| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Centre for Neuroscience, University of Alberta, Edmonton, Alberta T6G 2S2, Canada
Submitted 20 October 2003; accepted in final form 7 January 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
2 s), and enhanced by repeated stimulation. Brief high-frequency stimulation trains (0.5 s, 100 Hz, 1.5 x T) evoked even longer duration responses (5–10 s), with repeated bursts of activity that were similar to the repeated muscle spasms evoked in awake rats with stimulation trains or manual skin stimulation. Stimulation of a given dorsal root evoked long-lasting reflexes in both the ipsilateral and contralateral ventral roots. Long-lasting reflexes did not occur in the sacrocaudal spinal cord of acute spinal rats (S2 transection), which is similar to the areflexia seen in awake acute spinal rats. However, long-lasting reflexes could be made to occur in the acute spinal rat by altering K+ (7 mM) or Mg2+ (0 mM) concentrations, or by application of high doses of the neuromodulators norepinephrine (NE, >20 µM) or serotonin (5-HT, >20 µM). In chronic spinal rats, much lower doses of these neuromodulators (0.1 µM) enhanced the long-lasting reflexes, suggesting a denervation supersensitivity to 5-HT and NE following injury. Higher doses of NE or 5-HT produced a paradoxical inhibition of the long-lasting reflexes. The high dose inhibition by NE was mimicked by the 
2-adrenergic receptor agonist clonidine but not the 1-adrenergic receptor agonist methoxamine. In summary, the sacral spinal in vitro preparation offers a new approach to the study of spinal cord injury and analysis of antispastic drugs. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
; Young 1994). Although spasticity can severely impair residual motor function, it remains poorly understood, and standard antispastic drugs are often not well tolerated by patients (Penn 1990). To date, the study of spasticity has been restricted to human or whole adult animal in vivo preparations (Ashby and McCrea 1987; Heckman 1994; Taylor et al. 1997), because spinal cord injury usually occurs in adults, and spasticity can take months to be fully manifested. Thus one purpose of this paper was to develop and validate a new in vitro preparation for the study of spasticity in adult rats, where the whole chronically injured adult spinal cord is explanted and maintained in artificial cerebrospinal fluid (ACSF). This enables full control of the extracellular environment and thus detailed testing of the pharmacological basis of hyperreflexia associated with spasticity. To validate this in vitro preparation, we began by quantifying the ventral root hyperreflexia seen in vitro to demonstrate that it is similar to the hyperreflexia seen in the corresponding awake rats (Bennett et al. 1999a, 2004). Related intracellular motoneuron recordings are reported in Bennett et al. (2001c).
Recently, it is has been shown that a full spasticity syndrome develops in the tail musculature of rats about 1 mo after a low spinal transection (at the S2 sacral cord; chronic spinal state) (Bennett et al. 1999a). This tail spasticity is functionally relevant to spasticity after higher level injury in humans, because the anatomy, mechanics, and segmental control of tail muscles is similar to other proximal muscles, especially those of the axial muscles of the back and neck (Richmond and Loeb 1992; Richmond et al. 1992; Wada et al. 1994a,b), which can have major spasms after spinal cord injury in humans (Stauffer 1974). Furthermore, descending supraspinal and propriospinal innervation of the sacrocaudal spinal cord is similar to those at other spinal cord levels (Masson et al. 1991; Wada et al. 1993), and thus spinal cord injury has similar descending tract denervation. The hyperreflexia associated with tail spasticity has been quantified in the awake chronic spinal rat by examining the muscle reflexes (EMG) in response to stimulation of the caudal nerve trunk supplying the tail (Bennett et al. 1999a, 2004). The hallmark of these spastic reflexes is the emergence of a long-lasting reflex that is evoked by low threshold afferent stimulation, facilitated by repeated stimulation, and mediates the flexor/extensor spasms seen in these chronic spinal rats.
An interesting aspect of this low spinal model of spasticity is that the affected sacrocaudal spinal cord is small enough that it should, in principle, survive whole when explanted and maintained in vitro. Indeed, Long et al. (1988) have shown that the sacrocaudal spinal cord of normal adult rat can survive in vitro, although they used a sagittal hemisection to improve oxygenation. In the current experiments, we removed the whole sacrocaudal spinal cord of chronically injured adult rats, examined the ventral root reflex responses to dorsal root stimulation, and compared these reflexes to those seen in the awake chronic spinal rat (Bennett et al. 1999a, 2004). The results described in this paper indicate that the long-lasting reflexes were remarkably well preserved in vitro compared with in vivo, with a similar threshold, duration, and facilitation with repetition. Interestingly, we found that it was important to study the whole sacrocaudal cord with minimal tissue cutting during the preparation process. For example, an additional sagittal hemisection (method used by Long et al. 1988) eliminated the long-lasting reflexes, suggesting an important involvement of the contralateral and/or midline circuitry in the long-lasting spastic reflexes.
One characteristic of the awake acute spinal rat (as opposed to chronic spinal rat) that we found difficult to replicate in the in vitro preparation was the facilitation of low-threshold long-lasting reflexes by vigorous cutaneous stimulation (Bennett et al. 2004), mainly because pure cutaneous stimulation is difficult to evoke from dorsal root stimulation in vitro. However, we did find that by raising the excitability of the reflexes by altering extracellular K+ or Mg2+ concentrations, long-lasting reflexes appeared in the in vitro sacrocaudal cord of acute spinal rats (similar to the results of Long et al. 1988). Specific monoaminergic neuromodulators, such as norepinephrine (NE) or serotonin (5-HT), also facilitated the long-lasting reflexes in acute spinal rats. Interestingly, similar applications of monoaminergic neuromodulators in chronic spinal rats prolonged the already large long-lasting reflexes and had effects at much lower doses, suggesting major changes in sensitivity to these substances. Parts of this work have been presented in abstract form (Bennett et al. 1999b).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
). In the main experimental group, rats had a previous chronic transection at the S2 sacral spinal level 1–6 mo before (transection made in adult rats at >50 days of age). This produced pronounced spasticity in the tail musculature (chronic spinal group) (see Bennett et al. (1999a, 2004)). The sacrocaudal spinal cord was removed by transecting above the previous chronic transection so as not to re-injure the sacrocaudal cord (Fig. 1A). In control rats, there was no previous injury, and the transection during removal and preparation of the spinal cord was considered an acute injury (acute spinal group). All procedures were approved by the University of Alberta animal welfare committee.
|
CORD REMOVAL. Chronic spinal and normal rats were deeply anesthetized with urethane (0.18 g/100 g; maximum of 0.45 g per rat for rats over 250 g), and a T13–L6 dorsal laminectomy was performed to expose the lumbar and sacrocaudal spinal cord. The dura was opened, and modified ACSF (mACSF; at 21°C) was dripped liberally onto the cord for the remainder of the surgery. A suction line removed excess mACSF and blood. The rat was given pure oxygen to breathe, using a mask, for 5 min, until the dorsal vein of the cord appeared hyper-oxygenated (bright red). Following this, the carotids and jugular were cut to lower the blood pressure and minimize blood in the cord. The spinal cord was quickly removed by cutting the sacrocaudal roots as they traversed the L6 vertebra, transecting the lumbar cord, and gently lifting the cord caudally while cutting lumbar roots as they appeared. The cord was placed in a dissection dish containing oxygenated mACSF at 21°C. The rat was killed with a sodium pentobarbital overdose.
DISSECTION DISH. In the dissection dish, the long dorsal and ventral roots of the sacrocaudal cord were untangled, identified, and cut into a manageable 1–2 cm length in preparation for stimulation and recording. The excess spinal cord was cut away by re-transecting the cord just rostral to the S2 sacral spinal level (i.e., rostral to the chronic transection site). In some cases (not standard), the cord was also midsagittally hemisected (Long et al. 1988). The mACSF was changed to wash out the remaining urethane and tissue debris, and the cord was left to recover in the mACSF for 1 h.
ROOT IDENTIFICATION AND PREPARATION. In pilot experiments (4 rats), we traced the roots deep into the sacrum to find their entry points to identify them and their relation to the anatomy of the cord. The caudal roots were tightly attached to the terminae fibrum of the cord, which was characterized by the pigtail-like anterior artery that adhered to it (attached root shown in Fig. 1A, top). Moving more rostrally, the first root that was unattached to the terminae fibrum was usually the S4 sacral dorsal root (dorsal root arrow in Fig. 1A). Thus during the standard preparation for an in vitro experiment, this S4 root was used as a landmark, and the two separate roots rostral to this were assumed to be S3 and S2 (not shown in Fig. 1A for simplicity, although S2 lesion is shown). The relatively large caudal dorsal root adhering to the terminae fibrum was gently teased free, at which point it was usually possible to find an extremely fine caudal ventral root, which was also teased free. Both these caudal roots are referred to as Ca1, although they may also contain fibers from more caudal Ca2 and Ca3 roots. All the sacrocaudal dorsal roots are much larger in diameter than the ventral roots.
Although the sacral roots were unattached to the terminae fibrum, they adhered to the spinal cord for at least a segment before entering the dorsal horn (and for greater distances in older animals; >4 mo; data not shown in Fig. 1A). Because of their large size, they interfered with oxygenation of the dorsal sacrocaudal cord. Thus these roots were gently teased away from the cord, and this procedure markedly increased the length of time for which the long-lasting polysynaptic reflexes were viable (from 1 to >5 h; see RESULTS).
Recording chamber
After recovery in the dissection dish, the sacrocaudal cord was transferred to a recording chamber and bathed with normal ACSF flowing at 4–6 ml/min and maintained at 24–25°C. The cord was usually placed on its lateral side, with dorsal and ventral roots from one side upward and easily accessible for recording. The chamber had a narrow channel (5.5 mm wide) to hold the spinal cord, and this channel opened at one end to form a wider but shallow area for holding the long roots before mounting them for recording. ACSF flowed by gravity into the narrow channel near the rostral end of the cord and drained out through a wick placed in the wide shallow area near the end of the cord. The cord was held in place by pins at the rostral and caudal extremes and was supported on two layers of porous nappy paper pinned to a Sylgard bottom surface. The bottom of the narrow channel was angled to fit the tapering profile of the cord and thus allowed the ACSF level to be maintained as shallow as possible. Thus the total chamber volume was 0.25 ml, giving 20 bath changes/min at the 5 ml/min flow rate.
Root stimulation and recording
Ventral and dorsal roots were mounted for monopolar recording and stimulation on up to 8 chlorided silver wires suspended just above the ACSF at the root entry points. The root was wrapped around the wire in the air and quickly covered with a petroleum jelly/mineral oil mixture (1:1 by weight). Usually we mounted at least S4 and Ca1 dorsal and ventral roots on the left side of the cord and a further four roots from the remaining ipsilateral and contralateral roots. Three additional sliver wires were placed directly in the ACSF for the stimulation-return, monopolar recording reference, and instrumentation ground.
Dorsal roots were stimulated with 0.1-ms current pulses, usually at 0.1 mA [ 10 x threshold (10 x T); T = 0.009–0.01 mA; isoflex stimulator, AMPI]. The anode (+) was connected to the root, and the cathode (-) was connected to the nearby stimulation-return wire in the bath. Tests with stimulating one end of a root and recording the other end of the same root indicated that this simulation arrangement recruited the nerve effectively and did not cause any stray stimulation of nearby roots. Originally, we stimulated with the reverse polarity (cathode on the root and anode in bath), but found that this produced some anodal block (above 2–5 x T), which presumably occurred when the action potential propagated past the oil/water interface into the bath (which was positive). From the faster conduction times in our standard stimulation arrangement (anode on nerve) compared with the reverse polarity, we suppose that action potentials were initiated near the oil/water interface in our standard arrangement.
Ventral roots were recorded with a custom-built differential preamplifier (QT5), with one lead connected to the root and a second to the reference wire in the bath near the ventral roots. The root recordings were high-pass filtered at 100 Hz, low-pass filtered at 3 kHz, amplified by 5,000 times, and sampled at 6–10 kHz (Axoscope, Axon Instruments). Dorsal root stimulation was repeated more than five times with an inter-stimulation interval normally set at 20 s to give multiple ventral root reflexes for averaging. Before each experiment, we transferred a piece of unused dorsal root to the recording chamber and recorded one end while stimulating the other. This gave an estimate of the stimulation threshold (T; consistently 0.008–0.01 mA) and the conduction velocity (16–24 m/s at 25°C).
ACSF and drug applications
The normal ACSF used in the recording chamber had a composition (in mM) of 122 NaCl, 24 NaHCO3, 3 KCl, 2.5 CaCl2, 1 MgSO4, and 12 glucose mixed in distilled water (osmolarity of 298 mOsm). The ACSF was bubbled with 95% O2-5% CO2 to give a pH of 7.4. During dissection and recovery, mACSF was used to minimize neural and general metabolic activity. Originally, we made mACSF from normal ACSF mixed with urethane (Long et al. 1988), but found that ACSF with NaCl replaced by sucrose (at equal osmolarity; 0 mM NaCl, 214 mM sucrose) or ACSF made with kynurenic acid (an endogenous neuroprotectant and broad spectrum glutamate transmission blocker; Kekesi et al. 2002) worked better. The mACSF made with kynurenic acid became our standard method and yielded the most healthy preparations; its composition was (in mM) 118 NaCl, 24 NaHCO3, 1.5 CaCl2, 3 KCl, 5 MgCl2, 1.4 NaH2PO4, 1.3 MgSO4, 25 D-glucose, and 1 kynurenic acid (note the low Ca2+ and high Mg2+ used to further reduce excitotoxicity). Ventral root reflex responses were tested in normal ACSF and during manipulations of the ACSF, including removal of Mg2+ (0 mM MgCl2, 3.5 mM CaCl2), increase in K+ (7 mM KCl, 121 mM NaCl), and the addition of 5-HT, NE, L-DOPA, methoxamine, or clonidine (Sigma). We added the later drugs in increasing doses to give cumulative dose–response relations, with >15 min between each dose for steady-state responses to be reached at each dose.
Data analysis
Ventral root reflexes were quantified using custom software written in Matlab (Mathworks). The data were high-pass filtered at 800 Hz (1st order filter), low-pass filtered at 3 kHz (4th order Butterworth filter), and rectified. The rectified data were averaged over a 100- to 1,100-ms window poststimulus to estimate the tonic long-lasting reflex and over a 2- to 4-ms window to estimate the monosynaptic reflex. Background ventral root activity just before the stimulus was also averaged (in a 100-ms window). Such window averages were computed for each trial, and a grand mean and SD were computed across trials for each reflex component (5 trials/average). When there was no activity, the mean ventral root average was not zero, due to noise in the data that was rectified and thus not removed by averaging. This average noise signal was subtracted from all other measurements to remove its effect. Throughout text and figures, mean ± SD is shown. Statistical differences were computed with a Student t-test at the 95% confidence level.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
When the whole sacrocaudal spinal cord of normal rats was acutely transected and transferred to the in vitro recording chamber (transected at S2 level), long-lasting ventral root reflexes were never observed, regardless of the dorsal root stimulation intensity or rate (n = 22 rats; Fig. 1B). There was at times a brief short-latency reflex in the most caudal roots (n = 7/22), including a small monosynaptic response (0.33 ± 0.42 mV; Fig. 1B, top arrow), which was significantly smaller than the monosynaptic reflex in chronic spinal rats. This is consistent with the relatively areflexic state seen in the tail of awake acute sacral spinal rats or normal intact rats (see Fig. 1 in Bennett et al. 2004).
Chronic spinal state
In contrast, the sacrocaudal spinal cord of chronic spinal rats (>30 days after S2 transection, Fig. 1A) had large and long-lasting ventral root reflexes in response to dorsal root stimulation when tested in normal ACSF (Fig. 1, C and D; n = 30 rats). Again, this is consistent with the long-lasting reflexes seen in the tail muscles of awake chronic spinal rats (see Fig. 1, D and E, in Bennett et al. 2004). Furthermore, the awake chronic spinal rats that were most spastic prior to in vitro testing (top spasticity rating in Bennett et al. 2004) always had larger in vitro ventral root reflexes compared with less spastic rats.
The preservation of the long-lasting reflexes in vitro probably occurred because the spinal cord was relatively unaffected when transferred to the in vitro recording chamber as it was left whole and removed by cutting above the original transection. The main difference between the in vivo and the in vitro states was the temperature, which was 25°C in the latter. When the temperature was raised beyond 27°C, the reflexes disappeared irreversibly.
The reflexes generally increased and stabilized in the first 30 min after transferring the sacrocaudal cord to the recording chamber, after which they remained stable for hours. Over time, the reflexes on the most rostral roots (S3) disappeared first (after >5 h), presumably because of the poorer viability of the larger diameter spinal cord. Reflexes in more caudal roots lasted longer, for up to 18 h, although they gradually diminished with time.
The long-lasting reflex response to Ca1 dorsal root stimulation was always largest compared with other dorsal root stimulation. Thus we either stimulated the Ca1 root by itself (Fig. 2; n = 67 roots tested from 19 rats) or the ipsilateral S3, S4, and Ca1 dorsal roots together (Fig. 1; n = 52 roots from 14 rats), which in awake rats would be similar to stimulating the tip of the tail or the whole caudal nerve trunk, respectively (Ca1 innervates tip of the tail, Bennett et al. 2004). Either dorsal root stimulation arrangement evoked long-lasting reflexes in the Ca1 and S4 ventral roots (n = 99/117 roots) and occasionally in S3 roots (n = 13/32). The reflexes lasted for an average of 2.1 ± 0.47 s (Fig. 1D) and had similar components to those in the awake rat (Bennett et al. 2004): 1) a short latency initial peak, including a major polysynaptic component (double arrow in Fig. 1C); 2) a 50-ms pause in activity following the initial peak; and 3) a long-lasting tonic component (Fig. 1, C and D). The tonic component of the long-lasting reflex was quantified by averaging the rectified EMG over a 1-s window, starting 100 ms after the stimulus (see METHODS and Figs. 2, 3, 4) and is the main focus of this paper.
|
|
|
). The larger monosynaptic reflex might have been a result of the lower temperature, which is known to enhance synaptic transmission (Barron and Matthews 1938; Brooks et al. 1955; Pierau et al. 1976). The monosynaptic reflex recorded in vitro was significantly larger in chronic spinal rats than in acute spinal rats (1.9 compared with 0.33 mV).
|
skin afferents, as in the awake rat (Bennett et al. 2004). That is, it could be evoked by stimuli just above afferent threshold (which was 0.008–0.01 mA; n = 20/20 roots pairs tested at threshold; Fig. 2, B and A, top left). Likewise, the monosynaptic reflex was of low threshold (Fig. 2C). In most ventral roots, the long-lasting reflex got significantly larger with increasing stimulus intensity [up to 20 x T; n = 16/20; Fig. 2, A and B, solid symbols; group means, normalized to the maximum reflex (%max)] and thus was also in part due to higher threshold afferents. In other roots, high-threshold afferent stimulation had an inhibitory effect (n = 4/20), with the long-lasting reflex decreasing with intensities above 5 x T (example shown in Fig. 2B, open symbols).
ENHANCEMENT OF REFLEX WITH REPETITION. As in the awake rat (Bennett et al. 2004), the in vitro ventral root reflexes were enhanced by repeated stimulation (n = 36/39 roots tested). A 2-s interval was used to test this effect, since it produced the maximal enhancement in the awake rat, and the reflexes usually lasted for about 2 s, allowing interactions between successive stimuli (Fig. 3A). The long-lasting component built up in amplitude and duration with repetition (Fig. 3, A and C), although there was considerable variability in the responses (e.g., 2nd pulse in Fig. 3A). In contrast, the monosynaptic component was always depressed by this repeated simulation (Fig. 3, B and D), consistent with the effects of rate depression and associated presynaptic mechanisms (Thompson et al. 1998). Rate depression of the monosynaptic reflexes also occurred in the acute spinal rat (data not shown). A longer interval of 20 s produced no significant interactions between reflexes (neither monosynaptic nor long-lasting components; data not shown).
Reflexes evoked by stimulation trains
Recently, we have shown that motoneurons of chronic spinals rats have pronounced persistent inward currents (PICs; sodium and calcium currents) that amplify the synaptic currents, prolong the motoneuron discharge, and ultimately produce the long-lasting reflexes in the present sacrocaudal preparation (see DISCUSSION and Bennett et al. 2001c; Li and Bennett 2003; Li et al. 2004). However, because these PICs are only slowly activated, short stimuli, such as from a single dorsal root stimulation pulse, should not be as effective as longer stimuli in evoking PICs that produce long-lasting reflexes (Li and Bennett 2003; Li et al. 2004). Therefore we also examined the responses to 0.5-s-long 100-Hz stimulation trains at primary afferent intensity (1.5 x T). In all cases, stimulation trains evoked significantly longer duration responses (6.0 ± 2.8 s; Fig. 4A) than the responses to single shocks (2.1 ± 0.47 s; Fig. 4C; n = 37 roots tested; note that Fig. 4B is an expanded copy of Fig. 4A, on the same time scale as Fig. 4C). Furthermore, these long-lasting responses to stimulation trains were more robust than the single pulse responses, in that they occurred more reliably in all ventral roots (including S3), they were often largest at the lowest stimulation threshold (compare Fig. 4, B and E), and they continued to occur for longer periods as the preparation deteriorated (at >5 h). Also, they were more resistant to pharmacological manipulations (see below).
Interestingly, a brief low-threshold stimulation train (0.5 s) sometimes triggered two or more long-lasting bursts of activity (1–10 s each; 2 bursts shown in Fig. 4A; n = 10/37 roots), similar to the repeated spasms seen in awake chronic spinal rats (compare with Fig. 5 of Bennett et al. 2004). This likely occurred because these stimulation trains mimicked the large afferent barrage that occurred during the mechanical stimulation that was used to evoke spasms in the awake rat (Bennett et al. 1999a, 2004). This spasm-like activity was usually irregular, although at times there was rhythmic activity that occurred reciprocally in left and right roots (data not shown). Spasm-like bursting could also be evoked with higher threshold stimulation (10 x T; n = 12/37; Fig. 4D). Longer duration trains (2 s) were also effective, although very long stimuli (7 s) often produced inhibition of ongoing activity (data not shown). Stimulation trains never evoked long-lasting activity in acute spinal rats, regardless of the intensity or duration.
Contralateral long-lasting reflexes
The standard single dorsal root stimulation pulse also evoked long-lasting reflexes in the contralateral ventral roots (Fig. 5C; n = 28/30 rats). These contralateral reflexes were in most respects identical to those on the ipsilateral side (Fig. 5A): they had a short latency polysynaptic component (Fig. 5D) and were low threshold, enhanced with repetition, and altered by neuromodulators (see below). However, there was never a monosynaptic component in contralateral ventral root responses (Fig. 5D). Interestingly, the contralateral long-lasting reflexes (Fig. 5C) were significantly larger than the ipsilateral reflexes (Fig. 5A), with a given dorsal root stimulation (23 ± 16% increase). Likewise, a given ventral root responded with a significantly larger long-lasting reflex response to its contralateral dorsal root stimulation (Fig. 5, E and H) compared with its corresponding ipsilateral dorsal root (Fig. 5, F and G; n = 56 root pairs tested). These crossed reflexes were reciprocally organized, since stimulating both ipsi- and contralateral dorsal roots together gave a smaller response compared with individually (n = 4/4; data not shown). This crossed reciprocal organization functionally corresponds to the withdrawal reflexes seen in the awake chronic spinal rat, where the tail is flexed away from a cutaneous contact (Bennett et al. 2001a).
Lesions and localization of neurons involved in long-lasting reflex
The importance of the contralateral side of the spinal cord in the long-lasting reflexes was also made clear from lesion experiments. After the contralateral side was removed by a midsagittal hemisection (one-half the cord removed while cord was in mACSF; see METHODS), the long-lasting reflexes on the remaining ipsilateral roots were eliminated (n = 10/10 rats). In contrast, this sagittal hemisection caused a significant increase in the monosynaptic reflex (Harvey et al. 1999), so at least we know that the ipsilateral primary afferents and motoneurons were not injured by this procedure.
Other lesions also reduced or eliminated the long-lasting reflexes in chronic spinal rats. For example, minor injury to the dorsal cord during removal of the dorsal vein and attached pia (which we only did in initial trials to improve oxygenation) resulted in a preparation with weak long-lasting reflexes, even though the awake rat was initially very spastic (n = 3 rats). Thus the dorsal surface of the cord is important for the reflexes, and this is consistent with the additional finding that leaving the large sacral dorsal roots attached over the dorsolateral funiculus greatly reduced the long-lasting reflexes, presumably through interfering with oxygenation (see METHODS). Slightly stretching the caudal end of the cord during preparation could likewise reduce or eliminate the reflexes (n = 5 rats). In contrast, complete transections at the S3 level, below the chronic S2 lesion, did not disrupt the long-lasting reflexes (n = 3 rats). Thus it appears that the contralateral, dorsal, and caudal portions of the cord are minimally essential for the production of long-lasting reflexes.
|
In the awake acute spinal rat, some sustained reflex activity could be evoked in the tail muscles by high-threshold cutaneous/nociceptive afferent stimulation of the tip of the tail, and this could be used as a conditioning stimulation to facilitate the low-threshold, long-lasting reflexes (Bennett et al. 2004). In contrast, in the in vitro preparation, sustained ventral root activity could never be evoked in the acute spinal preparation, perhaps because it was not feasible to selectively stimulate the cutaneous afferents. However, we found that the low-threshold, long-lasting reflexes could be facilitated in vitro in acute spinal rats by pharmacological means. For example, simply raising the K+ to 7 mM (to generally depolarize cells and lower K+ currents) enabled long-lasting reflexes (n = 5; Fig. 6B), although these reflexes were significantly smaller and shorter than in the chronic spinal rat (see Fig. 6F; compare time scales in Figs. 6 and 7), and usually only in the caudal roots. Like-wise, increasing neuronal excitability, by lowering the Mg2+, also enabled a similar long-lasting reflex (n = 10; Fig. 6C; at times in all ventral roots, S3–Ca1), consistent with Long et al. (1988). However, these reflexes were significantly smaller than in the chronic spinal rat (Fig. 6F).
|
|
). However, these reflexes were variable and required large doses (Fig. 6, D and E) compared with chronic spinal rats. That is, 5-HT had no effect at doses below 20–30 µM, and above this dose, there were long-lasting reflexes, although they were much smaller than in the chronic spinal rat (20–100 µM; Fig. 6, D and F). These high doses usually evoked brief periods of spontaneous ventral root discharges (lasting 5 min, just after the application of 5-HT), and the long-lasting reflexes were still small during this spontaneous activity (Fig. 6D). Likewise, NE only facilitated long-lasting reflexes and spontaneous activity at high doses (20–100 µM). Furthermore, these reflexes could only be evoked by stimulation trains (Fig. 6E), and the single stimulation pulses never gave a long-lasting reflex (Fig. 6F). Facilitation of reflex activity in chronic spinal rat
In the chronic spinal rat, the long-lasting reflexes were significantly increased in duration by raising the K+ to 7 mM (compare Fig. 7, A and B), although the reflex amplitude measured in the first second of activity was not significantly increased (Fig. 7F; n = 7 roots tested). Interestingly, with high K+, the reflex responses to a single stimuli fluctuated markedly (Fig. 7B), with repeated spasm-like activity, as was seen in the awake rat, and as also seen with 100-Hz stimulation trains. When the excitability was raised by lowering the Mg2+, the long-lasting reflex duration was likewise increased (Fig. 7C), and in this case, the reflex amplitude was also significantly increased (Fig. 7F; n = 8).
The neuromodulatory transmitters 5-HT and NE also altered the long-lasting reflexes (Fig. 7, D and E), although with a complex dose dependency (Figs. 8 and 9; n = 13 and 12 roots tested with 5-HT and NE, respectively; L-DOPA acted similarly to NE; data not shown). The chronic spinal rat was much more sensitive to these substances than the acute spinal rat, with enhanced reflexes at doses orders of magnitude lower than the doses to evoke activity in acute spinal rats (0.1–1 µM; compare Figs. 6 and 7). With this low dose of 5-HT or NE the long-lasting reflex to a single pulse was significantly increased in amplitude (67%; Fig. 7F; see 
in Fig. 9, A and B) and duration (54%, Figs. 7 and 8C). Likewise, the long-lasting reflex responses to a stimulation train were significantly increased in amplitude and duration (Fig. 8D, see 
in Fig. 9, A and B). In contrast, the monosynaptic was not increased (see 
in Fig. 9, A and B).
|
|
). This high dose of NE (or 5-HT) evoked long periods of spontaneous activity (5–10 min), usually shortly after application (Fig. 8E). However, the reflexes were still inhibited during this activity (Fig. 8E, arrow), and even when this activity had subsided, the reflexes remained lowered (Fig. 8F). The monosynaptic reflexes were not blocked by high doses, although they were somewhat reduced in 30 µM NE (Fig. 9B, ). Furthermore, the associated long-lasting responses to a stimulation train were also not blocked by high doses of 5-HT or NE (no inhibition at high doses; Fig. 8, G compared with B), and instead, were significantly larger than control levels at all doses 
0.1 µM (Fig. 9, A and B, ; see DISCUSSION). Washout of these drugs took long periods to reverse the changes in reflexes (>1 h) and was thus unreliable.
The 1-adreneric receptor agonist methoxamine facilitated the long-lasting reflex to a single dorsal root pulse at all doses (n = 6 root pairs tested) in both duration (to 18.6 ± 11 from 2.1 ± 0.5 s; significant increase) and amplitude (Fig. 9C, ; amplitude change not significant due to high variability), consistent with its action on facilitating PICs on motoneurons (see DISCUSSION). No high-dose inhibition was seen, unlike with NE. Furthermore, methoxamine had no inhibitory effect on the monosynaptic reflex () and markedly facilitated the long-lasting response to a 100-Hz stimulation train (Fig. 9C, ; significant increase).
In contrast, at very low doses, the 2-adrenergic receptor agonist clonidine significantly reduced the long-lasting response to a single stimulation pulse (0.03 µM; n = 8 roots tested), and at higher doses (
0.1 µM), eliminated the reflex in most ventral roots tested (n = 6/8 roots; Fig. 9D, ), consistent with its action on blocking polysynaptic reflexes (see DISCUSSION). All doses of clonidine only marginally reduced the monosynaptic reflex (only significant reduction at 0.1 dose; Fig. 9D, ) and did not significantly reduce the long-lasting response to a stimulation train (Fig. 9D, ), with a moderate increase in this train-evoked reflex at the highest dose (Fig. 9D).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
, 2004; Burke et al. 1970; Kuhn and Macht 1948; Noth 1991; Powers et al. 1989), can be seen in vitro with ventral root recordings and dorsal root stimulation. With this new preparation, we have provided evidence that the long-lasting reflexes in these spastic rats are enhanced by neuromodulators such as 5-HT and have a major contralateral component. The mechanisms for these results are discussed further below. First, however, we will summarize the close parallels between the reflexes in the in vivo and in vitro preparations, and thus help validate the study of spinal cord injury and spasticity using this in vitro method, despite the many arbitrary factors of in vitro preparations such as ACSF composition, osmolarity, pH, tissue oxygenation, and temperature (Jiang et al. 1991; Kerkut and Bagust 1995). Comparison of in vivo and in vitro reflexes
The ventral root reflexes to dorsal root stimulation in the in vitro sacrocaudal spinal cord are very similar to the tail muscle reflexes seen with stimulation of the caudal nerve trunk of the tail in the awake rat, including their long duration, detailed reflex components (short latency initial peak, pause, and tonic discharge), involvement of the lowest threshold afferents, particular sensitivity to the most caudal tail afferents, and enhancement with repeated stimulation (Bennett et al. 2004). Likewise, in the acute spinal state, the reflexes recorded in vitro and in vivo are similar in that they are small and of short duration, and in these preparations, long-lasting reflexes can be facilitated by strong cutaneous stimulation or drug applications to enhance the excitability of the spinal cord. The one clear difference in the reflexes recorded in vitro is the presence of a moderately large monosynaptic reflex, which is not usually seen in the tail muscle responses of awake rats (Bennett et al. 2004). This may be related to the lower temperature required for tissue survival in vitro (25°C), which is known to increase synaptic transmission (Barron and Matthews 1938; Brooks et al. 1955; Pierau et al. 1976), although many other factors might increase the monosynaptic reflex in the sacrocaudal spinal cord, including damage to the contralateral or midline parts of the spinal cord (see RESULTS and Harvey et al. 1999).
Because the sacrocaudal cord mediating these reflexes in chronic spinal rats is at the end of the neural axis (Fig. 1A), it can be removed and transferred to a recording chamber without any additional injury to the cord, which most likely assures that the long-lasting reflexes to dorsal root stimulation are well preserved. Indeed, when additional injury is deliberately made to the dorsal, contralateral, or caudal cord (e.g., hemisection), the long-lasting reflexes are eliminated.
Role of contralateral spinal cord in spasticity
We had not anticipated the large contralateral reflexes seen in the in vitro preparation, although in retrospect, this was a mistake due to the special anatomy and function of the sacrocaudal spinal cord (Jankowska et al. 1978; Ritz et al. 1989, 1991, 2001). There is crossed segmental reflex control in the sacral cord, which is not seen at higher spinal segments, including crossed disynaptic inhibition, crossed monosynaptic excitation from Ia afferents, and crossed presynaptic inhibition of afferents (Akatani et al. 2002; Harvey et al. 1999; Jankowska et al. 1978; Wada and Shikaki 1999). Functionally, motoneurons to muscles on both sides of the tail must at times be reciprocally coordinated and at others co-contracted, as we have observed (Fig. 5), because the tail acts as a single axial appendage. For example, the muscles on the left side of the tail are activated to withdraw the tail from a noxious stimulation on the right side, whereas left and right ventral tail muscles are co-contracted to flex the tail down away from a dorsal stimulation, and both these responses occur in the chronic spinal rat (Bennett et al. 2001a). The finding that ablation of the contralateral cord, by sagittal hemisection, eliminates the long-lasting spastic reflexes further stresses the importance of crossed or midline circuitry in spasticity of axial musculature.
Neuromodulatory control over the long-lasting spastic reflex
Following chronic spinal cord injury, there are major changes in the neuronal excitability of the spinal cord caudal to the injury, which ultimately lead to long-lasting spastic reflexes (Bennett et al. 2001c; Li and Bennett 2003; Li et al. 2004). In particular, motoneurons exhibit a marked increase in excitability, as a result of the emergence of large voltage-gated PICs (Li and Bennett 2003), which are not seen in acute spinal rats. These PICs are regeneratively activated once a critical depolarization of the motoneuron occurs, and produce subthreshold plateau potentials and self-sustained firing that ultimately cause the several seconds long portion of the long-lasting reflex (Fig. 10 of Li et al. 2004). However, because these PICs have relatively slow kinetics, they are only activated by moderately long depolarizations (0.1–1 s, depending on amplitude, Bennett et al. 2001c; Li and Bennett 2003). Short depolarizations, such as from a monosynaptic excitatory postsynaptic potential (monosynaptic EPSP), are not sufficient to evoke full PICs or associated long-lasting reflexes. However, repeated high-frequency stimulation (100 Hz, 0.5-s trains) summates both short and long EPSPs, and these stimulation trains are particularly effective in evoking PICs and long-lasting reflexes (Bennett et al. 2001b; Li et al. 2004). This is consistent with the very large train-evoked long-lasting reflexes described in the RESULTS, and the facilitation of the train evoked reflexes by methoxamine, clonidine, NE, and 5-HT, all of which facilitate PICs/plateaus (Hounsgaard and Kiehn 1989; Lee and Heckman 1999).
Oddly enough, a single stimulation pulse to the caudal dorsal roots can activate PICs and a corresponding long-lasting reflex in chronic spinal rats (Bennett et al. 2001c; Li et al. 2004), despite the slow activation of PICs. This occurs because caudal dorsal root stimulation evokes a relatively prolonged polysynaptic EPSP (pEPSP; 0.5 s long) that can be seen when the PICs are blocked either pharmacologically or by hyperpolarization (Li et al. 2004). This pEPSP is sufficiently long to trigger the PICs in motoneurons of chronic spinal rats and thus triggers a several seconds long-lasting reflex response. Thus long-lasting reflexes triggered by a single stimulation shock require both pEPSPs and motoneuron PICs, whereas long-lasting reflexes triggered by a stimulation train do not require a pEPSP and only require PICs and a short EPSP that will summate. Therefore the present finding that the 2 agonist clonidine (or high-dose NE or 5-HT) blocks the long-lasting response to a single stimulation shock, but not a stimulation train, suggests that clonidine likely inhibits the pEPSP that triggers the PICs and long-lasting reflex, consistent with its known inhibitory action on polysynaptic pathways (Chau et al. 1998; Stone et al. 1998) and lack of inhibitory action on motoneuron plateaus/PICs (Conway et al. 1988). Clonidine does not block the monosynaptic reflex and facilitates PICs/plateaus (Conway et al. 1988), and thus in the presence of clonidine, a stimulation train still produces a prolonged synaptic excitation that likely activates a PIC and long-lasting reflex. In contrast, the 1 agonist methoxamine likely does not block the pEPSP, since it does not inhibit the single dorsal root response, and instead enhances it, likely by enhancing the PICs.
Supersensitivity to monoamines in chronic compared with acute spinal rats
The doses of 5-HT and NE required to facilitate the long-lasting reflexes in acute spinal rats were 100 times more than those in chronic spinal rats, and thus there is a remarkable supersensitivity to monoamines that occurs with long-term injury, as previously reported (Barbeau and Bedard 1981). The long-lasting reflexes induced by these monoamines in acute spinal rats likely resulted from induction of PICs (Bennett et al. 2001c), although apparently a higher dose was required in acute compared with chronic spinal rats to affect the PICs. In acute spinal cats, the noradrenergic precursor L-DOPA like-wise evokes late, long-lasting reflexes (FRA reflexes; Jankowska et al. 1967), and these are in part mediated by PICs/plateau potentials on motoneurons (Hounsgaard et al. 1988).
Summary and functional implications
Our results demonstrate that the long-lasting spastic reflex that emerges following spinal cord injury can be studied in vitro in the sacrocaudal cord; thus this preparation enables the detailed study of the pharmacological basis for spasticity and ultimately can be used to develop and test new antispastic drugs. Thus far, we have shown that the long-lasting spastic reflexes critically depend on the spontaneous emergence of plateau potentials (Bennett et al. 2001a,c) and associated PICs in motoneurons (L-type Ca2+ and persistent Na+ currents; Li and Bennett 2003) and emergence of polysynaptic EPSPs that are sufficiently long to activate PICs (Li et al. 2004). Furthermore, bath application of 5-HT, NE, or the 1-adrenergic receptor agonist methoxamine enhances the long-lasting reflexes at unusually low doses (supersensitivity, Figs. 7, 8, 9), perhaps due to a facilitation of plateaus/PICs.
Following complete spinal transection, 5-HT and NE levels below the injury are much reduced, but not eliminated (Newton and Hamill 1988), and thus may still play a major role in spasticity, given the pronounced supersensitivity to these neuromodulators. Furthermore, in incomplete spinal cord–injured patients, NE and 5-HT derived from spared descending pathways could further modulate spasticity and may thus in part explain the high variability of spastic reflexes in patients with spinal cord injury and observation that incomplete injury can produce a more exaggerated spastic syndrome. We have found that the neuromodulators 5-HT and NE can both facilitate and inhibit the long-lasting reflexes depending on the dose, with inhibition likely being mediated by a block of pEPSPs. Importantly, this inhibition can be mimicked by the 2-adrenergic receptor agonist clonidine, and thus the antispastic action of clonidine (Anderson et al. 1982) or related drugs such as tizanidine (Nance et al. 1997) may be used to block these polysynaptic pathways.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
GRANTS
Funding was provided by the Canadian Institutes of Health Research, Canadian Foundation for Innovation, and the Alberta Heritage Foundation for Medical Research.
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: D. Bennett, 513 HMRC, Div. of Neuroscience, Univ. of Alberta, Edmonton, Alberta T6G 2S2, Canada (E-mail: bennettd{at}ualberta.ca).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Anderson RJ, Brigham P, Cady WJ, and Kellar KJ. Role of alpha-2 adrenergic affinity in the action of a non-sedative antispasticity agent. Acta Neurol Scand 66: 248-258, 1982.[ISI][Medline]
Ashby P and McCrea DA. Neurophysiology of spinal spasticity. In: Handbook of the Spinal Cord, edited by Davidoff RA. New York: Marcel Dekker, 1987, p. 120-143.
Barbeau H and Bedard P. Denervation supersensitivity to 5-hydroxytryptophan in rats following spinal transection and 5,7-dihydroxytryptamine injection. Neuropharmacology 20: 611-616, 1981.[CrossRef][ISI][Medline]
Barron DH and Matthews BHC. Dorsal root reflexes. J Physiol 94: 26P-29P, 1938.
Bennett DJ, Gorassini M, Fouad K, Sanelli L, Han Y, and Cheng J. Spasticity in rats with sacral spinal cord injury. J Neurotrauma 16: 69-84, 1999a.[ISI][Medline]
Bennett DJ, Gorassini MA, and Siu M. In vitro preparation to study spasticity in chronic spinal rats. Soc Neurosci Abstr 25: 1394, 1999b.
Bennett DJ, Li Y, Harvey PJ, and Gorassini M. Evidence for plateau potentials in tail motoneurons of awake chronic spinal rats with spasticity. J Neurophysiol 86: 1972-1982, 2001a.
Bennett DJ, Li Y, and Sanelli L. Role of NMDA in spasticity following sacral spinal cord injury in rats. Soc Neurosci Abstr 31: 933.11, 2001b.
Bennett DJ, Li Y, and Siu M. Plateau potentials in sacrocaudal motoneurons of chronic spinal rats recorded in vitro. J Neurophysiol 86: 1955-1971, 2001c.
Bennett DJ, Sanelli L, Cooke C, Harvey PJ, and Gorassini MA. Spastic long-lasting reflexes in the awake rat after sacral spinal cord injury. J Neurophysiol 91: 2247-2258, 2004.
Brooks CM, Koizumi K, and Malcolm JL. Effects of changes in temperature on reactions of the spinal cord. J Neurophysiol 18: 205-216, 1955.
Burke D, Gillies JD, and Lance JW. The quadriceps stretch reflex in human spasticity. J Neurol Neurosurg Psychiat 33: 216-223, 1970.
Chau C, Barbeau H, and Rossignol S. Effects of intrathecal alpha1- and alpha2-noradrenergic agonists and norepinephrine on locomotion in chronic spinal cats. J Neurophysiol 79: 2941-2963, 1998.
Conway BA, Hultborn H, Kiehn O, and Mintz I. Plateau potentials in alpha-motoneurones induced by intravenous injection of L-dopa and clonidine in the spinal cat. J Physiol 405: 369-384, 1988.
Harvey PJ, Gorassini MA, Yakovenko S, Gritsenko V, and Bennett DJ. Contralateral presynaptic inhibition of the monosynaptic reflex in the sacral spinal cord. Soc Neurosci Abstr 25: 1394, 1999.
Heckman CJ. Alterations in synaptic input to motoneurons during partial spinal cord injury. Med Sci Sports Exerc 26: 1480-1490, 1994.[ISI][Medline]
Hounsgaard J, Hultborn H, Jespersen B, and Kiehn O. Bistability of alpha-motoneurones in the decerebrate cat and in the acute spinal cat after intravenous 5-hydroxytryptophan. J Physiol 405: 345-367, 1988.
Hounsgaard J and Kiehn O. Serotonin-induced bistability of turtle motoneurones caused by a nifedipine-sensitive calcium plateau potential. J Physiol 414: 265-282, 1989.
Jankowska E, Jukes MG, Lund S, and Lundberg A. The effect of DOPA on the spinal cord. 5. Reciprocal organization of pathways transmitting excitatory action to alpha motoneurones of flexors and extensors. Acta Physiol Scand 70: 369-388, 1967.
