Forty male SD (Sprague-Dawley) rats (180–200 g) were obtained from Beijing Viton Lihua Laboratory Animal Technology Co. (Beijing, China). The rats were housed in a strictly controlled, specific pathogen-free environment (five animals per cage, temperature 22 $\pm$ 2 °C, and relative humidity 55 $\pm$ 5%). All animals were maintained on a 12-hour light-dark cycle and had free access to rodent laboratory food and water. The handling of rats and experimental procedures in this study were reviewed and approved by the Laboratory Animal Ethics Committee of Jiaxing University Medical College (Protocol Number: JUMC2023-082). All surgeries were performed under sodium pentobarbital anesthesia, and every effort was made to minimize suffering. The rats were randomly assigned to four groups: normal rats (N) (n = 10), lumbar sympathectomized normal rats (NL) (n = 10), SNI model rats (S) (n = 10), and lumbar sympathectomized SNI model rats (SL) (n = 10).

SNI was used to induce cold sensitization in rats following nerve injury. After a few days of acclimatization feeding, the rats were subjected to surgical intervention, as described by Guida et al. [13], to induce cold allodynia. The SNI model was established by deeply anesthetizing the animal through an intraperitoneal injection of sodium pentobarbital (P3761, Sigma-Aldrich, Darmstadt, Germany) at a concentration of 35 mg/kg and shaving the right side below the pelvis, followed by the removal of skin and muscle of the hind limbs to expose the sciatic nerve and its terminal branches. The peroneal and tibial nerves were isolated from the fascia and transected distally after being ligated with a No. 5.0 silk thread. Incisions in the muscle and skin layers were closed using sutures. All surgeries were conducted under aseptic conditions. Following surgery, the rats were placed on a warm blanket and returned to their cages after recovering from anesthesia [14]. The success rate of the SNI model typically ranges from 70% to 90%. In this experiment, the successfully established SNI model rats were expected to exhibit pain-related behavioral responses, along with changes in body weight and reduced activity.

After the rats were deeply anesthetized with an intraperitoneal injection of sodium pentobarbital at a concentration of 35 mg/kg, they were positioned in the supine position. The rat’s abdominal area was shaved and disinfected with povidone-iodine, ensuring a sterile operating environment. A longitudinal incision of approximately 3–4 cm was made along the lower segment of the rat’s abdominal midline. The skin and subcutaneous tissues were carefully separated, avoiding damage to deeper organs. A longitudinal incision of approximately 3–4 cm was made in the lower midline of the rat’s abdomen. An abdominal retractor was used to expose the abdominal cavity, carefully separating the skin and subcutaneous tissue to avoid damage to deep organs. Landmarks such as the right kidney, inferior vena cava, and iliolumbar veins were identified. Dissect the fascia on the right side of the inferior vena cava to expose the lumbar sympathetic nerve adjacent to the abdominal aorta. Using fine forceps and microsurgical scissors, carefully excise the unilateral lumbar sympathetic chain, including the sympathetic ganglia from L2 to L4, while minimizing damage to nearby blood vessels and tissues. Postoperatively, absorbable sutures were used to close the abdominal muscles and skin layer by layer, ensuring the incision remained tension-free and minimizing the risk of infection.

The normal group (N) rats serve as the control group and undergo no nerve injury or surgical interventions. They are subjected only to standard acclimatization feeding and regular management. The SNI model group (S) rats undergo only the SNI surgical procedure. The lumbar sympathectomy normal group (NL) rats undergo only lumbar sympathectomy without the SNI procedure. The SNI-followed lumbar sympathectomy group (SL) rats first undergo the SNI procedure, followed by lumbar sympathectomy.

Before performing the behavioral study, the rats were individually placed in a transparent Plexiglas chamber on a raised mesh floor for 30 minutes to acclimate them to the environment. An electronic pain meter (Almemo 2450, AHLBORN, Holzkirchen, The Netherlands) was used to measure changes in the mechanical pain threshold (paw withdrawal threshold, in g) in different groups of rats. During the test, a wire probe was used to apply force vertically to the mid-plantar surface of the right hind paw of the rats; sudden retraction, licking, or violent shaking of the hind paw in response to the stimulus was considered a pain-like reaction and the pain threshold was recorded. The average paw withdrawal threshold was determined from every five measurements taken at 30-second intervals.

Cold pain and cold sensitivity in the rats were measured using a hot and cold plate analyzer (Beijing Zongshi Dichuang Science and Technology Development Co., Ltd., Beijing, China). The rats were placed in the experimental environment for 15–30 minutes to minimize interference from environmental factors. Changes in cold and hot pain over time in each group were assessed by recording the number of times the rats lifted their feet at 4 °C, with measurements repeated three times for each rat. The minimum tolerated temperatures of rats in each group were determined by downward gradient cooling, with 37 °C used as a benchmark, and the measurements were repeated three times for each rat.

Observe the rat’s response to acetone (Sigma-Aldrich, Catalog No. 650501) stimulation, which typically manifests as frequent lifting of the foot, licking, or scratching. These behaviors are associated with the rat’s sensitivity to cold stimuli, with stronger reactions indicating higher cold pain sensitivity. The number of lifts within 1 minute was recorded to assess the rat’s cold sensitivity, with measurements repeated three times for each rat.

Rats in the N, S, and SL groups were deeply anesthetized with sodium pentobarbital (50 mg/kg, intraperitoneally) on day 21, followed by myocardial perfusion with 0.9% saline (4 °C). After perfusion, segments L4–L6 of the dorsal horn of the spinal cord were harvested. Total RNA was extracted from rat spinal cord tissues using a Magzol kit (Magen Biotech, Catalog No. R480101, Guangzhou, China), and RNA yield and integrity were assessed using a K5500 microspectrophotometer (Beijing Kaiao, Beijing, China) and an Agilent 2200 TapeStation (Agilent Technologies, Santa Clara, CA, USA), respectively. The mRNA was enriched using oligo (dT) according to the instructions of NEBNext® Poly(A) mRNA Magnetic Isolation Module (New England Biolabs, Ipswich, MA, USA). Subsequently, the mRNA was fragmented to approximately 200 bp.

The RNA fragments underwent first- and second-strand cDNA synthesis, followed by junction ligation and low-cycle enrichment, according to the instructions of the NEBNext Ultra RNA Library Preparation Kit (New England Biolabs, Catalog No. E7530S, Ipswich, MA, USA). The purified library products were evaluated using an Agilent 2200 TapeStation and Qubit (Thermo Fisher Scientific, Catalog No. Q32852, Waltham, MA, USA), and the libraries were prepared using Illumina (San Diego, CA, USA) at RiboBio Co., Ltd. (Guangzhou, China) with a 150 bp paired-end approach for sequencing.

Raw FASTQ sequence data were processed with Trimmomatic software (version 0.36, https://github.com/usadellab/Trimmomatic). Specifically, end sequences with Phred quality scores below 20 were removed (software options: TRAILING, 20; SLIDING WINDOW, 4:15; MINLEN, 52) to standardize sequence lengths for the downstream clustering process. The paired-end reads were aligned to the rat reference genome (rn6) using HISAT2 (The University of Maryland, College Park, MD, USA). HTSeq (version 0.12.4, https://github.com/simon-anders/HTSeq) was used to count the reads mapped to each gene. Statistically significant DEGs were obtained using DESeq2 software (version 1.40.2) (https://bioconductor.org/packages/release/bioc/html/DESeq2.html) (by adjusting for p-values 0.05 and log2 (fold-change) $>$1). Finally, based on the transcript per million values of the DEGs in different groups, hierarchical clustering and principal component analyses were performed on the data from each group using the R language package plots (https://cran.r-project.org/web/packages/ggplot2/index.html). Different colors represent groupings, and the same colors clustered together indicate similar expression patterns, including similar functions or involvement in the same biological processes. In addition, we screened intersecting genes of the DEGs through pairwise comparisons as candidate genes using Venn plots.

DEGs were analyzed for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment using the “clusterProfiler” package (https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) in R Bioconductor software (version 3.19) (https://bioconductor.org/). To explore the correlation between these DEGs, protein-protein interaction networks were constructed using the STRING database (https://cn.string-db.org/) and visualized using Cytoscape software (version 3.10.1) (https://cytoscape.org/).

Six DEGs (four genes with upregulated expression according to transcriptomic analysis: small integral membrane protein 6 (Smim6), LOC102550818, LOC108349650, and family with sequence similarity 187 member a (FAM187A); and two genes with downregulated expression, epithelial mitogen gene (EPGN), and histone cluster 2 H3 family member C2 (Hist2h3c2)) were selected for quantitative reverse transcription polymerase chain reaction (qRT-PCR) validation. Total RNA was extracted from rapidly frozen spinal cord tissue using TRIzol reagent (Thermo Fisher Scientific, Catalog No. 15596026). RNA quality and concentration were determined using a NanoDrop-2000 spectrophotometer (Thermo Fisher Scientific). Reverse transcription was performed using SweScript All-in-One RT SuperMix for qPCR (One-Step cDNA Remover) (ServiceBio, Catalog No. G3330, Wuhan, China). The target gene mRNA was quantified using qRT-PCR with 2X SG Fast qPCR Master Mix (B639271, BBI, Roche, Basel, Switzerland) and 2.5 µM gene primers. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used for normalization. The Transcript Data®-Clark method was calculated using the 2^{-Δ⁢Δ⁢Ct} method to determine the expression levels of target genes relative to that of GAPDH. The primer details used for qRT-PCR validation are presented in Table 1.

Data are presented as means $\pm$ standard deviation and were analyzed using SPSS 22.0 (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). Multiple group comparisons were performed using one-way analysis of variance (ANOVA) followed by Tukey’s test for multiple comparisons. For results from behavioral tests and qRT-PCR, p-values less than 0.05 (p 0.05) were considered statistically significant.

We established a rat model of SNI by surgically inducing selective injury of the sciatic nerve branches. Lumbar sympathetic nerve resection was performed on day 5 post-SNI surgery for both the SL and NL groups. Mechanical pain thresholds in the hind paw and hot and cold pain sensitivity tests were conducted on the day before the SNI surgery (pre) and post-surgery days 2, 5, 8, 11, 13, 17, and 20. Notably, a lumbar sympathectomy was performed on SNI rats on day 6. We illustrated the timeline of the experimental procedure, showing the time points for the different behavioral tests and the surgical interventions (SNI surgery and lumbar sympathectomy) (Fig. 1A). We observed that, compared to the N group, the S group exhibited significantly lower mechanical pain thresholds (p 0.01, or p 0.001, Fig. 1B) and increased cold-induced anomalous pain (p 0.05, p 0.01, or p 0.001, Fig. 1C–E). These findings support the successful establishment of the rat SNI model. Pain behaviors in the S group began to emerge on the second postoperative day, peaking on the fifth day. Subsequently, these pain behaviors stabilized and remained consistent throughout the experimental period. Lumbar sympathectomy significantly alleviated the pain sensitivity induced by SNI, with significant differences observed between the S and SL groups (p 0.01, p 0.001, Fig. 1B) during the later postoperative period. The SL group showed significantly fewer cold pain responses, particularly in the later postoperative period, with highly significant differences between the S and SL groups (p 0.01 or p 0.001, Fig. 1C). Lumbar sympathectomy (SL group) demonstrated significantly lower cold sensitivity than the S group in the later postoperative period (p 0.01 or p 0.001, Fig. 1D). In the acetone test, the SL group demonstrated significantly fewer responses, with significant differences observed between the S and SL groups in the later postoperative period (p 0.01, p 0.001, Fig. 1E).

Fig. 1.

Lumbar sympathectomy attenuates abnormal pain induced by nerve injury in spared nerve injury (SNI) model rats. (A) Establishment of a rat model of SNI and experimental protocol for performing lumbar sympathectomy. (B) Mechanical withdrawal threshold (MWT). (C) Cold pain sensitivity. (D) Cold allodynia. (E) Acetone. Compared with the normal group (N), *p 0.05, **p 0.01, and ***p 0.001 indicate significant differences, and compared with the model group (S), ^#⁢#p 0.01 and ^#⁢#⁢#p 0.001 indicate substantial differences. n = 10. NL, lumbar sympathectomized normal rats; SL, lumbar sympathectomized SNI model rats.

In addition, compared with the normal group, the lumbar sympathetic nerve of normal rats was removed without affecting their perception of mechanical stimulation (p $>$ 0.05, Fig. 1B). Still, it did affect the perception of cold stimuli in their hind limbs to varying degrees (p 0.05, p 0.01, or p 0.001, Fig. 1C–E). These results suggest that lumbar sympathectomy improved mechanical pain, cold allodynia, and cold sensitivity in the ipsilateral hind paw of SNI rats.

To analyze the neurophysiological mechanisms involved in lumbar sympathectomy modulating cold sensitivity in the SNI rat model, we collected spinal cord segments L4–L6 from normal rats (N), SNI model rats (S), and lumbar sympathectomized SNI model rats (SL), and analyzed the transcriptional profiles of their spinal cords using poly(A)-seq. Each sequencing sample yielded approximately 50.3 million, 69.1 million, and 64.4 million raw reads in the N, S, and SL groups, respectively, with the average valid reads accounting for approximately 96.86%, 96.30%, and 96.78%, respectively. More than 94.90%, 94.99%, and 94.81% of the valid reads mapped to the rat genome, respectively (Table 2). The saturation curve increased with the number of reads and then flattened (Fig. 2A). The coverage homogeneity curve indicated that the transcripts covered the rat genome homogenously (Fig. 2C), confirming that a sufficient amount of data was detected. We performed hierarchical clustering and principal component analysis to evaluate changes in the spinal cord expression profiles among sample groups based on the transcripts per million data of all expressed genes obtained through sequencing. First, hierarchical clustering showed that the samples of rats in group S had high similarity and differed from those of rats in groups N and SL (Fig. 2B). In addition, principal component analysis showed that the expression profiles of rats in group S differed from those of rats in groups N and SL. In contrast, the spinal cord expression profiles of rats in groups N and SL partially overlapped (Fig. 3A). These results suggest that spinal cord gene expression was altered in SNI model rats compared with the normal group and that lumbar sympathectomy shifts the expression profile of model rats closer to that of normal rats.

Fig. 2.

Quality assessment of rat spinal cord tissue using poly(A)-seq. (A) Sequencing saturation results of each sample. (B) Hierarchical cluster analysis dendrogram for each sample. (C) Transcriptome gene sequencing frequency and gene location distribution of each sample.

Fig. 3.

Lumbar sympathectomy modulates differentially expressed genes in the spinal cord of SNI rat. (A) Principal component analysis of spinal cord expression profiles of normal rats (N), SNI model rats (S), and lumbar sympathectomized SNI model rats (SL). (B) Volcano plot of differentially expressed genes (DEGs) in the spinal cord tissues of rats in groups S and N. Red dots indicate upregulated DEGs, green dots indicate downregulated DEGs and grey dots indicate non-significant differences in DEGs. (C) Volcano plot of DEGs in spinal cord tissues of rats in groups S and SL. Red, green, and grey dots have the same meanings as those in (C). (D) Heatmap inset of hierarchical clustering analysis of DEGs in the spinal cord tissues of rats in groups S and N. (E) Heatmap inset of hierarchical clustering analysis of DEGs in the spinal cord tissues of rats in groups S and SL. (F) Genotype classification statistics of DEGs in the spinal cord tissues of rats in groups S and N. (G) Genotype classification statistics of DEGs in the spinal cord tissues of rats in groups S and SL. PCA, principal component analysis; miscRNA, miscellaneous RNA; tRNA, transfer RNA; miRNA, micro RNA; lncRNA,long non-coding RNA.

We screened for DEGs with log2 (fold-change) $>$1 and p 0.05. Compared with the normal group, 278 DEGs were identified in the gene expression profile of the SNI model rats, with 177 upregulated and 101 downregulated genes (Fig. 3B). Compared with the SNI model rats, 174 DEGs were identified in the gene expression profile of SNI rats after lumbar sympathectomy, with 69 upregulated and 105 downregulated genes (Fig. 3C). In addition, we visualized the expression of the DEGs using a heatmap and performed clustering analysis (Fig. 3D,E). The heatmap clustering results showed that the expression profiles of the DEGs in the original SNI model rats were similar to those of the normal group after lumbar sympathectomy. We also evaluated the genotype classification statistics for the differentially expressed genes (Fig. 3F,G). Compared with the normal group, DEGs in the SNI model group included 208 protein-coding genes (mRNAs), 54 long non-coding RNAs, and several other RNAs (e.g., miRNAs and tRNAs). Compared with the SNI model group, DEGs in the SL group included 121 protein-coding genes (mRNAs) and 46 long non-coding RNAs, and several other RNAs (e.g., miRNAs and miscRNAs). We analyzed differentially expressed mRNAs using KEGG and GO enrichment analyses. KEGG enrichment analysis of DEGs in the SNI model group revealed the top 30 pathways, including neutrophil extracellular trap formation, IL-17 signaling pathway, chemokine signaling pathway, B-cell receptor signaling pathway, and other immune response pathways (Fig. 4A). The top 30 GO entries included lymphocyte-mediated immunity, leukocyte-mediated immunity, inflammasome complex, immunoglobulin-mediated immune response, immune receptor activity, humoral immune response, chemokine activity, B cell-mediated immunity, and other immune-related biological processes and cellular molecular components (Fig. 4B). Similarly, significant enrichment of immune-related pathways and cellular components was observed in GO and KEGG enrichment analyses of differential genes between the SNI and SL groups (Fig. 4C,D). Collectively, these results demonstrate that SNI model rats exhibit differential gene expression in the spinal cord compared with normal rats and that lumbar sympathectomy can strongly modulate this differential expression in the SNI model.

Fig. 4.

Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment were performed to analyze differentially expressed mRNAs. (A,C) KEGG enrichment bubble chart showing common DEGs. (B,D) GO enrichment bubble chart showing common DEGs.

Using Venn plot analysis, we identified 46 intersecting DEGs between SNI model rats and SNI rats after lumbar sympathectomy (Fig. 5A). We constructed a protein-protein interaction network of the intersecting genes using STRING software (https://cn.string-db.org/) and found that 26 genes were involved in protein interactions, including guanine nucleotide-binding protein g(4) subunit (Gng4), C-X-C motif chemokine ligand 10 (also known as IP-10) (Cxcl10), Cluster of differentiation 74 (also known as invariant chain) (Cd74), and C-X-C chemokine receptor type 3(Cxcr3) (Fig. 5B). We used a heatmap to display the expression patterns of selected lncRNAs and mRNAs in three experimental groups (N, S, and SL), with hierarchical clustering distinguishing lncRNA (upper section) and mRNA (lower section). Some lncRNAs (LOC1025508018, LOC108349650) and mRNAs (FAM187A, Smim6, etc.) were upregulated in the S group, represented by red or yellow in the color bar, while some genes (EPGN, Hist2h3c2, etc.) were downregulated in the S group, shown in blue in the color bar, and returned to levels close to those in the N group in the SL group. This result suggests that these genes may play an essential role in the process of cold allodynia induced by nerve injury. The bar plot on the right shows the statistical significance of the differential expression of various genes, with the red dashed line indicating the p-value threshold (0.05). A total of 19 genes were statistically significant (p 0.05), suggesting that these genes may be involved in the regulation of cold sensitivity (Fig. 5C). We next presented the relative mRNA expression levels of six genes (EPGN, Hist2h3c2, Smim6, FAM187A, LOC108349650, and LOC102550818) across the three experimental groups: N, S, and SL. Compared to the N group, the S group showed a significant decrease in EPGN expression (p 0.01, Fig. 5D) and Hist2h3c2 (p 0.001, Fig. 5E). Conversely, the opposite result was observed for Smim6 (p 0.01, Fig. 5F), FAM187A (p 0.001, Fig. 5G), LOC108349650 (p 0.001, Fig. 5H), and LOC102550818 (p 0.001, Fig. 5I). The SL group partially restored the expression of all six relative mRNAs, showing significant differences compared to the S group (p 0.05 or p 0.001). These findings collectively suggest that lumbar sympathectomy exerts analgesic effects at the molecular level, partially restoring the altered gene expression associated with nerve injury.

Fig. 5.

Screening and identification of core genes responsible for the effect of lumbar sympathectomy in improving limb cold allodynia caused by nerve injury. (A) Venn diagram of intersecting genes of two sets of DEGs (N vs. S, S vs. SL). (B) Protein-protein interaction network constructed from common DEGs. Grey lines indicate interactions between the corresponding proteins of a gene; a larger circle indicates that the gene is important. (C) A heatmap and bar plot analysis of differentially expressed lncRNAs and mRNAs across three experimental groups (N, S, and SL) with three biological replicates each. Relative mRNA expression of (D) EPGN, (E) Hist2h3C2, (F) Smim6, (G) FAM187A, (H) LOC108349650, and (I) LOC102550818. Compared with group N, **p 0.01 and ***p 0.001 indicate significant differences. Compared with group S, ^#p 0.05 and ^#⁢#⁢#p 0.001 indicate significant differences. n = 6.