A total of 70 healthy adult female Sprague-Dawley rats (specific pathogen-free grade; age, 8 weeks; weight, 200±20 g) were purchased from Shanghai SIPPR-Bk Lab Animal Co., Ltd. (license no. SCXK: 2008-0016). The animals were housed at 23±2°C, with 12-h light-dark cycles and 50% humidity. The animals had free access to food and drink, and were adaptively housed for 1 week prior to experiments. The animals were caged individually after surgery. The experiments were approved by the Animal Care and Use Committee of Xi'an Jiaotong University Health Science Center. All efforts were made to minimize animal numbers and suffering.

A total of 20 rats, using a random number table, were randomly assigned to the sham operation group, and the remaining 50 rats underwent severe SCI modeling. During the operation, four rats died, which was within the acceptable limits of the ethical approval obtained for the present study, and two rats were excluded due to modeling failure. Forty rats were randomly selected among the ones with successful modeling, and randomized to the SCI and SNS groups (20 rats/group). The remaining four rats received the same nursing interventions (such as anti-infection treatment and assisted urination) as the rats with SCI. After the experiment was completed, the rats were sacrificed humanly.

The first day of the experiment was modeling. SNS intervention was performed for 14 days, from day 2 to day 15. On day 16, the time to first black stools was recorded (fasting started on day 15 and the animals had free access to water for 24 h). Dry weight of the fecal pellets was recorded. The animals were sacrificed on day 17.

Prior to the induction of the severe SCI model, the rats were deeply anesthetized by the intraperitoneal injection of 10% chloral hydrate (300 mg/kg). Thermal support was provided and the depth of anesthesia was monitored by the toe pinch method. The skin region was disinfected with iodophor and a median incision was made from T10 to T13. The spinous process and vertebral plates were exposed after the dissection of the superficial fascia and the removal of the T11-T12 vertebral plates. A bone window was generated and the spinal cord was exposed. SCI was induced by striking the exposed spinal cord using a 10 g weight falling from 60 mm with the New York University (NYU) Impactor device (W.M. Keck Center for Collaborative Neuroscience Rutgers, State University of New Jersey); this apparatus was used because it avoids the occurrence of heavy bleeding (44,45). Successful SCI modeling was determined according to the improved Basso, Beattie, Bresnahan locomotor rating (BBB) score (45). In the sham operation group, the spinal cord was exposed at T10-T13, but the NYU impactor was not used. In all rats, the wound was sutured within 5 min and covered with gentamicin ointment.

The improved BBB score was determined to evaluate hind limb motor function (45). The rats were placed in an open space before modeling (day 1), 24 h after modeling (day 2) and on day 16. The following parameters were observed independently by two observers: Joint motions of the hip, knee and ankle; hind limb weight-bearing condition; walking capability and hind limb-forelimb coordination; trunk stability; paw position; and tail movement. The scores were between 0 and 21 points, higher scores indicated increased improvement of the hind limb motor function.

From the day of modeling, all rats received intraperitoneal injections of gentamicin for 16 days at 5,000 U/kg once daily. The lower abdomen, perineum and hind limbs of the rats were cleaned daily, and passive movements of the hind limbs were performed daily. The Crede maneuver (46) was applied every 12 h to assist voiding: The SCI rats were held upright and gentle pressure was applied on the bulging bladder from top to bottom for assisting urination. Before voiding, the bladder was palpated and urinary retention was estimated based on the degree of bladder filling.

The passive movements were performed after Crede-assisted urination. The operator fixed the rat in the prone position with one hand and grasped the toes of one hind limb with the other hand. Then, the rat was pulled rearward and outwardly at a 45° angle with the spine, until the knee and ankle joints were in complete extension. After this, the hind limb was pushed towards the trunk in the opposite direction until the hip and knee joints were completely flexed and the ankle was completely dorsally flexed. Flexions were performed 60 times/min for 1 min. The same procedure was performed for the other hind limb. Passive movement of both hind limbs was performed once every 12 h. An observer blinded to grouping observed fur appearance, autonomic activity, food and water intake, defecation and urination and bodyweight during the study.

At 24 h after successful modeling, the rats in the SNS group were immobilized in the prone position. Needle electrodes were placed on bilateral S3 neural foramen. The electrodes were connected to a Myolito electrical stimulator (MTR+ Vertriebs GmbH). The stimulation settings were: Pulse width of 0.2 msec; frequency of 10 Hz; stimulation duration of 10 sec; intervals of 5 sec; and current of 2–3 mA. The appropriate stimulation degree was indicated by slight shivering of the tail, but without braying. Each session of SNS lasted 15 min, with one session per day for 14 days.

The first black feces were discharged at 8:00 a.m. on the 15th day of the experiment. Then, 10 rats from each group were randomly selected and fasted for 24 h, but with free access to water. At 8:00 a.m. on the 16th day of the experiment, these rats were given 2 ml of 100 g/l activated charcoal suspension by gavage. The time interval between charcoal gavage and the first black feces was recorded.

The dry weight of feces discharged within 24 h was recorded. In the remaining 10 rats/group, on the 16th day of experiment, feces discharged within 24 h (8:00 a.m.-8:00 a.m.) were collected, dried and weighed.

On the 17th day of the experiment, the rats were sacrificed by decapitation after intraperitoneal injection of 10% chloral hydrate (400 mg/kg), followed by exposure of the spinal cord at S2-4 (in accordance with the vertebral bodies L3-5). The fresh spinal cord tissue of these segments was rapidly harvested and split into two parts. The abdomen was cut open for the quick dissection of ~1 cm of distal colon, which was cut open along the longitudinal axis to rinse off colonic contents with physiological saline, and split into two parts. One part of fresh spinal and colon tissues were directly preserved in liquid nitrogen for ELISA, reverse transcription-quantitative PCR (RT-qPCR) and western blotting. The other parts of spinal cord and spread colon tissues were fixed in 4% paraformaldehyde for hematoxylin-eosin (H&E) staining and immunohistochemistry (IHC).

Hence, the samples of the colon and spinal tissues (20 rats/group) were both divided into two parts. One part was fixed with 4% paraformaldehyde, while the other part was stored in liquid nitrogen. Among the 20 samples fixed with 4% paraformaldehyde, five were randomly selected for H&E staining, five were randomly selected for IHC and the other 10 were stored for eventual future use. Among the 20 samples stored in liquid nitrogen, 10 were used for ELISA, five were used for RT-qPCR, four were used for western blotting, and the remaining one was stored for eventual future use.

Colon tissues of 5 rats in each group were fixed in 4% paraformaldehyde for 24 h at room temperature (23±2°C), followed by routine paraffin embedding, and sectioning at 4 µm. Tissue slides were subjected to routine H&E staining, as follows. Dewaxing was performed using xylene and a descending ethanol series (100, 100, 95, 95, 85 and 75%) for 5 min at each step. The sections were treated with hematoxylin for 5 min, 1% hydrochloric acid in ethanol for 2 sec, tap water for 10 min, and eosin for 3 min. The sections were dehydrated in an ascending alcohol gradient (75, 85, 95, 95, 100 and 100%) and xylene, 3 min each step. The sections were sealed with neutral gum. All steps were performed at room temperature (23±2°C). Pathological changes of colon tissues were observed in six different fields and photographed using a DMLS32 optical light microscope (magnification ×200; Leica Microsystems GmbH).

Colon and spinal cord tissues of 10 rats in each group were homogenized in PBS (weight:volume ratio of 1:9) and supernatants were extracted by centrifugation at 5,000 × g for 5–10 min at room temperature (23±2°C). The standards (24, 12, 6, 3, 1.5 and 0.75 ng/ml) from the 5-HT ELISA kit (cat. no. 201710; Nanjing Jiancheng Bioengineering Institute Co., Ltd.) and test samples (50 µl) were added into the wells. Subsequently, the samples were incubated with horseradish peroxidase-labeled 5-HT antibody (100 µl) for 60 min at 37°C, as per the kit instructions. Optical density values were measured at a wavelength of 450 nm.

Colon and spinal cord tissues from five rats in each group were fixed in 4% paraformaldehyde for >24 h at room temperature (23±2°C), followed by routine paraffin embedding and sectioning (5-µm thick sections). One slide with an intact tissue section from each animal was selected for IHC. For dewaxing and hydration, the sections were successively immersed in xylene, xylene, anhydrous ethanol, 95% ethanol, 85% ethanol and 70% ethanol for 5 min each, followed by rinsing with PBS three times for 3 min each. Endogenous peroxidase activity was blocked using 3% H₂O₂ for 10 min at room temperature, followed by rinsing with PBS three times for 3 min each. The sections were blocked with 5% goat serum (50–100 µl; cat. no. C0265, Beyotime Institute of Biotechnology) for 20 min at room temperature (23±2°C). The primary antibodies (all purchased from Abcam) included anti-5-HT3AR (cat. no. ab13897; 1:100) and anti-5-HT4R (cat. no. ab60359; 1:200). The sections were incubated with the primary antibody solution (50–100 µl) overnight at 4°C, followed by washing three times with PBS, 3 min each. The sections were incubated with horseradish-peroxidase goat anti-rabbit IgG (H+L) secondary antibody (cat. no. 111-035-003; 1:200; Jackson ImmunoResearch Laboratories, Inc.) at 37°C for 1 h. Chromogen detection was performed by incubating the sections with DAB solution for 15 min at room temperature (23±2°C). The sections were counterstained with hematoxylin for 10 min at room temperature (23±2°C) and washed with distilled water. Subsequently, the sections were immersed in 70, 85, 95 and 100% ethanol for 5 min each, and twice in xylene for 10 min each. Neutral gum was added to seal the slides. The sections were examined under a DMLS32 optical light microscope (Leica Microsystems GmbH). The positive staining was examined (magnification, ×40, followed by magnifications, ×100/x400). Six high-power fields (HPF; magnification, ×400) were randomly selected on every slide. The JD801 image analysis system (Nanjing Jiancheng Bioengineering Institute Co., Ltd.) was used to measure the average optical density (AOD) of the positive stained area. Image-Pro software (version 6.0.0.309; Media Cybernetics, Inc.) was used to count immunopositive cells in six HPFs (magnification, ×400; 0.1323 mm²) per animal.

Cryopreserved colon and spinal cord tissues from 5 rats in each group were used for RNA extraction with TRIpure reagent (cat. no. 2702026AX; Beijing Aidlab Biotechnologies Co., Ltd.). Total RNA (2 µg) was reverse transcribed to cDNA using a OneScript cDNA Synthesis kit (cat. no. G234; Applied Biological Materials, Inc.), according to the manufacturer's instructions. qPCR was performed using the EvaGreen Express 2X qPCR MasterMix-ROX kit (cat. no. MasterMix-ER; Applied Biological Materials). Fluorescence qPCR was carried out in a TL988-IV qPCR system (Tianlong Science and Technology Co., Ltd.), in triplicate. The primer pairs used were as follows: 5-HT3AR forward, 5′-GGCACCTGGTCCTAGACAGAA-3′ and reverse, 5′-GGTTTCCCATGGCTGAGCAGT-3′; 5-HT4R forward, 5′-CCGTTTCTCCTCATGGTGCT-3′ and reverse, 5′-AACATCTGGATCTGCTGGGC-3′; and β-actin forward, 5′-TGGGTATGGAATCCTGTGGCA-3′ and reverse, 5′-TGTTGGCATAGAGGTCTTTACGG-3′. The reaction conditions were: Initial denaturation at 95°C for 10 min; and 40 cycles of amplification at 95°C for 5 sec and 60°C for 10 sec. The issolution curves of the PCR products were analyzed. The Cq values were determined using the MED-TL software (version IV; Xi'an Tianlong Science and Technology Co., Ltd.). The relative mRNA expression levels were calculated using the 2^−ΔΔCq method (47) and normalized to the internal reference gene β-actin.

Cryopreserved colon and spinal cord tissues were used for western blotting. Proteins were purified with the Protein Extraction kit (cat. no. KGP250; Nanjing KeyGen Biotech Co., Ltd.). Total protein was quantified using a bicinchoninic acid assay and 170 µg protein/lane were separated via SDS-PAGE on 10% gels. The high mass of protein loaded per lane was due to the low expression levels of 5-HT3AR and 5-HT4R in the samples. The separated proteins were subsequently transferred onto nitrocellulose membranes and blocked for 1.5–2 h at room temperature with 5% skim milk powder. The membranes were incubated overnight on a shaking plate at 4°C with primary antibodies against: 5-HT3AR (1:500; cat. no. ab13897, Abcam), 5-HT4R (1:1,000; cat. no. ab60359; Abcam) and GAPDH (cat. no. E12-052; Nanjing EnoGene Biotech Co., Ltd.). After rinsing with TBST, the membranes were incubated for 1–2 h at room temperature with a goat anti-rabbit IgG-HRP secondary antibody (1:5,000; cat. no. E1WP318; Nanjing EnoGene Biotech Co., Ltd.). Protein bands were visualized using the ECL Chemiluminescence kit (cat. no. E1WP3132; Nanjing EnoGene Biotech Co., Ltd.), prior to scanning in the BOX chemiXR5 chemiluminescence imaging system (SynGene Europe). The band densities of the target proteins (5-HT3AR/5-HT4R) were measured using Gel-Pro Analyzer software (version 32; Meyer Instruments) with GAPDH as the loading control. The ratio of target protein vs. internal control was considered as the relative expression level of the target proteins.

SPSS 16.0 (SPSS, Inc.) was used for statistical analysis. Data were expressed as the mean ± SD. Analyses of normal distribution and homogeneity of variance were first conducted. For data with normal distribution, comparisons among multiple groups were carried out with one-way ANOVA, as appropriate, with the least significant difference post hoc test (for data with homogeneity of variance) or the post hoc Dunnett's T3 test (for data with heterogeneity of variance). For the BBB score, the Wilcoxon non-parametric test with Bonferroni's correction was used (same experimental group, different time points). For comparisons between unmatched data (same time point, different experimental group), the Mann-Whitney U-test with Bonferroni's correction was used to analyze the data. P<0.05 was considered to indicate a statistically significant difference.

The NYU impactor device was used to establish the rat model of severe thoracic SCI (44). In the sham group, hind limb movement was not affected by the operation, whereas in the SCI and SNS groups, hind limbs presented flaccid paralysis and complete loss of motor function at the early stage after modeling. Within 2 weeks after SCI modeling, the hip and ankle joints of hind limbs presented extensive motion in the SCI and SNS groups (data not shown).

On the 17th day of the experiment, the rats in the sham group showed bright fur, autonomic activity, food and water intake, defecation and urination, and stable body weight (even with some gain). In the SCI group, the rats showed fatigue, irritation and aggressiveness, loose and dim fur, emaciation, decreased autonomic activity, decreased food intake, decreased defecation, urinary retention, hematuria and urinary incontinence (in some rats), as well as muscular atrophy of various degrees in the hind limbs. In the SNS group, the rats showed an improved state and fur condition compared with the SCI group. They showed emaciation, decreased autonomic activity, decreased food intake, decreased defecation (although improved compared with the SCI group) and urinary retention (although milder than in the SCI group; data not shown as it was not an endpoint in the present study).

Before modeling, the BBB scores were 21 points for all rats. At 24 h after modeling and compared to the sham group, the BBB scores of the SCI and SNS groups were 0 (both P<0.01). On the 16th day, compared to the sham operation group, BBB scores of the SCI and SNS groups were decreased significantly (both P<0.01). Compared with the SCI group, the BBB score of the SNS group was higher (P<0.01; Fig. 1).

The effects of SCI and SNS on intestinal functions were observed by the characterization of the fecal pellets and the passage of black fecal pellets. On the 16th day, rats in the sham group showed unobstructed defecation, with fecal pellets of a long grainy shape, discharged as 2–3 consecutive pellets or a single pellet with soft and humid texture. On the other hand, the rats in the SCI group showed constipation, with fecal pellets showing a small grainy shape, discharged as single pellet with hard and dry texture, and the amount of defecation was significantly less than that of the sham group (P<0.01; Fig. 2A). In the SNS group, rats showed unobstructed defecation, with fecal pellets being slightly smaller than in the sham group and with long grainy shape, discharged as 2–3 consecutive pellets or single pellet with slightly hard texture, and the amount was larger than in the SCI group (P<0.01).

In addition, the time interval between the operation and subsequent first defecation was different among the three groups (P<0.01). Compared with the sham group, the discharge of the first black feces was significantly delayed in the SCI group (P<0.01). Compared with the SCI group, the time interval was significantly shortened in the SNS group (P<0.01; Fig. 2B).

Taken together, these results suggested that SCI significantly impaired intestinal function. SNS could restore, at least in part, intestinal function in SCI rats.

The rats were sacrificed and colon histopathological examination was carried out to determine the effects of SCI and SNS on colon tissues. As shown in Fig. 3, compared with the sham group, the SCI group showed mucosal erosion, lower number of glands, interstitial edema and prominent atrophy of the muscular layer. The SNS group displayed mild atrophy of the muscular layer and proper glands, mild interstitial edema and with colon histology similar to that of the sham group. These results suggested that SCI caused marked alterations to colon histology, while SNS could alleviate those changes, at least in part.

The levels of 5-HT in the colon and sacral cord were measured in rats with SCI, and those treated with SNS. As shown in Fig. 4, compared with the sham group, 5-HT expression levels in the colon and spinal cord tissues were decreased in the SCI group (both P<0.01). Compared with the SCI group, 5-HT expression levels were elevated in the SNS group (both P<0.01). However, when compared to the sham group, 5-HT content in the colon and spinal cord tissues was lower in the SNS group (P<0.01 and P<0.05 respectively). These results suggested that SNS could restore 5-HT expression after SCI.

IHC was performed to determine the effects of SCI and SNS on 5-HT3AR and 5-HT4R protein levels in the colonic myenteric nerve plexus, colonic mucosa, sacral intermediolateral nucleus and dorsal horn of sacral cord. As presented in Fig. 5A, positive staining for 5-HT3AR/5-HT4R was present in the colonic myenteric plexus (cytoplasmic and nuclear staining). In the colonic myenteric plexus, cells positive for 5-HT3AR and 5-HT4R were continuously and densely distributed in the sham group; positive cells were scattered (a few faintly stained cells) in the SCI group, but densely distributed in the SNS group. In the SCI group, the AOD values of 5-HT3AR and 5-HT4R staining were significantly lower in the colonic myenteric plexus compared with the sham group (P<0.01 and P<0.05, respectively). In the SNS group, the AOD values of 5-HT3AR and 5-HT4R staining were elevated in the colonic myenteric plexus vs. the SCI groups (P<0.01, P<0.05, respectively; Fig. 5A). Compared with the sham group, the numbers of 5-HT3AR positive cells in the colonic myenteric nerve plexus of the SCI and SNS groups were significantly lower (P<0.01). Compared with the sham group, the numbers of 5-HT4R positive cells in the colonic myenteric nerve plexus of the SCI group were significantly lower (P<0.01); however, there was no significant difference in the SNS group. Compared with the SCI group, the number of 5-HT3AR and 5-HT4R positive cells in the SNS group was significantly higher (P<0.01; Fig. 5A).

In the colonic mucosa (Fig. 5B), cells positive for 5-HT3AR and 5-HT4R were densely distributed and strongly stained in the sham group, while scattered and faintly stained in the SCI group, and more densely distributed and moderately stained in the SNS group. Compared with the sham group, the AOD values of 5-HT3AR and 5-HT4R staining were significantly decreased in the colonic mucosa of the SCI group (P<0.05). Compared with the SCI group, the AOD values of 5-HT3AR and 5-HT4R staining in the colonic mucosa were significantly elevated in the SNS group (P<0.01 and P<0.05, respectively; Fig. 5B). Compared with the sham group, the numbers of 5-HT3AR and 5-HT4R positive cells in the colonic mucosa of the SCI group were significantly lower (P<0.01); however, there was no significant difference in the SNS group. Compared with the SCI group, the number of 5-HT3AR and 5-HT4R positive cells in the SNS group was significantly higher (P<0.01; Fig. 5B).

As shown in Fig. 5C and D, in the sacral intermediolateral nucleus and the dorsal horn of the sacral cord, 5-HT3AR and 5-HT4R positive cells were moderately stained in the sham and SCI groups, while they were strongly stained in the SNS group. There were no differences in the AOD values and number of positive cells of 5-HT3AR and 5-HT4R between the sham and SCI groups (all P>0.05). The AOD values of 5-HT3AR and 5-HT4R were higher in the SNS group compared with the sham and SCI groups (all P<0.05), but there was no difference in the number of positive cells (all P>0.05; Fig. 5C and D). Taken together, these results suggested that SCI decreased 5-HT3AR and 5-HT4R protein expression in the colon of rats, while SNS appeared to promote 5-HTR expression to above-sham levels.

In order to validate the IHC results, RT-qPCR and western blotting were performed. Compared with the sham group, the relative expression levels of 5-HT3AR and 5-HT4R mRNA and protein were both downregulated in colon tissue (all P<0.05; Figs. 6A and B; 7A and B), but there was no significant difference in the sacral cord (both P>0.05) in the SCI group (Figs. 6C and D; 7C and D). In the SNS group, the relative 5-HT3AR mRNA (Fig. 6A) and protein levels (Fig. 7A) were elevated compared with the SCI group (all P<0.05) in the colon tissues. The relative 5-HT3AR and 5-HT4R mRNA (Fig. 6C and D) and protein levels (Fig. 7C and D) were both elevated in sacral cord tissues compared with the SCI group (all P<0.05). Compared with the SCI group, the relative expression levels of 5-HT3AR and 5-HT4R mRNA (Fig. 6) and protein (Fig. 7) in the colon and sacral cord tissues were elevated in the SNS group (all P<0.05). Taken together, these results validated the IHC results.