Male C57BL/6 mice (28–32 g) at the age of 8–10 weeks were obtained from the Laboratory Animal Center of Ningxia Medical University. Male mice were selected for two primary reasons: first, the occurrence of SCI is notably more prevalent among men compared to women, making male mice a suitable model (Barbiellini Amidei et al., 2022). second, sex differences are known to influence neuroinflammatory responses (Khaksari et al., 2017). To eliminate sex-related variability, experiments were performed exclusively on male subjects (Zeng et al., 2023). All mice were housed under controlled conditions at a constant temperature (21 ± 3 °C) and humidity (50% ± 5%), with a 12-h light/dark cycle.

The mice were randomly divided into two groups: the Sham group and the SCI group. Behavioral experiments were conducted on days 31, 32, and 34, including the open field test (OP) (n = 3), the tail suspension test (TS) (n = 3), and the sucrose preference test (SP) (n = 3), to assess whether the mice exhibited depressive-like behaviors. Hippocampal tissue was then harvested for downstream analyses. Total RNA was extracted and analyzed by reverse transcription-quantitative PCR (RT-PCR); the sample size for validating the four candidate genes identified through bioinformatics analysis was n = 6, whereas all other RT-PCR assays were conducted with n = 3. Additional assays included western blotting (WB, n = 3), Nissl staining (n = 3), hematoxylin–eosin (HE) staining (n = 3), and immunofluorescence staining (n = 3).

The SCI model was generated following established protocols as previously documented (Ni et al., 2019). The mice were subjected to anesthesia using isoflurane (2–4% for induction and 1.5% for maintenance) before performing a mid-back incision at the tenth thoracic level to expose the spinal cord. After performing a T10 laminectomy, a rod weighing 10 g was dropped from a height of 20 millimeters onto the spinal cord utilizing an impactor, with careful attention to prevent the application of excessive pressure. Subsequently, the suturing of the muscles and skin was carried out in a sequential manner, and manual bladder evacuation was conducted twice daily for 1 week until the restoration of normal bladder functionality was achieved. The body temperature of the animals were kept at a controlled temperature of 35 ± 1 °C utilizing a miniature heating pad. In the Sham group, a laminectomy procedure was conducted without inducing an SCI model.

Hippocampal tissue samples were obtained, and protein extraction was performed utilizing a Total Protein Extraction Kit (KeyGEN Bio TECH, Nanjing, Jiangsu, China) in alignment with the guidelines stipulated by the manufacturer. Subsequently, the protein concentrations were determined using a BCA protein assay kit (KeyGEN Bio TECH, Nanjing, Jiangsu, China). Identical volumes of protein samples were subjected to separation by SDS-PAGE using gels with concentrations ranging from 10 to 12%. The proteins underwent a transfer process onto polyvinylidene difluoride (PVDF) membranes. Blocking was conducted utilizing a freshly prepared solution of 5% skim milk for a duration of 1 to 2 h, succeeded by an overnight incubation at 4 °C with the primary antibodies: IκB-α (dilution 1:2000; Abcam, Cat. No. ab32518, Cambridege, Massachusetts, United States), p-IκB-α (dilution 1:1000; Abcam, Cat. No. ab92700, Cambridege, Massachusetts, United States), p-p65(dilution 1:1000; Invitrogen, Cat. No. MA5-15160, Carlsbad, California, United states), and beta actin (dilution 1:10000; Affinity, Cat. No. AF7018, Nanjing, Jiangsu, China). Finally, horseradish peroxidase (HRP)-conjugated secondary antibodies were introduced to the membranes and incubate at room temperature for 1 h. The visibility of the blotted protein bands was achieved through the application of an enhanced chemiluminescence (ECL) kit (KeyGEN Bio TECH, Nanjing, Jiangsu, China), and the quantitative assessment was performed utilizing ImageJ software (National Institutes of Health, Bethesda, MD, USA), with normalization against β-actin levels.

To validate the gene expression alterations identified through transcriptomic analysis, a RT-PCR assay was conducted. Total RNA was isolated from the hippocampus using TRIzol reagent.

The quality of the RNA was evaluated utilizing a Thermos Nanodrop Analyzer, with an A260/A280 ratio between 1.8 and 2.0 indicating good RNA purity. Reverse transcription was conducted utilizing a Light Cycle 96 Real-time PCR system. The relative expression levels of relative genes were determined employing the 2^-ΔΔCt methodology, utilizing β-actin as the internal control. Gene expression analyses were performed at three time points. The sequences of the primers can be found in Table 1.

The immunofluorescence staining protocol described by Zhang et al. (2025) was performed to achieve precise image visualization. At first, hippocampal brain sections were subjected to overnight incubation at 4 °C with a rabbit anti-IκB-α antibody (dilution 1:50; Abcam, Cambridge, Massachusetts, United States). Following a wash with phosphate-buffered saline, a mouse anti-Iba-1 antibody (dilution 1:100; Abcam, Cambridge, Massachusetts, United States) was utilized and incubated overnight. Subsequently, after an additional PBS wash, brain sections were permitted to incubate at ambient temperature for 1 h with goat anti-rabbit IgG (dilution 1:50; Proteintech, Wuhan, Hubei, China) and goat anti-mouse IgG (dilution 1:200; Proteintech, Wuhan, Hubei, China). Thereafter, the sections were allowed to incubate at room temperature for a duration of 5 min DAPI (Biosharp, Hefei, Anhui, China). Eventually, the sections that had been stained were inspected using a BX43F fluorescence microscope (Olympus, Japan), and images were obtained at 400 × magnification.

To evaluate neuronal cell death, Nissl staining was conducted following established protocols from previous studies. Brain specimens underwent a process of fixation, dehydration, embedding in paraffin, and were subsequently sectioned to a thickness of 10 μm. The tissue samples were subjected to a dewaxing process using xylene, repeating the procedure three times for a duration of 5 min each time. Subsequently, the specimens were treated sequentially with anhydrous ethanol for a duration of 5 min, 90% ethanol for 2 min, 70% ethanol for another 2 min, and distilled water for 2 min. The specimens underwent Nissl staining for a duration of 10 min, followed by two brief rinses with distilled water. Next, the specimens underwent a dehydration process using 95% ethanol, applied twice for a duration of 2 min per treatment. Subsequently, a clearing procedure was performed with xylene, repeated twice for 5 min each time. Finally, the specimens were mounted utilizing neutral gum. Under light microscopy, normal neurons displayed large cell bodies with abundant cytoplasm. Conversely, damaged neurons exhibited reduced cell body size and the presence of numerous vacuoles.

HE staining was conducted using an HE staining kit (G1120, Solarbio, Beijing, China). In summary, brain sections were stained with Mayer’s hematoxylin and eosin. After a 50-s exposure to eosin and subsequent dehydration in 95 and 100% ethanol, the sections were cleared with xylene and mounted. Images were captured by microscope (Olympus, Tokyo, Japan).

The OP is widely used in neurological research for its simplicity and effectiveness in assessing depression-like behaviors in murine models. In this experiment, mice were positioned at the center of a box with dimensions of 50 × 50 × 35 cm and were permitted to explore the environment for a duration of 5 min. The lighting condition in the experimental area was set at 30 lux. The behavioral patterns of each mouse were captured utilizing a video tracking system provided by RWD Life Science Co., Ltd.

For the TS, mice were hung by bands attached to the edges of a 20-cm-high platform. The duration of immobility during a 5-min assessment period was recorded utilizing a video recording device (RWD Life Science Co., Ltd.). Immobility was defined as the lack of movement, with the exception of whisker movements and respiratory activity.

The SP was conducted under controlled conditions. Prior to the main experiment, mice were given access to two containers, each filled with a 1% sucrose solution (Sinopharm, Shanghai, China) for a duration of 24 h. The two bottles were placed on opposite sides of the cage, ensuring that the distance from the nest area to each bottle was identical. Subsequently, one of the containers was substituted with pure water, allowing the mice to acclimatize for an additional 24-h period. Following this acclimation phase, the locations of the two containers were exchanged. Prior to the assessment, the mice underwent a 24-h deprivation of both food and water. After this deprivation, they were offered pre-determined quantities of the 1% sucrose solution and pure water during a 3-h testing interval. The residual volumes in each container were measured post-assessment to determine the sucrose preference percentage, which was calculated using the formula: sucrose preference (%) = sucrose consumption / (sucrose consumption + water consumption) × 100%.

Gene expression datasets GSE45006 and GSE183386 were downloaded from the GEO database using the search terms “Spinal cord injury” and “Depression.” The corresponding annotation platforms for these datasets are GPL1355 and GPL6247, respectively.

The GEO2R online platform was utilized to analyze raw data for the purpose of identifying DEGs between SCI and MDD. DEGs were filtered based on the criteria of |log₂FC| > 0.5 and p < 0.05, yielding a comprehensive list of genes (Supplementary Table S1). Volcano plots for DEGs in datasets GSE45006 and GSE183386 were accomplished utilizing the “Ggplot2” package within the R, whereas heatmaps were produced with the “ComplexHeatmap” package (Gu et al., 2016). The intersection of DEGs from GSE45006 and GSE183386 was identified, and Venn diagrams were constructed using the “Ggplot2”and “VennDiagram” packages.

To investigate the notably enriched biological functions and signaling pathways associated with these DEGs, the “ClusterProfiler” package (Wu et al., 2021) was used to perform GO and KEGG analyses on the common DEGs from GSE45006 and GSE183386. Bubble plots summarizing the enrichment analysis results were generated using the “Ggplot2” package.

The DEGs identified in the study were subsequently submitted to the STRING database to facilitate the construction of a PPI network through the use of the available integrated analysis tools. Subsequently, the results were imported into Cytoscape software, where the MMC algorithm in the cytoHubba plugin was used to identify the top four hub genes based on interaction scores.

The data are presented as the mean ± standard deviation (SD). Statistical differences were assessed using Student’s t-test for all comparisons in this study. All statistical analyses were conducted using GraphPad Prism (version 9.3). A p-value < 0.05 was considered statistically significant.

To evaluate the presence of depressive symptoms, we conducted standardized behavioral evaluations, including the OP, SP, and TS. As illustrated in Figure 1A, the SCI model was established on day 0, followed by OP, SP, and TS on days 31, 32, and 34, respectively. The OP results revealed significant behavioral changes, with SCI mice exhibiting reduced exploration of the center zone compared to Sham controls (Figures 1B,C). Although the total distance traveled did not significantly differ between the SCI and Sham groups during the OP, the SCI group showed a reduction in distance traveled (Supplementary Figure S1A). Furthermore, the average velocity of the SCI group was significantly lower in the OP (Supplementary Figure S1B). These findings suggest that the reduced exploration in the center zone during the OP may more accurately reflect the presence of depressive-like behaviors in the SCI group. The SP results indicated significant anhedonia in the SCI mice (Figure 1E). Similarly, in the TS, the SCI mice exhibited prolonged immobility following the injury (Figure 1D). Due to motor impairments in mice with SCI, we are unable to determine whether the increased immobility observed in the TS is caused by depressive-like behavior or impairments in motor function. Therefore, we combined the results from the OP and the SP to comprehensively assess whether the mice exhibited depressive-like behaviors. Ultimately, we concluded that mice exhibit significant depressive-like phenotypes following SCI.

To investigate the molecular factors associated with the comorbidity of SCI and depression, we compared the transcriptomes from SCI (GSE45006) with those from depression (GSE183386). Following normalization, as illustrated in Figures 2A,B for the GSE45006 and GSE183386 datasets, respectively, the distribution of the two sample sets met standard criteria, confirming that the microarray data were of high quality. Utilizing the “ggplot2” package, we identified 4,633 DEGs in the GSE45006 dataset and 38 DEGs in the GSE183386 dataset, with volcano plots illustrating these findings in Figures 2C,E, respectively. The “ComplexHeatmap” package was employed to generate heatmap visualizations of these DEGs, as shown in Figures 2D,F.

Intersection analysis of DEGs from both datasets using the ggplot2 and “VennDiagram” packages identified 16 co-expressed DEGs, illustrated in a Venn diagram (Figure 2G). These shared DEGs include Nfkbia (NF-κB inhibitor alpha); Fkbp5 (FK506-binding protein 5); Acer2 (alkaline ceramidase 2); Cd180 (CD180 molecule); Cdkn1a (cyclin-dependent kinase inhibitor 1a); Errfi1 (ERBB receptor feedback inhibitor 1); Ddit4 (DNA damage-inducible transcript 4); Gpr149 (G protein-coupled receptor 149); Mmp14 (matrix metallopeptidase 14); Pcdh20 (protocadherin 20); Pla2g3 (phospholipase A2 group III); Psmb9 (proteasome 20S subunit beta 9); Sgk1 (serum/glucocorticoid-regulated kinase 1); Pvalb (parvalbumin); Slc18a3 (solute carrier family 18 member a3); and Slc5a7 (solute carrier family 5 member 7).

Both GO and KEGG analyses confirmed a significant association between the 16 co-expressed DEGs and various signaling pathways (Figure 3). These DEGs were classified into three main functional categories: biological processes (BP), molecular functions (MF), and cellular components (CC). The key BP comprised “Response to glucocorticoid,” “Cellular response to glucocorticoid stimulus” and “Response to corticosteroid.” MF analysis identified “Heat shock protein binding” and “Ubiquitin protein ligase binding.” The CC category mainly included “Clathrin-coated endocytic vesicle,” “Neuromuscular junction,” and “Clathrin-coated vesicle.” KEGG pathway analysis pinpointed key inflammatory cascades, such as “PI3K-Akt signaling pathway.” Collectively, the data indicated that these DEGs predominantly participate in neuroinflammatory regulation.

To explore protein interactions, we constructed a PPI network consisting of 36 common genes, utilizing the STRING database. The PPI network analysis (confidence coefficient = 0.15) revealed 16 hub targets, with 28 edges, an average node degree of 3.5, and an average clustering coefficient of 0.633 (Figure 4A). We imported the PPI network into Cytoscape, visualized it using the Cytohubba plugin, and analyzed it with the MCC algorithm to evaluate its overall topological structure. Based on node significance, the four key hub genes identified were Nfkbia, Fkbp5, Sgk1, and Cdkn1a (Figure 4B). These candidate genes are annotated in the volcano plots (Figures 2C,E).

To confirm the transcriptomic findings, we performed RT-PCR and WB analyses on hippocampal tissues obtained from Sham and SCI mice. As shown in Figure 5A, SCI mice exhibited significant downregulation of Nfkbia mRNA compared to Sham controls, while Fkbp5, Sgk1, and Cdkn1a did not show significant changes (Figures 5B–D).

In terms of mechanism, canonical NF-κB activation involves phosphorylated IκB-α, which releases the NF-κB (p50/p65) complex, allowing it to translocate to the nucleus and initiate the transcription of pro-inflammatory cytokines such as IL-1β, IL-6, TNF-α. This process is counter-regulated by the resynthesis of IκB-α, encoded by Nfkbia (Taniguchi and Karin, 2018). Next, we evaluated the expression of IκB-α and phosphorylated IκB-α (p-IκB-α) using WB analysis. The WB results showed an increased p-IκB-α/IκB-α ratio in the SCI group compared to the Sham group (Figures 5E,F), indicating prolonged activation of NF-κB signaling. We then assessed the expression of p-p65 using WB analysis. Semi-quantitative analysis of p-p65, the active form of the NF-κB transcription factor, revealed a significant increase in SCI mice compared to Sham controls (Figure 5G). Taken together, these findings identify Nfkbia as a key regulator of the comorbidity between SCI and depression.

Supporting our molecular findings, RT-PCR analysis revealed increased levels of pro-inflammatory cytokines in SCI mice, including IL-1β (Figure 6A), IL-6 (Figure 6B), and TNF-α (Figure 6C), compared to Sham controls. The observed cytokine expression patterns correlate with the NF-κB activation profile shown in Figure 5G, highlighting the functional relationship between IκB-α phosphorylation and the production of inflammatory mediators.

Histopathological examination using HE staining revealed notable changes in the hippocampal CA1 region of mice with SCI-depression comorbidity, including partial neuronal loss, cytoplasmic vacuolation, and disrupted alignment of pyramidal cells (Figure 6D). Nissl staining demonstrated that Sham mice retained normal cytoarchitecture, exhibiting densely packed CA1 pyramidal neurons with intact Nissl bodies (Figure 6E). In contrast, SCI mice exhibited significant neurodegeneration characterized by disrupted neuronal arrangement, enlarged intercellular gaps, shrunken neuronal cell bodies, and a marked decrease in cytoplasmic Nissl body density. Immunofluorescence microscopy indicated that IκB-α co-localized with the microglia marker Iba-1 (Figure 6F, red arrowheads). Additionally, IκB-α levels were reduced in the SCI group compared to the Sham group, as highlighted by the red arrowheads in Figure 6F. This spatial distribution supports the increased p-IκB-α/IκB-α ratio observed in Figure 5F. Figure 6G illustrates the number of IκB-α and Iba-1 positive cells in both the SCI and Sham groups.

The prevalence of depression among individuals with SCI is about twice as high as that in the general population (Brakel and Hook, 2019). This trend is also evident in rodent models, where subjects with SCI demonstrate increased depressive-like behaviors, assessed through various experimental measures in this study, including the OP, SP, and TS. Although psychosocial and socioeconomic factors have traditionally been viewed as the main causes of depression after SCI, recent preclinical evidence indicates that biological mechanisms also significantly contribute to its development. In a study by Brakel et al., individuals with higher levels of pro-inflammatory cytokines both before and after SCI were more likely to develop depressive symptoms (Brakel et al., 2021). Similarly, the study found that ongoing peripheral and central inflammation significantly affects psychological health following SCI (Maldonado-Bouchard et al., 2016). Inflammation may serve as either a cause or a biomarker for the onset of depression-like behaviors after SCI.

Consistent with previous studies, we observed significant differences in pro-inflammatory cytokine expression in the SCI group (Wu et al., 2014b; Brakel and Hook, 2019; Santo et al., 2019). The levels of pro-inflammatory cytokines in the hippocampal tissue of the SCI group were elevated compared to those in the sham group. This suggests that inflammation extends beyond the primary injury site, as the brain also produces pro-inflammatory cytokines. Following SCI, the brain also expresses pro-inflammatory cytokines linked to activated microglia, which release cytotoxic and pro-inflammatory factors. This leads to neurotoxicity in the surrounding environment and triggers behavioral changes related to cell death (Wu et al., 2014a). Despite these findings, there are currently no specific treatment options available for depression following SCI. A thorough exploration of the molecular pathways involved in SCI-specific depression is essential for understanding the condition and identifying potential treatment options.

We identified potential co-expressed genes and pathways that interact with SCI and depression using bioinformatics analysis. We subsequently screened four hub genes from the protein–protein interaction (PPI) network: Nfkbia, Fkbp5, Sgk1, and Cdkn1a. RT-PCR analysis of these four genes showed that only Nfkbia mRNA was significantly downregulated in the SCI group compared to the sham group; this finding was further validated by WB analysis. In sepsis-associated encephalopathy, ALKBH5 attenuates inflammation by slowing the degradation of NFKBIA mRNA, thereby increasing NFKBIA protein levels (Ye et al., 2025). KDM5B enhances macrophage-mediated inflammatory responses by repressing Nfkbia transcription (Zhang et al., 2023). This suggests that Nfkbia is important for regulating inflammatory responses. Importantly, the Nfkbia gene encodes IκB-α, which maintains NF-κB in an inactive cytoplasmic state under resting conditions (Yu et al., 2020). Among the NF-κB dimer family, the p50/p65 heterodimer, the most common isoform in mammalian cells, serves as the primary driver of inflammatory signaling (Taniguchi and Karin, 2018). IκB molecules are phosphorylated in response to inflammatory cytokines, bacterial or viral products, and various forms of cellular stress (Israel, 2010). This process releases the p50/p65 heterodimer, enabling its translocation into the nucleus to activate the transcription of inflammatory genes. When NF-κB is activated, it may promote the expression of the IκB-α gene through a direct mechanism. This process ensures that cells can rapidly synthesize IκB-α protein, effectively regulating the activity of NF-κB (Sun et al., 1993). Additionally, studies have shown that the circadian rhythm gene NFIL3 can activate the NF-κB signaling pathway by inhibiting the transcription of Nfkbia (Yang et al., 2022). This finding further emphasizes the complexity of Nfkbia regulation and suggests that further research is needed to gain a deeper understanding of this mechanism. In our study, SCI mice exhibited: (1) an increased p-IκB-α/IκB-α ratio and elevated p-p65 levels, and (2) a reduction in the intensity of IκB-α immunofluorescence. These findings provide compelling evidence that the reduction of Nfkbia following SCI activates the IκB/NF-κB axis, thereby promoting neuroinflammation and ultimately mediating depression-like behaviors.

Investigations of neuroinflammatory pathways have yielded significant insights into the mechanisms driving the comorbidity of SCI and depression. Our experimental validation of NF-κB pathway activation (evidenced by an increased p-IκB-α/IκB-α ratio and enhanced p-p65 expression) provides direct molecular evidence that SCI-induced neuroinflammation underlies the comorbidity of depression. Although the KEGG analysis did not specifically indicate the involvement of the NF-κB pathway, the observed enrichment of the PI3K/Akt signaling pathway suggests a bidirectional regulatory interaction with NF-κB signaling. This finding is consistent with prior research showing mutual interactions between the PI3K/Akt and NF-κB pathway (Taniguchi and Karin, 2018). Specifically, the PI3K/Akt pathway promotes NF-κB nuclear translocation through IKK-dependent IκB phosphorylation. This pathway crosstalk creates an inflammatory amplification loop, consistent with the observed increases in cytokines (IL-1β, IL-6, TNF-α). Identifying these pathways lays a foundation for future research. This research aims to develop targeted interventions that address both the neuroinflammatory and psychological aspects of SCI-induced depression.

Histopathological examinations confirm the behavioral and molecular findings, revealing significant neuronal loss and alterations in the CA1 region of the hippocampus in SCI mice. The observed neurodegenerative changes were associated with behavioral outcomes, establishing a clear link between physical trauma and the neurobiological basis of depression. This finding is particularly relevant, as the hippocampus is crucial for mood regulation, and its dysfunction may exacerbate depressive symptoms in patients with SCI. These observations highlight the necessity for histopathological evaluations in future research to clarify the neurobiological mechanisms underlying SCI and depression, thereby guiding neuroprotective strategies that protect neuronal integrity and function in vulnerable populations.

This study identified key genes associated with depressive-like behaviors following SCI through bioinformatics and validated the potential of Nfkbia as a diagnostic and therapeutic biomarker for these behaviors through in vivo experiments, thereby providing a theoretical foundation for clinical diagnosis and treatment. However, there are some limitations to this study. First, our behavioral tests were conducted at only one late time point, which limits the breadth of the findings. Second, although we explored the role of Nfkbia in neuroinflammation and depression related to SCI, there was a lack of in-depth functional analysis. Additionally, we only investigated the IκB/NF-κB signaling pathway in microglia, leaving its presence in neurons or astrocytes unexplored. Finally, the sample size in this study was relatively small. Future studies will adopt a longitudinal design to characterize the temporal dynamics of SCI-associated depressive symptoms, deepen our functional understanding of Nfkbia, and systematically compare the contribution of the IκB/NF-κB pathway in microglia, astrocytes, and neurons to the behavioral phenotype. As our understanding of these interrelationships continues to deepen, targeted therapeutic strategies addressing both the physiological and psychological aspects of SCI may significantly improve patient outcomes and quality of life.