Bexarotene improves motor function after spinal cord injury in mice

Abstract

Spinal cord injury is a challenge in orthopedics because it causes irreversible damage to the central nervous system. Therefore, early treatment to prevent lesion expansion is crucial for the management of patients with spinal cord injury. Bexarotene, a type of retinoid, exerts therapeutic effects on patients with cutaneous T-cell lymphoma and Parkinson’s disease. Bexarotene has been proven to promote autophagy, but it has not been used in the treatment of spinal cord injury. To investigate the effects of bexarotene on spinal cord injury, we established a mouse model of T11–T12 spinal cord contusion and performed daily intraperitoneal injection of bexarotene for 5 consecutive days. We found that bexarotene effectively reduced the deposition of collagen and the number of pathological neurons in the injured spinal cord, increased the number of synapses of nerve cells, reduced oxidative stress, inhibited pyroptosis, promoted the recovery of motor function, and reduced death. Inhibition of autophagy with 3-methyladenine reversed the effects of bexarotene on spinal cord injury. Bexarotene enhanced the nuclear translocation of transcription factor E3, which further activated AMP-activated protein kinase-S-phase kinase-associated protein 2-coactivator-associated arginine methyltransferase 1 and AMP-activated protein kinase-mammalian target of rapamycin signaling pathways. Intravenous injection of transcription factor E3 shRNA or intraperitoneal injection of compound C, an AMP-activated protein kinase blocker, inhibited the effects of bexarotene. These findings suggest that bexarotene regulates nuclear translocation of transcription factor E3 through the AMP-activated protein kinase-S-phase kinase-associated protein 2-coactivator-associated arginine methyltransferase 1 and AMP-activated protein kinase-mammalian target of rapamycin signal pathways, promotes autophagy, decreases reactive oxygen species level, inhibits pyroptosis, and improves motor function after spinal cord injury.

Introduction

Spinal cord injury (SCI) has become a global health burden because of the lost productivity and high health care expenses of patients with SCI. Approximately 2.4 million cases of SCI were reported in 2016 (GBD 2016 Traumatic Brain Injury and Spinal Cord Injury Collaborators, 2019). Generally, there are two phases in SCI: primary and secondary (Ahuja et al., 2017). These two phases contribute to irreversible dysneuria in patients following SCI. The initial traumatic events lead to primary injury, which is often a determinant of injury severity. The secondary injury exacerbates neurological deficits and worsens outcomes because it not only mediates injured cell death but also leads to ambient neurocyte dysfunction (Hellenbrand et al., 2021). Inflammation, oxidative stress and cell death have been demonstrated to exert important effects in secondary injury. Inflammation, which plays key roles in the initial injury, leads to subsequent mechanisms resulting in severe injury in the spinal cord (Liau et al., 2020).

Autophagy is a conserved process in eukaryotes that plays an important role in balancing the intracellular environment (Dikic and Elazar, 2018). During autophagy, damaged organelles and protein aggregates are engulfed in autophagosomes, which fuse with lysosomes, and cytosolic waste is degraded (Füllgrabe et al., 2013). Cell debris is ultimately discharged from the cell by exocytosis. Autophagy has been confirmed to play a positive role in SCI (Lipinski and Wu, 2015). Autophagy is particularly important in neurons because neurons are terminally differentiated cells from which aggregated proteins and defective organelles must be removed. The effects of autophagic activity on the spinal cord posterior to SCI are still controversial, but the consensus is that autophagy contributes to pathological cell elimination, while excessive autophagy increases neuronal death because of the resulting autophagic cell death (Ray, 2020). Transcription factor E3 (TFE3) is a transcription factor that drives the expression of autophagy- and lysosome-related genes to promote the formation and fusion of autophagosomes and lysosomes (Raben and Puertollano, 2016). TFE3 plays an important role in stimulating autophagy after SCI and may be a treatment target for SCI (Zhou et al., 2020).

Pyroptosis signal pathways are crucial for neurological conditions such as SCI (McKenzie et al., 2020). Pyroptosis mediates cell death through the formation of membranous pores at the plasma membrane through which mature interleukin-1β (IL-1β) and interleukin-18 (IL-18) are released and cause a range of inflammatory reactions or even death in surrounding cells (Jorgensen and Miao, 2015). Pyroptosis is induced by various stimuli, including reactive oxygen species (ROS). ROS act as intermediate triggers to activate the NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome, exacerbating subsequent inflammatory cascades and cell damage pathways, such as pyroptosis (Bai et al., 2020). Previous studies showed that the stimulation of mitophagy, a form of autophagy, suppresses the production of ROS and inhibits pyroptosis (Yu et al., 2018; Lin et al., 2019). Mitochondria are the primary producers of ROS through the redox reaction. In pathological conditions such as SCI, excess redox reaction products and H₂O₂ at supraphysiological concentrations induce unwanted protein oxidation, change reaction patterns, and damage biomolecules; this condition is called oxidation distress (Rao et al., 2019; Wang et al., 2019). Therefore, inhibiting ROS production or reducing ROS levels to suppress pyroptosis has become a potential therapeutic strategy in SCI (Lv et al., 2019).

Bexarotene (Bex) is a selective agonist for retinoid X receptors (RXR) that is used to treat patients with cutaneous T-cell lymphoma. Bex has been confirmed to activate autophagy, but the mechanism is unclear. A previous study showed that RXRa, a subunit of RXR, promoted cell autophagy by regulating phospholipase D1, leading to inhibition of the long noncoding RNA LINC00511 (Shi et al., 2020). In nervous system disease, Bex has been demonstrated to mediate mitophagy, restore autophagy, and improve mitochondrial anomalies in impaired hippocampal neural stem cells (Martín-Maestro et al., 2019). Furthermore, specific concentrations of Bex alleviate amyloid β-induced endoplasmic reticulum stress and apoptosis, while excessive concentrations of Bex upregulated endoplasmic reticulum stress proteins and activated pro-apoptotic BAD (He et al., 2018a). In tumor cells, Bex has been demonstrated to dose-dependently induce oxidative stress, DNA damage, and apoptosis via the peroxisome proliferator-activated receptor-γ/nuclear factor-κB signaling pathway (Hacioglu et al., 2021). Thus far, no study has examined the effects of Bex on acute damage, such as SCI.

Our study explored the effects of Bex on SCI and investigated the effect of Bex on oxidative stress, autophagy, mitophagy, and pyroptosis following SCI. We further examined the signaling pathway by which Bex promotes autophagy in SCI.

Methods

Animals

Healthy adult C57BL/6J mice (female, 20–30 g, 6–7 weeks old) were acquired from the Experiment Animal Center of Wenzhou Medical University (Zhejiang, China; license No. SCXK (Zhe) 2015-0001). Mice were housed under normal experimental conditions (specific-pathogen-free level, 12-hour light/dark period, 21–25°C, 50–60% humidity, adequate food and water), with a maximum of five animals per cage. All animal-related experiments were performed following the China National Institutes of Health Guidelines for the Welfare and Usage of Laboratory Animals and were approved by the Animal Ethics Board of Wenzhou Medical University (approval No. wydw2017-0096) on July 19, 2017.

We randomly divided 180 mice into seven groups: sham (n = 30), SCI (n = 30), SCI + Bex (n = 30), SCI + Bex + 3MA (n = 30), SCI + Bex + scrambled shRNA (n = 24), SCI + Bex + TFE3 shRNA (n = 24), and SCI + Bex + dorsomorphin (n = 12) groups. All animals in the SCI groups were subjected to mild SCI. To examine the effect of Bex on severe SCI, we established two additional groups: severe SCI (n = 6) and severe SCI + Bex (n = 6).

Animal model of SCI

Mice were intraperitoneally injected with pentobarbital sodium (1%, 50 mg/kg, Shanghai Fuzhe Chemicals Ltd., Shanghai, China) for anesthesia. The lamina from the T11–T12 vertebrae was removed and the dura circle was exposed. Mild or severe spinal cord contusion injuries were performed as previously described (Byrnes et al., 2007). In brief, a bar (15 g and Φ 3.0 mm) was dropped down to the exposed spinal cord with a 30 (mild injury) or 100 (severe injury) kDyne force using a MASCIS impactor model III (W.M. Keck, HI, USA). After the injury, the muscle, fascia, and skin were sutured. Mice in the sham group were subjected to the same operation without SCI. Bladders were manually expressed three times per day after the operation.

Drugs and AAV vector administration

The experimental overview of the study is shown in Figure 1. The SCI + Bex group was treated by intraperitoneal injection of Bex (Cat# 153559-49-0, MedChemExpress, Monmouth Junction, NJ, USA; 1 mg/kg, 50 mg Bex dissolved in 250 mL dimethyl sulfoxide and 2250 mL normal saline) every day for 5 days after SCI. The same volume (1 mL) of normal saline was intraperitoneally injected into the mice in the sham and SCI groups. In the SCI + Bex + 3MA group and SCI + Bex + dorsomorphin group, 3-methyladenine (3MA; Cat# M9281, Sigma-Aldrich, St. Louis, MO, USA; 15 mg/kg, 300 mg 3MA dissolved in 1000 mL phosphate-buffered saline) or dorsomorphin (compound C; Cat# 703866405-64-3, MedChemExpress, Monmouth Junction, NJ, USA; 1.5 mg/kg, 60 mg dorsomorphin dissolved in 2000 mL normal saline) was intraperitoneally injected daily 30 minutes after Bex administration. In the SCI + Bex + scrambled shRNA and SCI + Bex + TFE3 shRNA groups, at 14 days before operation, animals were intravenously injected with 100 µL of phosphate-buffered saline containing 1 × 10¹⁰ packaged genome particles (AAV-TFE3 shRNA and scrambled shRNA, Shanghai GeneChem Chemical Technologies, Shanghai, China). After SCI, both groups received Bex as described above. All mice were sacrificed with an overdose of pentobarbital sodium 5 or 28 days after SCI.

Figure 1.

Flow chart of animal study.

Mild or severe spinal cord injury (SCI) was induced in mice. After operation, animals were administered the indicated interventions (groups 1–4, 7–8). In some experiments, animals were treated pre-operatively with control or TFE shRNA (groups 5 and 6). Animals were sacrificed on days 5 and 28 after injury for western blot, HE staining, Masson staining, immunofluorescence staining, ELISA, qPCR, and functional behavior assessments. 3MA: 3-Methyladenine; Bex: bexarotene; BMS: Basso mouse scale; CC: compound C; ELISA: enzyme-linked immunosorbent assay; HE: hematoxylin and eosin; NS: normal saline; qPCR: real-time polymerase chain reaction; ROS: reactive oxygen species; SCI: spinal cord injury; shRNA: short hairpin RNA; TFE3: transcription factor E3.

Functional behavior assessments

To measure locomotion and evaluate functional recovery, we performed the Basso mouse scale (BMS) score test on days 0, 1, 7, 13, 21, and 28 after SCI following a previous protocol (Basso et al., 2006). The total BMS score is 9, and a higher BMS score indicates better motor function. We also performed gait analysis. The footprints of the hind legs in mice were imaged 28 days after SCI (Wu et al., 2021). The stride length of the footprint was measured.

Hematoxylin-eosin and Masson staining

On day 28 after SCI, mice were sacrificed with an overdose of pentobarbital sodium (500 mg/kg) and perfused with normal saline. The spines were collected and soaked in 4% paraformaldehyde. After 24 hours, the spinal cord was stripped and trimmed to a 1-cm-long section centered on the SCI. After graded dehydration with ethanol, the spinal cords were embedded in paraffin and longitudinally sliced into paraffin sections (4 µm thickness) for hematoxylin-eosin (HE) staining and Masson staining following a previous protocol (Parra-Villamar et al., 2021). The HE staining kit (Cat# G1120) and Masson’s staining kit (Cat# G1346) were acquired from Solarbio Science & Technologies (Beijing, China). The proportion of the collagen deposition area in Masson-stained images of the spinal cord cross-section (at magnification 20×) was quantified by ImageJ ver.1.53a.

Western blot analysis, immunoprecipitation, and cell fractionation

Spinal cords (1.5 cm in length) were collected from mice on day 5 after SCI. Samples were lysed in radio immunoprecipitation assay buffer (Solarbio Science & Technologies, Beijing, China). Cytoplasmic and nuclear proteins were extracted using nuclear and cytoplasmic abstraction reagents (Cat# 78833, Thermo Fisher Scientific, Rockford, IL, USA). The protein concentrations were determined with a bicinchoninic acid assay protein analysis kit (Cat# 71023227, Thermo Fisher Scientific). Samples were separated by 12.5% gel electrophoresis, electrotransferred onto polyvinylidene fluoride membranes (Servicebio, Wuhan, China), and blocked in Western Blocking Buffer (Solarbio Science & Technologies). The membranes were probed with the following antibodies at 4°C overnight: rabbit polyclonal antibodies against adenosine 5′-monophosphate-activated protein kinase (AMPKα, 1:1000, Cell Signaling Technology, Beverly, MA, USA, Cat# 5832, RRID: AB_10624867), rabbit polyclonal antibodies against phospho-adenosine 5′-monophosphate-activated protein kinase (p-AMPKα, 1:1000, Cell Signaling Technology, Cat# 2531, RRID: AB_330330), rabbit polyclonal antibodies against Forkhead box O3 (FOXO3A, 1:1000, Affinity Biosciences, Cincinnati, OH, USA, Cat# AF6020, RRID: AB_2834954), rabbit polyclonal antibodies against phospho-FOXO3A (Ser253) (1:1000, Affinity Biosciences, Cat# AF3020, RRID: AB_2834427), rabbit polyclonal antibodies against mammalian target of rapamycin (mTOR, 1:1000, Cell Signaling Technology, Cat# 2983, RRID: AB_210562), rabbit polyclonal antibodies against p-mTOR (1:1000, Cell Signaling Technology, Cat# 5536, RRID: AB_10691552), rabbit polyclonal antibodies against S-phase kinase-associated protein 2 (SKP2/p45, 1:1000, Affinity Biosciences, Cat# AF0505, RRID: AB_2834158), rabbit polyclonal antibodies against coactivator-associated arginine methyltransferase 1 (CARM1, 1:1000, Affinity Biosciences, Cat# BF0658, RRID:AB_2833878), rabbit polyclonal antibodies against TFE3 (1:1000, Sigma-Aldrich, Cat# HPA023881, RRID: AB_1857931), rabbit polyclonal antibodies against gasdermin D–N (GSDMD-N, 1:1000, Affinity Biosciences, Cat# AF4013, RRID:AB_2846780), rabbit polyclonal antibodies against NLRP3 (1:1000, Cell Signaling Technology, Cat# 15101, RRID: AB_2722591), rabbit polyclonal antibodies against cleaved Caspase 1 (1:1000, Protein Technology Group, Rosemont, IL, USA, Cat# 22915-1-AP, RRID: AB_2876874), rabbit polyclonal antibodies against cleaved IL-1β (1:1000, ABclonal Technology, Woburn, MA, USA, Cat# A11369, RRID: AB_2758528), rabbit polyclonal antibodies against cleaved IL-18 (1:1000, Affinity Biosciences, Cat# DF6252, RRID:AB_2838218), rabbit polyclonal antibodies against apoptosis-associated speck-like protein containing a CARD (TMS1/ASC, 1:1000, Cell Signaling Technology, Cat# 67824, RRID: AB_2799736), rabbit polyclonal antibodies against BCL2/adenovirus E1B 19 kDa interacting protein 3 (BNIP3, 1:1000, Affinity Biosciences, Cat# DF8188, RRID: AB_2841498), rabbit polyclonal antibodies against BCL2/adenovirus E1B 19 kDa interacting protein 3 like (BNIP3L, 1:1000, Affinity Biosciences, Cat# DF8163, RRID: AB_2841486), rabbit polyclonal antibodies against Parkin (1:1000, Affinity Biosciences, Cat# AF0235, RRID: AB_2833410), rabbit polyclonal antibodies against vacuolar protein sorting 34 (VPS34, 1:1000, Protein Technology Group, Cat# 12452-1-AP, RRID: AB_2299709), rabbit polyclonal antibodies against Beclin1 (1:1000, Cell Signaling Technology, Cat# 3738, RRID: AB_490837), rabbit polyclonal antibodies against light chain 3 (LC3B, 1:500, Cell Signaling Technology, CAT# 3868, RRID: AB_2137707), mouse monoclonal (1B7) antibodies against p62/sequestosome 1 (p62; 1:1000, Abcam, Cambridge, UK, Cat# ab56416, RRID: AB_945626), rabbit polyclonal antibodies against Cathepsin D (CTSD, 1:1000, Protein Technology Group, Cat# 21327-1-AP, RRID: AB_10733646), rabbit polyclonal antibodies against histone protein H3 (H3, 1:1000, Protein Technology Group, Cat# 17168-1-AP, RRID: AB_2716755) and rabbit polyclonal antibodies against glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 1:1000, Protein Technology Group, Cat# 10494-1-AP, RRID: AB_2263076). The membranes were then incubated with secondary antibodies (1:3000) at room temperature for 120 minutes. Bands were visualized using the enhanced chemiluminescence Plus Reagent Kit (Beyotime Biotechnology, Beijing, China); the bands were photographed and examined with Image Lab 3.0 software (Bio-Rad, Hercules, CA, USA).

For immunoprecipitation, the samples were homogenized in lysis buffer and lysates were clarified by centrifugation for 20 minutes in a microcentrifuge at 4°C. After preclearing with 20 mL Protein A Sepharose beads for 1 hour at 4°C, tissue lysates were incubated with 5 µL anti-TFE3 antibody and rotated for 12 hours at 4°C. Protein A Sepharose beads (20 µL) were added to each sample and the samples were incubated for 2 hours at 4°C. The immune complex was washed three times with lysis buffer, heated in sample buffer, separated by SDS-PAGE, and immunoblotted on polyvinylidene fluoride membranes; the membranes were incubated for 12 hours with anti-TFE3 antibodies. Immunocomplexes were detected using an enhanced chemiluminescence system as described above.

Dihydroethidium staining

Spinal cords were collected from mice 5 days after SCI. The spinal cord tissues were gradient-dehydrated in 5%, 10%, 20%, and 30% sucrose solutions. Tissues were covered with optimal cutting temperature compound (SAKURA Bio, Tokyo, Japan) in a –20°C environment until the optimal cutting temperature compound had completely solidified. The frozen tissue was cut into 5 µm sections. The frozen sections were thawed and treated with an autofluorescence quencher for 10 minutes. Next, dihydroethidium (DHE) working solution was dripped onto the spinal cord tissue and the sections were incubated at 37°C for 1 hour. DHE (C21H21N3, Cat# 50102ES02) was purchased from Yeasen Biotechnology (Shanghai) Co., Ltd. (Shanghai, China). Sections were washed with phosphate-buffered saline for 5 minutes three times and imaged using a fluorescence microscope (Olympus, Tokyo, Japan).

Immunofluorescence staining

Spinal cord from mice on day 5 after SCI was paraffin-embedded and dissected into cross-sections. The cross-sections were deparaffinized, rehydrated, cleaned, and exposed to sodium citrate buffer at 95°C for 20 minutes. The slides were blocked with 10% bovine serum albumin (Boyun Biotech, Shanghai, China) for 1 hour and incubated overnight at 4°C with mouse monoclonal (1B7) antibodies against NeuN (1:10,000, Abcam, Cat# ab104224, RRID: AB_882786), GSDMD-N (1:200), cleaved Caspase-1 (1:200), LC3B (1:200), and p62 (1:200), rabbit monoclonal (EP1098Y) antibodies against synaptophysin (SYN, 1:200, Abcam, Cat# ab52636, RRID: AB_10711040), and rabbit monoclonal (EPR19691) antibodies against microtubule-associated protein 2 (MAP2, 1:200, Abcam, Cat# ab183830, RRID:AB_2895301) and TFE3 (1:200). Samples were washed three times for 15 minutes each at room temperature and incubated with the following secondary antibodies (1:200) at room temperature for 60 minutes: goat anti-rabbit IgG H&L (Alexa Fluor® 488) (Cat# ab150077, RRID: AB_2630356), goat polyclonal anti-rabbit IgG-H&L (Alexa Fluor® 647) (Cat# ab150079, RRID: AB_2722623), goat anti-mouse IgG H&L (Alexa Fluor® 488) (Cat# ab150113, RRID: AB_2576208) and goat anti-mouse IgG H&L (Alexa Fluor® 647) (Cat# ab150115, RRID: AB_2687948); all secondary antibodies were acquired from Abcam. Sections were counterstained in 4′,6-diamidino-2-phenylindole (Cat# 36308ES11, Yeasen Biotechnology, Shanghai, China) after three washes. Images of the ventral spinal cord from transverse sections were obtained and assessed under a fluorescence microscope in six fields taken at random in three random sections for each sample. Optical densities of expressions were quantified from the immunofluorescence images using ImageJ.

We used the following formula to calculate TFE3 nuclear translocation: percentage of TFE3 translocation into nucleus = number of cells with TFE3-positive nucleus/total number of TFE-3-positive cells. In other experiments, the number of synapses in contact with each neuron and the proportion of LC3B-positive neurons in immunofluorescence images of the ventral spinal cord were counted by three researchers, and the median values were calculated.

Enzyme-linked immunosorbent assay and malondialdehyde assay

Spinal cord from mice 5 days after SCI was homogenized and then frozen and thawed repeatedly in liquid nitrogen. The homogenate was centrifuged (10,000 × g, 4°C, 10 minutes), and the supernatant was extracted for subsequent analysis. Enzyme-linked immunosorbent assay (ELISA) was used to determine the levels of 8-hydroxy-2 deoxyguanosine (8-OHdG), advanced oxidation protein products (AOPP), cleaved Caspase-1, GSDMD-N, IL-1β, and IL-18 (all ELISA kits were purchased from Boyun Biotech). Thiobarbituric acid assays were used to determine malondialdehyde assay (MDA) levels (Boyun Biotech). A microplate (Beyotime Biotechnology) (calibration wavelength of 450 nm) was used to measure the optical density at 550 nm to quantify levels of 8-OHdG, AOPP, cleaved Caspase-1, GSDMD-N, IL-1β and IL-18 and 535 nm to quantify MDA levels.

Real-time polymerase chain reaction

Spinal cords from mice on day 5 after SCI were treated with TRIzol reagent (Boyun Biotech) to isolate RNA. Complementary DNA (cDNA) was synthesized with reverse using the Prime Script II 1^st Strand cDNA Synthetic Kit (6210B, Takara Bio Inc., Tokyo, Japan). Primers were developed with Primer Premier 5.0 (PREMIER Biosoft, San Francisco, CA, USA) in comparison with the mRNA sequences of VPS34, Beclin1, CTSD, SQSTM1, LC3, Bnip3, Bnip3l, and Prkn in GenBank (https://www.ncbi.nlm.nih.gov/genbank/); primers were synthesized by Nanjing Zoonbio Biological Technology Company (Nanjing, China). The primer sequences are shown in Table 1. Real-time PCR was performed with SYBR Premix Ex Taq (RR420A, Takara Bio Inc.). The reaction conditions were as follows: annealing at 65°C for 30 seconds, denaturation at 95°C for 30 seconds, and extension at 72°C for 45 seconds for 30 cycles. The signal was measured at 72°C. Beclin1, VPS34, CTSD, SQSTM1, LC3, Bnip3, Bnip3l, and Prkn mRNA levels were normalized to β-actin mRNA levels through the 2^–ΔΔCT method (Li et al., 2022).

Table 1.

Primer sequences

Primer	Sequences
VPS34	Forward: 5′-TAA CGT GGA GGC AGA TGG TT-3′
	Reverse: 5′-CAT GTG TCC TTG CCG ATG AG-3′
Beclin1	Forward: 5′-ATG GAG GGG TCT AAG GCG TC-3′
	Reverse: 5′-TGG GCT GTG GTA AGT AAT GGA-3′
SQSTM1	Forward: 5′-ACA ACC CGT GTT TCC TTT-3′
	Reverse: 5′-TGC CAC CTT TCA CTC ACT A-3′
CTSD	Forward: 5′-GGG CAT CCA GGT AGT TTT-3′
	Reverse: 5′-CGT CTT GCT GCT CAT TCT-3′
LC3	Forward: 5′-CTA CGC CTC CCA AGA AAC C-3′
	Reverse: 5′-AGA GCA ACC CGA ACA TGA CT-3′
Bnip3	Forward: 5′-GCT CCA AGA GTT CTC ACT GTG AC-3′
	Reverse: 5′-GTT TTT CTC GCC AAA GCT GTG GC-3′
Bnip3l	Forward: 5′-GCA TGA GGA AGA GTG GAG CCA T-3′
	Reverse: 5′-AAG GTG TGC TCA GTC GTT TTC CA-3′
Prkn	Forward: 5′- TCT TCC AGT GTA ACC ACC GTC-3′
	Reverse: 5′-GGC AGG GAG TAG CCA AGT T-3′
β-Actin	Forward: 5′-GGC TCC TAG CAC CAT GAA GA-3′
	Reverse: 5′-AGC TCA GTA ACA GTC CGC C-3′

Bnip3: BCL2/adenovirus E1B 19 kDa interacting protein 3; Bnip3l: BCL2/adenovirus E1B 19 kDa interacting protein 3 like; CTSD: cathepsin D; LC3: light chain 3; VPS34: vacuolar protein sorting 34.

Statistical analysis

No statistical methods were used to predetermine sample sizes; however, our sample sizes are similar to those reported in a previous publication (Lou et al., 2022). No animals or data points were excluded from the analysis. All assays were randomized and performed in a blinded manner. Statistical analyses were performed with SPSS version 26 (IBM Corp., Armonk, NY, USA). Data are expressed as the mean ± standard error of the mean (SEM). All data were normalized to the control to account for variation from unwanted sources. The significance of differences between groups was evaluated with one-way analysis of variance with the least significance difference post hoc test for groups with equal variances. The differences in survival between the groups were detected with the Kaplan-Meier analysis for survival curves. P < 0.05 indicated statistical significance.

Results

Bex ameliorates motor functional recovery after SCI

We established an SCI model in mice and examined the effects of Bex treatment on the functional recovery of animals after 28 days. We first examined the proportion of the collagen deposition area in the sagittal sections of the spinal cords by HE and Masson staining. While collagen deposition was significantly greater after SCI compared with sham operation, Bex treatment significantly reduced the proportion of the collagen deposition area (P < 0.01; Figure 2A and B). Immunofluorescence staining showed that MAP2 expression and the number of SYN-contacting ventral motor nerve cells were significantly greater in the SCI + Bex group compared with the SCI group after operation (P < 0.01; Figure 2C–E). Footprint analysis showed that while the stride length in the SCI group was significantly shorter than that in the sham group, the stride length in the SCI + Bex group was significantly increased compared with that of the SCI group (P < 0.01; Figure 2F and G). Analysis of BMS scores and survival curves revealed that mice in the SCI + Bex group recovered better lower limb motor function and had a lower death rate than those in the SCI group (Figure 2H and I).

Figure 2.

Bexarotene promotes functional recovery after spinal cord injury.

(A) Representative HE and Masson staining of spinal cord sections (sagittal) from mice in the sham, SCI, and SCI + Bex groups on day 28 after SCI. The dotted line indicates the collagen deposition area. (B) Quantification of the proportion of the collagen deposition area in sagittal spinal cord sections in A. (C) Representative images of immunofluorescence staining for MAP2 (indicating soma and dendrites of neurons, green), NeuN (indicating neurons, red), and SYN (indicating synapse of neurons, green) in spinal cord transverse sections. Scale bars: 1000 μm (A), 5 μm (C). (D, E) Quantification of the optical density of MAP2 in motor neurons and SYN-contacting synapse numbers in C. (F) Images of mouse footprints in gait analysis on day 28 after SCI. (G) Quantitative analysis of stride lengths in F. (H) Survival analysis of mice after SCI. (I) BMS scores at different time points after SCI. Data are expressed as the mean ± SEM (n = 6). **P < 0.01, vs. sham group; #P < 0.05, ##P < 0.01, vs. SCI group. One-way analysis of variance with the least significance difference post hoc test (B, D, E, G, I), or Kaplan-Meier analysis (H). Bex: Bexarotene; BMS: Basso Mouse Scale; DAPI: 4′,6-diamidino-2-phenylindole; HE: hematoxylin-eosin; IOD: integrated optical density; MAP2: microtubule-associated protein 2; SCI: spinal cord injury; SYN: synapse.

To further determine the effectiveness of Bex, we evaluated the effects of Bex treatment in the mouse model with severe SCI. There were no significant differences in BMS scores between the two groups in the first 28 days after operation; on day 28, the severe SCI + Bex group showed a significantly higher BMS score than the severe SCI group (P = 0.038; Additional Figure 1 ^{(174.1KB, tif)} ). Together these findings suggest that Bex ameliorates functional recovery after SCI.

Bex attenuates pyroptosis in the spinal cord after SCI

We examined the effects of Bex treatment on pyroptosis on the 5^th day after SCI. High levels of pyroptosis are observed in the spinal cord after SCI and are thought to suppress motor function recovery (Al Mamun et al., 2021). Therefore, we next explored whether Bex improved functional recovery in the spinal cord after SCI by attenuating pyroptosis. We examined the expressions of GSDMD-N and Caspase-1, two critical proteins involved in pyroptosis (Shi et al., 2017), by immunofluorescence staining (Figure 3A–D). The levels of GSDMD-N and cleaved Caspase-1 were significantly increased in the SCI group compared with the sham group; furthermore, the levels were markedly decreased in the SCI + Bex group compared with the SCI group (all P < 0.01, vs. sham group; Figure 3B and D). We next examined the levels of cleaved Caspase-1, GSDMD-N, IL-1β, and IL-18 (pyroptosis markers (Al Mamun et al., 2021)) in the spinal cord by ELISA. All levels were significantly suppressed after Bex administration (P < 0.01, vs. SCI group; Figure 3E–H), suggesting that pyroptosis was reduced. Western blot results for GSDMD-N, NLRP3, Caspase-1, IL-1β, IL-18, and ASC expression levels were consistent with the immunofluorescence staining and ELISA results (Figure 3I and J). The SCI + Bex group had markedly reduced levels of these markers compared with the SCI group. Together, these results indicated that Bex suppresses pyroptosis in the spinal cord after SCI.

Figure 3.

Bexarotene attenuates neuronal pyroptosis in the spinal cord after SCI.

(A) Immunofluorescence staining of Caspase 1 (pyroptosis-related marker, red) and NeuN (indicating neurons, green) (original magnification 30×) in spinal cord sections. The optical densities of Caspase 1 were markedly decreased in the SCI + Bex group and significantly increased in the SCI group. Scale bars: 25 μm. (B) Quantification of Caspase-1 in neurons of spinal cords in A. (C) Immunofluorescence staining of GSDMD-N (pyroptosis-related protein, green) and NeuN (indicating neurons, red) (original magnification 30×). Scale bars: 25 μm. (D) Quantification of GSDMD in neurons of spinal cords in C. (E–H) Evaluation of CASP1, GSDMD, IL-18, and IL-1β in the spinal cord by ELISA. (I) Western blot assay of GSDMD-N, NLRP3, Caspase-1, IL-1β, IL-18, and ASC. (J) Quantitative analyses of data from I; data were normalized to GAPDH. Data are expressed as the mean ± SEM (n = 6 mice per group). **P < 0.01, vs. sham group; ##P < 0.01, vs. SCI group (one-way analysis of variance with the least significance difference post hoc test). ASC: Apoptosis-associated speck-like protein containing a CARD; Bex: bexarotene; C-CASP-1: cleaved Caspase 1; DAPI: 4′,6-diamidino-2-phenylindole; ELISA: enzyme-linked immunosorbent assay; GAPDH: glyceraldehyde-3- phosphate dehydrogenase; GSDMD-N: gasdermin D-N; IL: interleukin; IOD: integrated optical density; NLRP3: NOD-like receptor thermal protein domain associated protein 3; SCI: spinal cord injury.

Bex enhances autophagy in the spinal cord after SCI

We next examined the effects of Bex treatment on autophagy on the 5^th day after SCI. We evaluated autophagy-related proteins and genes (p62, VPS34, Beclin1, CTSD, and LC3) in the spinal cord after SCI (Figure 4A–D). Immunofluorescence showed that the SCI group exhibited a greater proportion of LC3-positive cells than the sham group; after Bex treatment, the proportion of cells containing LC3-labelled autophagosomes was significantly greater than that in the SCI group or sham group (both P < 0.01). The levels of p62 in the SCI groups were significantly higher than that in the sham group (both P < 0.01); the level of p62 in the SCI + Bex group was markedly decreased compared with that in the SCI group (P < 0.01). Western blot analysis of spinal cord samples indicated that VPS34, p62, Beclin1, CTSD, and LC3B levels were significantly higher in the SCI group compared with the sham group; the levels of VPS34, Beclin1, CTSD, and LC3B levels were markedly increased and p62 was decreased in the SCI + Bex group compared with the SCI group (all P < 0.01; Figure 4E and F). The qPCR results indicated that the expressions of autophagy-related genes (VPS34, SQSTM1, Beclin1, CTSD and LC3) were higher in the SCI group compared with the sham group; the levels were higher in the SCI + Bex group than the SCI group (all P < 0.01; Figure 4G). These results demonstrate that autophagy was increased in the spinal cord after mild injury and Bex further upregulated autophagy in the SCI model.

Figure 4.

Bexarotene enhances autophagy in neurons following spinal cord injury.

(A, B) Immunofluorescence staining of LC3 (autophagy-related marker, green), NeuN (indicating neurons, red or green), and p62 (substrate of autophagy, red) (original magnification 30×). Scale bars: 25 μm. (C, D) Quantification of p62 and the percentage of LC3II-positive cells in motor neurons of spinal cord in A and B. (E) Western blot assay of VPS34, p62, Beclin1, CTSD, LAMP2, ATG5, and LC3-II in the different groups. (F) Quantitative analyses of data from E normalized to GAPDH. (G) The levels of autophagy-related genes in the spinal cord were determined by qPCR and normalized to those of β-actin. Data are expressed as the mean ± SEM (n = 6 mice per group). **P < 0.01, vs. sham group; ##P < 0.01, vs. SCI group (one-way analysis of variance with the least significance difference post hoc test). Bex: Bexarotene; CTSD: Cathepsin D; DAPI: 4′,6-diamidino-2-phenylindole; GAPDH: glyceraldehyde- 3-phosphate dehydrogenase; LC3B: light chain 3 B; qPCR: real-time polymerase chain reaction; SCI: spinal cord injury; VPS34: vacuolar protein sorting 34.

Suppression of autophagy reverses the effect of Bex on pyroptosis in the spinal cord after SCI

We next investigated whether Bex suppresses pyroptosis after SCI through enhancing autophagy. We administered 3MA, an autophagy inhibitor (Yang et al., 2018), to SCI model mice treated with Bex. We confirmed that the effect of Bex on enhancing autophagy was suppressed upon 3MA treatment, as shown by immunofluorescence staining for LC3 and p62 and western blot analysis of VPS34, p62, Beclin1, CTSD, and LC3II (Figure 5A–D). The SCI + Bex + 3MA group showed a significant decrease in the proportion of LC3II-positive cells and an increase in the level of p62 compared with the SCI + Bex group, as determined by immunofluorescence staining (all P < 0.01; Figure 5A–D). We next measured pyroptosis markers in the SCI + Bex and SCI + Bex + 3MA groups. Immunofluorescence staining showed that the immunopositivities of GSDMD-N and Caspase-1 were markedly increased in the SCI + Bex + 3MA group compared with levels in the SCI + Bex group (P < 0.01; Figure 5E–H). In western blot analysis: the SCI + Bex + 3MA group showed the same trend in terms of the levels for VPS34, p62, Beclin1, CTSD, and LC3II compared with levels in the SCI + Bex group (P < 0.01; Figure 5I and J). The SCI + Bex + 3MA group showed increased protein expression levels for GSDMD-N, NLRP3, Caspase-1, IL-1β, IL-18 and ASC compared with levels in the SCI + Bex group (P < 0.01; Figure 5K and L). These results indicate that 3MA inhibited the effect of Bex on reducing pyroptosis.

Figure 5.

Suppression of autophagy reverses the effect of Bexarotene on pyroptosis after SCI.

(A–H) Immunofluorescence staining of LC3 (autophagy-related marker, green; A, B), p62 (substrate of autophagy, red; C, D), C-CASP-1 (pyroptosis-related protein, red; E, F), and GSDMD (pyroptosis-related marker, green; G, H) in neurons (NeuN (indicating neurons, red or green)) (original magnification 30×). Scale bars: 20 μm. Quantitative analysis of the levels of GSDMD, Caspase-1, and p62 and the percentage of LC3-positive neurons shown in B, D, F, and H. (I–L) Western blot assay of pyroptosis-related proteins and autophagy-related proteins. Data in J and L were normalized to GAPDH. Data are expressed as the mean ± SEM (n = 6). **P < 0.01, vs. SCI + Bex group (one-way analysis of variance with the least significance difference post hoc test). ASC: Apoptosis-associated speck-like protein containing a CARD; Bex: bexarotene; C-CASP-1: cleaved Caspase 1; CTSD: Cathepsin D; DAPI: 4′,6-diamidino-2-phenylindole; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GSDMD-N: gasdermin D-N; IL: interleukin; IOD: integrated optical density; LC3B: light chain 3 B; NLRP3: NOD-like receptor thermal protein domain associated protein 3; SCI: spinal cord injury; VPS34: vacuolar protein sorting 34.

Bex reduces the levels of ROS by activating mitophagy in the spinal cord after SCI

We next explored other mechanisms by which Bex suppresses pyroptosis after SCI. After SCI, ROS are upregulated in the injury area; these proinflammatory factors mediate inflammatory mechanisms, such as pyroptosis-mediated cell death (David and Kroner, 2011; Shi et al., 2015). We next measured the levels of ROS by immunofluorescence staining and ELISA. The results showed that Bex significantly attenuated ROS accumulation after SCI, along with pyroptosis, which was reversed by 3MA (all P < 0.01, Figure 6A–D; P < 0.01 in SCI group vs. SCI + Bex group, P < 0.05 in SCI + Bex group vs. SCI + Bex + 3MA group, Figure 6E). ELISA measurement of the levels of 8-OHdG, AOPP, and MDA also indicated that ROS levels were markedly decreased in the SCI + Bex group compared with the SCI group (Figure 6F–H). Western blot assay further confirmed that Bex effectively ameliorates oxidative stress after SCI (Figure 6I and J).

Figure 6.

Bexarotene promotes mitophagy and decreases ROS levels after SCI.

(A) Immunofluorescence staining for GSDMD (pyroptosis-related marker, green), C-CASP1 (pyroptosis-related marker, red), NIX (mitophagy-related marker, red) and DHE (indicating ROS-positive cells, red) in neurons in the spinal cord (original magnification 30×). Scale bar: 25 μm. (B–E) Quantitative analysis of levels of GSDMD (B), C-CASP1 (C), NIX (D) and DHE (E) in A. (F–H) The levels of 8-OHdG and AOPP in the spinal cord were detected by ELISA, and the levels of MDA were detected by the thiobarbituric acid assay. (I, L) Western blot assay for pyroptosis-related and mitophagy-related proteins. Data were normalized to GAPDH. (M) The levels of mitophagy-related genes in the spinal cord were detected by qPCR and normalized to β-actin. Data are expressed as the mean ± SEM (n = 6 mice per group). *P < 0.05 and **P < 0.01, vs. SCI group; #P < 0.05 and ##P < 0.01, vs. SCI + Bex group (one-way analysis of variance with the least significance difference post hoc test). ASC: Apoptosis-associated speck-like protein containing a CARD; Bex: bexarotene; BNIP3: BCL2/adenovirus E1B 19 kDa interacting protein 3; C-CASP-1: cleaved Caspase 1; DAPI: 4′,6-diamidino-2-phenylindole; DHE: dihydroethidium; ELISA: enzyme-linked immunosorbent assay; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GSDMD-N: gasdermin D-N; IOD: integrated optical density; MDA: malondialdehyde; NIX/BNIP3L: BCL2/adenovirus E1B 19 kDa interacting protein 3 like; NLRP3: NOD-like receptor thermal protein domain associated protein 3; SCI: spinal cord injury.

We next examined the effect of Bex on mitophagy by the evaluation of mitophagy-related protein levels. As shown in Figure 6K–L, the levels of BNIP3, BNIP3 L, and Parkin (markers of mitophagy) were significantly higher in the SCI + Bex group than those in the SCI group, while they were significantly lower in the SCI + Bex + 3MA group compared with the SCI + Bex group. Significantly higher levels of pyroptosis-related markers (GSDMD-N, NLRP3, Caspase-1, IL-1β, IL-18 and ASC) were observed in the SCI and SCI + Bex + 3MA groups compared with the SCI + Bex group (Figure 6I). qPCR results also demonstrated that mitophagy-related gene levels were significantly increased in the SCI + Bex group compared with the SCI group and SCI + Bex + 3MA group (P < 0.01; Figure 6M). These results indicate that Bex may reduce the levels of ROS by activating mitophagy.

Inhibition of autophagy and mitophagy in the spinal cord reverses the effects of Bex on SCI

We further explored the effect of Bex on autophagy in the functional recovery of the spinal cord after SCI. The levels of MAP2 and number of SYN-positive synapses on ventral motor nerve cells were lower in the SCI + Bex + 3MA group compared with those in the SCI + Bex group (both P < 0.01; Figure 7A–D). The proportion of the collagen deposition area in the impaired spinal cord was markedly increased in the SCI + Bex + 3MA group compared with that in the SCI + Bex group (P < 0.01; Figure 7E and F). Footprint analysis showed a significantly shorter stride length in the SCI + Bex + 3MA group than that in the SCI + Bex group (P < 0.01; Figure 7G and H). The SCI + Bex group showed higher BMS scores and survival curves compared with those in the SCI + Bex + 3MA group (Figure 7I and J). These results demonstrated that autophagy and mitophagy play vital roles in SCI and that Bex improved functional recovery after SCI by influencing autophagy and mitophagy.

Figure 7.

Bexarotene promotes functional recovery following SCI via enhanced autophagy.

(A, B) Immunofluorescence staining for MAP2 (indicating soma and dendrites of neurons, green), NeuN (indicating neurons, red), and SYN (indicating synapse of neurons, green) in spinal cord transverse sections on day 28 after SCI (original magnification 30×). (C) Quantitative analysis of the level of MAP2 in A. (D) Quantitative analysis of the number of motor neuron–contacting synapses (synapses in contact with each neuron) in B. (E, F) Representative HE and Masson staining of sagittal spinal cord sections. The dotted line indicates the collagen deposition area. Scale bars: 25 μm (A, B) and 1000 μm (E). The proportion of the collagen deposition area in the sagittal spinal cord sections was quantitatively analyzed. (G, H) Images and quantitative analysis of mouse footprints on day 28 after spinal cord injury. (I) Survival analysis after SCI. (J) Basso mouse scale (BMS) scores at different time points after SCI. Data are expressed as the mean ± SEM (n = 12). *P < 0.05, **P < 0.01, vs. SCI + Bex group (one-way analysis of variance with the least significance difference post hoc test (C, D, F, H, J) or Kaplan-Meier analysis (I)). Bex: Bexarotene; BMS: Basso Mouse Scale; DAPI: 4′,6-diamidino-2-phenylindole; HE: hematoxylin-eosin; IOD: integrated optical density; MAP2: microtubule-associated protein 2; SCI: spinal cord injury; SYN: synapse.

Bex facilitates autophagy and mitophagy in the spinal cord by upregulating TFE3 and subsequently suppresses pyroptosis and oxidative stress in SCI

We next examined the mechanism by which Bex facilitates autophagy and mitophagy. We previously found that the TFE3 transcription factor promotes autophagy and mitophagy (Zhou et al., 2020; Lou et al., 2022). To determine whether Bex impacts autophagy and mitophagy in SCI through TFE3, we first examined TFE3 expression. Immunofluorescence staining and western blot showed that TFE3 was increased after SCI, and the increase was significantly increased in the SCI + Bex group (both P < 0.01; Figure 8). Additionally, the proportion of TFE3 that translocated into the nucleus was significantly increased in the SCI + Bex group compared with the sham and SCI groups (P < 0.01; Figure 8A–D).

Figure 8.

Bexarotene promotes the expression and nuclear translocation of TFE3 following SCI.

(A) Western blot assay of cytoplasmic and nuclear TFE3 expression. (B) Quantitative analysis of data from A normalized to GAPDH or H3. (C) Immunofluorescence staining for TFE3 (green) and NeuN (indicating neurons, red) in the spinal cord (original magnification 30×). The proportion of TFE3 that translocated into the nucleus was increased in the SCI + Bex group. Scale bar: 25 μm. (D) Quantitative analysis of the percentage of TFE3 translocated into the nucleus (top) and the total FE3 (bottom) in neurons in C. Data are expressed as the mean ± SEM (n = 6). **P < 0.01, vs. SCI + Bex group; ##P < 0.01, vs. SCI + Bex + 3MA group (one-way analysis of variance with the least significance difference post hoc test). Bex: Bexarotene; DAPI: 4′,6-diamidino-2-phenylindole; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; H3: histone protein H3; SCI: spinal cord injury; TFE3: transcription factor E3.

To further explore the role of TFE3 in autophagy and mitophagy induced by Bex, we used TFE3 shRNA to silence TFE3 expression. Immunofluorescence results showed that the percentage of LC3II-positive cells was markedly reduced in the Bex + TFE3 shRNA group compared with the SCI + Bex + scrambled shRNA group; there was no significant difference between the SCI + Bex and SCI + Bex + scrambled shRNA group. The level of p62 was significantly higher in the Bex + TFE3 shRNA group compared with the SCI + Bex + scrambled shRNA group; no significant difference was observed between the SCI + Bex group and SCI + Bex + scrambled shRNA group (both P < 0.01; Figure 9A–J). Western blot assay revealed decreases in VPS34, Beclin-1, CTSD, and LC3II and an increase in p62 in the SCI + Bex + TFE3 shRNA group compared with the SCI + Bex group and SCI + Bex + scrambled shRNA group; no significant differences were observed between the SCI + Bex group and SCI + Bex + scrambled shRNA group (Figure 9K–L). Immunofluorescence staining for CASP-1 and GSDMD indicated that these pyroptosis-related markers, which were expressed at relatively lower levels in the SCI + Bex and SCI + Bex + scrambled shRNA groups compared with the SCI group, were significantly increased in the SCI + Bex + TFE3 shRNA group; there was no significant difference between SCI + Bex and SCI + Bex + scrambled shRNA groups. The levels of GSDMD-N, NLRP3, Caspase-1, IL-1β, IL-18, and ASC were comparable in the three groups; levels in the SCI + Bex + TFE3 shRNA group were higher than those in the SCI + Bex group and SCI + Bex + scrambled shRNA groups (P < 0.01; Figure 9M and N). Immunofluorescence staining and western blotting of mitophagy-related proteins were consistent with the change in autophagy-related markers. All levels were significantly decreased after treatment with TFE3 shRNA compared with levels in SCI + Bex and SCI + Bex + scrambled shRNA groups (all P < 0.01, Figure 9A, G, O and P).

Figure 9.

Bexarotene inhibits pyroptosis and reduces the levels of ROS by enhancing autophagy and mitophagy by upregulating TFE3 in SCI.

(A) Immunofluorescence staining for TFE3 (green), NeuN (indicating neurons, red or green), LC3 (autophagy-related marker, green), p62 (substrate of autophagy, red), GSDMD-N (pyroptosis-related marker, green), Caspase-1 (pyroptosis-related marker, red), NIX (mitophagy-related marker, red), and DHE (indicating ROS-positive cells, red) in spinal cord sections (original magnification 30×). Scale bar: 25 μm. (B–H) Quantitative analysis of data from A. (I) Western blot assay of cytoplasmic and nuclear TFE3. (J) Quantitative analysis of nuclear and cytoplasmic TFE3 levels normalized to GAPDH and H3, respectively. (K–P) Western blot of autophagy-, pyroptosis- and mitophagy-related proteins, which were normalized to GAPDH. (Q–S) Levels of 8-OhDG, AOPP and MDA in the spinal cord. Data are expressed as the mean ± SEM (n = 6). **P < 0.01, vs. SCI + Bex group; ##P < 0.01, vs. SCI + Bex + scrambled shRNA group (one-way analysis of variance with the least significance difference post hoc test). ASC: Apoptosis-associated speck-like protein containing a CARD; Bex: bexarotene; BNIP3: BCL2/adenovirus E1B 19 kDa interacting protein 3; C-CASP-1: cleaved Caspase 1; CTSD: Cathepsin D; DAPI: 4′,6-diamidino-2-phenylindole; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GSDMD-N: gasdermin D-N; H3: histone protein H3; IOD: integrated optical density; LC3B: light chain 3 B; NIX/BNIP3L: BCL2/adenovirus E1B 19 kDa interacting protein 3 like; NLRP3: NOD-like receptor thermal protein domain associated protein 3; SCI: spinal cord injury; shRNA: short hairpin RNA; TFE3: transcription factor E3; VPS34: vacuolar protein sorting 34.

ROS was detected by immunofluorescent staining with DHE staining. The results showed that ROS accumulation was significantly increased in the SCI + Bex + TFE3 shRNA group compared with the SCI + Bex + scrambled shRNA group; no significant differences were observed between the SCI + Bex group and the SCI + Bex + scrambled shRNA group (all P < 0.01, vs. SCI + Bex and SCI + Bex + scrambled shRNA groups; Figure 9A). ELISA for 8-OHdG, AOPP, and MDA levels showed the same trend as the immunofluorescent staining results (all P < 0.01; Figure 9Q–S). These results indicated that TFE3 is required for Bex facilitation of mitophagy-mediated reduction of ROS accumulation.

We then examined the requirement for TFE3 in the effects of Bex on inducing recovery after SCI by HE and immunofluorescence staining. The proportions of the collagen deposition area in the sagittal sections of the spinal cords were markedly elevated in the SCI + Bex + TFE3 shRNA group compared with amounts in the SCI + Bex and SCI + Bex + scrambled shRNA groups (Additional Figure 2A ^{(981.3KB, tif)} and B ^{(981.3KB, tif)} ). TFE3 shRNA treatment also resulted in a significantly shortened stride length compared with Bex treatment alone, indicating that the recovery of motor function was reduced (Additional Figure 3A ^{(708.5KB, tif)} and B ^{(708.5KB, tif)} ). Immunofluorescence showed that the expression of MAP2 and the number of SYN-positive synapses on ventral motor nerve cells were significantly lower in the SCI + Bex + TFE3 shRNA group compared with levels in the SCI + Bex and SCI + Bex + scrambled shRNA groups (Additional Figure 2C ^{(981.3KB, tif)} –E ^{(981.3KB, tif)} ). Additionally, the BMS score after SCI and survival curves were reduced in the group with TFE3 silencing (Additional Figure 3C ^{(708.5KB, tif)} and D ^{(708.5KB, tif)} ).

Together, these results indicate that Bex increases the expression and nuclear translocation of TFE3, which leads to enhanced autophagy and mitophagy, reduced pyroptosis, and inhibition of oxidative stress. Furthermore, the functional recovery after SCI induced by Bex is facilitated by TFE3-induced mechanisms.

Bex activates TFE3 through the AMPK-mTOR and AMPK-SKP2-CARM1 signaling pathways in the spinal cord after SCI

Finally, we investigated the mechanism by which Bex activates TFE3. A previous study demonstrated that the AMPK-mTOR pathway increased the expression CARM1, which binds TFEB and promotes TFEB transcription in the nucleus (Shin et al., 2016). Notably, TFE3 and TFEB, which belong to the MiT/TFE family, share signal transduction networks (Miller et al., 2005). Therefore, we considered that TFE3 is also regulated by the AMPK pathway. To examine whether Bex activates TFE3 by the AMPK pathway, we performed western blot of cytoplasmic and nuclear fractions of spinal cord tissue samples to analyze the level of markers in the AMPK pathway. AMPK, p-AMPK, mTOR, and p-mTOR levels were detected in cytoplasmic fractions. Bex significantly increased the phosphorylation of AMPK and decreased the phosphorylation of mTOR compared with levels in the SCI group; compound C (an AMPK blocker) reversed the effects of Bex on p-AMPK and p-mTOR. The expression of AMPK and mTOR in the cytoplasm was not significantly different among the three groups (Figure 10A and B). These results indicated that Bex stimulates the AMPK-mTOR signaling pathway in the cytoplasm.

Figure 10.

Bexarotene activates TFE3 through the AMPK-mTOR and AMPK-SPK2- CARM1 signaling pathways in SCI.

(A) Western blot assay of the cytoplasmic expression levels of AMPK, p-AMPK, mTOR, and p-mTOR. (B) Quantification of AMPK, p-AMPK, mTOR, and p-mTOR from A, normalized to GADPH. (C) Western blot assay of the nuclear expression levels of AMPK, p-AMPK, FOXO3a, p-FOXO3a, SKP2, and CARM1. (D) Quantification of AMPK, p-AMPK, FOXO3a, p-FOXO3a, SKP2, and CARM1 from C, normalized to H3. (E, F) Immunoprecipitation images showing nuclear CARM1–TFE3 binding. Data are expressed as the mean ± SEM (n = 6 per group). **P < 0.01, vs. SCI group; ##P < 0.01, vs. SCI + Bex group. One-way analysis of variance with least significance difference post hoc test. (p-)AMPK: (phospho-) adenosine 5′-monophosphate-activated protein kinase; (p-)FOXO3a: (phospho-)forkhead box O3; (p-)mTOR: (phospho-)mammalian target of rapamycin; Bex: bexarotene; CARM1: coactivator-associated arginine methyltransferase 1; CC: compound C; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; H3: histone protein H3; SCI: spinal cord injury; SKP2: S-phase kinase-associated protein 2; TFE3: transcription factor E3.

In the signal pathway by which AMPK activates TFEB, AMPK-mediated phosphorylation of FOXO3a reduces the transcription of SKP2 and increases the protein expression of CARM1, which binds TFEB (Shin et al., 2016). We thus next analyzed the level of proteins in AMPK-SKP2-CARM1 signaling pathways in nuclear fractions. Higher levels of nuclear p-AMPK, p-FOXO3a, and CARM1 were observed in the Bex group than the SCI group or SCI + Bex + compound C group. SKP2 expression was reduced in the SCI + Bex group compared with SCI and SCI + Bex + compound C groups. The expression of AMPK and FOXO3a in the cytoplasm was not significantly different among the three groups (Figure 10C and D). Immunoprecipitation revealed increased binding of TFE3 to CARM1 in the Bex treatment group compared with SCI and SCI + Bex + compound C groups (Figure 10E and F). These results demonstrated that Bex activates the AMPK-SKP2-CARM1 signaling pathway.

Discussion

Here we provide new evidence that Bex, a synthetic product that structurally resembles retinoic acid compounds, improves spinal cord survival and recovery after SCI. We confirmed that Bex stimulates the AMPK-mTOR pathway in the cytoplasm and the AMPK-SKP2-CARM1 signaling pathway in the nucleus, promotes TFE3 translocation to the nucleus, activates autophagy and mitophagy, suppresses ROS production to alleviate oxidative stress, inhibits pyroptosis, and ultimately improves spinal cord rehabilitation after SCI. Our results in a severe SCI model demonstrated the functional improvement effectiveness of Bex in severe SCI. Functional improvement was observed at a late time point (28 days after operation) in the severe SCI model. Therefore, we used the mild SCI model in our subsequent research to more clearly show the effect of Bex in SCI.

Pyroptosis, a form of programmed necrosis, is mainly mediated by NOD-like receptors (NLRPs) in inflammatory bodies and GSDMD via the release of proinflammatory mediators such as IL-1β and IL-18 (Lu et al., 2020). The role of pyroptosis in SCI has been confirmed in recent years (Al Mamun et al., 2021). Our research demonstrated that Bex significantly suppressed pyroptosis-associated protein expression, indicating that Bex suppressed pyroptosis in the spinal cord after injury. Following Bex-mediated inhibition of pyroptosis, the recovery of the spinal cord after injury was significantly promoted. 3MA is an extensively used suppressor of autophagic activity because of its suppressive effect on class III PI3K (Miller et al., 2010). Notably, 3MA reversed the Bex-mediated suppression of pyroptosis. A previous study showed that autophagy suppressed pyroptosis via the IL-13 and Janus kinase 1 (JAK1)/signal transducers and activators of transcription 1 (STAT1) pathways in a mouse model of moderate traumatic brain injury (Gao et al., 2020). This is consistent with our current results. Thus, we conclude that Bex inhibits pyroptosis. In addition to Caspase-1 mediating pyroptosis in SCI, Caspase-4, -5 and -11 also mediate pyroptosis (Shi et al., 2015). We plan to explore whether Bex also regulates pyroptosis mediated by Caspase-4, -5 and -11 in further research.

An imbalance in ROS generation and available antioxidation defense leads to oxidative stress (Slimen et al., 2014). ROS are generated primarily during infection or under inflammatory conditions. Mitochondrial ROS are considered one of the most critical triggers of the NLRP3 inflammatory body, which subsequently mediates pyroptosis and other damage mechanisms (van Bruggen et al., 2010). With a persistent inflammatory response to SCI, the mitochondrial antioxidant defense system and reactive ROS production are increased (Yang and Zhang, 2021). Subsequently, excess ROS causes pyroptosis by activating the NLRP3 inflammasome. We investigated the effects of Bex on oxidative stress after SCI and found that Bex significantly inhibited oxidative stress after SCI by reducing ROS production and accumulation.

Autophagy is an evolutionarily conserved process in eukaryotic cells that involves the degradation and recycling of organelles, cytosol, and proteins (Dikic and Elazar, 2018). Autophagy has a complex role in SCI, and efforts have been made to determine the relationship between these processes (Kanno et al., 2011; Hao et al., 2013). The severity of SCI determines the activation or inhibition of autophagy: mild injury stimulates autophagic flux to remove impaired organelles, while severe injury results in the blockade of autophagic flux and subsequently leads to cell death (Zhou et al., 2017). We evaluated the levels of autophagic flux in SCI and found that autophagic flux was enhanced in mild spinal cord contusion injury. Furthermore, the levels of autophagic flux and autophagy marker proteins during autophagy progression were increased in the Bex group. Therefore, our results demonstrated that Bex promoted autophagy. Suppressing autophagic activity with 3MA induced the accumulation of ROS and activated pyroptosis. Mitophagy also plays an important in SCI. Mitophagy is a specific form of autophagy that selectively eliminates damaged mitochondria (Holt, 2010). Mitochondrial damage has been shown to increase the levels of mitochondrial ROS production (Voloboueva and Giffard, 2011). Neurons are postmitotic cells that need to survive throughout the lifetime of an organism, so they are particularly sensitive to mitochondrial dysfunction (Palikaras et al., 2018). During SCI, the outer membrane of mitochondria in neurons is destroyed by calcium influx, which leads to the release of ROS and proapoptotic proteins into the cytoplasm and ultimately contributes to cell death (Pivovarova and Andrews, 2010).

Bex is a retinoid X receptor-selective agonist that belongs to a group of compounds known as rexinoids. A previous study reported that Bex upregulates and directly binds the long non-coding RNA nuclear paraspeckle assembly transcript 1, which inhibits apoptosis and inflammation by capturing p53-induced death domain-containing protein 1, resulting in better functional recovery in mice after traumatic brain injury (Zhong et al., 2017). Our previous research also demonstrated that apoptosis plays a critical role in SCI, and inhibition of apoptosis is an effective method to treat SCI (Shi et al., 2021). Given these previous studies, we left out the assessment of apoptosis in the present study. Additionally, Bex has been reported to reduce M1 microglial polarization, microglia-derived proinflammatory cytokines, and the number of A1 astrocytes after controlled cortical impact (He et al., 2018b). We speculate that Bex may affect microglia and astrocytes in SCI and more research is required to confirm this possibility.

Bex was reported to induce cell death in ovarian cancer cells through Caspase-4-gasdermin E-mediated pyroptosis (Kobayashi et al., 2022). Our study is the first to explore the mechanism by which Bex reduces pyroptosis. We did not explore whether Bex directly mediates pyroptosis or whether other mechanisms are involved. Thus, we cannot conclude that Bex-mediated induction of autophagy is the main pathway through which Bex alleviates pyroptosis. Likewise, we cannot conclude that attenuating pyroptosis is the main way or only way by which Bex improves function after SCI. We speculate that there are other mechanisms through which Bex attenuates SCI damage, and we aim to explore this in future research.

We further examined the mechanism by which Bex promotes autophagy and mitophagy in SCI. Recent studies suggest that autophagy is regulate by transcriptional and posttranscriptional mechanisms, including the regulation of autophagy-related gene expression, which depends on the action of transcriptional regulators (Di Malta et al., 2019; Shu et al., 2020). The microphthalmia family of bHLH-LZ transcription factors (MiT/TFE) regulates autophagy and oxidative stress (Martina et al., 2014). Furthermore, TFE3 has been reported to be pivotal for autophagy (Zhou et al., 2020). Our results showed that TFE3 translocates into the nucleus after mild SCI. In cells silenced for TFE3, the influence of Bex on the activation of autophagy and mitophagy was reversed. In summary, our findings suggest that Bex activates autophagy and mitophagy through TFE3.

We further explored how Bex regulates the activity of TFE3. We observed the activation of the AMPK-mTOR-TFE3 nuclear pathway, which is consistent with the previous study (Martina et al., 2012). Activated AMPK promotes FOXO3a phosphorylation, reduces the transcription of SKP2, and increases the expression of CARM1, which binds TFE3 to increase TFE3 transcription. Our results suggest that the AMPK-mTOR and AMPK-SKP2-CARM1 pathways are stimulated by Bex in the cytoplasm and nucleus after SCI, respectively, ultimately promoting TFE3 nuclear translocation.

This study has several limitations. Jiang et al. (2017) noted that retinoid level regulation in the ovary exerts molecular control of ovarian development, steroidogenesis, and oocyte maturation, which influences the level of estrogen. Estrogen has been demonstrated to have neuroprotective effects after traumatic brain injury and SCI (Brotfain et al., 2016). After inducing SCI, we squeezed the mouse bladders, expelling urine to prevent bladder rupture. We chose female mice in this study because the urethra of male mice is too long for manual micturition, which would lead to urinary tract infection and a high death rate, interfering with the reliability of the results. We chose one sex of mice to avoid deviations from the use different sexes. We will further explore whether the effect of Bex is based on sex. Furthermore, we only demonstrated that Bex could ameliorate motor functional recovery in the spinal cord after SCI. The effect of Bex on the recovery of sensory function was not examined.

In conclusion, our results showed that Bex promoted TFE3 translocation into the nucleus through the AMPK-mTOR and AMPK-SKP2-CARM1 signaling pathways to enhance autophagy and mitophagy and then inhibit oxidative stress and pyroptosis in an SCI model. Autophagy and mitophagy suppress the production of ROS and then may suppress pyroptosis. Our results indicate that Bex treatment increases spinal cord viability after SCI. We believe that Bex will have potential clinical application value after the necessary clinical trials are conducted. Future studies should explore how Bex activates the AMPK-mTOR and AMPK-SKP2-CARM1 pathways and whether Bex promotes TFE3 nuclear translocation through other mechanisms, such as by directly influencing channel proteins.

Additional files:

Additional Figure 1 ^{(174.1KB, tif)} : Functional improvement by Bex treatment is relatively poor in severe SCI.

Additional Figure 1

Functional improvement by Bex treatment is relatively poor in severe SCI.

The data are expressed as the mean ± SEM (n = 6). *P < 0.05, **P < 0.01, vs. mild SCI + Bex group; #P < 0.05, vs. severe SCI + Bex group (one-way analysis of variance with least significance difference post hoc test). Bex: Bexarotene; BMS: Basso Mouse Scale; SCI: spinal cord injury.

Additional Figure 2 ^{(981.3KB, tif)} : Suppression of TFE3 reverses histological repair exerted by Bex after SCI.

Additional Figure 2

Suppression of TFE3 reverses histological repair exerted by Bex after SCI.

(A) Representative HE and Masson staining of sagittal spinal cord sections. The dotted line indicates the collagen deposition area. (B) Quantification of the proportion of collagen depositional in sagittal spinal cord amounts of sections in A. (C-F) Representative images of immunofluorescence staining for NeuN (indicating neurons, red)/SYN (indicating synapse of neurons, green) and MAP2 (indicating soma and dendrites of neurons, green) in spinal cords transverse sections on the 28th day after SCI (original magnification 30×). Scale bars: 1000 μm (A), 5 μm (C). Quantification of motor neuron-contacting synapse amounts and optical density of MAP2 in motor neurons. The data are expressed as the mean ± SEM (n = 6). **P < 0.01, vs. SCI + Bex group; ##P < 0.01, vs. SCI + Bex + scrambled shRNA group (one-way analysis of variance with least significance difference post hoc test). Bex: Bexarotene; DAPI: 4’,6-diamidino-2-phenylindole; HE: hematoxylin-eosin; MAP2: microtubule-associated protein 2; SCI: spinal cord injury; shRNA: short hairpin RNA; SYN: synapse; TFE3: transcription factor E3.

Additional Figure 3 ^{(708.5KB, tif)} : Suppression of TFE3 reverses motor function and survival rate exerted by Bexarotene after SCI.

Additional Figure 3

Suppression of TFE3 reverses motor function and survival rate exerted by Bex after SCI.

(A) Images of footprints on the 28th day after spinal cord injury. The stride length is similar in SCI + Bex group and SCI + Bex + scramble group, and much little in SCI + Bex + TFE3 shRNA group. (B) Corresponding quantification of step length. (C) BMS at different time points after SCI. (D) The analysis of survival rate of mice from respective groups after SCI. The data are expressed as the mean ± SEM (n = 6). *P < 0.05, **P < 0.01, vs. SCI + Bex group; ##P < 0.01, vs. SCI + Bex + scrambled shRNA group (one-way analysis of variance with least significance difference post hoc test (B, C) or Kaplan‒Meier-rank test (D)). Bex: Bexarotene; BMS: Basso Mouse Scale; SCI: spinal cord injury; shRNA: short hairpin RNA; TFE3: transcription factor E3.

Additional file 1: Open peer review report 1 ^{(85.5KB, pdf)} .