Altered distribution of interstitial cells and innervation in the rat urinary bladder following spinal cord injury

Abstract Changes in the distribution of interstitial cells (IC) are reportedly associated with dysfunctional bladder. This study investigated whether spinal cord injury (SCI) resulted in changes to IC subpopulations (vimentin-positive with the ultrastructural profile of IC), smooth muscle and nerves within the bladder wall and correlated cellular remodelling with functional properties. Bladders from SCI (T8/9 transection) and sham-operated rats 5 weeks post-injury were used for ex vivo pressure–volume experiments or processed for morphological analysis with transmission electron microscopy (TEM) and light/confocal microscopy. Pressure-volume relationships revealed low-pressure, hypercompliance in SCI bladders indicative of decompensation. Extensive networks of vimentin-positive IC were typical in sham lamina propria and detrusor but were markedly reduced post-SCI; semi-quantitative analysis showed significant reduction. Nerves labelled with anti-neurofilament and anti-vAChT were notably decreased post-SCI. TEM revealed lamina propria IC and detrusor IC which formed close synaptic-like contacts with vesicle-containing nerve varicosities in shams. Lamina propria and detrusor IC were ultrastructurally damaged post-SCI with retracted/lost cell processes and were adjacent to areas of cellular debris and neuronal degradation. Smooth muscle hypertrophy was common to SCI tissues. In conclusion, IC populations in bladder wall were decreased 5 weeks post-SCI, accompanied with reduced innervation, smooth muscle hypertrophy and increased compliance. These novel findings indicate that bladder wall remodelling post-SCI affects the integrity of interactions between smooth muscle, nerves and IC, with compromised IC populations. Correlation between IC reduction and a hypercompliant phenotype suggests that disruption to bladder IC contribute to pathophysiological processes underpinning the dysfunctional SCI bladder.


Introduction
Spinal cord injury leads to a typical neuropathic bladder exhibiting urinary retention, infections, altered contractility and detrusorsphincter dyssynergia (DSD). Early stage post-injury is characterized by an areflexic bladder followed by emergence of a spinal micturition reflex enabling limited voiding. Urinary retention, typical in acute and chronic SCI represents a significant clinical problem and is associated with high or low detrusor pressure with the former presenting serious complications such as renal failure. However, low-pressure retention, commonly attributed to reduced detrusor contractility is less amenable to surgical correction and thus persists longer [1,2]. The pathogenesis of low-pressure retention is unknown and it is important to determine if it is a failure of muscle contractility or a remodelling of tissues within the bladder wall.
The established rat model of neuropathic SCI bladder exhibits time-dependent changes in contractility and compliance coupled with alterations in bladder wall structure [3]. The remarkable complexity that exists within the bladder wall with structural interactions occurring between smooth muscle cells (SMC), nerves, IC, microvessels and the urothelium has been shown over the last decade. The novel IC have attracted particular interest and while their precise roles in bladder physiology have not been fully elucidated, they may modulate bladder smooth muscle activity [4, 5 for recent review]. Interestingly, changes in IC distribution are associated with dysfunctional bladder [6]. Overactive bladder (OAB) mucosal tissue from human and neurogenic rat bladder contain increased Cx43 labelling, associated with lamina propria IC [7,8] moreover, increased numbers of KIT-positive IC occur in overactive human detrusor [9]. Guinea-pig obstructed bladder exhibits enhanced sub-urothelial and sub-serosal IC [10,11]. Intriguingly, the underactive, unobstructed, distended bladders typical of megacystis-microcolon intestinal hypoperistalsis syndrome (MMIHS) apparently lack IC [12].
The aim of this study was to evaluate our original hypothesis that 'changes in bladder IC populations following SCI are associated with bladder dysfunction' by (1) investigating the distribution of IC in SCI rat bladder, (2) examining changes in smooth muscle and innervation post-SCI and (3) correlating the morphological findings with compliance, evaluated by ex vivo pressure-volume experiments.

Rat SCI model
Protocols complied with European Community Council Directive 86/609/EEC and French legislation under a French Ministry license to Neureva (France). Bladders from adult female Sprague-Dawley rats post-SCI (T8 laminectomy, isoflurane anaethesia 2% in O2, 0.6 L/min) or sham-operated controls (shams) were obtained from Neureva. Animals were assisted with bladder voiding for 2 weeks by abdominal compression twice daily until micturition reflexes emerged and were injected daily, for the first week with gentamicin (2 mg/kg). After 5 weeks, rats were weighed, killed by cervical dislocation and exsanguination; the bladders were removed, weighed and stored in Ca 2ϩ -free Krebs at 4ЊC. Data are expressed as mean Ϯ S.E.M.; statistical comparisons were made using Student's unpaired t-test with significance accepted at P Ͻ 0.05.

Immunohistochemistry and electron microscopy
Bladders were opened longitudinally and pinned to the Sylgard base of a dissecting dish. The mucosal layer containing urothelium and lamina propria was removed by sharp dissection to leave the underlying detrusor layer. Mucosal and detrusor tissue sheets were labelled with antibodies as previously described [13] and imaged with confocal microscopy (Nikon C1). Control tissues were prepared by (1) omission of all antibodies to control for autofluorescence; (2) omission of the primary antibody to control for non-specificity of the secondary antibody and (3) guinea-pig bladder was used as a positive control. Details of the antibodies used are shown in Table 1. Cell counts were made by two independent, blinded researchers on reconstructed z-stacks from non-overlapping fields of view. Vimentin-positive IC with at least two cytoplasmic projections from a central cellular soma were counted for each field of view and expressed per unit area (100 m 2 ). Images from slides stained with anti-neurofilament or anti-vAChT were reconstructed as z-stacks in Image J (NIH, Bethesda, MD, USA) software, the background was subtracted, image intensity (integrated density) was measured and expressed per 100 m 2 . Data are expressed as mean Ϯ S.E.M.; comparisons were made with Mann-Whitney test with significance accepted at P Ͻ 0.05.
Tissues fixed in 8% neutral buffered formalin were processed for wax histology. Sections (5 m) were deparaffinized in xylene, rehydrated in alcohol, washed in tap water and heated (pressure cooker) in Tris-EDTA (pH 9.0) for antigen retrieval. Endogenous peroxidase activity was blocked with 3% H2O2, sections were washed, blocked with 1% BSA in Tris buffer solution (TBS) and incubated in anti-vimentin overnight. Sections were then washed in TBS, incubated in rabbit antimouse horseradish peroxidase conjugate (1:200), washed, stained with diaminobenzidine tetrahydrochloride (DAB) and substrate chromogen system (DAKO, Cambridgeshire, UK) before counterstaining with haematoxylin (Autostainer, Leica) and viewing with light microscopy.
Bladders removed from sham (N ϭ 3) and SCI (N ϭ 3) rats immediately after killing were processed for TEM as previously described [14].

Results
Animals (N ϭ 21 in each group) were weighed before surgery and pre-sacrifice. Ex vivo bladders were weighed post-sacrifice ( Fig. 1). Significant differences were not found in animal mass pre-surgery (229.9 Ϯ 2.3 g sham versus 230.4 Ϯ 2.5 g SCI, P ϭ 0.87); however, post-surgery SCI animals were significantly lighter than

Pressure-volume relationships
Bladders were filled at a constant rate and the corresponding pressure changes recorded. Differences in the pressure-volume relationships ( Fig. 2) between sham controls and SCI bladders indicated that SCI bladders were more compliant than shams. Compliance was calculated from the secondary, linear part of pressure-volume relationships and was greater in SCI bladders (26.5 Ϯ 4.7 l/cmH2O, N ϭ 8, mean Ϯ S.D.) compared to control (2.9 Ϯ 0.9 l/cmH2O, N ϭ 9).  Whole-mount preparations of sham mucosa imaged with confocal microscopy showed networks of stellate-shaped, vimentinpositive cells (Fig. 4A and B), however SCI tissues had rounded cells, with truncated cellular branches which appeared unable to form networks (Fig. 4C). Dense networks of vimentin-positive stellate-shaped cells and elongated, bipolar cells at the edge of smooth muscle bundles, were found in sham detrusors (Fig. 5A, C and E). SCI detrusors had a striking reduction in vimentinpositive cells, containing only small patches of non-networked cells (Fig. 5B, D and F). At higher magnification, vimentin-positive cells in SCI detrusors lacked branched processes, becoming rounded or unipolar. Semi-quantitative analysis showed that

Visualisation of bladder nerves
The neuronal plexus labelled with anti-neurofilament found in sham lamina propria was markedly reduced post-SCI ( Fig. 6A and B). Fluorescence intensity (arbitrary units per 100 m 2 ) was reduced from 26,771 Ϯ 2621 in shams (n ϭ 18 micrographs, N ϭ 3 bladders) to 12,777 Ϯ 2150 (n ϭ 14 micrographs, N ϭ 3 bladders). Dense innervation in sham detrusor, typified by large nerve trunks bifurcating and becoming progressively smaller became regions of patchy denervation alongside morphologically disrupted nerves post-SCI ( Fig. 6C and D Fig. 7). Summary data are shown in Figure 7E.

Transmission electron microscopy
TEM revealed differences in the ultrastructural integrity of components of the bladder wall between shams and SCIs. Cells with the typical ultrastructural profile of bladder IC (discontinuous basal lamina, central nuclei, lateral branches, rough and smooth endoplasmic reticulum, mitochondria, caveolae and cytoplasmic vesicles; Refs. [14] and [15]) were abundant in the lamina propria of shams with frequent IC-IC contacts ( Fig. 8A and B). Recent publications indicate that tissues may contain cells distinct from fibroblasts, myofibroblasts or interstitial cells of Cajal (ICC), termed 'telocytes' [16,17]. Telocytes have been defined as having multiple cell projections, which are very thin (less than 0.2 m), characteristically long (may be Ͼ100 m) and having multiple dilations and caveolae. It is possible that the IC described in this study may encompass telocytes although their presence in bladder has not yet been reported. Lamina propria IC post-SCI had fewer cellular associations and a retraction/loss of cell branches was evident ( Fig. 8C and D). Sham detrusor (Fig. 9A-D) contained IC, which were frequently associated with nerve varicosities and SMC; significantly, IC made close contacts with nerve varicosities. Neuronal structures were either missing or too difficult to identify due to degradation in SCIs (Fig. 9E-H), which typically contained areas of cellular debris. Moreover, SMC also showed a degree of degradation. IC cell bodies, exhibiting significant degradation were found in the detrusor, these were more difficult to find post-SCI, consistent with the reduction in vimentin-positive cells seen in confocal micrographs. Ultrastructurally, damaged IC exhibited cytoplasmic vacuolization, swollen endoplasmic reticulum, apparent collapse of the cytoskeleton in the perinuclear area and some swollen mitochondria.

Discussion
The human SCI bladder exhibits early (acute) and late (chronic) changes in function, also exhibited by the spinal cord transected (T8/9 or T9/10) rat. After the acute areflexic phase, chronic SCI bladders develop voiding reflexes attributed to new spinal cord reflexes [18], which enable limited voiding. Simultaneous contraction of detrusor and urethral smooth muscle (detrusorsphincter dyssynergia) however, results in urinary retention, increased workload on the bladder and subsequent bladder hypertrophy [3]. This study focussed on the chronic SCI rat bladder, which is reported to have increased compliance and structural changes including smooth muscle hypertrophy [19,20]. Functional experiments demonstrated that the SCI bladders in the present study had the hypercompliant phenotype typical of chronic SCI. Interestingly, chronic SCI led to a marked reduction in IC sub-populations, labelled with the broad IC marker, anti-vimentin [13,15,21], which had been abundant in shams. Lamina propria IC may have been less affected than detrusor IC but their distribution and morphology were clearly altered post-SCI. TEM confirmed that IC cell bodies were present in the lamina propria post-SCI but that their morphological integrity and structural connections were compromised. The remarkable near-absence of IC in confocal images of post-SCI detrusor was confirmed with TEM which demonstrated occasional, ultrastructurally damaged IC. The ultrastructural damage to the bladder IC which included swollen mitochondria, cytoplasmic vacuolisation, enlarged endoplasmic reticulum is reminiscent of that reported for gastrointestinal ICC in Crohn's disease [22]. Recent evidence supports a correlation between numbers of IC and bladder activity in pathophysiological conditions [6]. Increased sub-urothelial and sub-serosal IC in (overactive) obstructed guinea-pig bladder [10,11] and a lack of IC in the non-contractile bladders of patients with MMIHS has been reported [12]. Increased Cx43 expression in rat mucosa 4 weeks post-SCI [7] and in neurogenic or idiopathic OABs [8,23] may indicate up-regulation in numbers of Cx43-positive IC. Biers et al. [9] reported increased KIT-positive IC in human neuropathic detrusor. If IC modulate detrusor activity in normal bladder and their up-regulation correlates with an OAB phenotype, then their loss in the low-pressure, hypercompliant chronic SCI bladder in this study would be consistent with a decompensated phenotype.
The patchy denervation observed in this study is in agreement with findings from human neurogenic (NDO) bladder biopsies [24,25]. Axonal degeneration observed by TEM in human NDO biopsies, notably exceeded the degree of smooth muscle degeneration within the same biopsy [26,27]. We also found remarkable areas of degeneration and cell debris throughout the detrusor (where neuronal structures would be typically located) adjacent to SMC, which themselves showed a degree of degeneration. Post-SCI decreases in vAChT-immunoreactivity in pelvic ganglia and bladder wall nerves have been reported to recover, accompanied with improvement in bladder function [28]. However, the significant decrease in detrusor vAChT varicosities post-SCI in this study indicated that at 5 weeks post-injury, recovery was incomplete. A link between innervation and maintenance of the IC phenotype has been demonstrated in gut where anterograde degeneration of vagal afferent innervation resulted in a concomitant degeneration of IC and nerves [29]. This may indicate that release of a neuronal trophic factor is necessary for maintenance of the IC phenotype; interestingly, stem cell factor, the natural ligand of the IC KIT receptor is present in gastrointestinal nerves [30]. This study demonstrated close synaptic-like contacts between nerve varicosities and IC in sham bladders, which were not detected post-SCI likely due to nerve degeneration and reductions in IC. These findings may suggest that bladder IC rely on neuronal release of trophic factors, for example KIT for their maintenance.
Alternative explanations for IC degeneration post-SCI may be related to detrusor SMC hypertrophy linked with increased levels of the neurotrophin, nerve growth factor (NGF). Increased mRNA and protein expression of NGF in chronic SCI rat bladder is Fig. 6 Confocal imaging of bladder neural networks. (A) Whole mount, flat sheet preparations of mucosa from sham bladders were labelled with anti-neurofilament and imaged with confocal microscopy. All images are reconstructions of optical sections. A neural plexus was observed in these sections comprising larger nerves, which bifurcated to increasingly smaller fibres. (B) In SCI animals, the neural plexus was significantly disrupted with an apparent reduction in nerves. (C) The detrusor region of sham control bladders was densely innervated with large nerve trunks subdividing into increasingly smaller nerves, which became varicosities (arrows) within the smooth muscle bundles. Mucosal tissues from SCI animals also had strikingly compromised neuronal distribution with only occasional patches of immunopositive nerves. (D) In SCI detrusor, the innervation was comparatively sparse with only patchy areas of nerves visible. associated with bladder hypertrophy [18] moreover, NGF is released from bladder SMC in vitro following mechanical stretch [31,32]. The association between bladder hypertrophy and increased NGF is common to bladder pathophysiologies including SCI, bladder inflammation/cystitis and urethral obstruction [33].
Interestingly, intrathecal neutralisation of NGF improved bladder function in SCI rats [34]. Taken together, hypertrophied smooth muscle in response to chronic SCI-induced DSD could release NGF resulting in IC degeneration. If so, our findings may indicate that NGF is an anti-survival factor for IC maintenance.  After SCI, damaged IC (dIC) were apparent, although their branched process and cell-cell contacts were less extensive than sham controls. dIC typically had cytoplasmic vacuolisation (v), swollen endoplasmic reticulum (arrows), apparent collapse of the cytoskeleton in the perinuclear area (arrowheads) and some swollen mitochondria (m). Blood vessels are visible (BV).