Urinary dysfunction after spinal cord injury: Comparing outcomes after thoracic spinal transection and contusion in the rat

Spinal cord


Introduction
Normal micturition relies on synchronised contraction and relaxation of the two functional units of the lower urinary tract (LUT): the urinary bladder and the urethra.Such coordination relies on intricate neuronal circuits involving afferent and efferent pathways as well as supraspinal centres.Spinal cord injuries (SCI) rostral to the lumbosacral spinal cord eliminate voluntary supraspinal control of micturition.SCI is followed by a period of little or no bladder reflex activity, termed spinal shock, with partial or full urinary retention during the immediate days/ weeks after spinal insult (Anderson et al., 2023;Bywater et al., 2018).Later, as neuroplastic mechanisms operating at the lumbosacral spinal cord are established, including sprouting of afferent fibres and establishment of new synaptic contacts, an alternative micturition reflex pathway, independent from supraspinal input, emerges (de Groat and Yoshimura, 2012;Wada et al., 2022).Micturition becomes involuntary and largely inefficient due to strong and frequent detrusor contractions, demonstrated by urodynamic evaluation, giving rise to neurogenic Abbreviations: CGRP, calcitonin gene-related peptide; DSD, detrusor sphincter dyssynergia; GAP43, growth-associated protein-43; LUT, lower urinary tract; NDO, neurogenic detrusor overactivity; SCI, spinal cord injury; TH, tyrosine hydroxylase; VAChT, vesicular acetylcholine transporter; SCT, spinal cord transection; SCC, spinal cord contusion; mSCC, mild contusion; sSCC, severe contusion.
detrusor overactivity (NDO) (Gajewski et al., 2018).NDO is often concurrent with detrusor sphincter dyssynergia (DSD), defined as a loss of coordination between the detrusor and the urethral sphincter (Stoffel, 2016).Consequently, the bladder may not be efficiently emptied, causing bladder enlargement, due to an accumulation of high volumes of residual urine, giving rise to dangerously high intravesical pressures (Wada et al., 2022;Wyndaele, 2016) which may be deleterious to the upper urinary tract.
The reorganisation of the lumbosacral neuronal pathways governing micturion following SCI is accompanied by noteworthy alterations in innervation and histology of LUT organs (Ferreira et al., 2022;Shimizu et al., 2023), including neuronal sprouting at the spinal cord (Frias et al., 2015;Oliveira et al., 2019;Vizzard, 1999), increase in neurotrophic factors secretion (Keefe et al., 2017;Ochodnicky et al., 2012;Vizzard, 2006), and changes in the proprieties of LUT afferents (de Groat and Yoshimura, 2006).Morphological changes in the bladder have also been described both in the mucosa (Apodaca et al., 2003;Birder, 2006;Kullmann et al., 2017) and detrusor muscle (Johnston et al., 2012;Oliveira et al., 2019).Alterations in the urethra in response to SCI are less documented, but we have recently demonstrated marked urethral epithelium reorganization accompanied by sphincter atrophy and a pronounced decrease in the expression of general innervation markers, particularly markers of sensory and noradrenergic nerve fibres (Ferreira et al., 2022).
Much of our knowledge about SCI-induced generated urinary dysfunction has been obtained in studies using spinal cord sectioning.Indeed, complete spinal cord transection (SCT) and spinal hemisection have been the most commonly used experimental SCI protocols (Ferreira et al., 2023;Sharif-Alhoseini et al., 2017;Verstappen et al., 2022), with complete SCT being traditionally prefered for many years in studies focusing exclusively on bladder function (Cheng et al., 1999;Cruz et al., 2006a;de Groat et al., 1990).In addition to allowing the identification of supraspinal contribution to the regulation of bladder function, complete SCT also has the advantage of being inexpensive and easily reproducible.However, complete spinal lesions are rarely seen in clinical practice and the translational value of SCT as a model of disease is debatable.Indeed, spinal contusions are the most frequent causes of human SCI and lesions are often incomplete (Liu et al., 2019;Theodore, 2016).Experimental contusions are currently seen as more relevant in SCI research (Blight, 2000;Sharif-Alhoseini et al., 2017) despite carrying some inter-trial variability (Cheriyan et al., 2014;Sharif-Alhoseini et al., 2017).While complete SCT has been classically used to study SCI-induced bladder dysfunction, as contusions become a more commonly used model, it is important to compare both experimental Fig. 1.Contusion apparatus.(A) The contusion system incorporates a rigid surface with attached forceps to stabilize the vertebrae of the animal during impact.The device is attached to a manipulator that allows adjustments of the position of the impactor tip in relation to the animal's exposed spinal cord.The weight drop device consists of a 20 cm cylindrical tube with 6 mm in diameter attached to the manipulator's arm.(B) The weight is a small steel bar (10 g, 4.5 cm long) with a 2.5 mm diameter tip.(C) The braking system is composed of a brake (a 2.8 cm long piece of plastic, adapted with a 0.5 cm lock in the centre placed at the end of the conducting cylinder, which is linked to the impactor weight and a counterweight (30 g).(D) The caudal and rostral vertebrae of the animal should be clamped, to ensure the animal is immobilised during impact.(E) For an appropriate impact, it is important to expose the spinal cord.(F) To validate the device, a plastic jar with acovering silicone membrane.A pressure transducer was inserted in the silicone cover and connected to a pressure transducer and an amplifier, to record the pressure of the membrane before and after weight drops.(G) The average impact force produced was greater when the metal rod was dropped from 10 cm high, compared to when the weight was dropped from 10 cm high.The standard deviation between trials was small, validating the reproducibility of the model.Data were compared with a t-test (**** p < 0.0001).
approaches in what concerns bladder function and changes in neuronal content at both the LUT and lumbosacral spinal cord levels.This would confirm if data obtained in SCT studies can be extended to models of spinal contusion, a matter we investigated in the present study.

Animals and drugs
In-house-bred female Wistar Han rats (200-250 g; n = 4-5 per experimental group; derived from Charles River Laboratories, France).To prioritize animal welfare and considering anatomical differences, including shorter urethra and absence of prostate, female animals were preferred due to the relative ease of manually emptying their bladders during spinal shock, when there is little or no bladder reflex activity able to promote voiding.While we recognize the use of male animals should be considered in terms of representation of human injuries, male SCI rodents are more prone to urinary retention, present higher risk of bladder infection, urethral blockages, and lethal kidney failure.Animals were maintained under a 12 h light/dark inverted schedule and controlled temperature and air humidity, with ad libitum access to food and water.Experimental procedures were carried out per the European Communities Council Directive 2010/63/EU, to ethical guidelines for investigating pain in animals and internal regulations of the Faculty of Medicine of Porto.The ARRIVE guidelines have been followed during the course of this study.Animals were randomly divided into four groups: SHAM (controls), complete spinal cord transection (SCT), mild contusion (mSCC) and severe contusion (sSCC).All efforts were made to reduce the number of animals used and their suffering.

Drugs
All surgeries were performed under deep anaesthesia, induced with an intraperitoneal (IP) injection of medetomidine (0.5 mg/Kg) and ketamine (75 mg/Kg).Anaesthesia was reverted with an intramuscular injection of atipamezole (1 mg/Kg).After surgery, all animals received 5 % saline-glucose (1 mL, IP) to compensate for blood loss and dehydration.Animals also received oral antibiotics (enrofloxacin; 5 mg/Kg).Antibiotic administration was initiated on the day of surgery and maintained for 8-10 days.For pain management, rats were treated with subcutaneous buprenorphine (0.02 mg/Kg) for 3-4 days after surgery and whenever deemed necessary, after that time point.For euthanasia, animals received an intraperitoneal injection of sodium pentobarbital (56 mg/g).

Spinal cord contusion device
The device used in the present study was constructed based on previously used setups (Lima et al., 2020;Vasconcelos et al., 2016).To ensure stability, the device was placed on a rigid and stable surface (Fig. 1A).As spinal trauma typically causes involuntary reflexes, this rigid surface has attached forceps to allow steadying of the vertebrae during impact (Fig. 1A).The device itself consists of a 20 cm cylindrical hollow tube (adapted from a 10 mL plastic pipette) attached to a manipulator's arm, which allows the adjustment of the set-up to the animal's position (Fig. 1B, 1D).The tube conducted the weight onto the exposed spinal cord.The weight consists of a 4.5 cm long metal cylinder, weighing 10 g, with a 2.5 mm diameter tip (Fig. 1B).The device used has an incorporated braking system, important to control the amount of time the weight's tip rests on the exposed spinal cord, avoiding dissimilar functional outcomes.The brake allows for a more controlled weight drop and consists of a 2.8 cm long piece of plastic, with a 0.5 cm lock in the centre, placed at the end of the conductive cylinder (Fig. 1C).A 30 g counterweight is also necessary to revert the weight's trajectory once it impacts the spinal cord.The three components (brake, weight, and counterweight) are linked by a series of cables that are lifted in a pivot at the top of the system (Fig. 1A).Once the weight is dropped via the conducting tube, placed on top of the exposed spinal cord (Fig. 1E) and its tip touches the spinal cord, the braking system is released, quickly lifting the counterweight, and retrieving the weight from the impacted area.The dwell time is controlled and constant between experiments.The reproducibility of the impacts produced by this device was assessed prior to animal SCI.A testing platform, consisting of a hollow plastic cylinder covered by a tense silicone membrane, was used (Fig. 1F).A pressure transducer and an amplifier were coupled to this setup and a total of 20 impacts on the rubber membrane were run, from two different heights.

Spinal cord injury surgery and post-operative care
After deep anaesthesia induction, animals were submitted to a laminectomy between T7-T10 vertebrae, for exposure and visualization of the T8/T9 spinal segments.
For complete SCT, as in other studies (Chambel et al., 2022;Cruz et al., 2006a;Cruz et al., 2006b), thoracic vertebrae were exposed and a laminectomy was performed.The dura mater was pierced and the T8/T9 spinal cord was sectioned with a scalpel.A small piece of sterile haemostatic sponge was placed between the retracted ends of the cord to prevent bleeding.The surgical wound was closed in two layers and the animals were placed under post-surgical observation and care.
To perform mSCC and sSCC, after laminectomy, the animals were placed on the prepared platform for spinal impact.The caudal and dorsal vertebrae were firmly clamped with the forceps placed 1 cm from the desired impact zone.The system was adjusted so that the conducting cylinder for weight drop was very close to the exposed spinal cord (Fig. 1E).The brake system was activated to prevent accidental weight drop and the counterweight was suspended.After confirming the position of the animal and the conducting cylinder, the weight was placed in the conducting cylinder and dropped from 5 or 10 cm high, to respectively produce a mild or severe contusion of the spinal cord at T8/T9.Once the impact occurred, the weight immediately retracted.The wound was cleaned and sutured in two layers, followed by post-surgical care.Control animals underwent sham surgery in which the surgical wound was sutured after laminectomy without lesioning the spinal cord.
Post-surgical care for all groups consisted of subcutaneous administration of 1 mL of 5 % glycosylated saline after suturing, and antibiotics and analgesics for 8-10 days and 3-4 days, respectively.In SCI (SCT, mSCC and sSCC) animals, to avoid urinary retention, bladders were manually emptied by abdominal compression for 4 weeks.The voided volume was recorded every three days for later comparison.

Cystometries and terminal handling
To evaluate bladder reflex activity, animals (sham and SCI) were submitted to cystometries under anaesthesia 4 weeks after surgery.Following deep anaesthesia with subcutaneous urethane (1.2 g/Kg), a suprapubic skin incision was made, and muscle bundles separated for bladder exposure.A 21-gauge needle was inserted into the bladder dome, and sterile saline infused for 1 h at a rate of 6 ml/h.Bladder contractions were recorded by a pressure transducer connected to the needle.During this period, animals were placed on a heating plate to maintain body temperature at 37 • C. Cystometrograms were analysed using LabScribe software (World Precision Instruments, Hertfordshire, UK).Bladder contractions were disregarded if the amplitude was less than 5 cm H 2 O.
Collected bladders were sectioned in transversal, 20 μm thick slices, while urethras were cut into 12 μm thick longitudinal slices.The L5-S1 spinal cord was sectioned in transversal, 20 μm thick slices and the lesion site in longitudinal, 20 μm thick slices.Sections were obtained in a Leica cryostat (Leica, Famalicão, Portugal) and collected in Superfrost Plus slides.Slides were stored at − 20 • C until further processing.

Tissue morphology of the lesioned spinal cord
To analyze morphologic alterations at the lesion site, serial longitudinal sections of the lesion site (T8-T9 level) were left at room temperature for 24 h before staining with formol-thionine.Sections were incubated with acidic acetone for 5 min, followed by a period of 30 min in 10 % thionine in a 10 % formol solution.Sections were then washed in distilled water and mounted with Histomount mounting medium.Representative images obtained from three longitudinal sections, from each animal, in which the lesion site was identified in its entire length were investigated and collected in a Zeiss Axioscope 40 microscope using Leica LAS EZ v3.1.0software (Leica Microsystems, Switzerland).The rostro-caudal extension of the lesioned spinal tissue was calculated with Image J. Results are presented as the percentage of the lesioned tissue in relation to the spinal cord area.

Fig. 2. Effects of experimental spinal cord contusion and transection on urinary function. (A)
As SCI strongly affects urinary function and there is little or no bladder activity during the initial stages after spinal lesion, all animals were submitted to daily abdominal compression for urine removal.During this process, urine volumes were recorded.Data are presented on a 3-day interval basis.Values were recorded in mildly contused animals (mSCC), severely contused animals (sSCC), and fully transected animals (SCT).In SHAM animals, as the bladder was fully functional, urine volumes were unsignificant and therefore not recorded.Arrows indicate peaks of bladder reflex contractions.

Immunofluorescence analysis
Alternate sections from the bladder, urethra, lumbosacral spinal cord, and T8/T9 spinal cord tissue (lesion site) were thawed, washed in phosphate-buffered saline (PBS) and in PBS containing 0.3 % triton-X 100 (PBST), and blocked with 10 % normal horse serum (NHS) in PBST for 2 h.Tissue sections were incubated for 72 h at 4 • C with primary antibodies in 2 % PBST (Table 1).After several washes with PBST, sections were incubated with Alexa™ fluorochrome-labelled secondary antibody in 2 % PBST (Invitrogen − ThermoFisher Scientific, Porto, Portugal) for 1 h at room temperature.After subsequent washing, sections were mounted using an anti-fade mounting medium (Slowfade® Gold Life Technologies) and observed with an epifluorescence microscope (Axioimager Z1, Ziss Z1 from Zeiss) using the AxioVision 4.6 software.
Immunofluorescence labelling intensity was quantified by densitometry using Image J.For bladder tissue, five representative transverse sections of each animal were photographed in two distinct zones: the mucosa and the detrusor muscle.Only sections containing the bladder and urethral lumen were considered.In the case of urethral tissue, three longitudinal sections containing the urethral sphincter and in which the lumen was easily discernible were selected and photographed in three distinct zones: urethral mucosa, internal urethral sphincter (IUS) and external urethral sphincter (EUS).Results are presented as an average of the values measured in the posterior (close to the vagina) and anterior side of the urethral wall.In the case of spinal cord L5-S1 sections, at least 5 sections per segment (15 in total) were photographed in the following areas: dorsal horn, intermediolateral grey matter, central canal (lamina X) and ventral horn.Briefly, the mean level and standard deviation of background immunofluorescence were obtained in the region of interest without visible staining for each section analysed using ROI analysis.The threshold level for positive pixels was set at a value of 5 standard deviations above the mean background level.The mean percentage of staining was then calculated by delimiting the region of interest in each section and using ROI analysis (Chambel et al., 2022).In any case, whenever the staining procedure damaged the tissue under analysis and the adequate number of sections was not achieved, the animal was excluded.Nevertheless, even in groups in which we had to eliminate animals, the total number of sections per experimental group analysed was never less than 45 bladder sections, 10 urethral sections and 15 transverse lumbosacral spinal cord sections.

Statistics
Statistical analysis was conducted using GraphPad 8.2 software.The average impact force produced by the contusion device was compared using an unpaired t-test.The volumes of daily residual urine were analysed by a two-way ANOVA, followed by Tukey's Multiple comparison test.Data obtained in immunostaining analysis were first tested for normality (Shapiro-Wilk normality test).Normally distributed data were analysed using one-way ANOVA, followed by Tukey's Multiple comparison test.Abnormally distributed data were assessed by a nonparametric test (Krustal-Wallisś test), followed by Dunnś multiple comparison test.In all cases, p < 0.05 was considered statistically significant.Data is presented as the mean ± standard deviation (SD).Further details can be found as supplementary information.

Contusion and evaluation of the reproducibility of impact
To compare the effects of complete spinal cord transection and spinal contusion in LUT function, we used the well-established protocol of complete SCT (Cruz et al., 2006a;Cruz et al., 2006b;Oliveira et al., 2019) and a customised contusion device based on the weight-drop method, respectively.To test for potential variability between consecutive weight-drop trials, a hollow plastic cylinder covered by a silicone membrane was constructed and connected to a pressure transducer and an amplifier (Fig. 1F).The weight was dropped 20 times from 5 cm high and another 20 times from 10 cm high.The values recorded were, respectively, 11.24 ± 0.78 cm H 2 O and 15.17 ± 1.260 cm H 2 O (Fig. 1G), showing a statistically significant increase in impact intensity depending on the height the weight is dropped.The variation coefficients were 0.020 (drop from 5 cm high) and 0.018 (drop from 10 cm high).

Bladder dysfunction after spinal contusion and spinal transection
Every three days, urine volumes collected after daily abdominal compressions were recorded (Fig. 2A).Spinal intact animals were not submitted to daily abdominal compression, as there was no urinary retention following sham manipulation of the spinal cord.Previous unpublished observations from our group demonstrated that sham animals have a basal urine volume of 0.52 ± 0.18 mL.In SCI animals, the volume of urine collected on day 1 post-surgery was already high in all SCI groups.In the SCT group, urine volumes gradually increased, compared to the first day post-lesion, and reached the maximum average value of 9.00 ± 2.49 mL on day 7 post-SCT.From this time point onwards, urine volumes remained elevated but followed a descending trend.A similar variation was seen in the sSCC group on day 7 (7.80 ± 0.97 mL), with a decrease being observed only on day 13 post-lesion.In the mSCC group, the increase in urine volume was not as prominent, and the maximum urine volume (5.44 ± 0.57 mL) was recorded on day 7 post-contusion and decreased from that day onwards, reaching the average normal value 22 days after spinal injury.At timepoints of 10 (p < 0.01 versus sSCC), 13 (p < 0.05 versus sSCC) and 25 days after injury, the urine volume of mSCC animals was significantly lower than the urine volume of sSCC animals.At timepoint 25, mSCC animals presented significantly less urine than SCT animals (p < 0.05 versus SCT).
Bladder contractility was assessed by cystometry under urethane anaesthesia, in intact (B-SHAM), mild contused (C-mSCC), severe contused (D-sSCC) and fully transected animals (E-SCT).Urodynamic parameters of the amplitude (F), frequency of bladder reflex contractions (G), Basal pressure (H), and peak pressure (I) of the overall experimental groups.The residual urine volumes were analysed by a two-way ANOVA followed by Turkeyś multiple comparison test.The urodynamic data was tested by a one-way ANOVA, followed by Turkeyś multiple comparison test (*p < 0.05, **p < 0.01 versus SHAM; #p < 0.05 versus mSCC).

(caption on next page)
A. Ferreira et al.

Tissue damage at the lesion site
To evaluate the degree of damage at the thoracic spinal cord produced by SCT, mSCC and sSCC, longitudinal spinal sections of the injury site were stained with formol-thionin, a classical staining method to study the cytoarchitecture of neuronal tissue (Donovick, 1974), that we have previously used to analyse SCI (Oliveira et al., 2019) (Fig. 3).In sham-manipulated animals, neuronal tissue was intact (Fig. 3A).In contrast, spinal tissue damage was evident in samples from all SCI animals (****p < 0.0001 versus SHAM).In the areas of injury, staining allowed the identification of cavities and tissue integrity.In animals submitted to contusion, cavities were small, randomly distributed, with some preservation of grey matter, particularly in sections from mSCC animals.In sections from SCT animals, the cavities were larger and connected, with a complete interruption of the grey matter (Fig. 3A-D).
In groups submitted to spinal contusion, the area of damaged tissue was evident (Fig. 3B, C and D; ****p < 0.0001 versus SHAM, for both mSCC and sSCC) and the rostro-caudal extension was directly proportional to the severity of the contusion.In the sSCC group, the percentage of lesioned tissue was 36.36 % ± 3.87 of the total area, significantly higher than what was observed in mSCC rats, in which the lesioned tissue represented 27.76 % ± 5.04 of the total area analysed (***p < 0.001 versus sSCC animals).In the SCT group, the lesioned area was restricted to the vicinity of severed ends of the cord, as the centre was occupied by scar tissue.The rostro-caudal extension was smaller than mSCC and sSCC and represented 17.8 % ± 3.7 of the total area analysed (Fig. 3D and 5E; ****p < 0.0001 versus SHAM; *p < 0.05 versus mSCC; ****p < 0.0001 versus sSCC).
To evaluate the completeness of the lesion and axonal sparing, longitudinal spinal sections of the injury site were immunostained against GFAP and 5HT, respectively.SHAM animals presented very faint GFAP immunolabelling and had a normal pattern of 5-HT positive axons, visible along the entire length of the section (Fig. 3F).Both mSCC and sSCC animals displayed well-defined glial scar borders, with evident GFAP immunostaining surrounding the lesion core (Fig. 3G and H).In these sections, staining against 5-HT showed that conserved axons are present in the borders of the lesion, with some remaining even in its core (Fig. 3G2, 3H2).In some cases, 5-HT positive fibres could be observed caudal to the lesion core, more frequent and visible in sections from mSCC (Fig. 3G3) rather than sSCC rats (Fig. 3G4).In animals submitted to complete transection, the glial scar was visibly more intense (Fig. 3I), and no signs of spared 5-HT immunolabelled profiles were observed in the lesion core (Fig. 3I2) or in locations caudal to injury (Fig. 3I3).

Neuronal sprouting at the bladder and lower urinary tract after spinal cord contusion and transection
It is known that SCT courses with increased axonal sprouting at the lumbosacral spinal cord and LUT (Chambel et al., 2022;Ferreira et al., 2022;Oliveira et al., 2019) but it was not clear if the same would happen after spinal contusion at thoracic segments.As before, growthassociated protein-43 (GAP43) expression was used as a marker of axonal sprouting (Benowitz and Routtenberg, 1997).
In the bladder, we found that GAP43 was absent in the mucosa of all SCI animals, compared to sham animals (**p < 0.01 versus SHAM) (Fig. 4A, D, E, F and G).In the detrusor muscle, the decreased expression of GAP43 was less evident, being only observed in sections from SCT animals when compared to sham animals (*p < 0.05 versus SHAM) (Fig. 4A, H, I, J and K).In the urethral mucosa, GAP43 immunostaining was abundant in sham animals, but strongly decreased in all injured animals, particularly in mSCC animals (*p < 0.05, **p < 0.01 versus SHAM; Fig. 4B, L, M, N and O).The urethral sphincter followed the same tendency, with the internal sphincter of all injured animals showing similar patterns of GAP43 downregulation, compared to sham animals (*p < 0.05, **p < 0.01 versus SHAM; Fig. 4B, L, M, N and O).Likewise, GAP43 expression was elevated in the striated muscle of the external urethral sphincter in sham animals and decreased after spinal injury, particularly in contusioned animals, although a tendency for a decrease in SCT animals was also observed (*p < 0.05 versus SHAM; Fig. 4P, Q, R,  and S).
GAP43 expression was also evaluated at the L5-S1 spinal cord level, the spinal segments receiving sensory input generated in the LUT (Morgan et al., 1981).We found GAP43 immunoreactivity in lamina I-II of the dorsal horn (Lam I-II), in the intermediolateral grey matter (IML), and laminae X (Lam X), the pivotal areas involved in LUT sensory processing and autonomic regulation.In lamina I-II, we found an upregulation of GAP43 levels, particularly in animals submitted to spinal contusion (**p < 0.01, ****p < 0.0001 versus SHAM Fig. 4C, T, U, V and  W).In sections from SCT animals, changes in GAP43 expression did not reach statistical difference, compared to sham animals, despite a tendency to increase, but did reach significance in comparison with mSCC (###p < 0.001 versus mSCC).In the IML, all SCI groups presented the same pattern of increased GAP43 expression, compared to sham animals (*p < 0.05, **p < 0.01 versus SHAM; Fig. 4C, X, Y Z and A1).In lamina X, no statistically significant differences between experimental groups were found (Fig. 4C, A2, A3, A4 and A5).

Sensory innervation
To examine changes in sensory innervation, we studied the expression of calcitonin-gene-related peptide (CGRP), a neuropeptide present in peptidergic sensory fibres (Snider and McMahon, 1998) and widely expressed by bladder afferents (Avelino et al., 2002).Analysis of CGRP immunostaining showed that this sensory marker was absent in the bladder mucosa of all SCI groups when compared to sham animals (****p < 0.0001 versus SHAM) (Fig. 5A, D, E, F and G).In the detrusor muscle, there was a reduction in CGRP levels in SCI animals (**p < 0.01 versus SHAM; Fig. 5A, H, I, J and K).
In the urethra, there was also an evident loss of CGRP expression in all SCI groups (**p < 0.01; ***p < 0.001 versus SHAM; Fig. 5B, L, M, N  and O).In the urethral sphincters, we found decreased CGRP levels in both the IUS and EUS.In the IUS, the pattern of CGRP denervation was Fig. 3. Effects of different types of spinal cord injury on the histology of the lesioned area.Representative images of formol-thionin stained lesion site tissue.(A -SHAM) In sections from sham-manipulated animals, there was no evidence of neuronal damage at the laminectomy site.(B ¡ mSCC) Sections from animals submitted to mild contusion (mSCC) presented a compacted lesion area, with well-defined borders and signs of preserved tissue.(D ¡ sSCC) In sections from animals submitted to severe contusion (sSCC) the lesion area was more widespread, with poorly defined borders that were accompanied by severe neuronal loss.(D ¡ SCT) In sections from animals submitted to complete spinal cord transection (SCT), there was total disruption of spinal tissue between the caudal and rostral ends of the injury site.Scale bars equal 50 μm.(E) Quantification of the area of lesioned tissue was performed using image J. Data were compared using One-Way ANOVA followed by Tukey's multiple comparison test (***p < 0.005; ****p < 0.0001).The completeness of the lesion existence of spared axons was analysed by immunofluorescence against Glial fibrillary acidic protein (GFAP) and serotonin (5-HT).(F) GFAP immunostaining was very faint in sections from SHAM while 5-HT immunoreactive profiles were evident in all the extension of the sections.(G) In sections from mSCC animals it was possible to identify the scar tissue, surrounded by rostral and caudal GFAP immunostaining with some 5-HT positive profiles in the lesion core (G1) as well as in more caudal regions (G2).(H) In sections from sSCC animals it was also possible to identify the scar tissue, with strong surrounding GFAP immunostaining and fewer 5-HT positive fibres in the lesion core (H1) and in more caudal regions of the analysed sections (H2).(I)The scar tissue was evident in sections from SCT animals, with intense GFAP immunlabelling and no signs of 5-HT immunoreactive fibres in the lesion core (I2) and more caudal regions (I3).similar to that observed in the mucosa (*p < 0.05 versus SHAM; Fig. 5B,  L, M, N and O).In the striated muscle of the EUS, decreased expression of CGRP was observed only in mSCC and SCT animals, when compared to SHAM (*p < 0.05, **p < 0.01 versus SHAM; Fig. 5B, P, Q, R and S).The expression of CGRP was also investigated in the L5-S1 spinal segments of SHAM and SCI animals.In the dorsal horn (Lam I-II), we found that CGRP levels were significantly increased in animals submitted to mSCC (*p < 0.05 versus SHAM; Fig. 5C, T, U, V, and W).CGRP content in SCT animals was significantly reduced when compared with mSCC (#p < 0.05 versus mSCC).In the IML, CGRP levels were also increased, reaching statistical significance in mSCC animals when compared with SHAM (**p < 0.01 versus SHAM) and the other SCI animals (#p < 0.05 versus sSCC; #p < 0.05 versus SCT) (Fig. 5C, X, Y, Z and A1).No alteration was found in the Lam X area (Fig. 5C, A2, A3, A4 and A5).

Cholinergic innervation: Expression of vesicular acetylcholine transporter (VAChT)
To investigate cholinergic innervation, the expression of vesicular acetylcholine transporter (VAChT) was studied in the LUT and lumbosacral cord.In the bladder, like CGRP, VAChT expression was absent in the bladder mucosa of SCI animals (Fig. 6A, D, E, F and G) when compared to SHAM animals (Fig. 6A and D), irrespective of the type of spinal injury.Likewise, we also found a dramatic reduction in VAChT levels in the detrusor muscle of SCI animals, particularly in sSCC and SCT (*p < 0.05; versus SHAM; Fig. 6A, H, I, J and K).In the urethra, there were also changes in VAChT levels.A marked reduction in the expression of this marker was found after spinal contusion or transection when compared to sham animals (****p < 0.0001 versus SHAM; Fig. 6B,  L, M, N and O).The urethral sphincter was affected, with evidence of downregulation of VAChT expression in both the IUS and EUS (****p < 0.0001 versus SHAM, **p < 0.01 versus SHAM; Fig. 6B, L, M, N and O; P,  Q, R and S).
In the L5-S1 spinal cord segments, the presence of VAChT was also analysed.VAChT was absent from laminae I-II but present in the ILG, lamina X and ventral horn of spinal intact animals.In the IML, VAChT expression was similar between SHAM and mSCC animals but decreased after sSCC and SCT (*p < 0.05, ***p < 0.001 versus SHAM; mSCC; Fig. 6C, T, U, V and W).The SCT group also presented decreased VAChT immunostaining, compared to mSCC (##p < 0.01 versus mSCC), being the group that presented the most evident reduction in VAChT expression.No alterations were found in the Lam X area (Fig. 6XY, Z and A1).In the ventral horn, VAChT expression was only reduced in SCT animals when compared to mSCC (#p < 0.05 versus mSCC; Fig. 6A2, A3, A4 and  A5).

Noradrenergic innervation: Expression of tyrosine hydroxylase (TH)
Noradrenergic innervation was assessed by investigating tyrosine hydroxylase (TH) expression.In the bladder, no TH-positive profiles were observed, with only a reduced number of fibres present in the vicinity of small-calibre blood vessels (Fig. 7C).In the urethra, the presence of noradrenergic fibres was detected in the urethral sphincter, but not in the mucosa.In the IUS, TH-positive fibres were decreased after SCI, particularly in mSCC animals (*p < 0.05 versus SHAM), while in sSCC and SCT animals, downregulation of TH expression did not reach statistical significance (Fig. 7A, D, E, F and G).The same tendency was found in the EUS, where the reduction in TH expression was statistically significant in mSCC and sSCC (*p < 0.05 versus SHAM), but not in SCT animals (Fig. 7A, H, I, J and K).
At the lumbosacral spinal cord level, expression of TH was found in laminae X, where levels of this noradrenergic marker were reduced after SCI, irrespective of the type of insult (**p < 0.01, ***p < 0.001 versus SHAM; Fig. 7B, L, M, N and O).

Discussion
A better understanding of the pathophysiological mechanisms of SCIinduced LUT dysfunction is critical to promote advances in medical treatment.In this context, the use of animal models has been essential, as they allow the understanding of complex mechanisms of SCI and how they may affect LUT activity (Ferreira et al., 2023;Sharif-Alhoseini et al., 2017).The most commonly used experimental SCI protocol has been the complete SCT, a highly reproducible experimental protocol, but with a debatable translational value, as the majority of SCI result from contusion injuries.Therefore, it is important to clarify the similarities and differences in LUT function and reorganization of neuronal pathways after spinal contusion and transection.This has been a matter under investigation for some years (Breyer et al., 2017;Mitsui et al., 2014;Pikov et al., 1998) and here we expanded previous observations, comparing spinal damage, LUT function, and changes in the LUT and lumbosacral spinal cord after mild and severe contusion and transection at T8/T9 spinal segment.

Design of a new contusion method: model advantages and validation
In this study, we used a simple-to-use and affordable set-up to produce spinal cord contusion.This setup is an adaptation of the classical weight-drop method used by others (Lima et al., 2021).The weight-drop method was the first described method to experimentally produce a spinal cord contusion in rodents (Allen, 1911).More recently, sophisticated automated impactors have been introduced, allowing the production of contusive spinal injuries in a controlled and reproducible manner (Rabchevsky et al., 2003;Sharif-Alhoseini et al., 2017).Yet, these systems are costly and involve maintenance costs, which may preclude their widespread use.Here, we produced a simple and inexpensive alternative based on the classical weight-drop method, which includes a braking component that ensures that the impactor tip is immediately retracted following impact.
An important concern when using our setup is its reproducibility.Although we did not have access to a force transducer, our results demonstrated a correlation between the height of the weight drop and the force produced on a silicone membrane, with low deviations between trials.Thus, the device used in this study can be used to produce reproducible spinal cord contusions of different severities by adjusting the height of the weight drop.This was confirmed by histological Fig. 4. Assessment of neuronal sprouting after mild and severe spinal cord contusion and spinal cord transection.Axonal sprouting was accessed by GAP43 immunostaining, a well-established marker of neuronal growth.Expression of GAP43 was quantified in the bladder (A), urethra (B) and L5-S1 spinal cord segments (C).In the bladder, GAP43 levels were studied in the bladder mucosa of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G), and a marked decrease was observed in sections from SCI animals.The analysis of the bladder detrusor of SHAM (H), mSCC (I), sSCC (J) and SCT animals (K) showed the same tendency of GAP43 downregulation.In the urethra, three zones were analysed: urethral mucosa and the internal (IUS) and external (EUS) sphincters.In the mucosa and IUS of all groups, SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), GAP43 was also downregulated in animals submitted to spinal contusion but not transection.In the EUS, in comparison with SHAM animals (P), GAP43 was also decreased in sections from mSCC (Q), sSCC (R) but not from SCT rats (S).GAP43 expression was also analysed in the lumbosacral spinal cord.The observation of lamina I-II areas of SHAM (T), mSCC (U), sSCC (V) and SCT animals (X) showed that mSCC animals present higher GAP43 levels than the remaining SCI groups.In the intermediolateral nucleus (IML) of SHAM (Z), mSCC (A1), sSCC (A2) and SCT animals (A3), we observed increased sprouting in all injured groups, irrespectively of the SCI model.In the Laminae X of intact (A3), mSCC (A4), sSCC (A5) and SCT animals (A6), no differences were found.Normally distributed data was compared using One-Way ANOVA followed by Tukey's multiple comparison tests.Non-normally distributed data was compared using Krustal Wallis test followed by Dunnś multiple comparison test (*p < 0.05, **p < 0.01; **** p < 0.0001 versus SHAM; ##p < 0.01 versus mSCC).
A. Ferreira et al.

(caption on next page)
A. Ferreira et al. analysis of the injured spinal tissue, as well as by analysing GFAP and 5-HT immunostaining.Accordingly, we observed an extensive area of tissue damage after spinal contusion, which was larger and more evident in animals submitted to a 10 cm-high weight drop (severe spinal contusion), compared to rats on which the weight was dropped from 5 cm high (mild spinal contusion).Interestingly, the area of damaged tissue was more constrained, with a more limited rostrocaudal distribution of lesioned tissue, in animals submitted to spinal transection, reflecting the limited injury induced by restricted insertion of the scalpel blade.However, a more restricted spread of damage rostrocaudally does not necessarily mean a less severe injury.In fact, analysis of GFAP and 5-HT immunolabelling showed axonal sparing after contusion, particularly after mild contusion, with some positive fibres in the core of the lesioned tissue and caudally to the injury.In SCT, these 5-HT immunoreactive fibres were only present in the rostral spinal cord in relation to the injury, and absent in the rest of the sections, indicating that SCT produces a complete and more severe injury of the cord.

Effect of different SCI models on bladder function
Spinal lesion is followed by spinal shock, during which the bladder has little or no reflex activity (Anderson et al., 2023;Bywater et al., 2018).During this period, SCI animals need to be submitted to daily abdominal compression for urine removal.Residual urine volumes were recorded every three days, and we observed an increase in urine volume in all SCI animals during the first week post-spinal trauma, which was reduced upon the emergence of spontaneous voiding.While a reduction in urine volume was found from post-injury day 7 onwards in SCT and mSCC animals, in sSCC this was only observed from post-injury day 13.This suggests that the development of spontaneous voiding is affected by the area of damaged tissue.Indeed, in sSCC animals the area of lesioned tissue seen after severe contusion was extensive, affecting more rostral and caudal segments beyond the injury site.This larger area of damage, not seen in SCT or mSCC animals, likely delays tissue healing and precludes neural recovery, resulting in more prolonged spinal shock.
Bladder function was also studied by cystometry in sham and SCI animals.As others (Breyer et al., 2017;Mitsui et al., 2014), we did not find any major differences between SCI animals and only observed higher basal and peak pressures in SCT animals in comparison to spinal intact rats.The lack of significant differences in bladder function after mild and severe spinal contusion and spinal transection suggest that, despite different degrees of damage to the spinal cord, injuries resulted in similar functional impairment of the lower urinary tract (Pikov et al., 1998).While others have performed awake cystometries (Breyer et al., 2017;Mitsui et al., 2014), in this study urethane was used, which could have affected our observations.This anaesthetic remains a suitable option to study bladder reflex activity, as others produce a more detrimental effect on bladder function (Aizawa and Fujita, 2023;Yaksh et al., 1986).It should, however, be recalled that urethane affects urethral function, leading to increased sphincter resistance and residual volume (Chang and Havton, 2008).This effect of urethane is likely more pronounced in SCT rats (Yoshiyama et al., 2013), which lack supraspinal control over bladder function, leading to higher basal and peak pressures than in sham rats.

Effect of different SCI models on LUT innervation
Neuronal sprouting is a key mechanism to NDO emergence in response to the loss of bulbospinal input, leading to the strengthening of spinal synapses and the establishment of new ones at the lumbosacral spinal cord, accompanied by changes in afferent excitability (Shimizu et al., 2023;Wada et al., 2022).Axonal sprouting can be assessed by analysing the expression of GAP43, a well-established marker of neuronal sprouting (Benowitz and Routtenberg, 1997).In the LUT, GAP43 was absent in the bladder and urethral mucosa of SCI animals, irrespective of the type of injury, consistent with a lack of difference in bladder dysfunction.In the detrusor, GAP43 was significantly decreased in SCT animals, with a non-significant tendency for reduction in animals submitted to spinal contusion.While this may suggest that there is no remodelling of detrusor innervation after spinal injury, this is unlikely as the bladder muscle undergoes major tissue remodelling after SCI (Haferkamp et al., 2003;Johnston et al., 2012), resulting in bladder hypertrophy (Mimata et al., 1993;Nagatomi et al., 2005) and more disperse innervation, which complicates the assessment of GAP43 expression.In the urethral sphincter, GAP43 was also reduced in the IUS and EUS of mSCC and sSCC, but not after SCT.This is in line with our previous study which has also analysed GAP43 levels in urethral sections from SCT rats at the same time point (Ferreira et al., 2022).While in that study we did not separately examine GAP43 in different zones of the urethral wall, we also did not find any significant differences in the IUS and EUS.Overall, our results indicate that there is neuroplasticity in the LUT, particularly in the mucosa of the bladder and urethra.This reduction in GAP43 levels may also reflect LUT denervation as seen in human (Drake et al., 2000) and rodent SCI (Johnston et al. 2012).
Changes in sensory innervation of the bladder and urethra were investigated by analysing CGRP levels, a neuropeptide abundantly expressed in bladder afferents (Avelino et al., 2002;de Rijk et al., 2024).We found a marked decrease in CGRP levels in all layers of the bladder and urethral walls.The reasons for this sensory denervation may only be speculated at present but it may arise during spinal shock.Indeed, sensory fibres are highly dependent on Nerve Growth Factor (NGF) (Cruz, 2014;Denk et al., 2017), which is produced by LUT smooth muscle cells during contraction (Clemow et al., 2000;Persson et al., 1997).As in spinal shock there is little or no bladder reflex activity (Anderson et al., 2023;Bywater et al., 2018) and loss of smooth muscle cells in the urethra after SCI (Ferreira et al., 2022), this may cause a reduction in NGF, which may lead to sensory denervation of the LUT.
Cholinergic innervation was evaluated by investigating the expression of VAChT.In the bladder mucosa and detrusor muscle, we found a similar reduction in VAChT levels in all SCI animals, irrespective of the type of injury and in agreement with previous studies (Breyer et al., 2017;Johnston et al., 2012;Takahara et al., 2007).In the urethra, VAChT immunoreactivity was also reduced in the mucosa, IUS and EUS.Loss of VAChT immunoreactivity suggests dysfunction of cholinergic neurotransmission in this organ (Roy and Green, 2019;Takahara et al., 2007).Moreover, the recovery of the expression of this marker may be Fig. 5. Calcitonin Gene-related peptide (CGRP) expression in the LUT and lumbosacral spinal cord after SCI.The distribution of sensory fibres was assessed by immunostaining against CGRP, a well-established marker of peptidergic afferents.The expression of CGRP was quantified in the bladder (A), urethra (B) and L5-S1 spinal cord segments (C).In the bladder, the mucosa of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G) was analysed and showed absence of CGRP after SCI, irrespective of the type of injury.Likewise, in the bladder detrusor of SHAM (H), mSCC (I), sSCC (J) and SCT animals (K) there was also CGRP downregulation.In the urethra, three zones were analysed: the urethral mucosa and the internal (IUS) and external (EUS) sphincters.In the mucosa and IUS of SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), CGRP was downregulated in all SCI groups.The EUS was also assessed in SHAM (P), mSCC (Q), sSCC (R) and SCT animals (S), where significant loss of CGRP immunostaining was only seen in sections from mSCC and SCT animals.CGRP was also evaluated at the lumbosacral spinal cord.The analysis of lamina I-II of SHAM (T), mSCC (U), sSCC (V) and SCT animals (X) showed that mSCC animals presented significantly higher levels of CGRP immunostaining, compared to SCI models.In the intermediolateral nucleus (IML) of SHAM (Z), mSCC (A1), sSCC (A2) and SCT animals (A3) CGRP upregulation was also observed.No differences were found in the Laminae X of SHAM (A3), mSCC (A4), sSCC (A5) and SCT animals (A6).Normally distributed data was compared using One-Way ANOVA followed by Tukey's multiple comparison tests.Non-normally distributed data was compared using Krustal Wallis test followed by Dunnś multiple comparison test (*p < 0.05, **p < 0.01; ***p < 0.001, **** p < 0.0001 versus SHAM; # p < 0.05 versus mSCC).Scale bars equal 50 μm.
A. Ferreira et al. associated with improved LUT function (Breyer et al., 2017), which was not the case here.In turn, noradrenergic LUT innervation was studied by examining TH immunoreactivity, an enzyme involved in catecholamine biosynthesis.TH-positive profiles were very scarce in the bladder, with small fibres present in the vicinity of blood vessels.The paucity of TH fibres in the bladder was consistent with studies from other authors (Gosling et al., 1999;Persyn et al., 2016;Watanabe and Yamamoto, 1979).In the urethra, TH-positive fibres were present in the IUS and EUS and reduced after spinal trauma.The decrease in VAChT and TH expression in the LUT may reflect either a reduction in the synthesis of these proteins or injury of the nerve fibres, arising after SCI, even if this represents an indirect insult to bladder nerve fibres.Accordingly, a reduction in the expression of pan-neuronal markers has been observed in the bladder and urethra after spinal cord transection (Takahara et al., 2007).In the bladder, we found a significant decrease in VAChT levels in the bladder mucosa of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G).VAChT was also decreased in the detrusor of SHAM (H), mSCC (I), sSCC (J) and SCT animals (K).In the urethral mucosa and the internal urethral sphincter (IUS) of SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), VAChT levels were also decreased, irrespective of the type of SCI.VAChT downregulation was also observed in the EUS of SHAM (P), mSCC (Q), sSCC (R) and SCT animals (S).Levels of VAChT were also studied in the lumbosacral spinal cord, which were found in the IML, laminae X and ventral horn.The analysis of the intermediolateral nucleus (IML) of SHAM (T), mSCC (U), sSCC (V) and SCT animals (W) showed that loss of VAChT immunostaining was only seen in sections from sSCC and SCT animals.No differences were found in the laminae X of SHAM (X), mSCC (Y), sSCC (Z) and SCT animals (A1).A similar response was observed in the ventral horn of SHAM (A2), mSCC (A3), sSCC (A4) and SCT animals (A5).Normally distributed data was compared using One-Way ANOVA followed by Tukey's multiple comparison tests.Non-normal data was compared using Krustal Wallis test followed by Dunnś multiple comparison test.Scale bars equal 50 μm.In the bladder TH-positive profiles were scarce and only present in the vicinity of blood vessels (C).In the urethra, as no immunoreactive fibres were seen in the mucosa, analysis of the internal urethral sphincter (IUS) of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G) showed a decrease in labelling only in mSCC animals.In the external urethral sphincter (EUS), assessed in SHAM (H), mSCC (I), sSCC (J) and SCT animals (K), loss of TH immunolabelling was found in sections from mSCC and SCC, but not in the SCT group.In the lumbosacral spinal cord, TH expression was present in the laminae X (Lam X) of SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), showing a strong decrease of TH immunolabelling in SCI groups.Data were compared using One-Way ANOVA followed by Tukey's multiple comparison test (*p < 0.01, **p < 0.01, ***p < 0.001 versus SHAM).Scale bars equal 50 μm.
A. Ferreira et al.

Effect of different SCI models on the lumbosacral spinal cord
The development of SCI-induced bladder dysfunction is accompanied by neuroplasticity events at the lumbosacral spinal cord that lead to NDO development (Shimizu et al., 2023).Considering the different areas of tissue damage at the thoracic trauma site after contusion and transection, it was important to investigate its consequences on the lumbosacral cord, where important changes have been reported after SCT (Chambel et al., 2022).As in the LUT, axonal sprouting was investigated by assessing GAP43 levels.This protein was present in the superficial dorsal horn, where it increased after spinal trauma, in agreement with previous observations (Chambel et al., 2022;Frias et al., 2015).In lamina I-II, sprouting was more prominent after mSCC, likely indicating enhanced axonal sprouting.GAP43 expression was also upregulated in the IML, but without significant differences between the SCI groups, suggesting that IML sprouting was less dependent on the severity of spinal lesion.The superficial dorsal horn and the IML areas are known to receive sensory input originating in the bladder (Cruz et al., 2005;Cruz et al., 1994;Fuller-Jackson et al., 2021) and, though we did not perform co-localization analysis, it is likely at least part of GAP43-positive profiles are peptidergic sensory afferents, particularly in SCT animals (Frias et al., 2015), as GAP43 increase was accompanied by CGRP upregulation in the same spinal areas, more evident in sections from mSCC.This abnormal axonal expansion of peptidergic afferents has been shown to take place due to time-and NGF-dependent fluctuations in axon guidance cues and their receptors (Chambel et al., 2022).
Expression of VAChT was also investigated in the spinal cord.This marker is present in cholinergic cell bodies and terminals, and its expression can be used as a measure of neuronal injury (Maeda et al., 2004;Takahara et al., 2007).Indeed, we found a downregulation of VAChT expression in the IML, where pre-ganglionic neurons involved in micturition control are located (de Groat et al., 2015;Karnup, 2021).VAChT was also decreased in motor neurons located in the ventral horn, the Onuf's nucleus (Schellino et al., 2020), which are known to project and control the urethral sphincter.VAChT was reduced after SCI, particularly after SCT.This reduction correlated with urinary impairment, suggesting that bladder dysfunction also reflects changes in cholinergic cell bodies and fibres in the lumbosacral spinal cord.Moreover, the lack of supraspinal input induced by SCT led to the most marked decrease in VAChT expression, highlighting the important contribution of supraspinal centres to micturition control (Fowler et al., 2008).
Expression of TH at the lumbosacral spinal cord was also studied.The presence of this enzyme is considered a marker of descending noradrenergic fibres originating in the brainstem (Heinricher et al., 2009;Tavares et al., 2021;Westlund et al., 1983).TH-immunoreactive fibres were only seen in laminae X, in contrast with other studies in which THpositive cells were observed in other areas of the spinal cord (Hou et al., 2016;Fuller-Jackson et al., 2021), possibly reflecting the use of different antibodies with distinct sensitivity.Nonetheless, TH levels were reduced in all SCI animals, particularly after complete transection.This reflects a lack of supraspinal input to the lumbosacral spinal cord, irrespective of the area of damaged tissue.
Like all studies, there are also limitations that should be addressed.While the majority of SCI patients are male (Chen et al., 2016), female rats were used, as in most experimental studies.The reason for this resides in anatomical differences between sexes, with males presenting prostate and longer urethras.Because SCI is followed by a period of little or no bladder reflex activity (Anderson et al., 2023;Bywater et al., 2018), animals need to be submitted to abdominal compression for urine removal until automatic micturition develops.Abdominal compression is more easily done in female than male rats, which presents an increased risk of obstructions, urinary infection and, eventually, kidney failure.We acknowledge that in future studies male animals should be considered for a proper representation of human SCI.Moreover, one should also take into consideration that traumatic lesions of the spinal cord (compression, contusion, transection, hemisection) constitute the most widely used animal models of SCI but spinal lesions can also arise from degenerative age-related disorders, including degenerative joint disease.Therefore, future experimental and translational studies should address these non-traumatic SCIs as they are under investigated and require additional scientific investment.

Conclusions
Human spinal cord injuries vary between patients and produce different outcomes, which is widely acknowledged amongst clinicians and researchers.Most SCIs are accompanied by urinary impairment, resulting from NDO and DSD.Experimental SCI models have classically used spinal transections to clarify pathophysiological mechanisms and devise new therapeutic strategies, despite spinal contusions being more frequent.It is, thus, important, to compare changes in LUT function and peripheral and central neuroplasticity following spinal transection and contusion.For several years, this issue has been investigated (Breyer et al., 2017;Mitsui et al., 2014;Pikov et al., 1998), and here, we built on earlier findings and expand observations to the urethra and lumbosacral spinal cord, a major hub for post-SCI neuroplasticity.To our knowledge, this is the first study comparing the effects of different models of SCI at both spinal cord and LUT levels that include the analysis of urethral tissue.
We devised an affordable setup that produces reproducible contusions and have validated its use.We compared LUT function and tissue integrity at the lesion site and found a correlation between injury severity and bladder dysfunction and the area of damaged tissue.At the LUT, mild and severe contusions and transection produced similar changes in innervation, whereas at the lumbosacral dorsal horn less severe injury was accompanied by axonal sprouting, particularly of peptidergic afferents, eventually linked to changes in sensory function.This data shows that changes in LUT innervation and dysfunction arising after spinal cord injury by contusion and transection are similar but result from different neuroplastic events at the lumbosacral spinal cord.This could have implications for the future development of therapeutic tools for urinary impairment after SCI.

Fig. 6 .
Fig.6.Expression of Vesicular Transporter of Acetylcholine (VAChT) after mild and severe spinal cord contusion and spinal cord transection.Cholinergic innervation was analysed by immunostaining against VAChT, a well-established marker of cholinergic fibres.The expression of CGRP was quantified in the bladder (A), urethra (B) and L5-S1 spinal cord segments (C).In the bladder, we found a significant decrease in VAChT levels in the bladder mucosa of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G).VAChT was also decreased in the detrusor of SHAM (H), mSCC (I), sSCC (J) and SCT animals (K).In the urethral mucosa and the internal urethral sphincter (IUS) of SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), VAChT levels were also decreased, irrespective of the type of SCI.VAChT downregulation was also observed in the EUS of SHAM (P), mSCC (Q), sSCC (R) and SCT animals (S).Levels of VAChT were also studied in the lumbosacral spinal cord, which were found in the IML, laminae X and ventral horn.The analysis of the intermediolateral nucleus (IML) of SHAM (T), mSCC (U), sSCC (V) and SCT animals (W) showed that loss of VAChT immunostaining was only seen in sections from sSCC and SCT animals.No differences were found in the laminae X of SHAM (X), mSCC (Y), sSCC (Z) and SCT animals (A1).A similar response was observed in the ventral horn of SHAM (A2), mSCC (A3), sSCC (A4) and SCT animals (A5).Normally distributed data was compared using One-Way ANOVA followed by Tukey's multiple comparison tests.Non-normal data was compared using Krustal

Fig. 7 .
Fig. 7. Expression of Tyrosine Hydroxylase (TH) after mild and severe spinal cord contusion and spinal cord transection.Noradrenergic fibres were analysed by immunostaining against TH, a well-established marker of noradrenergic fibres.TH-immunoreactive fibres were investigated in the urethra (A) and L5-S1 spinal cord segments (B).In the bladder TH-positive profiles were scarce and only present in the vicinity of blood vessels (C).In the urethra, as no immunoreactive fibres were seen in the mucosa, analysis of the internal urethral sphincter (IUS) of SHAM (D), mSCC (E), sSCC (F) and SCT animals (G) showed a decrease in labelling only in mSCC animals.In the external urethral sphincter (EUS), assessed in SHAM (H), mSCC (I), sSCC (J) and SCT animals (K), loss of TH immunolabelling was found in sections from mSCC and SCC, but not in the SCT group.In the lumbosacral spinal cord, TH expression was present in the laminae X (Lam X) of SHAM (L), mSCC (M), sSCC (N) and SCT animals (O), showing a strong decrease of TH immunolabelling in SCI groups.Data were compared using One-Way ANOVA followed by Tukey's

Table 1
Primary antibodies description.