The SLAC Laboratory (Shanghai, China) afforded us with Sprague Dawley rats. The rats were all 6-week-old males weighing between 180 and 200 g. The rats were given standard water and rat chow. Their ambient temperature was controlled at 24 ± 1°C and the light and dark cycle was 12/12 h. All experimental processes were authorized by the Ethics Committee of Scientific Research of Shanghai University of Sport.

The rats (n = 18) were grouped randomly into three groups: the Sham group (n = 6), the CCI group (n = 6), and the CCI-Swim group (n = 6). Adaptive feeding was conducted for 1 week. The rats in Sham and CCI groups were routinely fed and did not participate in swimming exercise, while rats in the CCI-Swim group were adapted to swim for 1 week preoperatively. Swimming time gradually increased from 10 min on day 1 to 60 min on day 6. On the third day after the CCI operation, rats in the CCI-Swim group were made to swim for a total of 19 sessions, and the swimming time was gradually increased. The first and second sessions involved swimming for 30 min each day, the third and fourth sessions involved swimming for 40 min each day, and the fifth and sixth sessions involved swimming for 50 min each day. In the seventh to ninth sessions, the rats swam for 60 min each time (Figure 1). This swimming training program is based on a previously published animal swimming exercise program that was improved by our research group. The rats were made to swim in a plastic box (82 cm × 60 cm × 59 cm) at a temperature of 35–37°C. After swimming, the rats are caught and dried with a heat blower (Almeida et al., 2015).

We performed CCI on the sciatic nerve of the rats on the basis of a previous study (Jaggi et al., 2011). First, we anesthetized the animal with 5% isoflurane. Then, the skin of the rats’ right lower limb was cut, and the muscle and connective tissue were bluntly dissected to locate the sciatic nerve. Four chromic catguts (4.0 silk) were ligated on the sciatic nerve at an interval of approximately 1 mm. The sciatic nerve of rats in the Sham group was exposed but not ligated. The skin was then sewn with 5.0 silk sutures, and the rats were allowed a recovery period after the operation.

We performed mechanical withdrawal threshold (MWT) tests before and after CCI at 3, 7, 14, 21, and 28 days. The rats were placed in transparent glass boxes for 20 min to acclimatize before each behavioral test. The MWT test was performed by harmless stimulation to the posterior plantar of the rats. Von Frey wires (4–180 g, Aesthesio, Danmic Global, United States) were applied to the right hind paw, and the intensity of the Von Frey wires increased from small to large until the rats exhibited the desired behavior. Each rat was tested at least three times, at 5 min intervals.

On the 28th day after CCI operation, three rats in each group (i.e., Sham, CCI, and CCI-Swim groups) were killed, and the spinal cord was collected for histopathological analysis. The spinal cord was fixed in 4% paraformaldehyde solution at 4°C for 90 min. After phosphate-buffered saline (PBS) cleaning for 15 min, the samples were then respectively soaked in 15 and 30% sucrose solution for 24 and 48 h. The tissue samples were placed in compound compounds (SAKURA, 4583), sliced (14 μm thick), and dyed with hematoxylin and eosin (HE). The slices were rinsed with PBS for 5 min, soaked in hematoxylin staining solution (Servicebio, G1005-1) for 4 min, and then washed under running water for 20 min. After washing, 1% hydrochloric acid solution was used for differentiation for 3 s. Rinsing was repeated for 15 min. The slides were kept in 0.1% eosin staining solution (Servicebio, G1005-2) for 3 min and then soaked in 85, 95, and 100% alcohol solutions for 1 min in sequence. After soaking in xylene I and xylene II for 1 min, the tissue samples were sealed with neutral resin and then observed under an Olympus BX53 microscope (Tokyo, Japan).

On the 28th day after surgery, the remaining rats (n = 9, three in each group) were anesthetized with 3% pentobarbital sodium and instilled with saline (250 mL, 4°C) via the aorta. The corresponding positions of L4–L6 were located on the surgical side of the spine, and the spinal canal was opened to remove the L4–L6 spinal dorsal horn. TRIzol reagent was utilized to extract total RNA from the spinal dorsal horn. RNA quantification and qualification were conducted prior to sequencing. We established sequencing libraries through NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (NEB, United States) and appended the index code to attribute sequence of each sample. Then, the PCR products were purified by AMPure XP system, and the library quality was evaluated on Agilent Bioanalyzer 2100 system. Finally, the library preparation was sequenced by Illumina Hiseq platform.

The DESeq2 R package (1.10.1) was utilized to analyze the identified DEGs. DEG selection criteria is P-values < 0.05 and | log2 Fold Change (FC) | > 1.

Gene Ontology (GO) analysis was executed by the clusterProfiler R package. The KEGG database allows for genome deciphering, which enables a better understanding and analysis of the interactions, reactions, and relational networks of biomolecules produced by sequencing experimental techniques.¹ The ClusterProfiler R package was utilized to detect DEGs enrichment in KEGG pathways. The significant enrichment criterion for GO analysis and KEGG pathway analyses was P < 0.05.

Quantitative real-time PCR (qRT-PCR) was used to detect the DEGs to verify the accuracy of the sequencing results. Total RNA was reverse transcribed into cDNA by Evo M-MLV RT Kit with gDNA Clean for qPCR II (Cat. No. AG11711). The qRT-PCR was executed by SYBR^® Green Premix Pro Taq HS qPCR Kit II (Cat. No. AG11702) and run in a real-time system (Thermo Fisher Scientific, CA, United States). U6 and β-actin were identified as internal controls for lncRNAs and mRNAs, respectively, and the 2^–ΔΔCt approach was applied to describe relative expression ratios. The specific primer sequences in this work are listed Table 1.

Two-way repeated-measures analysis of variance with Tukey’s post hoc analysis was conducted to evaluate the results of weight measurement and behavioral tests. The results of qRT-PCR were determined by independent-sample t-tests. All of the data were analyzed by Graphpad Prism 9.0 and presented as mean ± standard error of the mean (SEM). The standard of statistical significance was P < 0.05.

In this study, the NP model of rats involved CCI on the right hind limb. MWT test was performed in all three groups before and after operation (Figure 2). No significant difference in body weight among the three groups. Contrast to rats in the Sham group, animals in CCI group showed higher mechanical allodynia sensitivity from day 3 to 28 after CCI surgery; the CCI-Swim group also exhibited higher MWT on days 3, 7, and 14 after surgery. MWT was significantly improved in the CCI-Swim group on the day 21 and 28 after operation in contrast to the CCI group.

The HE staining results showed that the spinal dorsal horn neurons in the Sham group were intact, regular, and orderly and that the nucleolus in vivo was distinct. In the CCI group, the spinal dorsal horn neurons were damaged and denatured, with pyknosis, fragmentation, and dissolution of the nucleus, as well as loose surrounding tissues. These features indicate that the CCI model had been successfully prepared. Compared with the CCI group, the CCI-Swim group showed a slightly irregular arrangement of sensory neurons in the spinal dorsal horn and relatively dense surrounding tissues. In general, the histological features of the spinal cord in this group were better than those of the CCI group overall (Figure 3).

The quality of the RNA-seq results was evaluated. Comparison of the Sham and CCI groups (Figure 4A) revealed a total of 734 DE lncRNAs, among which 385 lncRNAs were upregulated and 349 were downregulated. The top 20 DE lncRNAs in the Sham vs CCI group are shown in Table 2. There were 758 DE lncRNAs were noted in the CCI-Swim vs CCI group (Figure 4B) and the top 20 DE are listed in Table 3. The Venn diagram indicated 306 intersecting lncRNAs in the two comparisons (Figure 5). Hierarchical cluster analysis revealed the clustering patterns of the three groups of DE lncRNAs (Figure 6).

Figure 7 shows differences in mRNA expression among the three groups. Compared with the Sham group, 442 mRNAs were significantly changed in the CCI group, including 286 upregulated mRNAs and 156 downregulated mRNAs (Figure 7A). The top 20 DE mRNAs in the Sham and CCI groups are listed in Table 4. A total of 401 DE mRNAs were observed in the CCI-Swim and CCI groups (126 upregulated and 275 downregulated; Figure 7B). The top 20 DE mRNAs in the CCI vs CCI-Swim group are detailed in Table 5. The Venn diagram obtained showed 173 intersecting mRNAs, which are listed in Figure 8. Hierarchical clustering analysis of the three groups of DE mRNAs revealed their clustering patterns (Figure 9).

Gene Ontology analysis was conducted on the target genes with significant changes in DE lncRNAs. Between the CCI and Sham groups, the target genes were enriched in biological process (BP) terms such as muscle structure development, positive regulation of BP, and muscle organ development (P < 0.05). The cellular component (CC) terms were significantly related to Z disc, I band, and sarcomere (P < 0.05). Molecular functions (MFs) were enriched in terms such as acid phosphatase activity, structural constituent of muscle, and protein binding (P < 0.05). The target genes of DE lncRNAs in the CCI-Swim and CCI groups were concentrated in BP terms such as muscle system process, positive regulation of immune system process, and positive regulation of immune response (P < 0.05). The most enriched CCs were Z disc, I band, and sarcomere (P < 0.05). MF terms were focused on acid phosphatase activity, structural constituent of muscle, and phosphatase activity (P < 0.05).

Gene Ontology analysis was also applied to the DE mRNAs of two groups, and the results are shown in Figure 10 (P < 0.05). DE mRNAs in the CCI and Sham groups were enriched in BP terms such as positive T cell selection, negative T cell selection, and membrane fusion; DE mRNAs were also enriched in CC terms such as proteasome core complex, beta-subunit complex, proteasome core complex, and COP9 signalosome (P < 0.05). The most enriched MF terms were sarcosine oxidase activity, oxidoreductase activity, and threonine-type endopeptidase activity (P < 0.05). The DEGs of mRNAs in the CCI vs CCI-Swim group were focused on BPs terms such as nucleoside triphosphate biosynthetic process, sodium ion export from cell, and establishment or maintenance of transmembrane electrochemical gradient (P < 0.05). The CC terms were enriched in the sodium:potassium-exchanging ATPase complex. Finally, DE mRNAs were significantly enriched in MF terms such as cation-transporting ATPase activity, ATPase activity, coupled to transmembrane movement of ions, and sodium:potassium-exchanging ATPase activity (P < 0.05; Figure 11).

Figure 12 shows the KEGG pathways of the target genes of DE lncRNAs in the CCI vs Sham groups. The top five KEGG pathways were hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy, ribosome, asthma, and PPAR signaling pathway (P < 0.05). In the CCI-Swim vs CCI groups, the top five enriched KEGG pathways were HCM, dilated cardiomyopathy, olfactory transduction, ribosome, and cytokine–cytokine receptor interaction (P < 0.05). For mRNAs, DEGs in the CCI and Sham groups focused on aldosterone-regulated sodium reabsorption and B cell receptor signaling pathway. DE mRNAs in the CCI-Swim and CCI groups were significantly enriched in the KEGG pathways of metabolic pathways, oxidative phosphorylation, Parkinson’s disease, proximal tubule bicarbonate reclamation, glycosylphosphatidylinositol–anchor biosynthesis, and cAMP signaling pathway (P < 0.05; Figure 13).

We verified four lncRNAs (i.e., XLOC_274480, XLOC_105980, XLOC_137372, and AABR07047899.1) and four mRNAs (i.e., C3, Sgk1, Dnah7, and Vwa3a) to demonstrate the accuracy of the RNA-seq results. According to Figures 14, 15, the expression changes of the eight DEGs were consistent with the RNA-seq result.