Male Sprague Dawley (SD) rats (n=60; weight, 200±20 g; age, 4-6 weeks) were purchased from Jiangsu Laboratory Animal Centre and housed under standard conditions (12-h dark/light cycle; temperature, 23±2˚C; ad libitum access to food and water; relative humidity, 40-60%). All procedures were approved by the Animal Care and Use Committee of Xiamen University (approval no. 20210401; Xiamen, China). Morbidity was used as the humane endpoint in the present study. These endpoints were applied following the criteria delineated in the ‘Guidelines for Endpoints in Animal Study Proposals’ at the Zhongshan Hospital Affiliated to Xiamen University (15). The humane endpoints included: i) The animal exhibited symptoms including, but not limited to, a lack of responsiveness to manual stimulation, and/or immobility and an inability to eat or drink; ii) laboured breathing and cyanosis; iii) diarrhoea or urinary incontinence; iv) severe, rapid weight loss and emaciation (maximum 20% of body weight from baseline); v) impaired mobility; vi) other situations where a veterinarian had determined that euthanasia was necessary.

Primary neurons were cultured as previously described (16). Briefly, a pregnant SD rat was anaesthetized by intraperitoneal injection of 50 mg/kg pentobarbital sodium and rapidly sacrificed using CO₂ inhalation (50% chamber volume/min). Subsequently, the cerebral cortex of the six embryos (embryonic day 16) was cut into ~1-mm pieces and isolated using trypsin (Sangon Biotech Co., Ltd.) for 20 min at 37˚C. After being centrifuged (300 x g; 5 min; room temperature), cells were resuspended and seeded into poly-D-lysin-coated plates (Corning, Inc.). Cells were then cultured in Neurobasal-A medium (Gibco; Thermo Fisher Scientific, Inc.) supplemented with B-27^™ supplement (Gibco; Thermo Fisher Scientific, Inc.) and 100 U/ml penicillin/streptomycin (Sangon Biotech Co., Ltd.) at 37˚C with 5% CO₂. The medium was replaced every 2 days and primary neurons were harvested after 10 days of culture.

The primary neurons were plated into a 6-well plate (5x10⁵ cells/well) and were divided into two groups: The H₂O₂ group, in which primary neurons were treated with 100 µM H₂O₂ (Beijing Solarbio Science & Technology Co., Ltd.) (17,18); the normal group, in which the primary neurons were cultured without H₂O₂ stimulation. After 24 h of stimulation at 37˚C, cells were harvested for reverse transcription-quantitative PCR (RT-qPCR), western blotting and apoptosis analysis.

The FGF23 overexpression adenovirus-associated virus (AAV; oeFGF23), short hairpin (sh)RNA AAV (shFGF23) and negative control (NCs) AAVs (oeNC and shNC) were purchased from Shanghai GenePharma Co., Ltd. Briefly, primary neurons were seeded into 6-well plates (5x10⁵ cells/well) and infected (multiplicity of infection, 50) with overexpression NC AAV (oeNC), oeFGF23, shRNA NC AAV (shNC) or shFGF23 at 37˚C for 24 h. The sense sequences for shNC and shFGF23 were: 5'-GTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAAC-3' and 5'-GGAACAGCTATCACCTACATTCAAGAGATGTAGGTGATAGCTGTTCC-3', respectively. Uninfected primary neurons were used as control cells. After being cultured for 48 h at 37˚C, cells in all groups were stimulated with 100 µM H₂O₂ for 24 h at 37˚C and harvested for RT-qPCR, western blotting and apoptosis analysis.

oeNC- or oeFGF23-infected cells were seeded into 6-well plates (5x10⁵ cells/well) and divided into the following groups: OeNC, cells infected with oeNC; oeFGF23, cells infected with oeFGF23; LY294002, cells infected with oeNC and incubated with 10 µM LY294002 (PI3K inhibitor; MedChemExpress) at 37˚C for 12 h (19); and oeFGF23 + LY294002, cells infected with oeFGF23 and incubated with 10 µM LY294002 at 37˚C for 12 h. Primary neurons without infection or inhibitor treatment were used as the control group. The primary neurons in all groups were then stimulated with 100 µM H₂O₂ for 24 h at 37˚C. Finally, the cells were harvested for RT-qPCR, western blotting and apoptosis analysis.

RT-qPCR was performed to assess FGF23 mRNA expression levels in primary neurons. Briefly, total RNA was extracted from 1x10⁶ primary neurons using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Subsequently, RNA was reverse transcribed into cDNA using a PrimeScript^™ RT reagent kit (Takara Bio, Inc.) according to manufacturer's protocols. qPCR amplification was performed using a TB Green^® Fast qPCR mix kit (Takara Bio, Inc.), and relative mRNA expression levels of FGF23 were calculated using the 2^-ΔΔCq method (20). The primer sequences were as follows: FGF23 forward, 5'-TGGCCATGTAGACGGAACAC-3' and reverse, 5'-GGCCCCTATTATCACTACGGAG-3'; and GAPDH forward, 5'-CAAGTTCAACGGCACAGTCAAG-3' and reverse, 5'-ACATACTCAGCACCAGCATCAC-3'. The thermocycling conditions were as follows: 95˚C for 30 sec (one cycle), followed by 40 cycles at 95˚C for 5 sec and 61˚C for 20 sec.

The apoptotic rate of primary neurons was detected using an Annexin V-FITC/PI Cell Apoptosis Detection Kit (Beyotime Institute of Biotechnology). Briefly, infected or treated cells were washed twice with precooled PBS and adjusted to 2x10⁶ cells/ml. Subsequently, 5 µl Annexin V-FITC and 5 µl PI were added to the cell suspension for 10 min at room temperature. Flow cytometric analysis was carried out using the FACSCanto II flow cytometer (BD Biosciences). The data were analysed using FlowJo 7.6.1 (BD Biosciences).

The SCI rat model was constructed as previously described (21). The SD rats were anaesthetized by intraperitoneal injection of 50 mg/kg pentobarbital sodium and a laminectomy was performed at the T9 vertebral section to expose the spinal cord. Subsequently, the spinal cord was contused with a force of 200 kDyne using the Infinite Horizon Impactor (Precision Systems & Instrumentation). Injection of AAV or PBS at the site of injury was performed after the surgery. The rats were divided into five groups: Sham group (n=12), which received only laminectomy without contusion and 10 µl PBS injection; SCI group (n=12), which received SCI surgery and 10 µl PBS injection; SCI + oeNC group (n=12), which received SCI surgery and 10 µl oeNC injection; SCI + oeFGF23 group (n=12), which received SCI surgery and 10 µl oeFGF23 injection; and SCI + oeFGF23 + LY294002 group (n=12), which received SCI surgery, 10 µl oeFGF23 injection and LY294002 injection (0.3 mg/kg/day; intravenous) (22). At 7 days post-operation, rats were anaesthetized by intraperitoneal injection of 50 mg/kg pentobarbital sodium and rapidly sacrificed using CO₂ asphyxiation (50% chamber volume/min) and spinal cord lesion tissues (n=6/group) were harvested for western blotting. In the remaining rats, Basso-Beattie-Bresnahan (BBB) scoring (23) was performed to assess the locomotion recovery at 1, 3, 7, 10, 14, 21 and 28 days after surgery (n=6/group). At 28 days post-surgery, the rats were anaesthetized by intraperitoneal injection of 50 mg/kg pentobarbital sodium and rapidly sacrificed using CO₂ asphyxiation (50% chamber volume/min), and spinal cord lesion tissues were collected for H&E and TUNEL staining, and ELISA. The death of the rats was confirmed following the AVMA Guidelines for the Euthanasia of Animals: 2020 Edition, including lack of pulse, breathing, corneal reflex and response to firm toe pinch; inability to hear respiratory sounds and heartbeat by use of a stethoscope; greying of the mucous membranes; and rigor mortis (24).

The spinal cord lesion tissues were fixed with 4% paraformaldehyde (Beyotime Institute of Biotechnology) for 24 h at room temperature, embedded in paraffin and cut into sections (4 µm). H&E staining was accomplished using the H&E staining kit (Wuhan Servicebio Technology Co., Ltd.) according to the manufacturer's instructions. The images were captured using a light microscope (Olympus Corporation). TUNEL staining was performed using a One Step TUNEL Apoptosis Assay Kit (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. The images were observed using an inverted fluorescence microscope (Olympus Corporation).

The spinal cord lesion tissues were lysed in RIPA buffer (Beyotime Institute of Biotechnology) and the supernatant was collected by centrifugation (10,000 x g; 15 min; 4˚C). Subsequently, the supernatant was quantified using a BCA kit (Beijing Solarbio Science & Technology Co., Ltd.). Rat TNF-α (cat. no. D731168), IL-1β (cat. no. D731007) and IL-6 (cat. no. D731010) ELISA kits (Sangon Biotech Co., Ltd.) were then used according to the manufacturer's instructions. The absorbance at 450 nm was assessed using a microplate reader (BioTek Instruments, Inc.).

The primary neurons or spinal cord lesion tissues were lysed using RIPA buffer. After quantification using a BCA kit, a total of 50 µg protein/lane was separated by SDS-PAGE on precast 10% gels (Willget) and transferred onto nitrocellulose membranes (Wuhan Servicebio Technology Co., Ltd.). The membranes were then blocked with 5% BSA (Wuhan Servicebio Technology Co., Ltd.) at 37˚C for 1.5 h and incubated overnight at 4˚C with the following primary antibodies: Anti-FGF23 (cat. no. DF3596; 1:500; Affinity Biosciences), anti-cleaved (C)-caspase3 (cat. no. AF7022; 1:1,000; Affinity Biosciences), anti-Bcl-2 (cat. no. AF6139; 1:1,000; Affinity Biosciences), anti-phosphorylated (p)-PI3K (cat. no. ab182651; 1:2,000; Abcam), anti-PI3K (cat. no. ab191606; 1:2,000; Abcam), anti-p-AKT (cat. no. ab38449; 1:2,000; Abcam), anti-AKT (cat. no. ab8805; 1:2,000; Abcam) or anti-GAPDH (cat. no. T0004; 1:4,000; Affinity Biosciences). Following incubation with HRP-conjugated secondary antibodies for 1 h at 37˚C (cat. nos. S0001and S0002; 1:10,000; Affinity Biosciences), the protein bands were visualized using an Hypersensitive ECL Chemiluminescence Kit (Wuhan Servicebio Technology Co., Ltd.) and semi-quantified using ImageJ software (version 1.52; National Institutes of Health) using GAPDH as the loading control.

Quantitative data are presented as the mean ± standard deviation in triplicate and were analysed using GraphPad Prism software (Version 7.0; Dotmatics). Unpaired Student's t-test was used for comparisons between two groups. BBB scoring was analysed using Kruskal-Wallis test with Dunn's post hoc test. For other indexes, one-way ANOVA followed by Tukey's multiple comparisons test was used for comparisons among three or more groups. P<0.05 was considered to indicate a statistically significant difference.

Primary neurons were treated with H₂O₂ to establish a cellular model of SCI (17). The results revealed that the apoptotic rate was higher in the H₂O₂ group compared with that in the normal group (P<0.001; Fig. 1A and B). Furthermore, C-caspase3 expression levels were higher (P<0.01), whereas Bcl-2 expression levels were lower (P<0.05) in the H₂O₂ group compared with those in the normal group (Fig. 1C and D). Additionally, the protein (P<0.05; Fig. 1E and F) and mRNA expression levels of FGF23 (P<0.05; Fig. 1G) were decreased in the H₂O₂ group compared with those in the normal group; the protein expression ratios of p-PI3K/PI3K and p-AKT/AKT were both decreased in the H₂O₂ group compared with those in the normal group (both P<0.01; Fig. 1H and I).

The mRNA and protein expression levels of FGF23 were higher in the oeFGF23 group compared with those in the oeNC group (both P<0.001), whereas the mRNA and protein expression levels of FGF23 were lower in the shFGF23 group compared with those in the shNC group (both P<0.05) (Fig. 2A-C), which implied successful infection. The apoptotic rate was lower in the oeFGF23 group compared with that in the oeNC group (P<0.01), whereas it was higher in the shFGF23 group compared with that in the shNC group (P<0.05) (Fig. 2D and E). Additionally, the protein expression levels of C-caspase3 were lower in the oeFGF23 group compared with those in the oeNC group and higher in the shFGF23 group compared with those in the shNC group (both P<0.05), whereas Bcl-2 showed the opposite trend (oeFGF23 vs. oeNC, P<0.01; shFGF23 vs. shNC, P<0.05) (Fig. 2F and G).

Regarding PI3K/AKT signalling, the p-PI3K/PI3K (P<0.001) and p-AKT/AKT (P<0.01) protein ratios were higher in the oeFGF23 group compared with those in the oeNC group, whereas the relative protein expression levels of p-PI3K/PI3K (P<0.05) and p-AKT/AKT (P<0.05) were lower in the shFGF23 group compared with those in the shNC group (Fig. 3).

The p-PI3K/PI3K and p-AKT/AKT protein ratios were decreased in the LY294002 group compared with those in the oeNC group (both P<0.01); they were also lower in the oeFGF23 + LY294002 group compared with those in the oeFGF23 group, which implied that LY294002 decreased the effect of FGF23 on PI3K/AKT signalling (both P<0.001) (Fig. 4A and B). FGF23 mRNA and protein expression levels did not differ between the LY294002 and oeNC groups nor between the oeFGF23 + LY294002 and oeFGF23 groups, which indicated that LY294002 did not affect FGF23 expression (both P>0.05; Fig. 4C-E).

The apoptotic rate was higher in the LY294002 group compared with that in the oeNC group and this was also increased in the oeFGF23 + LY294002 group compared with that in the oeFGF23 group (both P<0.01; Fig. 5A and B). In addition, the protein expression levels of C-caspase3 were increased, whereas those of Bcl-2 were decreased in the LY294002 group compared with those in the oeNC group (both P<0.05) and in the oeFGF23 + LY294002 group compared with those in the oeFGF23 group (both P<0.01) (Fig. 5C and D). These results indicated that LY294002 reduced the effect of FGF23 on neuronal apoptosis.

FGF23 protein expression was decreased in rats in the SCI model group compared with those in the sham group (P<0.05), whereas it was increased in the SCI + oeFGF23 group compared with that in the SCI + oeNC group (P<0.001) (Fig. 6A and B). Additionally, the protein expression levels of FGF23 did not differ significantly between the SCI + oeFGF23 + LY294002 and SCI + oeFGF23 groups (P>0.05). Furthermore, the p-PI3K/PI3K and p-AKT/AKT protein expression ratios were increased in the SCI + oeFGF23 group compared with those in the SCI + oeNC group (both P<0.05; Fig. 6C and D). The p-PI3K/PI3K (P<0.01) and p-AKT/AKT (P<0.05) ratios were decreased in the SCI + oeFGF23 + LY294002 group compared with those in the SCI + oeFGF23 group, which indicated that LY294002 decreased the effect of FGF23 on PI3K/AKT signalling in SCI model rats.

Moreover, laceration and inflammatory cell infiltration were reduced in the SCI + oeFGF23 group compared with those in the SCI + oeNC group, whereas they increased in the SCI + oeFGF23 + LY294002 group compared with those in the SCI + oeFGF23 group (Fig. 7A). The BBB score (on day 28) was increased in the SCI + oeFGF23 group compared with that in the SCI + oeNC group (P<0.05) but decreased in the SCI + oeFGF23 + LY294002 group compared with that in the SCI + oeFGF23 group (P<0.01) (Fig. 7B). Furthermore, TNF-α (P<0.01; Fig. 7C) and IL-1β (P<0.05; Fig. 7D) levels were lower, whereas IL-6 levels were not altered (P>0.05; Fig. 7E) in the SCI + oeFGF23 group compared with those in the SCI + oeNC group. The TNF-α (P<0.001), IL-1β (P<0.01) and IL-6 (P<0.05) levels were higher in the SCI + oeFGF23 + LY294002 group compared with those in the SCI + oeFGF23 group. Overall, FGF23 overexpression suppressed tissue injury and inflammation, and improved locomotion recovery in rats with SCI, which was reversed by LY294002.

The number of TUNEL-positive cells was decreased in the SCI + oeFGF23 group compared with that in the SCI + oeNC group (P<0.05); however, the number of TUNEL-positive cells was increased in the SCI + oeFGF23 + LY294002 group compared with that in the SCI + oeFGF23 group (P<0.001; Fig. S1). This finding indicated that FGF23 suppressed apoptosis in rats with SCI, which was reversed by LY294002.