| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, Emory University, School of Medicine, Atlanta, Georgia
Submitted 17 December 2006; accepted in final form 22 January 2007
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
; Porter 1993; Young 1994). Despite their importance in the adult, we know very little about their incorporation into circuits during development. Inhibitory interneurons have been far less studied than excitatory projection neurons because of a lack of access to the locally projecting axons of the many different classes of inhibitory interneurons. As a result, our understanding of excitatory neuronal development is significantly advanced compared with that for inhibitory neurons; for example, several kinds of synaptic plasticity are better understood for excitatory neurons (long-term potentiation, long-term depression, synaptic refinement).
Synaptic refinement is the process of pruning initially exuberant synaptic projections. This kind of plasticity is thought to be important in allowing developing circuits to adjust to their environment. This process has been studied extensively in excitatory projection neurons, including the neuromuscular junction (synapse elimination), projections throughout the visual system, and many other excitatory projection systems (Campbell and Shatz 1992; Kasthuri and Lichtman 2003; O'Brien et al. 1978; Redfern 1970; Reh and Constantine-Paton 1984; Simon and O'Leary 1992; Sretavan and Shatz 1987; Sur et al. 1984; Tello 1917; Thompson et al. 1979). In the developing mammalian spinal cord, synaptic refinement and axon elimination were extensively demonstrated for corticospinal projections (Bates and Killackey 1984; Cabana and Martin 1985; Li and Martin 2000; Luo and O'Leary 2005; Martin 2005; Stanfield and O'Leary 1985) and refinement was postulated for dorsal root afferent projections into the cord (Kudo and Yamada 1985; Saito 1979). A synaptic reorganization that is likely the result of synaptic refinement was previously described for one excitatory spinal interneuron that projects to parasympathetic preganglionic neurons (Araki and de Groat 1997).
Until recently it was unknown whether any inhibitory neurons experienced synaptic refinement. We now know that one set of inhibitory projection neurons that forms tonotopic maps in the auditory system undergoes synaptic refinement at a stage when their transmitters [
-aminobutyric acid (GABA) and glycine] are depolarizing (Kandler and Gillespie 2005; Kim and Kandler 2003; Sanes and Friauf 2000; Sanes and Siverls 1991). It is unknown whether this is a general feature of inhibitory neuron development and would therefore also occur in locally projecting inhibitory interneurons. However, it is clear that local inhibitory circuits profoundly influence the development of their networks (Hensch 2005, 2004).
We recently identified an accessible inhibitory interneuron (R-interneuron) in the chick embryo spinal cord, which is the avian homologue of the mammalian Renshaw cell (Wenner and O'Donovan 1999) (Fig. 1, schematic). At embryonic day 10 (E10) R-interneurons receive monosynaptic input from motoneurons, mediated predominantly by nicotinic receptors. These interneurons make direct projections back onto motoneurons in the same and adjacent segments and are GABAergic (possible small glycinergic contribution). R-interneuron axons can project multiple segments through the ventrolateral funiculus (VLF) and are located in a nucleus dorsomedial to the lateral motor column at both E7 and E10 (Xu et al. 2005). This disynaptic circuit first forms by E7 and the pharmacology of the R-interneuron circuitry is largely unchanged from this point to E15; however, the caudal spread of the output of the R-interneuron circuit appears to change (Xu et al. 2005). Because GABA is depolarizing at these stages the R-interneuron is excitatory and as a result plays a distinct functional role to that in the adult. During development the cell and its circuit are involved in the normal initiation of the spontaneous network activity observed in the cord at embryonic ages (Wenner and O'Donovan 2001).
|
|
|
METHODS |
|---|
|
Experiments were performed on White Leghorn chick embryos aged E8–E15. Embryos were staged according to Hamburger and Hamilton (1951). This report refers to developmental age in days of incubation (i.e., embryonic day 8 or E8) corresponding to the Hamburger–Hamilton staging criteria, irrespective of whether this was the actual incubation period. Three developmental stages were used in the study: early (E8, stage 34, shortly after the onset of circuit formation at a period when the response was reliably detectable), middle (E10/11, stage 36/37 when previous studies were carried out, referred to as E10), and late (E14/15, stage 40/41, the latest stage when the response can still be reliably observed, referred to as E15).
Embryos were decapitated and eviscerated under continuous superfusion with oxygenated Tyrode's solution [concentration (in mM): NaCl 139, KCl 2.9, NaHCO3 17, glucose 12.2, CaCl2 3, MgCl2 1] cooled to about 15°C. A ventral laminectomy was performed and the spinal cord, including thoracic to sacral segments, was freed from the underlying dorsal lamina and the isolated cord was left overnight to recover at 17°C. The next morning the central portion of the dorsal roots was cut so that we could selectively activate motoneurons after stimulation of the spinal nerve (referred to throughout this report as ventral root stimulation). The preparation was warmed to room temperature and transferred to a recording chamber. Tyrode's solution was superfused at a constant temperature of 26–28°C. Tight-fitting glass suction electrodes were used to record and stimulate the ventral roots as described previously (O'Donovan 1987, 1989; Xu and Wenner 2005; Fig. 1, schematic). These ventral root recordings represent the electrotonically degraded motoneuron population potentials as shown originally in the 1940s (Brooks et al. 1948; Eccles 1946) and later using glass suction electrodes as sucrose gap recordings (Brink et al. 1981; Luscher et al. 1979; Roberts and Wallis 1978), and more recently in the in vitro preparation (Butt and Kiehn 2003; O'Donovan 1987; Wenner and O'Donovan 1999). The signals were amplified (1,000x), filtered (DC to 0.3–5 kHz), and digitally recorded using Axograph acquisition software (Axon Instruments) onto a Macintosh computer. Further analyses of the data were performed off-line.
Amplitudes of the responses in ventral root recordings after ventral root stimulation were normalized to the amplitude of the root potential produced during an episode of spontaneous network activity, where motoneurons throughout the rostrocaudal extent of the lumbosacral cord experience potentials of 20–30 mV in amplitude (O'Donovan 1999). This served to normalize each root recording, thereby accommodating for variability in the fit of the root in the electrode or to accommodate for a partially damaged root. Normalized values were still variable as they were subject to changes in the excitability of the network or variability in the stimulated root. To reduce this variance the normalized amplitudes were expressed as a percentage of the normalized response in the root adjacent to the stimulated root, which had the largest response (normalization process described in greater detail in Xu and Wenner 2005).
Whole cell recordings were obtained from antidromically identified motoneurons (Xu and Wenner 2005). Meninges were removed from the ventral surface overlying LS1–LS4 to allow for electrode penetration. Whole cell electrodes [K-gluconate solution concentration (in mM): NaCl 10, K-gluconate 94, KCl 36, HEPES 10, EGTA 1.1, CaCl2 0.1, MgCl2 1, Na2ATP 1] containing 5 µM QX-314 to block action potentials (usually after 10 min of dialysis) were then targeted to the lateral half of one side of the cord. Recordings were obtained from motoneurons using whole cell electrodes (5–15 M
). To reduce variability in potential responses after ventral root stimulation, resting membrane potential was maintained at –70 ± 2 mV. It is unlikely that variability was introduced based on variable dialysis of the cell because the predominantly dendritic chloride-mediated conductances are not believed to be significantly influenced by the patch solution (Chub and O'Donovan 2001; Chub et al. 2006). Only cells with resting potentials more hyperpolarized than –40 mV, strong network driven potentials, and expression of antidromic action potentials after ventral root stimulation were accepted into the database.
Optical recordings
To visualize the interneurons activated after stimulation of either ventral roots (R-interneurons) or the VLF (most interneurons) we isolated the spinal cord from thoracic segment 5 to lumbosacral segment 8 and retrogradely loaded many interneurons with calcium-sensitive dye (Ca2+ Green-1 dextran 10,000 MW; Molecular Probes). Interneurons were retrogradely labeled through their axonal projections into the VLF, which was drawn into a suction electrode containing roughly 20% wt/vol of the dye dissolved in distilled water containing 0.2% Triton X-100 detergent (O'Donovan et al. 1993). This configuration was left overnight to allow retrograde transport of the dye back to interneuronal cell bodies. We removed the dorsal half of the cord to better visualize the dye-loaded cells in a ventral half preparation, where the R-interneuron circuit is located. This was accomplished by removing the dorsal and lateral pia mater (along with dorsal roots) and the cord was then pinned to a Sylgard block, ventral side up. A vibratome blade (Leica VT 1000) was then positioned at the ventral surface of the cord. To create consistent ventral half preparations the blade was moved laterally away from the cord and then descended 230 µm (at E8) or 300 µm (at E10) toward the dorsal part of the cord and a horizontal cut was performed along the cord's rostrocaudal axis. This generated approximately equal dorsal and ventral half pieces. The ventral-half preparation (Fig. 5, schematic), along with intact ventral roots, was then transferred to the recording chamber and held in position with an anchor. Ventral roots were then drawn into suction electrodes for motoneuron stimulation and recording.
|
|
|
|
F). These were generated by subtracting a 30-frame average obtained before the activity (background image) from an average of eight consecutive frames during the peak of R-interneuron activity or a network-driven activity. During episodes of network activity virtually all labeled neurons showed changes in fluorescence or became optically active (O'Donovan et al. 1994). A median filter (1-pixel radius) was used to better resolve optical activity. To better identify the active cells these difference images were stretched across their dynamic range (i.e., contrast enhanced; if pixel intensity values of optically active cells ranged from 0 to 20, these values were then stretched to occupy the entire 256-value range of the lookup table). These processed images were then used to draw regions of interest (ROIs) corresponding to optically active interneurons. By increasing or decreasing the stretch across the dynamic range we could resolve additional ROIs in the image. This provided an initial set of ROIs (R-interneurons or interneurons) that were adjusted (ROIs were added or removed) by monitoring the ROIs during playback of the video, which provided the most compelling evidence (see supplemental videos of spontaneous network activity at E8 and E10).1 ROIs were accepted as R-interneurons (after ventral root stimulation) or optically active interneurons (during network activity) if these ROIs, of 
70 contiguous pixels (to distinguish individual cells from a part of the neuron—dendrite, axon), displayed an increased intensity of 2 SD over the mean background (O'Donovan et al. 1994). For acceptance into the database, preparations had to demonstrate multiple optically active R-interneurons that provided single-cell resolution after stimulus trains (10x, 50 Hz) to the ventral roots (Fig. 5). In this way we could be relatively confident that the dorsal–ventral cut was close enough to the R-interneuron population and that the stimulated ventral root was intact. Statistical analysis
Student's t-tests or one-way ANOVAs followed by Tukey post hoc tests were performed on the data to detect significant differences (JMP software), unless mentioned otherwise. For sequential antagonist applications in optical imaging a paired t-test was used. Means and corresponding SEs are presented.
|
|
RESULTS |
|---|
|
VENTRAL ROOT RECORDING. Previously, we reported preliminary evidence suggesting that caudal projections of the R-interneuron circuitry may become more locally focused between E8 and E15, although this was not quantified. Here we extend these studies by stimulating a ventral root and quantifying the responses in adjacent ventral roots, either rostral or caudal to the stimulated root. These recordings allow us to sample the inputs of the entire population of motoneurons within a segment and provide a synchronous comparison of the output of the activated R-interneurons to motoneurons in different rostral or caudal segments.
Single suprathreshold stimuli (30 µA, 0.5 ms) were delivered to the LS6 ventral root while recording LS5, -4, -3, and -2 ventral roots through tight-fitting glass suction electrodes and averaging the responses (Fig. 1A, n = 3–20 recordings/average). The amplitudes of these responses were normalized twice (see METHODS for normalization procedure). In this way we were assessing the strength of the circuit's rostral projection to different ventral roots. The LS3 ventral root response after LS6 stimulation became significantly weaker from E8 to E15 (Fig. 1; 39.8 ± 2.1% at E8; 17.6 ± 6.6% at E15). At E8 the responses at LS5 and LS4 were similar, but by E15 the LS4 response appeared to be reduced compared with that of LS5, although this did not reach significance (P = 0.09, one-tailed t-test). In contrast, the already weak projection to LS2 did not appear to change. Therefore a relative weakening of the rostral projection of the R-interneuron circuitry was observed from E8 to E15, but not in all roots.
We also quantified changes in the caudal projection of the R-interneuron circuit by stimulating LS2 and recording LS3–LS6 ventral roots (Fig. 2). In addition, to better define the period of circuit reorganization, we tested the circuit at E8, E10, and E15. After LS2 stimulation, the normalized LS4 root response was 111.5 ± 13.4% of the value for LS3 at E8, suggesting that the circuit projected to LS4 motoneurons with the same or greater strength as it did to LS3 motoneurons, despite the greater distance. The projection to LS5 and LS6 was sequentially weaker. However, later in development the LS4 root response was significantly weaker than the LS3 response (50.3 ± 7.4% at E10; 61.1 ± 6.1% at E15). Because of the normalization process it was possible that the LS4 responses were reduced relative to LS3, not because of changes in the R-interneuron circuit but because the episode potential amplitude increased for LS4 relative to LS3 at the later stages. This did not appear to be the case because amplitudes of the episode potentials were not significantly different for the different roots at any of the ages and, if anything, decreased for LS4 and LS5 at the later stages (LS3:LS4:LS5—E8, 1,033 ± 312:954 ± 244:920 ± 288 µV; E10, 864 ± 191:562 ± 80:532 ± 148 µV; E15, 329 ± 62:243 ± 34:285 ± 68 µV). As in the rostral projection of the circuit, the relative strength of the projection four segments away (caudally) did not change with development. These data show that developmental changes in the rostrocaudal projection of the R-interneuron circuitry occurred in a specific manner. The findings demonstrate that there was a functional reorganization of the circuitry underlying the ventral root response. To better understand the mechanisms underlying the reorganization we recorded intracellularly from motoneurons.
WHOLE CELL RECORDING OF MOTONEURONS.
The functional reorganization of the R-interneuron circuit could occur through changes in the number of motoneurons receiving the disynaptic input and/or in the strength of the inputs to each motoneuron. To examine how the circuit achieved the functional reorganization we obtained whole cell recordings from motoneurons, which provided information not only about the strength of R-interneuron inputs, but also the frequency of motoneurons receiving the disynaptic input. The clearest example of reorganization described earlier was observed after LS2 stimulation in the potentials produced in LS3 and LS4 ventral roots between E8 and E10. To test how functional reorganization occurred in this part of the circuit, we recorded from antidromically identified LS3 and LS4 motoneurons at E8 and E10, while stimulating the LS2 ventral root. It is likely that the potentials produced were mediated predominantly by a disynaptic circuit (motoneuron to R-interneuron and R-interneuron back to motoneuron) (Wenner and O'Donovan 1999; Xu et al. 2005). In neither whole cell nor ventral root recordings did we ever see potentials whose onset was <10 ms after the stimulation of adjacent ventral roots, suggesting that there were no direct monosynaptic connections between these motoneurons that could have contributed to the observed potential. The amplitudes of the potentials were measured at a latency corresponding to the peak of the potential recorded in the ventral root (about 100 ms). This was typically also the peak of the potential in the intracellular record.
We observed that after LS2 stimulation the average potential in LS3 motoneurons receiving input was significantly increased from E8 to E10 (Fig. 3, A and B; for number of motoneurons recorded see figure legend or Table 1). This change cannot be explained by changes in the input resistance of the motoneurons, which decreased from E8 to E10 (661 ± 85 to 457 ± 61 M; n = 18, n = 6). The amplitude of the LS2-evoked potential in LS4 motoneurons did not change from E8 to E10 (Fig. 3B). Therefore after LS2 stimulation, there was an increase in the potentials of LS3 compared with LS4 motoneurons.
|
We sought to determine how common this pattern of relative strengthening of the more local circuitry was. We therefore stimulated the LS3 and LS4 ventral roots while recording from LS3 or LS4 motoneurons. The strongest inputs were, at both ages, from the most local stimulus (i.e., LS3 motoneurons received the strongest input from LS3 stimulation; Fig. 4A). However, in neither segment was there a significant change from E8 to E10 after stimulation of either ventral root. If anything the LS4-evoked response got weaker in LS4 motoneurons at E10 (Fig. 4B). Next we looked at the frequency of LS3 or LS4 motoneurons receiving input after stimulation of LS3 or LS4 ventral roots (Table 1). We observed a dramatic increase in the frequency of LS3 motoneurons receiving LS4-evoked input by E10. On the other hand, the frequency of other inputs to these motoneurons did not change from E8 to E10. Collectively, these data demonstrate that disynaptic projections after LS2 stimulation are strengthened for adjacent segments compared with more distant segments (LS3 vs. LS4) by E10. However, this local strengthening was not observed uniformly throughout the rostrocaudal extent of the cord.
Optical recordings show distribution and percentage of activated R-interneurons at E8 and E10
The ventral root and motoneuron recordings from the LS3 and LS4 segment demonstrated that the disynaptic circuit undergoes a functional rearrangement. After LS2 stimulation, by E10 the circuit had weakened its projection to LS4 compared with LS3 motoneurons. Because we were assessing a disynaptic circuit, changes could have occurred in the recruitment of R-interneurons and/or in the R-interneuron projection to motoneurons. To better understand where in the circuit changes occurred, we tested the recurrent connection to R-interneurons. This was accomplished using calcium imaging of interneuron cell bodies by retrogradely labeling many interneurons from their axonal projections into the VLF (see METHODS). An increase in the fluorescence of calcium dye–labeled neurons is a strong indication that a neuron is experiencing spiking activity (O'Donovan et al. 1993, 1994). Although our technique labeled many different species of interneuron, we could selectively activate R-interneurons in a particular segment by delivering stimulus trains to the associated ventral root. A column of cells dorsomedial to motoneurons became optically active after ventral root stimulus trains (Wenner and O'Donovan 1999, 2001). We monitored optically active interneurons in the LS2 or LS3 segment. Experiments were carried out in the presence of bicuculline to block the depolarizing output of the R-interneuron and to prevent the possible recruitment of other neurons. This technique allows us to selectively image R-interneurons (Wenner and O'Donovan 2001).
At both E8 (Fig. 5, A–D) and E10 (Fig. 5, E–H, supplemental video) optical recordings showed that R-interneurons in LS2 or LS3 could be activated only by stimulating the LS2 or LS3 ventral root, respectively (eight of eight ventral root stimulations, five preparations at E8; eight of eight ventral root stimulations, five preparations at E10). Even when R-interneurons were very close to each other at the LS2/LS3 border (within about 100 µm; Fig. 5, C and G), only one root was capable of driving an R-interneuron to become optically active, never both, suggesting that the motoneurons activated R-interneurons only in the same segment. Moreover, we never observed R-interneurons after stimulation of both roots that had not been observed by one of the roots individually (spatial facilitation) and the optical signals produced were no larger than stimulation of the individual effective root. This suggests that stimulation of the ineffective root did not provide significant subthreshold input. Together the results imply that motoneuron recurrent collaterals are limited in their rostrocaudal projection to their own segment at both E8 and E10 and argue against a functional rostrocaudal refinement of the recurrent collateral at these stages.
Although the functional rostral and caudal spread of the motoneuron recurrent collateral did not appear to change during this developmental period, it was still possible that within a segment the number of activated R-interneurons after stimulation of that segment's ventral root was changed between E8 and E10. To address this possibility we counted the number of optically active R-interneurons in a segment after stimulation of that segment's ventral root and marked a region of interest (ROI) for each R-interneuron (Fig. 5, B and F, arrowheads). To account for variability in the labeling procedure we normalized the number of optically identified R-interneurons (ventral root stimulation) to the number of optically active interneurons during an episode of network activity when virtually all the labeled neurons become active and exhibit changes in fluorescence (O'Donovan et al. 1994; Fig. 5, D and H; see METHODS for a detailed explanation of technique). The number of optically active neurons during episodes of network activity was higher at E8, possibly explained by the fact that cells were smaller at this stage and the tissue was more translucent, so more interneurons are sampled in our focal plane (Fig. 5, D and H). Across all E8 preparations, 41 of the 217 optically active interneurons in the focal plane were R-interneurons (five segments, four preparations) whereas at E10, 54 of the 147 optically active interneurons were R-interneurons (six segments, four preparations). When comparing the average percentage of R-interneurons from the E8 (19.8 ± 2.9%) and E10 (35.5 ± 5.8%) preparations we saw that a significantly higher proportion of optically active cells were R-interneurons at E10 (P < 0.05, two-tailed t-test). This increased R-interneuron population could result from an increase in the strength of existing motoneuron recurrent collateral connections and/or newly formed connections at the later stage. In summary, by E10 more R-interneurons were recruited and their projections to LS3 motoneurons were stronger, yet fewer LS4 motoneurons received this R-interneuron input.
Optical recordings show glutamatergic component to the recurrent activation of R-interneurons at both E8 and E10
A recent report suggested that the ventral root–evoked ventral root response was mediated by glutamatergic, GABAergic, and nicotinic transmission. To determine whether motoneuron-evoked activation of R-interneurons was partly dependent on glutamatergic transmission, we made optical recordings of R-interneurons as described earlier at E8 and E10. Further, this allowed us to test whether a developmental change in the glutamatergic component of R-interneuron activation could play a role in the circuit reorganization. We monitored the average pixel value of ROIs before, 30 min after the addition of, and 45 min after the washout of the 
-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptor [6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), 50 µM] and N-methyl-D-aspartate (NMDA) receptor [D-2-amino-5-phosphonopentanoic acid (AP5), 50 µM] antagonists (Fig. 6; in the presence of bicuculline throughout). We found that after addition of the glutamate receptor antagonists to the bath all the ROIs continued to exhibit detectable signals, although the intensity of the optical signal was reduced at both ages to a similar extent (see Fig. 6 legend). On washout, the signals were stronger than before glutamatergic antagonists were added (see Fig. 6 legend). This finding suggests that reduced optical signals were not the result of bleaching and are consistent with the compensatory nature of the cord excitability after transmitter antagonist application (Chub and O'Donovan 1998). At E10, optical signals were reduced in either CNQX or AP5, each to a smaller extent than when both antagonists were present (CNQX: 90.8 ± 1.7%, n = 23, P < 0.0001; AP5: 87.9 ± 2.7%, n = 17, P < 0.005). The findings suggest that the recurrent activation of R-interneurons has a glutamatergic component. Further, the extent of this glutamatergic contribution to the recurrent activation of R-interneurons did not change between E8 and E10 and thus was not likely to underlie the reorganization of the circuit.
|
|
|
DISCUSSION |
|---|
|
Mechanisms underlying inhibitory circuit reorganization
Ventral root–evoked ventral root responses demonstrated that caudal and rostral projections of the R-interneuron circuit became more concentrated in nearby segments later in development. The observation that the circuit modifications occurred in both directions makes it unlikely that the reorganization was dependent on different maturational states along the rostrocaudal axis of the cord. Intracellular motoneuron recordings obtained after LS2 ventral root stimulation demonstrated a change in the relative strength of the response in LS3 versus LS4 motoneurons by E10. We determined that fewer LS4 motoneurons and more LS3 motoneurons received the disynaptic response at E10. This suggests that one of the ways the circuit had been modified was through changes in the types of motoneurons (LS3 vs. LS4) to which the activated R-interneurons projected. In addition, from E8 to E10 the amplitude of the disynaptic potential became larger in LS3 motoneurons but not LS4 motoneurons. Because motoneuron input resistance was smaller at E10 the disynaptic currents were likely even more increased for LS3 motoneurons and slightly increased for LS4 motoneurons.
Circuit reorganization could also come about as a result of changes in the recruitment of R-interneurons after motoneuron stimulation. Previous reports in the neonatal cat demonstrated that there is a developmental elimination of synaptic boutons of the motoneuronal recurrent collateral, but this refinement process did not reduce the rostrocaudal extent of recurrent boutons (Cullheim and Ulfhake 1982; Remahl et al. 1985). Using calcium imaging techniques to assess the number and location of R-interneurons, we demonstrated that motoneuron recurrent collaterals were capable of recruiting R-interneurons only in the same segment as the activated motoneurons at both E8 and E10, consistent with a previous anatomical study showing that the recurrent collateral did not extend rostrocaudally (Velumian and Poliakova 1992).
Although there was no change in the rostrocaudal recruitment of R-interneurons, we did find that within a segment, more R-interneurons were recruited (higher percentage of optically active interneurons) after ventral root stimulation at E10. This is consistent with the idea that the recurrent motoneuronal connections to R-interneurons continue to form after E8, one day after motoneurons are first able to activate R-interneurons (Xu and Wenner 2005). An alternative possibility is that similar numbers of R-interneurons are recruited at both ages, but that our imaging technique was unable to pick up some R-interneurons that spike at E8, possibly because they do not fire enough action potentials to produce a detectable calcium transient. Such a possibility would suggest that the R-interneuron population is close to our threshold for optically detecting their activity. We do not favor this possibility because blockade of an excitatory (glutamatergic) component of the synaptic drive to R-interneurons resulted in a reduction in the intensity of the optical signal, yet all of the optically active cells were still detectable. This suggests that the R-interneurons were not close to the threshold for detection. Several observations argue against the possibility that more R-interneurons project into the VLF by E10 and are therefore retrogradely labeled with the calcium dye: there are fewer total optically active interneurons during network activity at E10, many interneurons are known to project several segments by E4.5 (Oppenheim et al. 1988), and whole cell recordings of R-interneurons were more commonly obtained at E10 than at E8 (Xu and Wenner 2005).
Although several transmitters are known to mediate the disynaptic R-interneuron circuit (nicotinic, glutamatergic, GABAergic), it is unlikely that the circuit reorganization could be explained by developmental changes in the relative contributions of these different transmitters. No clear changes in transmitters were observed in the ventral root–evoked ventral root potential from E8 to E15 in a previous study (Xu et al. 2005). The study did find a small glutamatergic contribution to the ventral root response, consistent with previous work suggesting that the motoneuron input to Renshaw cells in the mouse is partly glutamatergic (Mentis et al. 2005; Nishimaru et al. 2005). It remained possible that the motoneuronal recurrent activation of R-interneurons in the chick embryo also had a glutamatergic component and that its relative contribution could change during development. In the current study we found that indeed there is a small glutamatergic contribution to the activation of most R-interneurons, but that it does not change from E8 to E10, and therefore is unlikely to contribute to the reorganization.
Taken together, these results are consistent with a reorganization of the R-interneuron circuit between E8 and E10, where more R-interneurons are recruited after ventral root stimulation at E10. This increased population then projects more strongly to each LS3 motoneuron (larger-amplitude potentials) and projects to more LS3 motoneurons. These R-interneurons, during the same period, project to fewer LS4 motoneurons (see following text).
Potential synaptic refinement of the GABAergic R-interneuron
Although recruiting more R-interneurons at E10 would be expected to lead to stronger projections to LS3 motoneurons, the observation that fewer LS4 motoneurons received input from these activated R-interneurons at E10 was surprising. This finding suggests that there was an elimination of functional R-interneuron synaptic projections to LS4 motoneurons, rather than the loss of a subpopulation of R-interneurons that make the longer range projection. The results make it likely that R-interneuron projections to some LS4 motoneurons were functionally silenced between E8 and E10. This finding suggests that inhibitory interneurons within a network can undergo a functional synaptic refinement. Additional studies will be necessary to determine the period and anatomical extent of this refinement.
Much like the studies of excitatory projection neurons relaying a topographic sensory map, recent studies suggest that inhibitory projection neurons that relay a tonotopic map in the auditory system undergo a functional refinement, followed by a structural pruning later in development (Kandler and Gillespie 2005; Kim and Kandler 2003; Sanes and Friauf 2000; Sanes and Siverls 1991). Functional refinement in the auditory system, as appears to be the case for the R-interneuron, occurs at a developmental stage when GABA and glycine are depolarizing as a result of high intracellular chloride concentration (Kandler and Gillespie 2005; Kim and Kandler 2003). Therefore inhibitory refinement may be a general phenomenon that shares certain underlying mechanisms with excitatory refinement (depolarization leading to calcium entry).
Functions of circuit reorganization
R-interneuron circuit reorganization would likely affect the expression of spontaneous network activity that is experienced by the embryonic cord. R-interneurons are known to play an important role in the initiation of this spontaneous network activity (Wenner and O'Donovan 2001), which is important in the maturation of synaptic strength, axon pathfinding, and limb development (Gonzalez-Islas and Wenner 2006; Hall and Herring 1990; Hanson and Landmesser 2004; Jarvis et al. 1996; Persson 1983; Roufa and Martonosi 1981; Ruano-Gil et al. 1978; Toutant et al. 1979).
Another consequence of the R-interneuron circuit reorganization is that the circuit is more locally focused later in development. The pattern of rostrocaudal projections of the homologous Renshaw cell circuit in the adult cat is such that the strongest projections are those closer to the activated motoneurons, a phenomenon termed the proximity principal (Eccles et al. 1954, 1961; McCurdy and Hamm 1994; Renshaw 1941; Thomas and Wilson 1967). The Renshaw circuit achieves its rostrocaudal spread by the Renshaw axonal projection through the VLF (Hultborn et al. 1971; Jankowska and Smith 1973; McCurdy and Hamm 1994; Ryall et al. 1971). Similarly, R-interneurons project rostrocaudally to motoneurons in adjacent segments through the VLF (Wenner and O'Donovan 1999; Xu and Wenner 2005). The R-interneuron circuit reorganization found in this study could contribute to a more mature proximally projecting circuit, like that of the adult cat Renshaw circuit.
Although R-interneuron circuit reorganization does appear to concentrate the circuit's projections closer to the activated motoneurons when stimulating at LS2 or LS6, the totality of our results suggest a more complex phenomenon. For instance, from E8 to E10 there was no change in the response (frequency receiving input or amplitude of potential) in LS3 motoneurons to LS3 ventral root stimulation (same segment, Fig. 4). During the same period there was a dramatic increase in the proportion of LS3 motoneurons that received LS4 input (more distant, Table 1). Therefore a general rule that assumes all R-interneurons are the same—in that they all weaken their more distant projections and strengthen their closer connections—is overly simplistic. Therefore a more complex set of rules must be at play. It is possible that the reorganization (strengthening or weakening) could serve to sharpen the connections between functionally related R-interneurons and motoneurons from E8 to E10 and could act to remove functional mismatches in interneuronal connectivity as was previously proposed (Glover 2000; Mears and Frank 1997; Seebach and Ziskind-Conhaim 1994). For example, it may be that LS2 stimulation activates a population of R-interneurons that are most related to adductor motoneurons, and at E8 these project to adductor (thigh) and femorotibialis (leg extensor) motoneurons in the LS3 segment, and in the LS4 segment to iliofibularis (leg flexor) motoneurons, among others. By E10 the connections to adductor and femorotibialis motoneurons might be strengthened, whereas the connections to LS4 iliofibularis motoneurons are weakened or silenced. Because we did not identify motoneurons by the muscle they innervated in this study, but rather by the segment they occupied, we were not in a position to assess connectivity across functional subclasses. Alternatively, it is possible that rostrocaudal focusing does not occur across different motoneuron species, but occurs within a particular motor pool and would thus be dependent on the unique rostrocaudal location of that motor pool. Regardless, we were able to determine that circuit reorganization does occur from E8 to E10, and it will be important in future experiments to test whether certain motoneuron species lose or strengthen their R-interneuron inputs. Any of the above-mentioned changes would affect interneuronal circuits and therefore influence locomotor activity in the more mature animal.
Another aspect of the reorganization that is inconsistent with a simple proximity-based sharpening is the observation that long-range (four-segment) R-interneuron projections did not appear to weaken in the ventral root recordings. Renshaw cells are thought to project only one or two segments at most in the cat and mouse. Recent reports, in the developing mouse and rat, suggested the possibility that there are different classes of interneurons that receive direct input from motoneurons and that some project multiple segments and may use a different transmitter than Renshaw cells (Hanson and Landmesser 2003; Machacek and Hochman 2006; Mentis et al. 2005). It is possible that the interneurons mediating multisegment ventral root responses in the chick embryo are similar to the cells described in the developing rat and/or mouse. These longer projecting neurons would have distinct functions from the classical Renshaw cells, and have been proposed to provide a feedback excitation, and would maintain ongoing activity in the more mature system (Machacek and Hochman 2006). In the chick embryo, these long projecting cells may not undergo refinement because it may be important to coordinate the activity of some cells that are separated by multiple segments in the more mature cord. Alternatively, if these multisegmental projections are glutamatergic (Machacek and Hochman 2006; Mentis et al. 2005) then they may undergo refinement, but at a later stage in development when other excitatory projections in the spinal cord refine.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
1 The online version of this article contains supplemental data. 
Address for reprint requests and other correspondence: P. Wenner, Department of Physiology, Room 601, Whitehead Bldg., Emory University, School of Medicine, Atlanta, GA, 30340 (E-mail pwenner{at}emory.edu)
|
|
REFERENCES |
|---|
|
Bates CA, Killackey HP. The emergence of a discretely distributed pattern of corticospinal projection neurons. Brain Res 315: 265–273, 1984.[Medline]
Brink E, Jankowska E, McCrea D, Skoog B. Use of sucrose gap for recording postsynaptic population potentials evoked by single interneurones in spinal motoneurones. Brain Res 223: 165–169, 1981.[CrossRef][Web of Science][Medline]
Brooks C, Eccles JC, Malcolm JL. Synaptic potentials of inhibited motoneurons. J Neurophysiol 11: 417–430, 1948.
Butt SJ, Kiehn O. Functional identification of interneurons responsible for left-right coordination of hindlimbs in mammals. Neuron 38: 953–963, 2003.[CrossRef][Web of Science][Medline]
Cabana T, Martin GF. Corticospinal development in the North-American opossum: evidence for a sequence in the growth of cortical axons in the spinal cord and for transient projections. Brain Res 355: 69–80, 1985.[Medline]
Campbell G, Shatz CJ. Synapses formed by identified retinogeniculate axons during the segregation of eye input. J Neurosci 12: 1847–1858, 1992.[Abstract]
Chub N, Mentis GZ, O'Donovan MJ. Chloride-sensitive MEQ fluorescence in chick embryo motoneurons following manipulations of chloride and during spontaneous network activity. J Neurophysiol 95: 323–330, 2006.
Chub N, O'Donovan MJ. Blockade and recovery of spontaneous rhythmic activity after application of neurotransmitter antagonists to spinal networks of the chick embryo. J Neurosci 18: 294–306, 1998.
Chub N, O'Donovan MJ. Post-episode depression of GABAergic transmission in spinal neurons of the chick embryo. J Neurophysiol 85: 2166–2176, 2001.
Cullheim S, Ulfhake B. Evidence for a postnatal elimination of terminal arborizations and synaptic boutons of recurrent motor axon collaterals in the cat. Brain Res 281: 234–237, 1982.[Medline]
Eccles JC. Synaptic potentials of motoneurons. J Neurophysiol 9: 87–120, 1946.
Eccles JC, Eccles RM, Iggo A, Ito M. Distribution of recurrent inhibition among motoneurones. J Physiol (Paris) 159: 479–499, 1961.
Eccles JC, Fatt P, Koketsu K. Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurons. J Physiol 126: 524–562, 1954.
Glover JC. Development of specific connectivity between premotor neurons and motoneurons in the brain stem and spinal cord. Physiol Rev 80: 615–647, 2000.
Gonzalez-Islas C, Wenner P. Spontaneous network activity in the embryonic spinal cord regulates AMPAergic and GABAergic synaptic strength. Neuron 49: 563–575, 2006.[CrossRef][Web of Science][Medline]
Hall BK, Herring SW. Paralysis and growth of the musculoskeletal system in the embryonic chick. J Morphol 206: 45–56, 1990.[CrossRef][Web of Science][Medline]
Hamburger V, Hamilton HC. A series of normal stages in the development of the chick embryo. J Morphol 88: 49–92, 1951.[CrossRef][Web of Science]
Hanson MG, Landmesser LT. Characterization of the circuits that generate spontaneous episodes of activity in the early embryonic mouse spinal cord. J Neurosci 23: 587–600, 2003.
Hanson MG, Landmesser LT. Normal patterns of spontaneous activity are required for correct motor axon guidance and the expression of specific guidance molecules. Neuron 43: 687–701, 2004.[CrossRef][Web of Science][Medline]
Hensch TK. Critical period regulation. Annu Rev Neurosci 27: 549–579, 2004.[CrossRef][Web of Science][Medline]
Hensch TK. Critical period plasticity in local cortical circuits. Nat Rev Neurosci 6: 877–888, 2005.[CrossRef][Web of Science][Medline]
Hultborn H, Jankowska E, Lindstrom S. Relative contribution from different nerves to recurrent depression of Ia IPSPs in motoneurones. J Physiol 215: 637–664, 1971.
Jankowska E, Smith DO. Antidromic activation of Renshaw cells and their axonal projections. Acta Physiol Scand 88: 198–214, 1973.[Web of Science][Medline]
Jarvis JC, Sutherland H, Mayne CN, Gilroy SJ, Salmons S. Induction of a fast-oxidative phenotype by chronic muscle stimulation: mechanical and biochemical studies. Am J Physiol Cell Physiol 270: C306–C312, 1996.
Kandler K, Gillespie DC. Developmental refinement of inhibitory sound-localization circuits. Trends Neurosci 28: 290–296, 2005.[CrossRef][Web of Science][Medline]
Kasthuri N, Lichtman JW. The role of neuronal identity in synaptic competition. Nature 424: 426–430, 2003.[CrossRef][Medline]
Kim G, Kandler K. Elimination and strengthening of glycinergic/GABAergic connections during tonotopic map formation. Nat Neurosci 6: 282–290, 2003.[CrossRef][Web of Science][Medline]
Kudo N, Yamada T. Development of the monosynaptic stretch reflex in the rat: an in vitro study. J Physiol 369: 127–144, 1985.
Li Q, Martin JH. Postnatal development of differential projections from the caudal and rostral motor cortex subregions. Exp Brain Res 134: 187–198, 2000.[CrossRef][Web of Science][Medline]
Luo L, O'Leary DD. Axon retraction and degeneration in development and disease. Annu Rev Neurosci 28: 127–156, 2005.[CrossRef][Web of Science][Medline]
Luscher HR, Ruenzel P, Fetz E, Henneman E. Postsynaptic population potentials recorded from ventral roots perfused with isotonic sucrose: connections of groups Ia and II spindle afferent fibers with large populations of motoneurons. J Neurophysiol 42: 1146–1164, 1979.
Machacek DW, Hochman S. Noradrenaline unmasks novel self-reinforcing motor circuits within the mammalian spinal cord. J Neurosci 26: 5920–5928, 2006.
Martin JH. The corticospinal system: from development to motor control. Neuroscientist 11: 161–173, 2005.
McCormick DA, Contreras D. On the cellular and network bases of epileptic seizures. Annu Rev Physiol 63: 815–846, 2001.[CrossRef][Web of Science][Medline]
McCurdy ML, Hamm TM. Topography of recurrent inhibitory postsynaptic potentials between individual motoneurons in the cat. J Neurophysiol 72: 214–226, 1994.
Mears SC, Frank E. Formation of specific monosynaptic connections between muscle spindle afferents and motoneurons in the mouse. J Neurosci 17: 3128–3135, 1997.
Mentis GZ, Alvarez FJ, Bonnot A, Richards DS, Gonzalez-Forero D, Zerda R, O'Donovan MJ. Noncholinergic excitatory actions of motoneurons in the neonatal mammalian spinal cord. Proc Natl Acad Sci USA 102: 7344–7349, 2005.
Nishimaru H, Restrepo CE, Ryge J, Yanagawa Y, Kiehn O. Mammalian motor neurons corelease glutamate and acetylcholine at central synapses. Proc Natl Acad Sci USA 102: 5245–5249, 2005.
O'Brien RA, Ostberg AJ, Vrbova G. Observations on the elimination of polyneuronal innervation in developing mammalian skeletal muscle. J Physiol 282: 571–582, 1978.
O'Donovan M, Ho S, Yee W. Calcium imaging of rhythmic network activity in the developing spinal cord of the chick embryo. J Neurosci 14: 6354–6369, 1994.[Abstract]
O'Donovan MJ. Developmental approaches to the analysis of vertebrate central pattern generators. J Neurosci Methods 21: 275–286, 1987.[CrossRef][Web of Science][Medline]
O'Donovan MJ. Motor activity in the isolated spinal cord of the chick embryo: synaptic drive and firing pattern of single motoneurons. J Neurosci 9: 943–958, 1989.[Abstract]
O'Donovan MJ. The origin of spontaneous activity in developing networks of the vertebrate nervous system. Curr Opin Neurobiol 9: 94–104, 1999.[CrossRef][Web of Science][Medline]
O'Donovan MJ, Ho S, Sholomenko G, Yee W. Real-time imaging of neurons retrogradely and anterogradely labelled with calcium-sensitive dyes. J Neurosci Methods 46: 91–106, 1993.[CrossRef][Web of Science][Medline]
Oppenheim RW, Shneiderman A, Shimizu I, Yaginuma H. Onset and development of intersegmental projections in the chick embryo spinal cord. J Comp Neurol 275: 159–180, 1988.[CrossRef][Web of Science][Medline]
Persson M. The role of movements in the development of sutural and diarthrodial joints tested by long-term paralysis of chick embryos. J Anat 137: 591–599, 1983.[Web of Science][Medline]
Porter RJ. The absence epilepsies. Epilepsia 34: S42–S48, 1993.[Web of Science][Medline]
Redfern PA. Neuromuscular transmission in new-born rats. J Physiol 209: 701–709, 1970.
Reh TA, Constantine-Paton M. Retinal ganglion cell terminals change their projection sites during larval development of Rana pipiens. J Neurosci 4: 442–457, 1984.[Abstract]
Remahl S, Cullheim S, Ulfhake B. Dimensions and branching patterns of triceps surae alpha-motor axons and their recurrent axon collaterals in the spinal cord during the postnatal development of the cat. Brain Res 355: 193–200, 1985.[Medline]
Renshaw B. Influence of discharge of motoneurons upon excitation of neighbouring motoneurons. J Neurophysiol 4: 235–243, 1941.
Roberts MH, Wallis DI. Dorsal and ventral root potentials recorded in vivo by the sucrose-gap method. J Physiol 277: 42P–43P, 1978.[Medline]
Roufa D, Martonosi AN. Effect of curare on the development of chicken embryo skeletal muscle in ovo. Biochem Pharmacol 30: 1501–1505, 1981.[CrossRef][Web of Science][Medline]
Ruano-Gil D, Nardi-Vilardaga J, Tejedo-Mateu A. Influence of extrinsic factors on the development of the articular system. Acta Anat (Basel) 101: 36–44, 1978.[Web of Science][Medline]
Ryall RW, Piercey MF, Polosa C. Intersegmental and intrasegmental distribution of mutual inhibition of Renshaw cells. J Neurophysiol 34: 700–707, 1971.
Saito K. Development of spinal reflexes in the rat fetus studied in vitro. J Physiol 294: 581–594, 1979.
Sanes DH, Friauf E. Development and influence of inhibition in the lateral superior olivary nucleus. Hear Res 147: 46–58, 2000.[CrossRef][Web of Science][Medline]
Sanes DH, Siverls V. Development and specificity of inhibitory terminal arborizations in the central nervous system. J Neurobiol 22: 837–854, 1991.[CrossRef][Web of Science][Medline]
Seebach BS, Ziskind-Conhaim L. Formation of transient inappropriate sensorimotor synapses in developing rat spinal cords. J Neurosci 14: 4520–4528, 1994.[Abstract]
Simon DK, O'Leary DD. Development of topographic order in the mammalian retinocollicular projection. J Neurosci 12: 1212–1232, 1992.[Abstract]
Sretavan DW, Shatz CJ. Axon trajectories and pattern of terminal arborization during the prenatal development of the cat's retinogeniculate pathway. J Comp Neurol 255: 386–400, 1987.[CrossRef][Web of Science][Medline]
Stanfield BB, O'Leary DD. The transient corticospinal projection from the occipital cortex during the postnatal development of the rat. J Comp Neurol 238: 236–248, 1985.[CrossRef][Web of Science][Medline]
Sur M, Weller RE, Sherman SM. Development of X- and Y-cell retinogeniculate terminations in kittens. Nature 310: 246–249, 1984.[CrossRef][Medline]
Tello JF. Genisis de los terminosiones nerviosis motrices y sensitivas. Trab Lab Invest Biol Univ Madrid 15: 101–199, 1917.
Thomas RC, Wilson VJ. Recurrent interactions between motoneurons of known location in the cervical cord of the cat. J Neurophysiol 30: 661–674, 1967.
Thompson W, Kuffler DP, Jansen JK. The effect of prolonged, reversible block of nerve impulses on the elimination of polyneuronal innervation of new-born rat skeletal muscle fibers. Neuroscience 4: 271–281, 1979.[CrossRef][Web of Science][Medline]
Toutant JP, Toutant MN, Renaud D, Le Douarin GH. Enzymatic differentiation of muscle fibre types in embryonic latissimus dorsii of the chick: effects of spinal cord stimulation. Cell Differ 8: 375–382, 1979.[CrossRef][Web of Science][Medline]
Velumian AA, Poliakova LA. The demonstration of recurrent motor axon collaterals in the chick embryo by using horseradish peroxidase axonal transport. Zh Evol Biokhim Fiziol 28: 240–243, 1992.[Medline]
Wenner P, O'Donovan MJ. Identification of an interneuronal population that mediates recurrent inhibition of motoneurons in the developing chick spinal cord. J Neurosci 19: 7557–7567, 1999.
Wenner P, O'Donovan MJ. Mechanisms that initiate spontaneous network activity in the developing chick spinal cord. J Neurophysiol 86: 1481–1498, 2001.
Xu H, Whelan PJ, Wenner P. Development of an inhibitory interneuronal circuit in the embryonic spinal cord. J Neurophysiol 93: 2922–2933, 2005.
Young RR. Spasticity: a review. Neurology 44: S12–S20, 1994.[Medline]
This article has been cited by other articles:
|
|
 |
|
  Z. Wang, L. Li, M. Goulding, and E. Frank Early Postnatal Development of Reciprocal Ia Inhibition in the Murine Spinal Cord J Neurophysiol, July 1, 2008; 100(1): 185 - 196. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
