al. 1998) did not significantly affect the variance of EIN-evoked EPSPs (P > 0.05, n = 11; data not shown). In DNQX, substance P significantly reduced the mean intraburst synaptic variance from 0.16 to 0.08 mV and the interburst synaptic variance from 0.17 to 0.08 mV (n = 6 of 11 connections; P < 0.05; data not shown). This proportion of number of connections did not differ significantly to the number of connections in which the synaptic variability was reduced in the absence of DNQX (P > 0.05, 2), which suggested against a role for non-NMDA receptors in the modulation of the synaptic variance. Blocking NMDA receptors with AP5 (100 µM; Parker et al. 1998) also did not significantly affect the intraburst or interburst synaptic variance (n = 6 of 7; P > 0.05; Fig. 6Ci). However, substance P failed to reduce the synaptic variance in any connection in AP5 (n = 7, P < 0.05, 2; Fig. 6, Ci and Cii), which suggested that the effect was NMDA dependent.
The NMDA dependence of the reduced synaptic variance was examined further by studying the effects of substance P on postsynaptic depolarizations evoked by pressure application of glutamate or NMDA. Pressure application of NMDA and glutamate was performed in TTX (1.5 µM) to ensure that the depolarizations were caused by direct postsynaptic effects on the cell that was recorded from. Substance P increased the amplitude of NMDA and glutamate-evoked depolarizations and significantly reduced the variance of the depolarization amplitude (glutamate: n = 4 of 5; NMDA: n = 4 of 5; F-test, P < 0.05; Fig. 7, A and B). The effects of substance P were also examined on TTX-resistant spontaneous miniature EPSPs (mEPSPs), which are assumed to reflect postsynaptic responses to the contents of single spontaneously released synaptic vesicles. Substance P increased the amplitude and significantly reduced the variance of the mEPSP amplitude (0.024 mV in control and 0.011 mV in substance P, P < 0.05, F-test; Fig. 7, C and D).
|
), and could either reflect differences in the proportion of the total vesicle content released from single vesicles (Klingauf et al. 1998) or differences in vesicle size (a twofold difference in vesicle size could result in an eightfold difference in transmitter content and in postsynaptic response). A role for quantal variability was examined using electron microscopy to measure the diameter of synaptic vesicle profiles in control cords and in cords 30 min after substance P application. As in other systems (Harris and Sultan 1995), vesicle diameters varied considerably in individual terminals. While there was no significant difference in the median diameter of either docked or nondocked vesicles at asymmetric synapses that contained round vesicles (assumed to be glutamatergic synapses; Peters et al. 1991) in control (Fig. 8A), there was a significant decrease in the variance of vesicle diameters in substance P–treated terminals. For docked vesicles, the variance decreased from 2.88 nm in control to 2.14 nm in substance P (P < 0.001, F-test), and for undocked vesicles from 2.84 to 2.54 nm (P = 0.007, F-test; Fig. 8, A, C, and D). In contrast, substance P had no significant effect on the median diameter or variance of vesicle diameters at symmetrical synapses that contained flattened or pleomorphic vesicles (assumed to represent inhibitory synapses; Peters et al. 1991). The reduction in the variance of synaptic vesicle diameters by substance P was thus specific to excitatory synapses.
|
|
|
DISCUSSION |
|---|
|
). Substance P reduced the variability of EIN spiking and of EIN-evoked EPSPs. Because long-term recordings from these interneurons are rare, the duration of the modulatory effects is unknown. The effects were unchanged 30 min to 1 h after substance P application (Fig. 3C). Because other cellular and synaptic effects of substance P have recovered at these times (Parker and Grillner 1998, 1999; Parker et al. 1998; Svensson et al. 2002), the effects shown here may be relatively long-lasting.
In common with other analyses, synaptic variability was examined by measuring the PSP variance (Aradi and Soltesz 2002; Földy et al. 2004, 2005). However, changes in the PSP amplitude can influence the synaptic variance (Fig. 5), and this could have complicated the interpretation of the effects shown here. An alternative approach is to use the CV (SD PSP amplitude/mean PSP amplitude), which has the advantage of normalizing the variability to the mean. However, the CV can only be used for stationary (i.e., activity-independent) effects: in nonstationary data sets, changes in the mean amplitude could result in changes in CV even when the SD was unaffected. While changes in PSP amplitudes could also have influenced the synaptic variance, this can simply be ruled out for three of the four connections examined. There were no significant effects of substance P on the amplitude of SiIN- or EScIN-evoked PSPs, which means that the changes in synaptic variance at these connections could not be caused by changes in the PSP amplitudes. These connections thus provide the first experimental examples of changes in the variance independently of changes in the mean PSP amplitude (Aradi and Soltesz 2002). The reduction in the variance of EIN-evoked EPSPs was not caused by a change in the EPSP amplitude, because the EPSP amplitude was increased by substance P, an effect that should have increased the PSP variance. However, the significant reduction of the IScIN-evoked IPSP amplitude by substance P could have contributed to the reduced variance at this connection.
Effects on cellular and synaptic variability
Substance P reduced the variability of subthreshold locomotor-related depolarizations and spiking in motor neurons during network activity. However, it did not reduce the variability of motor neuron spiking when it was evoked by depolarizing current pulses. This suggests that the effects of substance P on spiking during network activity were not caused by effects on the intrinsic properties of motor neurons but to a reduction in variability of the excitatory drive that the motor neurons received. This could be caused by the reduced variability of EIN spiking or by the reduced variance of EIN- and EScIN-evoked EPSPs.
In contrast to its effects on excitatory synaptic inputs, substance P increased the variability of SiIN-evoked IPSPs. The relevance of this effect is currently unknown. Simulated increases in inhibitory variance can significantly affect postsynaptic excitability (Aradi and Soltesz 2002), the nature of the effect depending on the amplitude and time course of the inhibitory input and on the membrane potential and excitability of the postsynaptic cell (Aradi et al. 2002). Changes in inhibitory synaptic variance could thus result in an increase or decrease in network excitability, depending on the current state of the network and of network cellular properties. Because highly regular inputs can make systems less flexible (Peters 2000; Sterling 2004; Wang and Jung 2002), an increase in the variance of SiIN-evoked IPSPs may also help to maintain network flexibility by compensating for the reduced variability of excitatory synaptic inputs.
In contrast to motor neurons, substance P reduced the variability of EIN spiking in response to depolarizing current pulses. This suggested an effect of substance P on the intrinsic properties of the EINs. The mechanisms underlying the modulation of EIN spiking have not been examined in any detail here. Substance P did not affect the KCa–mediated afterhyperpolarization in the EINs, suggesting against an effect on this conductance. An IA potassium current influences spiking and the regularity of network activity in the lamprey (Hess and El Manira 2001), whereas a slow potassium conductance (IKs) influences cellular excitability and the regularity of network activity in Xenopus (Dale and Kuenzi 1997). The role of these conductances in the cellular effects of substance P remains to be studied.
Mechanisms of the synaptic modulation
Multiple presynaptic or postsynaptic mechanisms could contribute to the reduced variance of EIN-evoked EPSPs. Postsynaptic mechanisms include dendritic conductances (Cash and Yuste 1999; Reyes 2001) or the number, saturation, or desensitization of transmitter receptors (Jones and Westbrook 1996; Oleskevich et al. 1999). Postsynaptic voltage or AMPA receptor desensitization do not affect EIN inputs during spike trains (Parker 2000a). However, a postsynaptic influence on the reduced variability was supported by its NMDA dependence and by the reduced variability of postsynaptic responses to pressure applied glutamate and NMDA.
Presynaptic effects on the synaptic variance could reflect differences in the number of vesicles released by each action potential. The PSP amplitude is the product of the release probability (p), the number of vesicles or release sites (n), and the quantal amplitude (q): the PSP amplitude = npq and the PSP variance = q2np(1 – p). A reduction in q or n will thus reduce the synaptic variance, but there is a parabolic relationship between the synaptic variance and p (Clements and Silver 2000). Substance P reduces p, increases the number of release sites and docked vesicles, and increases the quantal amplitude (Bevan and Parker 2004), effects that are not consistent with a reduction in the variability of the quantal content.
While changes in the quantal content probably do not account for the reduced synaptic variance, a potential presynaptic component was suggested by the reduced variability of vesicle diameters in glutamatergic terminals. Because vesicle size and transmitter content are correlated (Franks et al. 2003; Karunanithi et al. 2002), different sized vesicles could release different amounts of transmitter, thus generating quantal variability. While a direct link between changes in vesicle diameter and synaptic variability is lacking, a reduction in the variability of vesicle diameters (possibly caused by effects on vesicular glutamate transporters; Daniels et al. 2006) could have contributed to the reduced synaptic variability. However, because the intraburst or interburst variability could occur independently of each other, the effect on vesicle diameters alone cannot account for the reduced variance in all connections, because it should have affected the intraburst and interburst variance simultaneously. As with the other effects of substance P (Parker 2006), the changes in synaptic variance presumably reflect the influence of several parallel mechanisms.
Functional relevance
The functional role of cellular and synaptic variability is still not well understood. Is it simply an epiphenomenon that reflects the problems of precisely regulating multiple interacting elements or is it specifically designed to influence network outputs? Cellular, synaptic, and network variability clearly exists, and it can have significant effects. While variability is attracting increasing attention (Aradi and Soltesz 2002; Aradi et al. 2002; Prinz et al. 2004), it will be difficult to study experimentally without tools that allow the variance to be altered without changing mean values (Aradi and Soltesz 2002).
Despite the assumption that highly regular activity is the normal condition of rhythmic locomotor networks, their outputs often vary (Horn et al. 2004; Parker et al. 1998; Pearlstein et al. 2005; Szucs et al. 2005; Zhang and Grillner 2000). In reduced preparations this may reflect the absence of components needed to regulate network activity (e.g., sensory feedback). Alternatively, functionally appropriate outputs may still be generated under most conditions (Prinz et al. 2004), making it unnecessary or uneconomical to remove the variability. Some degree of variability may also be essential to normal network function (Peters 2000; Sterling 2004; Wang and Jung 2002). The effects of substance P shown here suggest that, under some conditions, variability can be targeted for modulation. The effects of substance P in lamprey only occur in migratory adult animals and not in larval or sexually mature adults (Parker 2000b; D. Parker and T. Gilbey, unpublished observations). This suggests a role for substance P during migration, where the increased frequency and improved regularity of network activity may facilitate swimming over long distances.
We know that substance P reduces the variability of the locomotor network output (Parker et al. 1998). We have shown here that this is associated with a reduction in the variability of locomotor-related depolarizations and of motor neuron spiking during network activity and with a reduction in the variability of glutamatergic synaptic transmission and spiking in the EINs in quiescent preparations. While the reduced variability of EIN cellular and synaptic activity is consistent with the reduced variability during network activity, it is difficult to causally link cellular and synaptic effects in quiescent preparation to changes during network activity (Parker 2006). This will require that synaptic variability is examined at connections between locomotor network interneurons (ideally during network activity) and that specific changes in variance are related directly to changes in the network output (Parker 2006).
It is generally difficult to examine individual synaptic inputs when the network is active (Parker, unpublished observations; Manor et al. 1999), and even harder to change variance independently of effects on mean values (Aradi and Soltesz 2002). We have tried to link cellular, synaptic, and network effects of substance P by correlating effects at different levels. Over a series of studies, we have linked metaplastic changes in EIN-evoked EPSPs to substance P–evoked changes in the network burst frequency (Bevan and Parker 2004; Parker and Grillner 1999; Parker et al. 1998). Because the effect on the network burst regularity has not been studied previously at the cellular and synaptic levels, the description of variability in this study is an essential initial step in this analysis. Potential links between the synaptic effects of substance P and the change in the network output are supported by the PKA dependence of both the reduced variability of EIN-evoked EPSPs and of network activity (Parker et al. 1998). In addition, because substance P did not affect the regularity of current injection–evoked spiking in motor neurons, the improved regularity of spiking during network activity presumably reflected changes in the excitatory drive caused by the reduced variability of EIN spiking and EIN-evoked EPSPs. However, further analyses will be needed to link these effects in quiescent preparations to changes during network activity.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: D. Parker, Dept. of Physiology, Development, and Neuroscience, Univ. of Cambridge, Downing St., Cambridge CB2 3EJ, UK (E-mail djp27{at}cam.ac.uk)
|
|
REFERENCES |
|---|
|
Aradi I, Santhakumar V, Soltesz I. Impact of heterogeneous perisomatic IPSC populations on pyramidal cell firing rates. J Neurophysiol 91: 2849–2858, 2004.
Aradi I, Soltesz I. Modulation of network behaviour by changes in variance in interneuronal properties. J Physiol 538: 227–251, 2002.
Barnes T, Kubota Y, Hu D, Jin D, Graybiel A. Activity of striatal neurons reflects dynamic encoding and recoding of procedural memories. Nature 437: 1158–1161, 2005.[CrossRef][Medline]
Barthe J-Y, Clarac F. Modulation of the spinal network for locomotion by substance P in the neonatal rat. Exp Brain Res 115: 485–492, 1997.[CrossRef][Web of Science][Medline]
Bevan S, Parker D. Metaplastic facilitation and ultrastructural changes in synaptic properties are associated with long-term modulation of the lamprey locomotor network. J Neurosci 24: 9458–9468, 2004.
Blitz DM, Regehr WG. Retinogeniculate synaptic properties controlling spike number and timing in relay neurons. J Neurophysiol 90: 2438–2450, 2003.
Brezina V, Proekt A, Weiss K. Cycle-to-cycle variability as an optimal behavioral strategy. Neurocomputing 69: 1120–1124, 2006.[CrossRef][Web of Science]
Buchanan J. Electrophysiological properties of identified classes of lamprey spinal neurons. J Neurophysiol 70: 2313–2325, 1993.
Buchanan J. Contributions of identifiable neurons and neuron classes to lamprey vertebrate neurobiology. Prog Neurobiol 63: 441–466, 2001.[CrossRef][Web of Science][Medline]
Buchanan J, Cohen A. Activities of identified interneurons, motoneurons and muscle fibers during fictive swimming in the lamprey and the effects of reticulospinal and dorsal cell stimulation. J Neurophysiol 47: 948–960, 1982.
Buchanan J, Grillner S. A new class of small inhibitory interneurones in the lamprey spinal cord. Brain Res 438: 404–407, 1988.[CrossRef][Web of Science][Medline]
Buchanan J, Grillner S, Cullheim S, Risling M. Identification of excitatory interneurons contributing to generation of locomotion in lamprey: structure, pharmacology, and function. J Neurophysiol 62: 59–69, 1989.
Buchanan J, Kasicki S. Activities of spinal neurons during brain stem-dependent fictive swimming in lamprey. J Neurophysiol 73: 80–87, 1995.
Buchman T. The community of the self. Nature 420: 246–251, 2002.[CrossRef][Medline]
Calvin W, Stevens C. Synaptic noise and other sources of randomness in motoneuron interspike intervals. J Neurophsyiol 31: 574–587, 1968.
Cash S, Yuste R. Linear summation of excitatory inputs by CA1 pyramidal neurons. Neuron 22: 383–394, 1999.[CrossRef][Web of Science][Medline]
Clements J, Silver R. Unveiling synaptic plasticity: a new graphical and analytical approach. Trends Neurosci 23: 105–113, 2000.[CrossRef][Web of Science][Medline]
Cohen A, Wallen P. The neuronal correlate of locomotion in fish. "Fictive swimming" induced in an in vitro preparation of the lamprey spinal cord. Exp Brain Res 41: 11–18, 1980.[Web of Science][Medline]
Colliver TL, Pyott SJ, Achalabun M, Ewing AG. VMAT-mediated changes in quantal size and vesicular volume. J Neurosci 20: 5276–5282, 2000.
Dale N, Kuenzi F. Ionic currents, transmitters and models of motor pattern generators. Curr Opin Neurobiol 7: 790–796, 1997.[CrossRef][Web of Science][Medline]
Daniels R, Collins C, Chen K, Gelfand M, Featherstone D, DiAntonio A. A single vesicular glutamate transporter is sufficient to fill a synaptic vesicle. Neuron 49: 11–16, 2006.[CrossRef][Web of Science][Medline]
Földy C, Aradi I, Howard A, Soltesz I. Diversity beyond variance: modulation of firing rates and network coherence by GABAergic subpopulations. Eur J Neurosci 19: 119–130, 2004.[CrossRef][Web of Science][Medline]
Földy C, Dyhrfjeld-Johnsen J, Soltesz I. Structure of cortical microcircuit theory. J Physiol 562: 47–54, 2005.
Franks KM, Stevens CF, Sejnowski TJ. Independent sources of quantal variability at single glutamatergic synapses. J Neurosci 23: 3186–3195, 2003.
Grillner S, Deliagina T, Ekeberg O, El Manira A, Hill R, Lansner A, Orlovsky G, Wallen P. Neural networks that co-ordinate locomotion and body orientation in lamprey. Trends Neurosci 18: 270–279, 1995.[CrossRef][Web of Science][Medline]
Harris C, Wolpert D. Signal-dependent noise determines motor planning. Nature 394: 780–784, 1998.[CrossRef][Medline]
Harris K, Sultan P. Variation in the number, location and size of synaptic vesicles provides an anatomical basis for the nonuniform probability of release at hippocampal CA1 synapses. Neuropharmacology 34: 1387–1395, 1995.[CrossRef][Web of Science][Medline]
Harsch A, Robinson HPC. Postsynaptic variability of firing in rat cortical neurons: the roles of input synchronization and synaptic NMDA receptor conductance. J Neurosci 20: 6181–6192, 2000.
Hess D, El Manira A. Characterization of a high-voltage-activated IA current with a role in spike timing and locomotor pattern generation. Proc Natl Acad Sci 98: 5276–5281, 2001.
Horn CC, Zhurov Y, Orekhova IV, Proekt A, Kupfermann I, Weiss KR, Brezina V. Cycle-to-cycle variability of neuromuscular activity in aplysia feeding behavior. J Neurophysiol 92: 157–180, 2004.
Jones M, Westbrook G. The impact of receptor desensitisation on fast synaptic transmission. Trends Neurosci 19: 96–101, 1996.[CrossRef][Web of Science][Medline]
Karunanithi S, Marin L, Wong K, Atwood HL. Quantal size and variation determined by vesicle size in normal and mutant drosophila glutamatergic synapses. J Neurosci 22: 10267–10276, 2002.
Klingauf J, Kavalali E, Tsien R. Kinetics and regulation of fast endocytosis at hippocampal synapses. Nature 394: 581–585, 1998.[CrossRef][Medline]
Manor Y, Nadim F, Epstein S, Ritt J, Marder E, Kopell N. Network oscillations generated by balancing graded asymmetric reciprocal inhibiton in passive neurons. J Neurosci 19: 2765–2779, 1999.
McLachlan E, Martin A. Non-linear summation of frog end-plate potentials in the frog and mouse. J Physiol 311: 307–324, 1981.
Newell K, Corcos D. Issues in variability and motor control. In: Variability and Motor Control, edited by Newell KM, Corocs DM. Champaign, IL: Human Kinetics Publishers, 1993, p. 1–12.
Ohta Y, Dubuc R, Grillner S. A new population of neurons with crossed axons in the lamprey spinal cord. Brain Res 564: 143–148, 1991.[CrossRef][Web of Science][Medline]
Oleskevich S, Alvarez F, Walmsley B. Glycinergic miniature synaptic currents and receptor cluster sizes differ between spinal cord interneurons. J Neurophysiol 82: 312–319, 1999.
Parker D. Activity and calcium-dependent mechanisms maintain reliable interneuron synaptic transmission in a rhythmic neural network. J Neurosci 20: 1754–1766, 2000a.
Parker D. Spinal-cord plasticity: independent and interactive effects of neuromodulator and activity-dependent plasticity. Mol Neurobiol 22: 55–80, 2000b.[CrossRef][Web of Science][Medline]
Parker D. Activity-dependent feedforward inhibition modulates synaptic transmission in a spinal locomotor network. J Neurosci 23: 11085–11093, 2003a.
Parker D. Variable properties in a single class of excitatory spinal synapse. J Neurosci 23: 3154–3163, 2003b.
Parker D. Complexities and uncertainties of neuronal network function. Philos Trans R Soc B Biol Sci 361: 81–99, 2006.
Parker D, Grillner S. Cellular and synaptic modulation underlying substance P-mediated plasticity of the lamprey locomotor network. J Neurosci 18: 8095–8110, 1998.
Parker D, Grillner S. Activity-dependent metaplasticity of inhibitory and excitatory synaptic transmission in the lamprey spinal cord locomotor network. J Neurosci 19: 1647–1656, 1999.
Parker D, Zhang W, Grillner S. Substance P modulates NMDA responses and causes long-term protein synthesis-dependent modulation of the lamprey locomotor network. J Neurosci 18: 4800–4813, 1998.
Pearlstein E, Mabrouk FB, Pflieger J, Vinay L. Serotonin refines the locomotor-related alternations in the in vitro neonatal rat spinal cord. Eur J Neurosci 21: 1338–1346, 2005.[CrossRef][Web of Science][Medline]
Peters A, Palay S, Webster H. Fine Structure of the Nervous System: Neurons and Their Supporting Cells, 3rd ed. New York: Oxford, 1991.
Peters R. Congestive heart failure: the case for decreased variability as a unifying theme. J Clin Basic Cardiol 3: 111–113, 2000.
Prinz A, Bucher D, Marder E. Similar network activity from disparate circuit parameters. Nat Neurosci 7: 1345–1352, 2004.[CrossRef][Web of Science][Medline]
Reyes A. Influence of dendritic conductances on the input-output properties of neurons. Annu Rev Neurosci 24: 653–675, 2001.[CrossRef][Web of Science][Medline]
Santhakumar V, Soltesz I. Plasticity of interneuronal species diversity and parameter variance in neurological diseases. Trends Neurosci 27: 504–510, 2004.[CrossRef][Web of Science][Medline]
Stein R, Gossen ER, Jones K. Neuronal variability: noise or part of the signal? Nat Rev Neurosci 6: 389–397, 2005.[Web of Science][Medline]
Sterling P. Principles of allostasis: optimal design, predictive regulation, pathophysiology and rational therapeutics. In: Allostasis, Homeostasis, and the Costs of Adaptation, edited by Schulkin J. Cambridge: Cambridge University Press, UK, 2004, p. 17–65.
Svensson E, Grillner S, Parker D. Synaptically evoked membrane potential oscillations induced by substance P in lamprey motor neurons. J Neurophysiol 87: 113–121, 2002.
Szucs A, Abarbanel H, Rabinovich M, Selverston A. Dopamine modulation of spike dynamics in bursting neurons. Eur J Neurosci 21: 763–772, 2005.[CrossRef][Web of Science][Medline]
Thomson A. Molecular frequency filters at central synapses. Prog Neurobiol 62: 159–196, 2000.[CrossRef][Web of Science][Medline]
Wallén P, Williams T. Fictive locomotion in the lamprey spinal cord in vitro compared with swimming in the intact and spinal animal. J Physiol 347: 225–239, 1984.
Wang H, Jung R. Variability analyses suggest that supraspino–spinal interactions provide dynamic stability in motor control. Brain Res 930: 83–100, 2002.[CrossRef][Web of Science][Medline]
Zador A. Impact of synaptic unreliability on the information transmitted by spiking neurons. J Neurophysiol 79: 1219–1229, 1998.
Zhang W, Grillner S. The spinal 5-HT system contributes to the generation of fictive locomotion in lamprey. Brain Res 879: 188–192, 2000.[CrossRef][Web of Science][Medline]
Zucker R, Regehr W. Short-term synaptic plasticity. Annu Rev Physiol 64: 355–405, 2002.[CrossRef][Web of Science][Medline]
This article has been cited by other articles:
|
|
 |
|
  M. Huss, A. Lansner, P. Wallen, A. El Manira, S. Grillner, and J. H. Kotaleski Roles of Ionic Currents in Lamprey CPG Neurons: A Modeling Study J Neurophysiol, April 1, 2007; 97(4): 2696 - 2711. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
