Coordination of the Bladder Detrusor and the External Urethral Sphincter in a Rat Model of Spinal Cord Injury: Effect of Injury Severity

Severity of spinal cord injury

At the time of contusion, the readings from the weight-drop apparatus provided information about the velocity of impact and the extent and duration of spinal cord compression (Gruner, 1992). From these data, the compression rate was calculated as the ratio of cord compression depth over time because this parameter was established previously to be the most sensitive measure of the initial impact (Constantini and Young, 1994). The average compression rate was found to be significantly different among the three SCI groups (Fig.1A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Varying degrees of contusive SCI were produced by dropping the 10 gm weight from the heights of 12.5, 25, and 50 mm.A, The compression rate calculated from weight-drop device measures was different among the three SCI groups. B, C, The hindlimb and locomotion behavioral tests done at 8 weeks after SCI indicated a significantly different level of somatic sensorimotor recovery as quantified by BBB (B) and CBS (C) in the 12.5 mm group compared with the more severe 25 and 50 mm groups. One asteriskindicates a significant difference from the 12.5 mm SCI group, andtwo asterisks indicate a difference between 25 and 50 mm groups, based on p < 0.05 in Tukey's post hoc test after ANOVA.

Behavioral tests of hindlimb function were done during the recovery period after SCI (day 1 and weekly through 8 weeks). As described previously (Gale et al., 1985; Basso et al., 1996), hindlimb functional deficits were maximal at day 1 and then showed partial recovery over the next several weeks with a plateau seen by 3 weeks after injury. At day 1 the three injury groups were indistinguishable, but by 1 week the 12.5 mm group demonstrated increased recovery compared with the 25 and 50 mm groups (data not shown). Hindlimb functional recovery in the small number of rats in the 25 mm group (n = 4) was not significantly different from that in the 50 mm group at any week after injury, although a nonsignificant trend toward enhanced recovery was seen on weeks 3–8. Chronic (week 8) behavioral measures of overall hindlimb functional deficits (Fig.1C, CBS) and of locomotion recovery (Fig.1B, BBB) showed significant differences between the 12.5 mm group, which attained weight-bearing locomotion, and the 25 and 50 mm groups, which on average did not. This is an important distinction, because voluntary overground locomotion with full weight support requires supraspinal control (Grillner and Dubuc, 1988; Mori et al., 1999).

After the behavioral and urodynamic evaluations on week 8, the chronic injury sites were examined for histopathology. As described previously for this and other models of contusion injury in the rat (Noble and Wrathall, 1985; Basso et al., 1996; Beattie et al., 1997), cross sections of the injury epicenter (region of maximal damage) demonstrated a central lesion devoid of normal spinal cord tissue surrounded by a peripheral rim of preserved white matter (Fig.2). White matter immediately adjacent to the central lesion appeared very lightly stained. More peripherally located white matter was closer to normal in its eriochrome cyanine staining. On the basis of a previous evaluation of residual white matter 8 weeks after SCI at the electron microscopy level (Wrathall et al., 1998), we interpreted the reduced myelin staining to be caused by a combination of axonal loss, which is greatest near the central cavity, as well as hypomyelination of the surviving axons. The amount of this residual myelinated white matter appeared to be less in the more severe injury groups. We quantified the area of residual white matter in lateral and ventral funiculi at the injury epicenter (Fig. 2,bottom right), because these funiculi have been shown previously to contain both the ascending and descending micturition-related pathways (Loewy et al., 1979; Fedirchuk and Shefchyk, 1991). The 12.5 mm group retained significantly greater residual white matter than did the more severely injured groups. They had more than triple the residual white matter of the 25 mm group and ∼10 times that of the 50 mm group. The residual white matter in the lateral and ventral funiculi of the 12.5 mm group was ∼20% of that in uninjured control rats. Taken together, the biomechanical, behavioral, and morphometric measures of SCI severity indicate that the 12.5 mm group of animals sustained significantly less initial traumatic injury and exhibited less neurological impairment and chronic white matter loss at the lesion site than did the 25 and 50 mm groups.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

White matter sparing at the SCI epicenter.Top, Bottom Left, Photomicrographs of representative sections through the lesion epicenter from the three SCI groups and from an uninjured control stained with eriochrome cyanine to label myelin. The dorsal, lateral, and ventral funicular white matter of the normal spinal cord is heavily stained, whereas little myelin staining is seen in the central gray matter. The cross-sectional profiles of the injured spinal cords are reduced in diameter. The center of the injured cords contains cavities and an abnormal loose network of cells, but no myelin staining is apparent. A peripheral rim of residual white matter is seen. Myelin staining is present but reduced compared with normal white matter, consistent with the chronic hypomyelination of residual axons. Bottom Right, The average areas of myelinated white matter from the ventral and lateral funicular zones at the lesion epicenter in the SCI groups. One asteriskindicates a significant difference from the 12.5 mm SCI group, based onp < 0.05 in Tukey's post hoc test after ANOVA. Scale bar, 250 μm.

Effect of injury severity on spontaneous LUT function

Initially after SCI the rats were not capable of spontaneous micturition, and their bladders were manually expressed twice daily. The volume of expressed urine was measured each time, and the data were used to estimate the initiation of spontaneous LUT function after SCI. In all rats the volume of manually expressed urine increased over the first few days after SCI and then decreased as spontaneous micturition was reestablished. On the basis of our previous data on increased bladder size with time after SCI (Pikov et al., 1998) and a limited study using metabolic cages (V. Pikov, unpublished observations), the increase in volume of expressed urine during the first 4 d after injury was interpreted as resulting from increased bladder size in the absence of spontaneous micturition. The subsequent decrease in manually expressed urine was interpreted to indicate the initiation of spontaneous micturition. Because of the maintenance of an enlarged bladder and the reduced voiding efficiency of SCI rats (Pikov et al., 1998), the volume of urine that could be expressed remained higher than that in uninjured rats even after spontaneous micturition was established. This was especially evident in the 25 and 50 mm groups. Thus, the injury groups demonstrated differences in that the more severely injured rats retained a higher residual urine volume and required a longer time period before displaying evidence of spontaneous micturition (Fig. 3). The average time for initiation of spontaneous micturition was significantly longer in the 50 mm group than in the 12.5 mm group (Table1). Chronically, at 8 weeks after SCI, bladder weights and volumes were significantly increased in the 25 and 50 mm groups compared with uninjured controls (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Time course of recovery of spontaneous voiding. Urinary bladders were expressed every 12 hr, and the collected urine volume was measured. The height of thebars indicates urine volume within one SE above and below the group mean at the specified time point. Compared with 25 and 50 mm SCI animals, the 12.5 mm SCI animals had progressively less urine collected with time after SCI.

Effect of injury severity on chronic urodynamic function

Urodynamic evaluation chronically at 8 weeks after SCI demonstrated that injury severity was correlated with the degree of recovery of detrusor–EUS coordination. Uninjured animals had large bladder detrusor contractions (Fig.4A; Table 1), which were accompanied by increased amplitude in EUS EMG (Fig.4D). The bladder contraction often had two incremental pressure increases, with the first intravesical pressure increase corresponding to a maximal increase in EUS activity and extensive voiding. A second increase in intravesical pressure (overshoot) was often seen and coincided with cessation of EUS activation and little or no voiding. Unlike uninjured animals, all SCI animals lacked the intravesical pressure overshoot and demonstrated somewhat decreased bladder detrusor contractions in response to bladder filling (Fig. 4B,C; Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Urodynamic analysis of detrusor–EUS coordination in representative uninjured (A, D, G, J), 12.5 mm SCI (B, E, H, K), and 50 mm SCI (C, F, I, L) animals. A–C, Bladder intravesical pressure (IVP) recordings during one voiding cycle. Solid horizontal lines in A and Bindicate the duration of stream-like voiding, and the dashed line in C indicates the drop-by-drop voiding.D–F, EUS EMG recordings, showing an activation of EUS EMG in relationship to the voiding cycle in uninjured and 12.5 mm SCI animals but not in the 50 mm SCI animal. G–I, Power spectrum analysis of EUS EMG activity as a function of time. A broad band of frequencies (5–40 Hz) shows an increased power during the voiding phase in uninjured and 12.5 mm SCI animals but not in the 50 mm SCI animal. J–L, Peak detection analysis of the ESA. Peaks were detected in 1 dsec intervals. An increase in ESA (dESA) occurred at the time corresponding to the voiding phase in uninjured and 12.5 mm SCI animals, but there was no change in the level of ESA in the 50 mm SCI animal.

The time–frequency analysis of EUS EMG yielded important quantitative parameters, such as mean EMG activity, mean EMG power frequency, and mean EMG-spiking activity (ESA) (Fig. 4). The timing of the coordinated EUS activation was unaltered by the SCI (data not shown) in the animals in which a detectable increase in EMG parameters was found. The EMG parameters were compared between the contraction and filling phases. All of the rats in the 12.5 mm SCI group exhibited bladder contractions leading to stream-like voiding (Fig. 4B) and an increase in the amplitude of EUS activity at the time of detrusor contractions (Fig. 4E). In contrast, in the animals from the 25 and 50 mm groups, the urine was released in a dropwise manner independent of detrusor contractions (Fig. 4C). Animals in these groups showed little or no increase in EUS EMG amplitude during detrusor contractions (Fig. 4F).

Changes in EUS EMG were evaluated during the progression of the micturition cycle by power spectrum analysis using the fast Fourier transform algorithm (Fig. 4G–I). During the voiding phase as compared with the filling phase, the power of the 5–40 Hz EMG frequency band was increased in both uninjured (Fig. 4G) and 12.5 mm SCI animals (Fig. 4H) but not in the 25 and 50 mm SCI animals (Fig. 4I). Peak detection analysis (Fig. 4J–L) demonstrated the increase in the ESA (dESA) during voiding in uninjured (Fig. 4J) and 12.5 mm SCI animals (Fig. 4K) but not in more severely injured animals, as illustrated by the representative animal from the 50 mm SCI group (Fig. 4L). The animals from the 12.5 mm SCI groups demonstrated dESA values that were not different from those of the uninjured group, whereas 25 and 50 mm SCI groups exhibited progressively decreased dESAs (Fig.5A). Among the studied EUS EMG parameters, the dESA was most sensitive to SCI and was therefore used in describing detrusor–EUS coordination in subsequent analyses.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Detrusor–EUS coordination was measured by the increase in EUS EMG during voiding (dESA). The value of dESA was correlated to SCI severity (A), to white matter sparing at the injury epicenter (T8 spinal segment) (B), to the amount of serotonin immunoreactivity (5HT-ir) in the DL nucleus (C), and to CRF immunoreactivity (CRF-ir) in the SPN (D). Correlation coefficients are shown at thetop left corner of B–D. InA, symbols indicate a significant difference from the uninjured group (one asterisk) or from both the uninjured and 12.5 mm groups (two asterisks). In B–D, the correlation coefficients are significant (p < 0.001).

Identification of spinal motoneurons innervating the bladder and the EUS

Motoneurons innervating the EUS in the rat are located in the DL nucleus of the L6–S1 ventral horn (Schroder, 1980), and those innervating the bladder are part of the SPN (Nadelhaft and Booth, 1984). To confirm the location of DL motoneurons at the lateral portion of the ventral horn (Schroder, 1980), we injected the retrograde tracer Fluororuby in the EUS using locations similar to those in which we placed EUS EMG wires. The general histological appearance of the spinal cord sections was traced and Fluororuby labeled, and nonlabeled ventral horn motoneurons were identified (Fig.6A, middle,C). Motoneurons that were retrogradely labeled from the EUS with Fluororuby were located in the lateral ventral horn, in the DL nucleus, over a length of ∼7 mm from the L5–L6 junction to the L6–S1 junction. At S1, however, the morphologically identifiable DL was no longer present as a nucleus (Fig. 6B,left), and no retrogradely labeled motoneurons were present in the ventral horn (Fig. 6B, middle).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

A, B, Bright-field photomicrographs and their drawings show sections through the DL nucleus at L6 (A) and the SPN at S1 (B).Left, Photomicrographs show neutral red-stained neurons.Middle, Tracings of these photomicrographs show the gray matter contour and outlines of motoneurons in the ventral horn that were backlabeled with the retrograde tracer Fluororuby (closed circles) or had no labeling (open circles).Right, Photomicrographs show the pseudorabies virus tracing at the corresponding levels of spinal cord. C, A fluorescence photomicrograph of the area indicated by thecircle in A (left) shows Fluororuby deposits in DL motoneurons that were retrogradely labeled from the EUS. D, A bright-field photomicrograph taken from the area indicated by the box in B(right) shows neurons in the SPN that were retrogradely labeled with pseudorabies virus from the bladder. Scale bars: A, B, 500 μm; C, D, 20 μm.

A separate animal was injected with pseudorabies virus into the bladder wall for transneuronal tracing as described previously (Pikov et al., 1998) to identify the location of bladder-projecting motoneurons in relationship to DL motoneurons. Sections through the L6 level (Fig.6A, right) indicated a lack of pseudorabies virus labeling in DL motoneurons. However, at the S1 level there was strong labeling in the intermediolateral gray in the location of motoneurons of the SPN (Fig. 6B, right,boxed area, D).

By coexamining the neutral red-stained sections with adjacent Fluororuby sections in the DL (Fig. 6A, circled area, left) and with the pseudorabies virus staining in the SPN (Fig. 6B, boxed area,left), the criteria for morphological identification of these areas were made. In subsequent immunohistochemical and in situ hybridization studies, the DL and SPN nuclei were identified on the basis of the neutral red-staining patterns.

Relationship between detrusor–EUS coordination and preservation of long tracts after SCI

White matter sparing at the epicenter was significantly correlated with dESA (Fig. 5B), indicating that the degree of general supraspinal connectivity was an important factor in recovery of detrusor–EUS coordination. The preservation of descending supraspinal connections believed to be involved in the control of detrusor–EUS coordination was further probed by immunohistochemistry for long tract-descending innervation of neurons in the DL and SPN. A portion of the supraspinal-descending control innervates the DL directly (Ding et al., 1995; Hermann et al., 1998). We therefore evaluated the supraspinal terminations around the EUS-innervating motoneurons in this area using serotonin immunoreactivity, because brainstem serotonergic pathways are involved in the control of micturition (McMahon and Spillane, 1982; Durant and Yaksh, 1988; Fukuda and Koga, 1991; Thor and Katofiasc, 1995; Espey et al., 1998).

The number of neutral red-stained motoneurons in the DL was unaffected by SCI: 9.7 ± 1.9, 9.8 ± 2.2, 10.2 ± 2.7, and 9.8 ± 2.0 neurons per section for uninjured and 12.5, 25, and 50 mm SCI groups, respectively. Similarly, the size of DL motoneurons, expressed as the cell body area, was similar in normal and SCI rats: 1113 ± 178 μm² compared with 1177 ± 153, 1157 ± 137, and 1251 ± 155 μm² in 12.5, 25, and 50 mm SCI groups, respectively. In uninjured rats, strong serotonin immunoreactivity was seen as a punctate stain surrounding DL motoneuronal cell bodies and their processes (Fig. 7A). All DL motoneurons demonstrated serotonergic innervation. DL motoneurons at 8 weeks after 12.5 mm SCI were also consistently associated with serotonergic innervation (Fig. 7B), and the density of innervation was similar to that of uninjured controls (103 ± 9%) based on comparison of injured and control sections on the same slide. All but 1 of 38 DL neurons analyzed in the 25 mm SCI group had associated serotonin immunoreactivity, but the density (positive pixels per unit area) was significantly reduced to 25 ± 19% of controls. In the 50 mm SCI groups many DL neurons (91 of 159) did not have associated serotonin immunoreactivity (Fig. 7C), and for those that did the density was lower than that in uninjured controls (24 ± 18%). The average density of serotonin immunoreactivity associated with DL motoneurons in individual rats was positively correlated to their detrusor–EUS coordination as measured by dESA (Fig. 5C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Immunoreactivity for descending projections to lumbosacral areas involved in micturition. A–C, Bright-field photomicrographs showing serotonin immunoreactivity in the DL. High levels of the immunoreactivity are seen in uninjured (A) and 12.5 mm SCI (B) animals, and virtually no labeling is seen in the 50 mm SCI animal (C). D–F, Bright-field photomicrographs showing CRF immunoreactivity in the SPN. Progressively decreased levels of immunoreactivity are seen in uninjured (D), 12.5 mm SCI (E), and 50 mm SCI (F) animals. Scale bar, 20 μm.

An important descending supraspinal projection to the SPN was identified using CRF immunoreactivity (Fig. 7D–F), which labels micturition- and distal colon-related pathways (Valentino et al., 1996, 2000). In uninjured rats, intense CRF immunoreactivity was associated with the SPN (Fig. 7D), and it was reduced in SCI animals (Fig. 7E,F). Quantification of CRF immunoreactivity showed an injury severity-dependent reduction chronically after SCI. As a percentage of pixel density of CRF immunoreactivity in uninjured control sections on the same slide, the SCI groups averaged 64 ± 3, 33 ± 11, and 21 ± 2% for the 12.5, 25, and 50 mm groups, respectively. The density of CRF immunoreactivity in the SPN was positively correlated with the degree of detrusor–EUS coordination as measured by dESA (Fig.5D).

In situ hybridization analysis of glutamate receptor subunit mRNA in DL motoneurons

In situ hybridization with oligonucleotide probes for NMDA NR1, NR2A, and NR2B and AMPA GluR1, GluR2, GluR3, and GluR4 subunit mRNAs indicated expression of these mRNAs in DL motoneurons of uninjured controls (Fig. 8). Quantification of grains showed no difference in the pattern of mRNA expressed between the uninjured animals and the animals from the 12.5 mm SCI group that demonstrated a normal level of detrusor–EUS coordination (Table 2). Significant increases were seen, however, in mRNA for GluR2 and NR2A in both the 25 and 50 mm SCI groups compared with uninjured controls or the 12.5 mm SCI group (Table 2). Regression analyses indicated a lack of significant correlation between levels of GluR2 or NR2A and dESA (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

Bright-field photomicrographs showing neutral red-stained DL motoneurons overlaid with autoradiography grains of NMDA subunit NR2A mRNA (left) and AMPA subunit GluR2 mRNA (right). Top, An uninjured animal shows moderate levels of mRNA expression.Middle, A 12.5 mm SCI animal has similar levels of mRNA expression. Bottom, A 50 mm SCI animal shows high expression of both NR2A and GluR2 subunit mRNAs. Scale bar, 20 μm.

Relationship between recovery of LUT function and hindlimb function after graded SCI

Injury severity had a generally similar effect on recovery of LUT function and hindlimb function. The more severe 25 and 50 mm SCI groups demonstrated a similar time course of recovery of spontaneous voiding (Fig. 3; Table 1), a similar incomplete recovery of detrusor–EUS coordination (Fig. 5A), and a similar chronic hindlimb functional deficit as measured by the CBS (Fig. 1C). Importantly, most of these animals did not recover weight-bearing locomotion, as measured by the BBB (Fig. 1B), which is believed to require supraspinal control (Grillner and Dubuc, 1988;Mori et al., 1999). In contrast, the 12.5 mm SCI group recovered coordinated weight-bearing locomotion (Fig. 1B) and virtually normal detrusor–EUS coordination (Fig. 5A). This recovery was associated with sparing of 20% of the white matter at the injury epicenter and normal chronic levels of serotonin immunoreactivity (Fig. 6B) and glutamate receptor subunit mRNA expression associated with DL motoneurons (Table 2). The relationship between the extent of chronic LUT function and chronic hindlimb function was further examined by linear regression analysis. There was a significant correlation between CBS and dESA (r² = 0.73; p< 0.001) and between BBB and dESA (r² = 0.71; p< 0.001).

Evaluating detrusor–EUS coordination

For this study we applied a method of digital data analysis used previously for surface EMG data from limb muscles (Feng and Mak, 1998) to characterize quantitatively detrusor–EUS coordination in the rat. This allowed flexibility for postacquisition data filtering and analyses in time–frequency domains. The time–frequency analysis yielded important quantitative parameters, such as mean EMG activity, mean power frequency, and mean ESA. Coordinated EMG activation during the bladder contraction phase was studied in terms of timing and the amount of change in the calculated EMG parameters relative to the filling phase. The former was unaltered, and the latter was significantly affected by SCI, with ESA being the parameter that was most sensitive to injury.

Role of supraspinal projections in detrusor–EUS coordination after SCI

There are several sites in the brainstem that are implicated in detrusor–EUS coordination. The EUS-controlling motoneurons, located in the DL nucleus of L6–S1 (Schroder, 1980), receive direct and indirect (via spinal interneurons) supraspinal projections mostly from Barrington's nucleus (pontine micturition center) (Vizzard et al., 1995; Nadelhaft and Vera, 1996; Marson, 1997), the D-region just ventral to Barrington's nucleus (Ding et al., 1995), and ventrolateral pontine periaqueductal gray (Marson, 1997; Ding et al., 1998; Matsuura et al., 1998). Barrington's nucleus is likely a primary micturition-controlling center (de Groat, 1990). Other brainstem nuclei with identified connections to bladder and EUS pathways are the raphe magnus, raphe pallidus, parapyramidal medullary reticular formation, subcoeruleus pars α, locus coeruleus, and the A5 and A7 nuclei (Vizzard et al., 1995; Marson, 1997).

Of these, cells in the raphe nuclei and nucleus paragigantocellularis in the medullary reticular formation produce serotonin (Marson, 1997). We, therefore, used serotonin as a marker of direct supraspinal projections to DL (Ramirez-Leon et al., 1994; Tang et al., 1998). Previous studies in our laboratory and the work of others have established that serotonin immunoreactivity distal to the injury site is a good indicator of both SCI severity and the degree of recovery of somatic sensorimotor function (Faden et al., 1988; Wrathall et al., 1994; Teng and Wrathall, 1997).

CRF is expressed in cells in Barrington's nucleus that project to the SPN (Valentino et al., 1996, 2000). The SPN contains preganglionic motoneurons that innervate the bladder and proximal urethra, as well as interneurons that convey information from Barrington's nucleus to DL motoneurons (Loewy et al., 1979; Nadelhaft and Booth, 1984; Nadelhaft et al., 1986; Pascual et al., 1993; Ding et al., 1997). We used CRF immunoreactivity to assess the micturition-related supraspinal connections to the SPN.

Both of the markers of supraspinal control of micturition that we used proved to be correlated to the degree of chronic detrusor–EUS coordination. Combined, these measures indicate that recovery of detrusor–EUS coordination is associated with sparing of supraspinal projections to areas in the lumbosacral spinal cord that control bladder and EUS function. An interesting distinction is that even the most mildly injured group (12.5 mm SCI) demonstrated significant reductions in CRF innervation of the SPN, which averaged 64% of normal at 8 weeks after injury. These same rats demonstrated normal (103%) levels of serotonin immunoreactivity associated with DL motoneurons. Approximately 80% of white matter is lost at the injury site in these animals and would be expected to include loss of serotonergic fibers from the brainstem. Indeed, unpublished data from our laboratory indicate a 40% reduction in the amount of serotonin immunoreactivity at 5 d after injury in 12.5 mm SCI rats. Therefore, sprouting and/or synaptic plasticity of spared serotonergic fibers may occur during recovery of detrusor–EUS coordination after injury to compensate for lost fibers.

Role of glutamatergic transmission in normal micturition and after SCI

Glutamate is widely distributed in both spinal and supraspinal circuits and thus cannot be readily used as a marker of supraspinal projections controlling micturition. The descending projections from Barrington's nucleus (Liu et al., 1995) described above and likely others involved in control of micturition use glutamate as a primary neurotransmitter (de Groat et al., 1998), with other substances like serotonin and CRF serving modulatory roles (Yoshiyama et al., 1994b; Pavcovich and Valentino, 1995; Hokfelt et al., 2000). It is important, therefore, to find alternative ways of studying the glutamatergic system in terms of its involvement in functional recovery after SCI. As the first effort in this direction, we investigated the pattern of ionotropic glutamate receptor subunit mRNA expression in DL motoneurons. A major finding was that the subunit expression levels in animals that recovered detrusor–EUS coordination were similar to those of uninjured controls, suggesting either that SCI and subsequent recovery did not involve the alteration in NMDA or AMPA subunit receptor patterns in DL motoneurons or that during recovery the receptor subunit patterns returned to normal values after an initial perturbation. In contrast, animals without recovery of detrusor–EUS coordination showed elevated levels of specific NMDA and AMPA subunits. If these changes in chronic mRNA levels are assumed to produce altered functional glutamate receptors, these alterations may be related to the aberrant hyperactivity of these motoneurons as seen in the spontaneous spastic activity chronically in lumbosacral motoneurons after SCI (Hiersemenzel et al., 2000; Little et al., 2000).

It should be noted that the absence of altered mRNA patterns in DL motoneurons in the rats that recovered detrusor–EUS coordination does not preclude a role of glutamatergic systems in this functional recovery. For example, significant alterations may take place in glutamatergic sensory, brainstem, and/or spinal cord interneuronal portions of the pathways involved or may involve other potential alterations in DL motoneurons that we did not study, such as altered expression of subunit splice variants that can occur after SCI (Prybylowski et al., 2001).

Adaptive plasticity in control of micturition: lessons from development

The establishment of supraspinal control of micturition during postnatal development involves adaptive synaptic plasticity. Studies by de Groat and colleagues (Kruse et al., 1993; de Groat et al., 1998) indicate that shortly after birth, animals undergo a switch from spinal to supraspinal control mediating the detrusor–EUS voiding reflex that persists through adulthood. This switch involves an activity-dependent competition between spinal and supraspinal inputs at individual synapses. It is possible therefore that recovery of detrusor–EUS coordination after incomplete SCI involves a similar adaptive activity-dependent mechanism. Continued recovery of voluntary EUS function weeks and months after SCI in humans (Schurch, 1999) suggests that there is an ongoing process of greater use of spared control. Because normal EUS control is mediated mostly by glutamatergic neurotransmission (de Groat et al., 1998), an improved coordination chronically after SCI may involve an adaptation in glutamatergic neurotransmission in the function of preserved supraspinal control. This hypothesis will be tested in our future experiments.

In conclusion, our results indicate that the recovery of detrusor–EUS coordination after SCI occurs in an injury severity-dependent manner. This functional recovery may offer a suitable model system to study the adaptive plasticity of residual supraspinal connections after SCI. Thorough pharmacological investigation of the key neurotransmitters and their receptors involved in the normal and recovered detrusor–EUS coordination may provide clues regarding natural mechanisms of recovery of supraspinal control. This information may serve as the basis of pharmacological and/or other strategies to enhance functional recovery after SCI.