The mechanism by which hyperbaric oxygen treatment alleviates spinal cord injury: genome-wide transcriptome analysis.

Accumulating studies have demonstrated that hyperbaric oxygen (HBO) treatment alleviates spinal cord injury (SCI). However, the underlying mechanism by which HBO alleviates SCI remains to be elucidated. In this study, we performed genome-wide transcriptional profiling of the spinal cord between SCI mice and mice that received HBO treatment by high-throughput RNA sequencing at 1 week after SCI. We also compared genome-wide transcriptional profiles from SCI mice and sham-operated mice. We found 76 differentially co-expressed genes in sham-operated mice, SCI mice, and HBO-treated SCI mice. Using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis, we identified the biological characteristics of these differentially expressed genes from the perspectives of cell component, biological process, and molecular function. We also found enriched functional pathways including ferroptosis, calcium signaling pathway, serotonergic synapse, hypoxia-inducible factor-1 signaling pathway, cholinergic synapse, and neuroactive ligand-receptor interaction. We performed quantitative reverse transcription-polymerase chain reaction and validated that HBO treatment decreased the expression of Hspb1 (heat shock protein beta 1), Hmox1 (heme oxygenase 1), Ftl1 (ferritin light polypeptide 1), Tnc (tenascin C) and Igfbp3 (insulin-like growth factor binding protein 3) and increased the expression of Slc5a7 (solute carrier family 5 choline transporter member 7) after SCI. These results revealed the genome-wide transcriptional profile of the injured spinal cord after HBO treatment. Our findings contribute to a better understanding of the mechanism by which HBO treats SCI and may provide new targets for SCI intervention.

Introduction

Spinal cord injury (SCI) is an urgent medical problem with a high disability rate. According to epidemiological statistics, the incidence of traumatic SCI is 10.5 cases per 100,000 people globally, with an estimated 768,473 new cases worldwide each year (Kumar et al., 2018). SCI is characterized by a primary injury followed by a secondary injury. Mechanical trauma directly causes the primary injury with spinal cord tissue destruction, bleeding, and necrosis. The secondary injury results from events that follow the primary injury, including local tissue edema, ischemia, hypoxia, inflammation, apoptosis, and abnormal metabolism, leading to severe cell injury and death, which last several days to weeks (Kwon et al., 2004; Brennan and Popovich, 2018; Wang et al., 2019). Because the primary injury is irreversible, most therapeutic strategies for SCI focus on interfering with or reducing the processes and mechanisms of secondary injury.

Hyperbaric oxygen (HBO) is the therapy of inhaling pure oxygen exposed to ambient pressure in more than one absolute atmosphere (Thom, 2011). HBO is a safe and non-invasive physical treatment and has been used for SCI treatment. Numerous studies have demonstrated that HBO treatment can mitigate secondary SCI by boosting blood oxygen partial pressure, increasing oxygen diffusion distance, inhibiting lipid peroxidation, relieving inflammatory responses, and reducing cell apoptosis (Sun et al., 2019; Ying et al., 2019; Chen et al., 2021). However, most of these studies have only identified several molecules or a single signaling pathway involved in the mechanisms of HBO treatment. An in-depth and comprehensive research of the underlying mechanisms of HBO treatment in SCI is lacking.

Several studies have performed genome-wide transcriptome profiling after SCI. Wang et al. (2021) identified differentially expressed lncRNAs and miRNAs in post-acute traumatic SCI by high-throughput sequencing and uncovered a regulatory axis in spinal fibrotic scars. Wu et al. (2019) investigated differential circRNA expression profiles following SCI in rats, and the results provide insight into the transcriptional regulation of the involved genes. However, there is no report on the gene transcriptome of the injured spinal cord following HBO treatment.

To comprehensively investigate the mechanisms by which HBO treatment alleviates SCI, we analyzed the differentially expressed genes of the spinal cord in mice at 1 week after injury using genome-wide transcriptome analysis.

Materials and Methods

Animals

Seventy-two C57BL/6J female mice (aged 7–9 weeks, weighing 19–22 g) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd (No. SCXK (Jing) 2016-0006). Animals were housed in a specific pathogen-free environment with stable temperature 24 ± 2°C and humidity 60 ± 10% in a 12-hour light/dark cycle. All mice were housed in individual cages and had free access to food and water. The mice were free to move around between measurements at each study time point. All animal procedures were approved by the Experimental Animal Ethics Committee of Beijing Chaoyang Hospital, Capital Medical University (approval No. 2020-3-18-98) on March 25, 2020. This study was reported in accordance with the ARRIVE 2.0 guidelines (Animal Research: Reporting of In Vivo Experiments) (Percie du Sert et al., 2020).

Prior to surgery, all mice were randomly divided into three experimental groups using the random number table method: (1) sham group (SH, n = 24): exposed spinal cords were not subject to a contusion injury or HBO treatment; (2) spinal cord injury group (SCI, n = 24): exposed spinal cords were subject to a contusion injury; and (3) HBO treatment group (HBO, n = 24): exposed spinal cords were subject to a contusion injury followed by HBO treatment.

SCI model establishment

A moderate contusive SCI model was established by the Multicenter Animal Spinal Cord Injury Study (MASCIS) Impactor mode III (W. M. Keck Center for Collaborative Neuroscience, Piscataway, NJ, USA) following a previous study (Sun et al., 2018). First, mice were anesthetized by intraperitoneal injection of 2% pentobarbital (75 mg/kg, Merck KGaA, Darmstadt, Germany) and T10 laminectomy was performed to expose the dura mater. The spine was fixed, and the MASCIS impactor was used to strike the exposed dura mater (a 10-g weight was allowed to fall freely from a height of 6.25 mm above the exposed dura mater) to create a SCI contusion. Signs of successful SCI modeling included edema and hemorrhage in the spinal cord tissue, intact dura, the wagging tail reflex, retraction of the lower limbs and body flutter, and flaccid paralysis of both lower extremities. For recovery from anesthesia and surgery, mice were fed and given postoperative care. Manual bladder evacuation was performed twice daily until the animals regained full reflexive bladder control.

HBO treatment

Mice in the HBO group were placed into an experimental HBO chamber (volume: 0.196 m3, 701 Research Institute, Wuhan, China) at 6 hours after injury and administered HBO treatment. The chamber was flushed using pure oxygen for 5 minutes to replace CO2. The pressure in the chamber was kept at 2.0 atmospheres absolute (ATA) through compressing oxygen for 10 minutes for inhaling 100% O2 for 60 minutes, followed by decompression to normal baric air for 10 minutes. HBO treatment was administered once daily over a period of 7 days to 7 weeks. Mice from the SH and SCI groups inhaled 21% O2 at normal atmospheric pressure in laboratory postoperatively.

Behavioral tests

The hind limb motor function of animals was evaluated based on the Basso Mouse Scale (BMS) as described in a previous study (Basso et al., 2006). The BMS scoring system ranges from 0 (complete paralysis) to 9 points (completely normal). Nine mice in each group were evaluated within 2 hours of the end of HBO treatment, every 2 days in the first week and weekly thereafter until 7 weeks. All behavioral analyses were performed by two investigators blinded to the experiments. The mean score of the two scores were used as the BMS score of the sample.

Tissue processing for histological analysis

To investigate the influence of HBO treatment on spinal cord histomorphology after SCI, the spinal cord was harvested and stained with hematoxylin-eosin (HE) on the 7th day after surgery. Six mice of each group were anesthetized as described above and perfused transcardially with 4% paraformaldehyde. The spinal cord section containing the injury epicenter (approximately 6 mm long) was removed and fixed with 10% neutral buffered formaldehyde. The pathological samples were stored at 4°C for 1 week and embedded in paraffin. The blocks were serially cut horizontally at 5 µm both sides symmetrically from the center of the injury using a pathology slicer (Shanghai Leica Instrument Co., Ltd., Shanghai, China). The slices were stained with HE (Wuhan Servicebio Technology Co., Ltd., Wuhan, Hubei Province, China) and examined by two pathologists blinded to the experiment using light microscopy (Nikon Corporation, Tokyo, Japan). Histopathological injury was scored based on the structure of grey matter and white matter, hemorrhage, cellular edema, necrosis, inflammatory cell infiltration, neuron number, and formation of vesicles according to our previously published method (Liu et al., 2014). Briefly, the severity of injury was graded for each variable (0–4 points): no injury, 0; injury to 25% of the field, 1; injury to 55% of the field, 2; injury to 75% of the field, 3; and diffuse injury, 4.

RNA extraction and transcriptome sequencing

We performed RNA sequencing (RNA-seq) on spinal cord tissue from different samples (SH, SCI, and HBO) on the 7th day after surgery and conducted gene expression analysis. The TransZol Up Plus RNA Kit (Quan Si Gold Biotechnology, Beijing, China) was used to extract total RNA from the spinal cord (6 mm in length, containing the injury epicenter) of each group (n = 3) following the manufacturer’s instructions. The TruSeq® RNA Sample Preparation Kit v2 (Illumina, San Diego, CA, USA) was used for isolation of polyadenylated mRNA with oligo-dT beads, second strand cDNA synthesis, and next generation sequencing library preparation. The AMPure XP system (Beckman Coulter, Beverly, CA, USA) was used to purify the products of libraries. We assessed the quantity and quality of each library using a Qubit®2.0 Fluorimeter (Thermo Fisher Scientific, Waltham, MA, USA) and Agilent 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA, USA), respectively. Shanghai Biotechnology Corporation (Shanghai, China) conducted the library creation and transcriptome sequencing on an Illumina Hissed 2500 platform (Illumina).

Data analysis for gene expression

Sequencing raw reads were filtered by Seqtk before mapping to the human GRCh38 reference genome using Hisat2 (version 2.0.4) (Kim et al., 2015; http://daehwankimlab.github.io/hisat2/). After genome mapping, Stringtie (version 1.3.0) (Pertea et al., 2016; http://ccb.jhu.edu/software/stringtie/) was used to generate fragments per kilobase of exon model per million mapped reads (FPKM) values for known gene models. EdgeR was used to identify differentially expressed genes. P-value in multiple tests was set by the false discovery rate (FDR). Fold-changes were estimated according to the FPKM in each sample. The differentially co-expressed genes among the three groups were selected using Venny (https://bioinfogp.cnb.csic.es/tools/venny/) using the following filter criteria: FDR ≤ 0.05 and fold change ≥ 1.5. Heatmap of differentially co-expressed genes among the three groups was made by pheatmap (R package). The transcriptome data were deposited in the gene expression omnibus (GEO) database under the accession number GSE185301 (https://www.ncbi.nlm.nih.gov/gds/?term=GSE185301).

Function and pathway enrichment analyses

Annotation of differentially expressed genes including cell component, biological process, and molecular function was obtained from the Gene Ontology (GO) database (http://geneontology.org/). Pathway analysis was performed with Kyoto Encyclopedia of Genes and Genomes (KEGG) database (https://www.kegg.jp/). GO and KEGG enrichment analyses were performed using cluster Profiler (R package) (Yu et al., 2012). The reciprocity network of differentially expressed genes was analyzed using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) software (http://string-db.org/) (Pertea et al., 2015). Shanghai Biotechnology Corp (Shanghai, China) helped conduct these analyses. P value < 0.05 was set as a cutoff.

Quantitative reverse transcription-polymerase chain reaction

Spinal cords (6 mm in length, encompassing the lesion epicenter) were collected from six mice of each group on the 7th day after injury and stored at –80°C. Total RNA was isolated from the frozen injured spinal cord using Trizol reagent (Thermo Fisher Scientific), and cDNA was transcribed from 1.0 µg of RNA using M-MLV reverse transcriptase and dNTP (10 mM) (Takara Biotechnology, Dalian, China). The accession number in GenBank database and primers of each gene are listed in . Quantitative reverse transcription-polymerase chain reaction was performed using TB Green® Premix Ex Taq™ II (Takara Biotechnology) detected by Roche Light Cycler® 480II Real-Time PCR System (Roche, Switzerland). The reaction mixture contained 12.5 µL Premix Taq, 1.0 µL cDNA, 10 μM primer F/R, and 10.75 µL dd H2O. The quantitative real-time PCR conditions were as follows: denaturation at 95°C for 5 minutes, followed by 40 cycles of 95°C for 10 seconds, 60°C for 30 seconds, and 72°C for 30 seconds. The PCR products were analyzed by electrophoresis on 1% agarose gels. The relative expression of each mRNA was calculated using the 2–ΔΔCt method (Pan et al., 2017).

Statistical analysis

No statistical methods were used to predetermine sample size; however, our sample sizes were similar to those reported in a previous publication (Hervera et al., 2018). Data were analyzed using SPSS 19.0 software (IBM Corp., Armonk, NY, USA). Data were expressed as the mean ± standard deviation. A two-way repeated analysis of variance with Bonferroni’s post hoc comparisons test was carried out to analyze the differences in BMS score between the groups over time. One-way analysis of variance with the least significant difference post hoc test was used to analyze other data. All tests were two sided. A value of P < 0.05 was considered statistically significant.

Results

HBO treatment improves locomotor function after SCI

As shown in , SCI model mice in both SCI and HBO groups exhibited total paralysis of the hind limbs with 0 BMS score at 1 day after injury. Gradual hind limb function recovery of mice was observed in both groups over time and peaked at 6 weeks after injury. Hind limb function of mice recovered more quickly in the HBO group. The BMS score of mice in the HBO group was significantly higher at each time point compared with the SCI group except for 1 day after injury (P < 0.01).

Effect of HBO treatment on functional recovery following SCI.

BMS score of mice in SH, SCI and HBO groups at indicated time points after SCI. The BMS scoring system ranges from 0 (complete paralysis) to 9 points (completely normal). Data are expressed as the mean ± standard deviation. n = 9 mice per group. The experiments were repeated three times. **P < 0.01, vs. SCI group (two-way repeated analysis of variance followed by Bonferroni’s post hoc test). BMS: Basso Mouse Scale; HBO: spinal cord injury + hyperbaric oxygen treatment; SCI: spinal cord injury; SH: sham.

HBO treatment alleviates histopathological injury after SCI

HE staining showed that the spinal cord tissue structure of the SH group was intact, characterized by non-injury necrosis or hemorrhage, clear boundary between grey and white matter, orderly arrangement of cells, clearly visible neurons and glial cells, and intact cell membrane and nucleus (). In the SCI group, spinal cord tissue displayed a disordered structure, extensive cell edema, a large amount of inflammatory cell infiltration, hemorrhage, widened cell space, formation of cyst or vesicles, significantly reduced number of neurons, severe tissue necrosis, and blurred boundary between grey and white matter (). After HBO treatment, these injuries were remarkably alleviated (). As shown in , the histopathological score was significantly increased after SCI compared with the score of the SH group (P < 0.01), whereas the HBO group showed a significantly reduced the histopathological score compared with the SCI group (P < 0.01).

Effect of HBO treatment on spinal cord histopathology following SCI.

(A–C) Representative HE staining images of the spinal cord in the SH (A), SCI (B), and HBO (C) groups. Spinal cord tissue displayed disorder structure, cell edema, inflammatory cell infiltration, hemorrhage, widened cell space, cyst formation, and reduced neuron number after SCI. HBO treatment lessened these injuries. Black arrows indicate neurons; white arrows indicate glial cells. Scale bars: 500 μm (left) and 50 μm (right) in A–C. (D) Histopathological scores of spinal cord tissue in the three groups. The histopathological score was significantly increased after SCI and the score was significantly decreased after HBO treatment. n = 6 mice per group. The experiments were repeated three times. Data are expressed as the mean ± standard deviation. **P < 0.01, vs. SH group; ##P < 0.01, vs. SCI group (one-way analysis of variance followed by the least significant difference post hoc test). HBO: Spinal cord injury + hyperbaric oxygen treatment; SCI: spinal cord injury; SH: sham.

Identification of differentially expressed genes

We performed a pairwise comparison among the three groups to identify the differentially expressed genes. Differentially expressed genes were defined as genes with a greater than 1.5-fold change and P < 0.05. On the basis of the criterion, there were 6207 differentially expressed genes between the SCI and SH groups, 5497 differentially expressed genes between the HBO and SH groups, and 645 differentially expressed genes between the HBO and SCI groups. A total of 76 differentially co-expressed genes were identified, including 67 upregulated and 9 downregulated genes (). The 76 differentially co-expressed genes are listed in . The dynamic changes of the 76 differentially expressed genes among the three groups are shown by the heatmap and cluster dendrogram in .

RNA-sequencing analysis of differentially expressed genes.

(A) Venn diagrams show significantly up- and down-regulated differentially expressed genes among the three experimental groups. Sixty-seven differentially expressed genes were upregulated after SCI, and the upregulation was diminished by HBO treatment. Nine differentially expressed genes were down-regulated after SCI, and these genes were increased after HBO treatment. P < 0.05 (edgeR analysis). (B) A clustered heat map shows the dynamic changes of 76 differentially expressed mRNAs among the three groups (green: low expression; red: high expression). n = 3 biological replicates. HBO: Spinal cord injury + hyperbaric oxygen treatment; SCI: spinal cord injury; SH: sham.

Bioinformatics analysis of differentially expressed genes

We analyzed the 76 differentially expressed genes using GO analysis. As shown in , the differentially expressed genes were mainly distributed in cell part (26.5%), organelle (21.8%), membrane (11.5%), and macromolecular complex (10.3%). shows the molecular function analysis of these 76 differentially expressed genes. The top function was binding (42.3% of all genes), followed by catalytic activity (19.5%), signal transducer activity (8.1%), and molecular transducer activity (7.4%). The differentially expressed genes were involved in different biological processes, such as single organism process (10.3%), cellular process (10.1%), response to stimulus (6.6%), and developmental process (6.6%) (.

Bioinformatics analysis of differentially expressed genes.

(A–C) Seventy-six differentially expressed genes were classified as cellular component (A), molecular function (B), and biological process (C) by GO analysis. (D) Functional enrichment analysis of differentially expressed genes relating to pathways by KEGG pathway analysis. Ferroptosis, calcium signaling pathway, serotonergic synapse, HIF-1 signaling pathway, cholinergic synapse, and neuroactive ligand-receptor interaction were the significantly enriched functional pathways. (E) STRING software shows the functional linkage between differentially expressed genes. GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; STRING: Search Tool for the Retrieval of Interacting Genes/Proteins.

KEGG pathway enrichment analysis of the integrated differentially expressed genes revealed ferroptosis, calcium signaling pathway, serotonergic synapse, HIF-1 signaling pathway, cholinergic synapse, and neuroactive ligand-receptor interaction as the significantly enriched functional pathways (). shows the close interrelationships between these differentially expressed genes.

Validation of differentially expressed genes

HBO treatment has been shown to involve oxidative stress and cell death (Pan et al., 2018; Hedetoft et al., 2021). After comprehensive analysis of the fold change and bioinformatics data of the differentially expressed genes, we evaluated the biological functions of genes related to oxidative stress and nerve cell death and identified five upregulated genes (Hspb1, Hmox1, Ftl1, Tnc, and Igfbp3) and one down-regulated gene (Slc5a7) for validation by qPCR using an independent sample set. As shown in , the expressions of Hspb1, Hmox1, Ftl1, Tnc, and Igfbp3 in the spinal cord were upregulated in the SCI group compared with the SH group (P < 0.01). The expressions of Hspb1, Hmox1, Ftl1, Tnc, and Igfbp3 were significantly reduced in the HBO group compared with the SCI group (P < 0.05 or P < 0.01). The expression of Slc5a7 was significantly lower in the SCI and HBO groups than that in the SH group (P < 0.01). The expression of Slc5a7 was significantly increased in the HBO group compared with the SCI group (P < 0.01). Trends in expression of these six genes were consistent with transcriptional analysis.

Detection of differentially expressed genes using qRT-PCR.

Six differentially expressed genes were selected for validation by qRT-PCR. (A) Hspb1 (heat shock protein beta 1), (B) Hmox1 (heme oxygenase 1), (C) Ftl1 (ferritin light polypeptide 1), (D) Tnc (tenascin C), (E) Igfbp3 (insulin like growth factor binding protein 3), and (F) Slc5a7 (Solute carrier family 5 choline transporter member 7). n = 6 mice per group. The experiments were repeated three times. Data are expressed as the mean ± standard deviation. *P < 0.05, **P < 0.01, vs. SH group; #P < 0.05, ##P < 0.01, vs. SCI group (one-way analysis of variance followed by the least significant difference post hoc test). HBO: Spinal cord injury + hyperbaric oxygen treatment; qRT-PCR: quantitative reverse transcription-polymerase chain reaction; SCI: spinal cord injury; SH: sham.

Discussion

In the current study, we established a clinically relevant model of contusive SCI in mice and our results showed that HBO treatment significantly alleviated pathological changes of damaged spinal cord and improved motor function of mice after SCI. To the best of our knowledge, this is the first study to examine transcriptional profiles of SCI in response to HBO through transcriptome sequencing. We not only reveal the HBO treatment-associated gene profile, but we also identified several oxidative stress- and cell death-related genes such as Hspb1, Hmox1, Ftl1, Tnc, Igfbp3 and Slc5a7 with altered expression in response to HBO. Further qPCR analysis validated our results and confirmed that the expression of these six genes were significantly different between HBO and the other two groups.

Similar to our previous studies (Liu et al., 2014, 2015), the histopathological scores of spinal cord were significantly reduced and BMS scores of hind limb were significantly increased by HBO treatment. In this study, we investigated the effect of HBO treatment on hind limb motor function for 7 weeks and did not limit our analysis to the acute phase of SCI. The recovery of limb function was fast in the first 2 weeks after SCI, gradually slowed down, and reached a peak at the 6th week. The BMS score at the 7th week was almost the same as that at the 6th week. These results suggested that the role of HBO in promoting the recovery of limb function was more obvious at the early stage of SCI, and HBO treatment should be performed as soon as possible after injury.

To date, the HBO treatment protocol following SCI has not been established. In one study, the pressure was 2–3 ATA, and each oxygen inhalation lasted 60–90 minutes (Patel et al., 2017). In China, the pressure of HBO treatment for SCI was mostly 2 ATA in the clinic (Huang et al., 2021). Hence, we selected 2 ATA and 60 minutes oxygen inhalation in this study. Moreover, most assessments regarding HBO treatment for SCI have been based on imaging, histopathology, specific molecules, and signaling pathways (Pan et al., 2018; Sun et al., 2018; Meng et al., 2019). However, genome-wide analysis of the injured spinal cord before and after HBO treatment has not been reported. In this study, we identified differentially expressed genes in damaged spinal cord from the three groups using high-throughput RNA sequencing at 1 week after SCI. Sequencing data showed that HBO treatment resulted in differential expression of 76 co-expressed genes compared with the other two groups. Functional GO and KEGG analyses indicated that some differentially expressed genes might play crucial regulatory roles in several pathways including binding, cell ferroptosis, and calcium signaling pathway. These findings suggested the possibility that HBO treatment may reduce SCI through regulating the expression of these genes.

We identified several differentially expressed genes related to oxidative stress and nerve cell death. The Hspb1 (heat shock protein beta1) gene encodes HSPB1. HSPB1, also known as HSP27 because of its relative molecular weight of 27 kDa, is one of the well-studied members of the small heat shock protein family. HSPB1 was originally identified as an intracellular molecular chaperone that maintains the normal cellular protein homeostasis in response to various stresses (Nakamoto and Vigh, 2007). Under normal condition, HSP expression is low in the central nervous system, but stressors such as ischemia, injury, and neurodegenerative diseases can induce its expression (Trivedi, 2007). The expression of HSP27 was increased after SCI and played a remarkable role in the resistance to oxidative stress, inflammation, and apoptosis after SCI (Nasouti et al., 2019; Zhou et al., 2019). Konda et al. (1996) reported that HBO treatment weakened the induction of HSP27 in the gerbil hippocampus after ischemia. Another study by Huang et al. (2014) showed that HBO preconditioning protected spinal neurons from oxidative injury through up-regulating HSPs (HSP32). Although the effect of HBO on HSPs is controversial, in our study, Hspb1 gene expression was induced by SCI, and HBO treatment inhibited the expression of Hspb1 on day 7 following injury. Our results indicated that HSPB1 may be involved in the process of HBO treatment ameliorating secondary SCI.

Hmox1 is a key cytoprotective gene and encodes the heme oxygenase-1 (HO-1) enzyme. HO-1 contributes to cytoprotection after injury (Abraham and Kappas, 2008). Activation of the nuclear factor-erythroid 2-related factor 2 (Nrf2)/HO-1 signaling pathway exerts an anti-oxidative stress response after SCI (Liu et al., 2018, 2020). Other studies reported that HBO treatment induced the expression of HO-1 in experimental and clinical contexts (Hedetoft et al., 2021; Nesovic Ostojic et al., 2021). Lin et al. (2019) showed that the expression of HO-1 increased greatly from 1 to 3 days after SCI. In the present study, upregulated expression of Hmox1 gene was observed on the 7th day after SCI. Unexpectedly, Hmox1 gene expression was downregulated in the HBO group compared with the SCI group. This may be because of the examined time point. Our findings suggest HBO treatment may attenuate SCI through regulating the expression of Hmox1 gene.

Ftl1 encoding ferritin protein is involved in ferroptosis. Ftl1 is involved in many diseases through multiple pathways such as oxidative stress and the NF-κB pathway (Bertoli et al., 2019). Cao et al. (2021) confirmed that the upregulation of Ftl1 activated ferroptosis in mouse hippocampus after chronic unpredictable mild stress. Some scholars have found that ferroptosis is an important step in the process of secondary SCI and inhibiting ferroptosis could improve the limb motor function of SCI rats (Yao et al., 2019; Zhang et al., 2019). Our results showed that the most enriched KEGG pathway was ferroptosis, the expression of Ftl1 gene was upregulated after SCI, and HBO treatment inhibited upregulation of the Ftl1 gene. Therefore, we speculated that HBO treatment inhibits ferroptosis after SCI by influencing the expression of Ftl1. To the best of our knowledge, this is the first study to report the relationship between HBO and the Ftl1 gene.

Extracellular matrix glycoprotein TNC (tenascin-C), which is encoded by the Tnc gene, is a large glycoprotein with several distinct domains. TNC binds to different types of cellular receptors and extracellular matrix via its structurally distinct domains, thereby exhibiting functional diversity (Taylor et al., 1989). Although early increased tenascin-C is necessary for regeneration and wound healing (Sumioka et al., 2013), sustained increases are deleterious. Different animal disease models such as models for SCI, Alzheimer’s disease, and subarachnoid hemorrhage indicated that Tnc was strongly upregulated after injury and exacerbated CNS inflammation. Furthermore, Tnc deletion or silencing could inhibit neuropathology, inflammation, and apoptosis (Hashimoto et al., 2005; Xie et al., 2013; Tong et al., 2020). Moreover, the axonal plasticity and growth in the injury site of Tnc depletion mice was enhanced after SCI (Schreiber et al., 2013). In this study, the expression of Tnc was significantly upregulated after SCI, and HBO treatment remarkably reduced its expression. Our results may partially explain the mechanism by which HBO treatment alleviates SCI.

As the major binding protein of insulin-like growth factor-1 (IGF-1), IGFBP-3 (insulin-like growth factor binding protein-3), encoded by the Igfbp3 gene, inhibits part of the neuroprotective effects of IGF-I by preventing IGF-I from binding to its receptor, therefore increasing cell apoptosis (Jogie-Brahim et al., 2009). IGFBP-3 concentration is increased in Alzheimer’s disease, which may contribute to neuronal degeneration (Rensink et al., 2002; Johansson et al., 2013). In addition, IGFBP-3 mRNA was moderately induced in reactive microglia and neurons of the injured hippocampus following hypoxic/ischemic injury (Beilharz et al., 1998). However, changes in Igfbp3 expression in spinal cord following SCI have not been reported. Our results suggested that HBO treatment may reduce secondary SCI through inhibiting the expression of Igfbp3.

The Slc5a7 (solute carrier family 5 choline transporter member 7) gene which encodes the presynaptic sodium-dependent high-affinity choline transporter (CHT: SLC5A7) is a fundamental determinant of neurotransmitter acetylcholine signaling (Iwamoto et al., 2016). Slc5a7–/– mice have very short survival because of motor paralysis and respiratory failure (Ferguson et al., 2004). SLC5A7 expressed in brain microvascular endothelial cells allows nutrients to selectively cross the blood-brain barrier (Inazu, 2019). Our results indicated that HBO treatment increased Slc5a7 gene expression, which may be helpful for alleviating SCI.

There are certain limitations to our study. First, the transcriptional analysis of differentially expressed genes is limited to one single time point, and analysis at multiple time points after injury is needed. Second, the molecular mechanism by which these differentially expressed genes influence SCI after HBO was not clarified, and further studies are needed.

In conclusion, our study uncovered a genome-wide expression profile of injured spinal cord following HBO treatment and identified that several genes including Hspb1, Hmox1, Ftl1, Tnc, Igfbp3, and Slc5a7 may be involved in the process of HBO treatment alleviating SCI. These results contribute to a better understanding of the mechanism by which HBO treats SCI and may identify new therapeutic targets for SCI.

Additional file:

Additional Table 1: Common differentially expressed genes in the three groups.

Additional Table 1

Common differentially expressed genes in the three groups

Gene ID(Ensembl database)	Gene name	HBO/SH		HBO/SCI		SCI/SH
		Fold change	P-value	Fold change	P-value	Fold change	P-value
ENSMUSG00000020427	Igfbp3	5.57	4.40E-27	0.57	0.002169	9.81	5.84E-27
ENSMUSG00000025650	Col7a1	7.85	7.87E-23	0.55	0.015444	14.39	1.53E-20
ENSMUSG00000005413	Hmox1	4.88	2.49E-19	0.42	0.001936	11.70	5.60E-16
ENSMUSG00000028364	Tnc	4.34	3.13E-19	0.58	0.006950	7.49	5.35E-17
ENSMUSG00000042379	Esm1	10.99	4.00E-16	0.62	0.042197	17.70	1.17E-21
ENSMUSG00000050708	Ftl1	3.40	5.42E-14	0.61	0.010553	5.61	2.68E-26
ENSMUSG00000062382	Gm10116	3.96	1.55E-13	0.62	0.024333	6.43	2.18E-26
ENSMUSG00000025473	Adam8	5.34	6.15E-13	0.60	0.040891	8.90	2.19E-16
ENSMUSG00000030208	Emp1	4.16	6.44E-10	0.58	0.017200	7.20	9.89E-26
ENSMUSG00000026938	Fcna	7.34	2.89E-09	0.55	0.020858	13.32	6.95E-14
ENSMUSG00000026208	Des	4.15	3.40E-09	0.33	0.003008	12.68	1.14E-10
ENSMUSG00000026228	Htr2b	5.97	3.83E-09	0.52	0.005843	11.47	5.26E-20
ENSMUSG00000029373	Pf4	6.14	3.88E-09	0.54	0.017480	11.46	2.36E-16
ENSMUSG00000068220	Lgals1	3.28	5.20E-09	0.60	0.015061	5.48	5.66E-15
ENSMUSG00000030787	Lyve1	6.97	2.06E-08	0.56	0.036853	12.54	7.27E-10
ENSMUSG00000050370	Ch25h	4.09	2.27E-08	0.64	0.033780	6.37	1.03E-12
ENSMUSG00000094338	Hist1h2bl	4.61	5.96E-08	0.60	0.043563	7.69	1.34E-11
ENSMUSG00000016356	Col20a1	2.58	1.00E-07	0.64	0.000405	4.02	1.73E-16
ENSMUSG00000032085	Tagln	3.10	3.32E-07	0.63	0.022431	4.92	1.25E-09
ENSMUSG00000006219	Fblim1	3.38	5.65E-07	0.65	0.037851	5.19	1.63E-19
ENSMUSG00000001025	S100a6	2.26	1.35E-06	0.62	0.018226	3.67	1.58E-11
ENSMUSG00000020377	Ltc4s	2.97	4.01E-06	0.66	0.044896	4.49	1.64E-11
ENSMUSG00000028464	Tpm2	2.21	6.88E-06	0.53	0.045645	4.20	1.02E-05
ENSMUSG00000030867	Plk1	2.89	7.46E-06	0.54	0.013798	5.34	5.11E-11
ENSMUSG00000105954	Gm42793	3.81	9.82E-06	0.54	0.044318	7.04	1.03E-10
ENSMUSG00000025701	Alox5	2.46	6.32E-05	0.60	0.008067	4.09	4.72E-12
ENSMUSG00000025270	Alas2	1.97	0.000132	0.56	0.029812	3.50	3.65E-06
ENSMUSG00000055775	Myh8	5.29	0.000156	0.21	0.005358	24.99	9.25E-08
ENSMUSG00000069300	Hist1h2bj	2.48	0.000209	0.52	0.003955	4.79	8.12E-12
ENSMUSG00000031097	Tnni2	5.04	0.000272	0.36	0.037796	14.10	4.45E-05
ENSMUSG00000051747	Ttn	3.40	0.000415	0.18	0.003850	19.10	2.04E-07
ENSMUSG00000000901	Mmp11	2.30	0.000483	0.58	0.026619	3.95	1.02E-07
ENSMUSG00000049699	Ucn2	5.11	0.000553	0.46	0.027000	11.02	7.36E-07
ENSMUSG00000037725	Ckap2	1.86	0.000623	0.65	0.025747	2.87	2.35E-08
ENSMUSG00000060470	Adgrg3	3.78	0.000658	0.49	0.011925	7.68	5.28E-09
ENSMUSG00000052305	Hbb-bs	1.50	0.000741	0.43	0.001689	3.49	2.59E-06
ENSMUSG00000034457	Eda2r	2.18	0.000841	0.57	0.015200	3.79	5.56E-09
ENSMUSG00000068744	Psrc1	2.28	0.001477	0.61	0.045358	3.77	1.03E-06
ENSMUSG00000037139	Myom3	4.86	0.001778	0.27	0.017661	17.74	2.81E-06
ENSMUSG00000017300	Tnnc2	5.61	0.001799	0.32	0.030543	17.31	7.03E-06
ENSMUSG00000027559	Car3	2.59	0.001862	0.38	0.036141	6.76	5.02E-05
ENSMUSG00000004951	Hspb1	1.87	0.001919	0.64	0.021352	2.94	6.38E-08
ENSMUSG00000026407	Cacna1s	4.63	0.002017	0.35	0.044762	13.26	0.000101
ENSMUSG00000060093	Hist1h4a	2.21	0.002602	0.53	0.014364	4.13	2.89E-06
ENSMUSG00000030730	Atp2a1	3.77	0.002754	0.22	0.021344	17.33	6.99E-06
ENSMUSG00000049134	Nrap	1.76	0.003589	0.38	0.008269	4.61	5.69E-05
ENSMUSG00000026950	Neb	2.86	0.003815	0.15	0.001650	19.08	2.88E-07
ENSMUSG00000028195	Cyr61	1.62	0.004419	0.65	0.018046	2.48	1.31E-06
ENSMUSG00000020061	Mybpc1	1.66	0.005029	0.41	0.003256	4.02	2.58E-06
ENSMUSG00000031972	Acta1	3.68	0.005198	0.17	0.008071	21.66	9.24E-07
ENSMUSG00000061723	Tnnt3	3.97	0.006093	0.31	0.027373	12.83	2.54E-05
ENSMUSG00000027022	Xirp2	3.25	0.006723	0.10	0.000941	31.83	1.52E-06
ENSMUSG00000037966	Ninj1	1.53	0.007674	0.62	0.013293	2.46	2.35E-06
ENSMUSG00000034413	Neurl1b	1.89	0.011064	0.53	0.013293	3.54	1.25E-06
ENSMUSG00000030905	Crym	2.22	0.011362	0.38	0.000416	5.81	1.03E-07
ENSMUSG00000098132	Rassf10	1.62	0.011696	0.60	0.003982	2.70	2.28E-08
ENSMUSG00000015468	Notch4	1.61	0.012422	0.63	0.018171	2.57	1.57E-05
ENSMUSG00000073940	Hbb-bt	1.51	0.015031	0.47	0.006031	3.20	1.33E-05
ENSMUSG00000050234	Gja4	1.77	0.016063	0.64	0.048328	2.76	9.55E-06
ENSMUSG00000020475	Pgam2	8.72	0.017905	0.20	0.010411	43.88	3.05E-06
ENSMUSG00000046295	Ankle1	1.91	0.017993	0.59	0.031139	3.25	5.75E-06
ENSMUSG00000043085	Tmem82	1.98	0.022230	0.48	0.012129	4.17	8.02E-06
ENSMUSG00000021367	Edn1	1.74	0.025682	0.55	0.020034	3.16	8.17E-06
ENSMUSG00000022769	Sdf2l1	1.53	0.031715	0.65	0.024352	2.36	4.93E-06
ENSMUSG00000006221	Hspb7	1.89	0.036322	0.22	0.001287	8.58	1.92E-05
ENSMUSG00000022218	Tgm1	1.50	0.040740	0.59	0.000973	2.54	7.40E-06
ENSMUSG00000067455	Hist1h4j	1.98	0.045598	0.52	0.034938	3.83	7.29E-06
ENSMUSG00000096458	Moap1	0.22	2.51E-06	35.58	8.03E-14	0.01	5.14E-28
ENSMUSG00000045349	Sh2d5	0.37	1.09E-05	1.73	0.016555	0.21	4.10E-14
ENSMUSG00000056423	Uts2b	0.22	1.57E-05	3.76	0.044830	0.06	7.51E-12
ENSMUSG00000023945	Slc5a7	0.50	0.000199	1.63	0.027411	0.31	1.25E-11
ENSMUSG00000021680	Crhbp	0.46	0.001344	1.89	0.046309	0.24	1.57E-08
ENSMUSG00000005268	Prlr	0.49	0.002342	3.47	2.40E-05	0.14	6.64E-25
ENSMUSG00000031285	Dcx	0.66	0.007058	1.51	0.021917	0.44	9.26E-07
ENSMUSG00000062561	Gm10118	0.64	0.016936	1.56	0.036242	0.41	1.57E-06
ENSMUSG00000025529	Zfp711	0.66	0.024679	1.66	0.011823	0.40	1.30E-07

Gene ID comes from Ensembl database. EdgeR was used to analyse the differentially expressed genes among the three groups. HBO: Hyperbaric oxygen; SCI: spinal cord injury; SH: sham.

Table 1

Primer sequences and accession number in GenBank database for qRT-PCR

Gene	Primer sequence (5’–3’)	Product size (bp)	Accession number
Hspb1	F: AAG GCA GGA CGA ACA TGG CTA CA	246	NM_007393.5
	R: GGC TTC TAC TTG GCT CCA GAC TGT
Hmox1	F: GAC CGC CTT CCT GCT CAA CAT T	105	NM_010442.2
	R: CTC TGA CGA AGT GAC GCC ATC TG
Ftl1	F: GGC AAC CAT CTG ACC AAC CTC C	122	NM_010240.2
	R: GCC TCC TAG TCG TGC TTG AGA GT
Tnc	F: TTG GCT TGG ACT GGA TAA CCT GAG	244	NM_001369211.1
	R: GTT GGT GAT GGC TGA GTC TGT GT
Igfbp3	F: CAG TAG ATG CTC CGT GCC ACA TAA	159	NM_008343.2
	R: TTC ACT CAG TTC ACC ACA CTC ACA
Slc5a7	F: TCG CAA GGC ACA GTG AAG AGA A	158	NM_022025.4
	R: CGG ACT GGA ATC AAC ATC AAG GAG
β-actin	F: GAG ATT ACT GCT CTG GCT CCT A	150	NM_007393.5
	R: GGA CTC ATC GTA CTC CTG CTT G

F: Forward; Ftl1: ferritin light polypeptide1; Hmox1: heme oxygenase1; Hspb1: heat shock protein beta1; Igfbp3: insulin-like growth factor binding protein3; qRT-PCR: quantitative reverse transcription-polymerase chain reaction; R: reverse; Slc5a7: solute carrier family 5 choline transporter member 7; Tnc: tenascin C.