Mitochondrial metabolism restoration via Tramiprosate suppresses mitochondrial ROS-driven foamy macrophage senescence post spinal cord injury

Abstract

Background

Myelin debris (MD) engulfment-induced foamy macrophage formation is a core neuropathology following spinal cord injury (SCI). The accumulation of these foamy macrophages within the injured foci sustains neuroinflammation, impeding long-term neuroregeneration and functional recovery. However, the mechanism underlying macrophage deterioration post-foaming remains elusive.

Methods

MD-induced foamy macrophage and SCI model were used to investigated the role of Tramiprosate (TMP) in vivo and in vitro. Histological staining and functional assessments (gait analysis, Basso Mouse Scale, and motor evoked potentials) were conducted to evaluate the therapeutic effects of TMP on SCI. Quantitative PCR, western blotting, flow cytometry, immunofluorescence, seahorse assay and transmission electron microscopy were used to investigate the senescence and mitochondria function in foamy macrophages. RNA sequencing revealed TMP's role in restoring mitochondrial metabolism. And we injected AAV-shRNA to examine the potential molecular mechanism of TMP.

Results

The current study reveals that lipid droplet-laden foamy macrophages exhibit mitochondrial dysfunction and a senescent phenotype, characterized by increased secretion of matrix metalloproteinases and proinflammatory cytokines. Restoring mitochondrial metabolism via TMP—via upregulation of Shmt2—inhibits mitochondrial reactive oxygen species (mtROS) and mitochondrial DNA (mtDNA) leakage. This reduces oxidative damage to nuclear DNA and suppresses the cyclic GMP-AMP synthase (cGAS)-mediated inflammatory response, thereby eliminating senescence in foamy macrophages.

Conclusions

Our work demonstrates that TMP is a potential therapeutic agent targeting mitochondrial dysfunction-induced macrophage senescence post SCI.

The Translational Potential of this Article

This study investigates the mechanisms underlying macrophage senescence following SCI and identifies TMP as a potential therapeutic agent to mitigate this process. Importantly, TMP is a taurine analogue with established blood–brain barrier permeability and a favorable safety profile in prior clinical investigations for neurodegenerative diseases. These characteristics support its potential treatment strategy for SCI.

Graphical abstract

The schematic diagram illustrates the effect of TMP in inhibiting the inflammatory response driven by senescent macrophages and facilitating functional recovery after SCI. Following SCI, a significant quantity of MD, due to demyelination, is phagocytosed by BMDMs, resulting in the formation of foamy macrophages. The buildup of lipid droplets in these foamy macrophages disrupts mitochondrial function, leading to the production of excessive reactive oxygen species (ROS) and inducing DNA damage. The release of mitochondrial DNA triggers the concurrent activation of the cGAS-STING signalling pathway. This sequence of events leads to the senescence of foamy macrophages and enhances the production of the SASP. TMP enhances intracellular GSH and NADPH via the overexpression of Shmt2, therefore suppressing mtROS generation and mtDNA leakage. As a result, DNA damage and the activation of the cGAS pathway are diminished, leading to a reduction in senescence-induced inflammation and facilitating functional recovery. All figures were generated using FigDraw (www.figdraw.com).

1. Introduction

Traumatic spinal cord injury (SCI), characterized by violence-induced neuropathy of the central nervous system (CNS), typically results from accidents, falls, and ballistic injuries [1]. Post-SCI demyelination generates substantial myelin debris (MD), creating a unique lipid-rich and neuroregeneration-inhibited microenvironment [2,3]. Professional phagocytic cells, primarily bone marrow-derived macrophages (BMDMs) and resident microglia, are capable of clearing MD. However, microglia exhibit limited recruitment to the SCI epicenter and possess inferior phagocytic efficiency compared to BMDMs; thus, MD clearance at injury sites depends predominantly on BMDMs [4]. Unfortunately, this essential process for neuroregeneration paradoxically transforms BMDMs into lipid droplet-laden foamy macrophages. These foamy cells subsequently secrete pro-inflammatory cytokines, driving secondary neuroinflammation [[5], [6], [7]] Nevertheless, the mechanisms by which foamy macrophages mediate post-SCI neuroinflammation remain elusive.

Cellular senescence represents a stable state of irreversible loss of proliferative capacity, often induced by stress or chronic stimuli, and is characterized by mitochondrial dysfunction and the secretion of pro-inflammatory molecules, including the senescence-associated secretory phenotype (SASP) [8,9]. Dysregulated lipid metabolism is implicated in age-related pathologies and may accelerate senescence [10,11]. Foamy macrophages—a hallmark of atherosclerosis—exhibit senescent phenotypes and contribute detrimentally throughout disease progression [12]. Within the CNS, senescent glia accumulates lipid droplets and display mitochondrial dysfunction [13], acting as key drivers of pathological neuroinflammation and functional impairment [14]. It remains unclear whether cellular senescence underlies foamy macrophage-mediated neuroinflammation post-SCI.

Taurine (2-aminoethanesulfonic acid) ranks among the most abundant amino acids in muscle and brain tissue [15]. Evidence indicates that taurine counteracts senescence by suppressing senescence markers, mitigating mitochondrial dysfunction and DNA damage, thereby promoting healthspan extension [16]. Beyond its anti-aging effects, taurine also attenuates senescence in pancreatic β-cells and hepatocytes following organ injury [17,18]. Tramiprosate (TMP, 3-aminopropanesulfonic acid), a taurine homolog primarily enriched in algae, shares neuroprotective properties but crosses the blood–brain barrier (BBB) more efficiently than taurine (Fig. S1A). However, the pharmacological effects of TMP in SCI are largely unexplored.

We hypothesized that TMP, like taurine, exerts mitochondrial protective and anti-senescence effects, thereby reducing senescent foamy macrophages and inflammation post-SCI. Our results demonstrate that TMP mitigates mitochondrial dysfunction-mediated senescence and neuroinflammation in foamy macrophages by preserving serine hydroxymethyltransferase 2 (Shmt2)-dependent mitochondrial respiratory metabolism. This intervention ultimately improves hindlimb motor function in SCI mice.

2. Materials and methods

2.1. Experimental SCI mouse model

Adult male C57BL/6J mice (8 weeks old, average weight 20 g were purchased from Huachuang (Taizhou, Jiangsu, China). The mice were anesthetized via intraperitoneal injection of tris-ethanol. After disinfection of the back with iodine, the spinal cord was exposed at the T10 level. A 5-g spinal cord impactor (68099Ⅱ-S-M, RWD, Shenzhen, China) was used to strike the spinal cord from a height of 5 cm. Successful SCI was characterized by the presence of a hematoma in the spinal cord, tail flaccidity, and hindlimb paralysis. For the 7 days following SCI, SCI mice received TMP (HY-14602, MedChemExpress, Shanghai, China) dissolved in sterile 0.9 % saline via intraperitoneal injection, while the Sham and SCI groups received an equal volume of normal saline. The dosing regimens experiments were selected based on previous studies investigating TMP in central nervous system injury models [19], as well as our preliminary dose-exploration experiments. These studies confirmed the efficacy of TMP in reducing neuroinflammation and showed no observable adverse effects within the selected dose range.

2.2. Histological analyses

After euthanasia, the mice were perfused with pre-cooled phosphate-buffered saline (PBS) and 4 % paraformaldehyde (PFA). The spinal cord was then immersed in 4 % PFA for 24 h and embedded in paraffin and cut into 5-μm sections.

• ●Hematoxylin and eosin(H&E) Staining: Paraffin sections were deparaffinized in xylene and dehydrated with ethanol. The sections were then stained with HE staining solution (G1076, Servicebio, Wuhan, China) according to the manufacturer's instruction.
• ●Nissl Staining: Nissl staining was performed using Nissl staining reagent (G1036, Servicebio). The sections were stained with 0.5 % toluidine blue for 5 min to visualize Nissl bodies in blue.
• ●Luxol Fast Blue (LFB) Staining: To examine the pathological changes of myelin sheaths after SCI, LFB staining was performed using LFB staining reagent (G1030, Servicebio).

All images were captured using a microscope (Thunder DMI8, Leica, Germany).

2.3. Locomotor analysis

• ●Basso Mouse Scale (BMS) Score: Hindlimb locomotor function was assessed by two researchers blinded to the experimental groups at 1,3,7,14, and 28 dpi [20].
• ●Gait Trajectory Measurement: Twenty-eight days after spinal cord injury, following a 1-h acclimation period in the testing arena, the forelimbs and hindlimbs of the mice were coated with red and blue ink respectively. The mice were then placed on a paper-covered runway to walk along a straight path, allowing for the assessment of hindlimb stride length and stride width.
• ●Motor Evoked Potentials (MEPs): MEPs of the gastrocnemius muscle in the hindlimbs were recorded using exercise-induced EMG. MEPs (PLC01, KAHA Science, USA) were elicited using single stimuli (2 mA, 0.2 ms, 1Hz), and the amplitude and latency of the MEPs were quantified to assess hindlimb nerve conduction function in mice.

2.4. Preparation of myelin debris (MD)

The mice were sacrificed, and their brains were homogenized and layered over between 0.32 M and 0.83 M sucrose solution. Using the sucrose density gradient method, the brains were centrifuged at 100,000×g for 45 min at 4 °C, and MD was collected from the interface between the two sucrose layers. After weighing, the MD was resuspended in sterile PBS (KGL2210, KeyGEN, Nanjing, China) to a final concentration of 100 mg/mL (Figure S1B).

2.5. Cell isolation and culture

• ●To obtain primary BMDMs, bone marrow was isolated from 8-week-old mice and resuspended in dulbecco's modified eagle medium (DMEM; KeyGEN) medium supplemented with 30 ng/mL macrophage colony-stimulating factor (M-CSF, HY-P7085, MedChemExpress). The cells were cultured for 6–7 days before being stimulated with 100 μg/mL MD homogenate (MD Homo) for 24 h (Fig. S1B).
• ●To isolate neurons, the cerebral cortex of 16–18-day-old fetal mice was dissected, cut into small pieces, and digested with papain at 37 °C for 15–30 min. The suspension was filtered and centrifuged at 1000 rpm for 5 min. The pellet was resuspended in DMEM supplemented with 10 % fetal bovine serum and plated onto poly-D-lysine-coated six-well plates. The medium was replaced with DMEM containing 1 % GlutaMAX (A5873601, Thermo Fisher, MA, USA) and 2 % B27 (17504044, Thermo Fisher, MA, USA) after cell adhesion.

2.6. Bulk RNA sequencing (RNA-seq) and bioinformatics analysis

After treating BMDMs with MD Homo for 24 h, RNA was extracted using TRIzol reagent (YFXM0011, YiFeiXue Biotechnology, Nanjing, China). RNA sequencing was performed using an Illumina HiSeq 4000 sequencer (Biomarker Technologies, Beijing, China) on RNA-Seq libraries. Gene expression was quantified using Tophat and Cufflinks. Gene expression differences were analyzed for biological enrichment and gene pathways using the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. Gene heatmaps were generated to visualize differences in gene expression. Differentially expressed genes were defined as those with fold changes ≥2 and P-values <0.01.

2.7. NADPH assay

Measurements were performed using the NADPH Detection Kit (S0179, Beyotime, Shanghai, China) according to the manufacturer's protocol. Cells were centrifuged to collect the pellets. The supernatant was then incubated at 60 °C to deplete NADP⁺. After cooling, the supernatant was mixed with the G6PDH working reagent and incubated at 37 °C for 10 min. The developing solution was subsequently added and incubated at 37 °C. Absorbance was measured at a wavelength of 450 nm, and the NADPH content for each group was calculated using the standard curve.

2.8. Glutathione (GSH) and oxidized glutathione (GSSG) assay

The levels of GSH and GSSG were determined using the GSH and GSSG Detection Kit (S0053, Beyotime). Cells were washed and centrifuged to collect the pellets. After protein removal, the samples were subjected to two rapid freeze–thaw cycles. The total GSH content was measured in the supernatant. The GSH removal auxiliary solution was then added to the remaining sample, which was incubated for 1 h to measure the GSSG content. Absorbance was measured at a wavelength of 412 nm. The GSH and GSSG content for each group was calculated using the standard curve.

2.9. Senescence-associated β-galactosidase (SA-β-gal) staining

SA-β-gal staining was conducted using the staining kit (G1580, Solarbio, Beijing, China). BMDMs were fixed in 4 % PFA at room temperature for 15 min, followed by the addition of SA-β-gal staining solution. The BMDMs were then incubated in a 37 °C incubator overnight. SA-β-gal⁺ cells, which were stained blue, were observed and counted as senescent cells under a light microscope.

2.10. Co-culture of foamy macrophages and neurons

To investigate the effect of foamy macrophages on neuronal survival, a co-culture strategy was employed. The culture medium of foamy cells was carefully transferred to the neuronal culture, and the impact of foamy macrophages on neuronal survival was subsequently observed (Fig. S1C).

2.11. Quantitative real-time PCR (qPCR)

Total RNA was extracted from cells and spinal cord tissue using Trizol reagent (YFXM0011, YiFeiXue Biotechnology). The RNA was then converted to cDNA using the HiScript II Q RT SuperMix reverse transcription kit (R122-01, Vazyme, Nanjing, China). Subsequently, qPCR reactions were performed using AceQ qPCR SYBR Green Master Mix (Q111-02, Vazyme) in a 7500 real-time PCR system (Applied Biosystems, USA). The mRNAs of the target genes were normalized to the β-actin gene. The results were performed using the △△CT method. The primers used are listed in Table 1.

Table 1.

Primers of interest for qPCR.

Gene Name	Forward Sequences (5′-3′)	Reverse Sequences (5′-3′)
IL-1β	GCAACTGTTCCTGAACTCAACT	ATCTTTTGGGGTCCGTCAACT
TNF-α	CCCTCACACTCAGATCATCTTCT	GCTACGACGTGGCCTACAG
IL-1α	CGAAGACTACAGTTCTGCCATT	GACGTTTCAGAGGTTCTCAGAG
IL-6	TAGTCCTTCCTACCCCAATTTCC	TTGGTCCTTAGCCACTCCTTC
MMP-3	ACATGGAGACTTTGTCCCTTTTG	TTGGCTGAGTGGTAGAGTCCC
MMP-9	CTGGACAGCCAGACACTAAAG	CTCGCGGCAAGTCTTCAGAG
β-actin	GGCTCTATTCCCCTCCATCG	CCACTTGCTAACAATGCCATCT

2.12. Measurement of cytosolic mtDNA

To quantify cytosolic mtDNA levels, cells were collected when they reached the desired density. The supernatant was transferred to lysis buffer containing proteinase K. DNA extraction buffer was then added in a 1:1 ratio with the lysate. Subsequently, 3M sodium acetate was added, followed by an equal volume of anhydrous ethanol. The DNA concentration was determined, and PCR was performed to quantify the copy number of the mt ND1 gene. The primers were: forward primer, GCACCTACCCTATCACTC, and reverse primer, TTGTTTGGGCTACGGCTC.

2.13. Western blotting (WB)

For protein extraction, cells were lysed using a protein extraction kit (KGB5303, KeyGEN). Protein concentration was determined using an enhanced BCA protein quantification kit (KGB2101, KeyGEN). Proteins were separated by sodium dodecyl sulfate - polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes. Specific primary antibodies were added and incubated at 4 °C overnight, secondary antibodies were then probed at room temperature. Proteins were visualized using the G: Box chemiluminescence imaging system (Syngene, Cambridge, UK). The antibodies used are listed in Table 2.

Table 2.

Antibodies of interest in the study.

Antibodies Name #Cat.No.	Source	Species	Application	Dilution rate
anti-iNOS antibody ab15323	Abcam	Rabbit	WB	1:250
anti-COX-2 antibody #12282	Cell Signaling Technology	Rabbit	WB	1:1000
anti-γH2AX antibody sc-517348	Santa Cruz Biotechnology	Mouse	WB	1:200
anti-p53 antibody 2524s	Cell Signaling Technology	Mouse	WB	1:1000
anti-p21 antibody sc-6246	Santa Cruz Biotechnology	Mouse	WB	1:200
anti-Shmt2 antibody 11099-1-AP	Proteintech	Rabbit	WB	1:1000
anti-cGAS antibody A8335	Abclonal	Rabbit	WB	1:1000
HRP-conjugated Beta Actin Monoclonal antibody #HRP-60008	Proteintech	Mouse	WB	1:10000
Goat Anti-Rabbit IgG Secondary antibody (H + L), HRP#YFSA02	YIFEIXUE BioTech	Goat	WB	1:10000
Goat Anti-Mouse IgG Secondary antibody (H + L), HRP #YFSA01	YIFEIXUE BioTech	Goat	WB	1:10000
anti-γH2AX antibody sc-517348	Santa Cruz Biotechnology	Mouse	IF	1:50
anti-p21 antibody sc-6246	Santa Cruz Biotechnology	Mouse	IF	1:50
anti-Shmt2 antibody 11099-1-AP	Proteintech	Rabbit	IF	1:50
anti-cGAS antibody A8335	Abclonal	Rabbit	IF	1:50
anti-8OHdG antibody HY-P81140	MedChemExpress	Rabbit	IF	1:100
anti-F4/80 antibody 29414-1-AP	Proteintech	Rabbit	IF	1:500
Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 405 A31553	Invitrogen	Goat	IF	1:200
Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™405 A31556	Invitrogen	Goat	IF	1:200
Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488 A21206	Invitrogen	Donkey	IF	1:200
Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488 A11001	Invitrogen	Goat	IF	1:200
Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 555 A21428	Invitrogen	Goat	IF	1:200
Donkey anti-Mouse IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 555 A31570	Invitrogen	Donkey	IF	1:200
Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 647 A31573	Invitrogen	Donkey	IF	1:200
Donkey anti-Mouse IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 647 A31571	Invitrogen	Donkey	IF	1:200

2.14. Immunofluorescence (IF)

Cells were washed with PBS and fixed with 4 % PFA for 15 min. Paraffin sections were deparaffinized and dehydrated, followed by antigen retrieval and blocking. Cells or tissues were incubated overnight with primary antibodies at 4 °C. After washing, fluorescence-conjugated secondary antibodies were added and incubated for 1 h in the dark. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) reagent (G1012, Servicebio). Fluorescence images were captured using a microscope (Thunder DMI8, Leica, Germany), and colocalization was analyzed using ImageJ software. The antibodies used are listed in Table 2.

2.15. Transmission electron microscopy (TEM) observation

Cells were digested and centrifuged to form 1 mm³ clumps. Samples were fixed at 4 °C with 2.5 % glutaraldehyde (Ted Pella, CA, USA) and 0.1 M PBS (pH 7.4). The cell pellet was washed three times with 0.1 M PBS at room temperature, followed by fixation in 1 % tetrazolium in the dark for 2 h. Subsequently, the sample was dehydrated sequentially with ethanol and acetone, each step lasting 15 min. After embedding, the sample was cured in an oven. Ultra-thin sections were prepared using an ultramicrotome. The grids were stained in the dark with a double stain consisting of 2 % uranyl acetate and lead citrate. Observations were conducted using a transmission electron microscope (HT7800, Hitachi, Japan) at 80 kV.

2.16. Oxygen consumption rate (OCR) measurement

To measure OCR, 1.5 × 10⁴ cells were seeded into an XF-96 plate and incubated in a 37 °C, 5 % CO₂ incubator for 24 h. Before the experiment, the sensor cassette was placed in XF96 calibration solution in a CO₂-free incubator for 4 h. Before measurement, the medium was replaced with glucose-free and sodium bicarbonates-absent DMEM. We performed OCR analysis using the XF96 analyzer (Agilent, CA, USA). After measuring the basal OCR, oligomycin, FCCP, and rotenone/antimycin A were successively injected into the plate. OCR data were normalized to cell number and analyzed using WAVE software (Agilent).

2.17. Measurement of mitochondrial superoxide

To detect superoxide in mitochondria, BMDMs were stained with 1 μM MitoSOX (M36008, Invitrogen, USA) at 37 °C. Following staining, the cells were analyzed using a flow cytometer (FACSVerse 8, BD Biosciences, NJ, USA). Flow data were analyzed using FlowJo 10.8.1 software. For immunofluorescence, photographs were acquired using a confocal microscope (STELLARIS 5, Leica).

2.18. Measurement of mitochondrial permeability transition pore (mPTP)

The mPTP was assessed using the mPTP reagent kit (C2009S, Beyotime. Briefly, BMDMs were digested and resuspended in a solution containing Calcein AM and a fluorescence quenching working solution. Following incubation, the cells were centrifuged and resuspended in cold PBS containing 1 % fetal bovine serum. We then performed flow cytometry analysis and analyzed the data using FlowJo 10.8.1 software.

2.19. Apoptosis detection

Cells were collected and stained with 5 μL of calmodulin-fluorescein isothiocyanate (V-FITC) and propidium iodide (PI). After incubation with the detection reagent (C1062S, Beyotime), the cells were washed and resuspended in binding buffer, and analyzed immediately using a flow cytometer. During analysis, cells exhibiting necrotic features, characterized by diffuse PI staining and strong fluorescent signals, were distinguished from apoptotic cells, which were identified as V-FITC-positive/PI-negative.

2.20. Assessment of mitochondrial membrane potential

To measure mitochondrial membrane potential, BMDMs were stained with 10 μg/mL JC-1 (C2003S, Beyotime) at 37 °C for 35 min. Subsequently, the cells were imaged under a confocal microscope (Leica).

2.21. Cell transfection

BMDMs were transfected with LV-NC and LV-ShShmt2 (10⁸ TU/mL) in the presence of 1 × HitransG A (GeneChem, Shanghai, China) at 37 °C for 12 h, the medium then was replaced with complete culture medium for a 60 h-culture.

2.22. Adeno-associated virus (AAV) production and delivery

A short hairpin RNA (shRNA) sequence targeting Shmt2 was cloned into a vector under the control of the F4/80 promoter, which ensures specific expression in macrophages. The construct was packaged into an adeno-associated viral vector (AAV-shShmt2, GeneChem, Shanghai, China) for Shmt2 knockdown. A mismatched shRNA sequence was used as a negative control (AAV-NC). For systemic AAV delivery, mice were anesthetized, and their tails were disinfected with iodine. Using a 30-gauge syringe, 200 μL of viral solution (1.33 × 10¹² vg/mL) was injected into the lateral tail vein of the mice, enabling specific transduction of macrophages via the F4/80 promoter.

2.23. Statistical analysis

Data are exhibited as mean ± SEM. Cell samples were from at least three independent biological replicates, while tissue samples were from at least six. Comparisons of more than two groups used one-way ANOVA followed by Tukey's post hoc test. p < 0.05 was considered significant.

3. Results

3.1. TMP encourages recovery of tissue and function following SCI

To assess the therapeutic potential of TMP, we first evaluated its effect on locomotor function recovery at 28 days post injury (dpi). H&E staining revealed a significantly smaller lesion area in SCI mice treated with 50 mg/kg TMP compared to untreated controls or the 25 mg/kg TMP group (Fig. 1A and Fig. S2A). Furthermore, mice receiving 50 mg/kg TMP exhibited reduced neuronal loss and diminished demyelination compared to untreated or 25 mg/kg TMP-treated mice (Fig. 1B and Fig. S2B and C). Electromyography (EMG) demonstrated that TMP administration significantly increased MEP amplitude in SCI mice. Notably, only the 50 mg/kg TMP dose significantly decreased MEP latency relative to untreated SCI mice (Fig. 1D–F). Behaviorally, TMP treatment significantly improved BMS scores across all doses, encompassing weight support, hind paw position, and overall hindlimb coordination (Fig. 1G). However, gait analysis indicated that only the 50 mg/kg TMP dose effectively increased stride length and reduced stride width compared to untreated mice (Fig. 1H–J). Collectively, these findings indicate that 50 mg/kg TMP promotes significant recovery of histological outcomes and hindlimb motor function following SCI.

Fig. 1.

TMP promotes tissue and functional recovery after SCI in mice. A) Representative images of HE staining from longitudinal sections centered on the injury core at 28 dpi in SCI mice treated with 25 mg/kg or 50 mg/kg TMP and control mice, n = 6; scale bars: a = 200 μm, b = 50 μm. B) Representative images of Nissl staining in the four groups of mice; scale bars: a = 200 μm, b = 20 μm, n = 6. C) Representative images of LFB staining in the four groups, scale bar = 200 μm, n = 6. D) Gastrocnemius muscle action potentials of the hind limbs in the four groups at 28 dpi, n = 6. E) Quantitative analysis of action potential amplitude in the four groups, n = 6. F) Quantitative analysis of action potential latency in the four groups, n = 6. G) BMS scores of the four groups, n = 6. H-J) Representative footprint images and analysis of the four groups, n = 6. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.2. TMP inhibits macrophage senescence and inflammation following SCI

We next investigated the role of TMP in cellular senescence within the injured spinal cord at 7 dpi. Western blot (WB) analysis revealed significant upregulation of senescence markers and pro-inflammatory mediators in SCI mice. TMP treatment significantly suppressed the expression of key inflammatory markers, including iNOS and COX-2, in a concentration-dependent manner (Fig. 2A–C). Compared to the 25 mg/kg dose, administration of 50 mg/kg TMP more effectively reduced expression of the DNA damage marker γH2A.X and key senescence regulators p53 and p21 (Fig. 2A–D-F). Consistent with these findings, transcriptional levels of multiple SASP factors—including IL-6, IL-1α, IL-1β, TNF-α, MMP-3, and MMP-9—were markedly elevated at 7 dpi. Treatment with 50 mg/kg TMP significantly attenuated this SASP upregulation compared to the 25 mg/kg dose (Fig. 2G). Immunofluorescence (IF) staining further demonstrated that 50 mg/kg TMP significantly reduced the co-localization intensity of the macrophage marker F4/80 with γH2A.X at the lesion site compared to 25 mg/kg TMP (Fig. 2H). Similarly, 50 mg/kg TMP significantly decreased the co-localization of F4/80 with p21 (Fig. 2I). Collectively, these results indicate that TMP suppresses DNA damage response and senescence in macrophages after SCI at 7 dpi. Given its superior efficacy at 50 mg/kg, this concentration was selected for subsequent in vivo experiments.

Fig. 2.

TMP inhibits macrophage senescence and inflammatory activation after SCI. A) WB analysis of protein levels of iNOS, COX-2, γH2A.X, p53, and p21 in spinal cord tissues of SCI mice treated with 25 mg/kg or 50 mg/kg TMP and control mice after SCI. β-actin (ACTB) served as a control, n = 6. B-F) Quantitative analysis of protein levels of iNOS, COX-2, γH2A.X, p53, and p21 in the four groups, respectively, n = 6. G) Heat map of relative mRNA levels of SASP in spinal cord tissues of the four groups at 7 dpi, n = 6. H) Representative panoramic IF images of the macrophage marker F4/80 (green) and γH2A.X (red) at 7 dpi, with corresponding colocalization analysis curves from the boxed region, scale bar = 312.5 μm, n = 6. I) Representative panoramic IF images of the macrophage marker F4/80 (green) and p21 (red) at 7 dpi, with corresponding colocalization analysis curves from the boxed region, scale bar = 312.5 μm, n = 6. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.3. TMP attenuates MD-induced foamy macrophage senescence

To explore the direct inhibitory role of TMP on MD-induced macrophage senescence, BMDMs were treated with TMP at concentrations of 2, 20, and 200 μg/mL. SA-β-gal staining exhibited a marked increase in SA-β-gal⁺ cells in MD-induced BMDMs. TMP treatment remarkably reduced the proportion of SA-β-gal⁺ cells, with maximal inhibition observed at 200 μg/mL (Fig. 3A and B). These findings indicate that MD-induced foamy macrophages exhibit a senescent phenotype, which is effectively suppressed by TMP. WB analysis demonstrated that TMP dose-dependently reduced expression levels of inflammatory markers iNOS and COX-2 in foamy macrophages (Fig. 3C–E). TMP treatment also significantly suppressed expression of γH2A.X and senescence regulators p53 and p21 (Fig. 3C–F-H). Consistent with these results, qPCR analysis revealed that TMP treatment attenuated the MD-induced transcriptional upregulation of SASP factors, including IL-6, IL-1α, IL-1β, TNF-α, MMP-3, and MMP-9. Notably, SASP expression in macrophages treated with 200 μg/mL TMP was nearly equivalent to baseline control levels (Fig. 3I). IF staining confirmed that 200 μg/mL TMP significantly reduced senescence-associated cells of γH2A.X (Fig. 3J, K) and p21 (Fig. 3J–L) in MD-induced BMDMs. These results demonstrate that phagocytosis of MD induces DNA damage and senescence in macrophages, while TMP effectively attenuates these processes and associated inflammation. Given its maximal efficacy at 200 μg/mL, this concentration was selected for subsequent in vitro experiments.

Fig. 3.

TMP alleviates the phagocytosis-mediated senescence of foamy macrophages due to MD. A) Representative images of SA- β -gal staining in BMDMs treated with control, TMP (2, 20, 200 μg/ml) and MD Homo, bar = 100 μm, n = 3. B) Quantitative analysis of the percentage of SA-β-gal + macrophages in total macrophages, n = 3. C) WB analysis of protein levels of iNOS, COX-2, γH2A.X, p53, and p21 in five groups of cells with β-actin as a control, n = 3. D-H) Quantitative analysis of protein levels of iNOS, COX-2, γH2A.X, p53, and p21 in the five groups, respectively, n = 3. I) Heat map of relative mRNA levels of SASP in the five groups, n = 3. J) Representative IF images of γH2A.X (green) and DAPI, p21 (red), and DAPI in the five groups, scale bar = 10 μm, n = 3. K) Percentage of positive γH2A.X cells in the five groups, n = 3. L) Percentage of positive p21 cells in the five groups, n = 3. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.4. TMP boosts Shmt2 expression to protect macrophage DNA from oxidative damage

To investigate the molecular mechanisms by which TMP regulates the senescence of foamy macrophages, we performed RNA sequencing. Analysis revealed that foamy macrophage differentiation is associated with immune inflammation and dysregulated lipid metabolism (Fig. S2F). We initially hypothesized that TMP mediates its effects on senescence by modulating lipid droplet accumulation and subsequent immune inflammation. However, BODIPY staining showed that the proportion of lipid-laden BMDMs was not reduced by TMP treatment (Fig. S2D and E). Transcriptomic analysis further revealed pronounced alterations in the expression of genes governing mitochondrial respiration, redox homeostasis, immune inflammation, oxidative stress, DNA repair, and cellular senescence following TMP treatment. Notably, TMP significantly upregulated the expression of Shmt2, encoding the mitochondrial enzyme Shmt2, which is implicated in oxidative stress responses (Figure S2F and Fig. 4A and, B). WB analysis displayed increased Shmt2 protein expression in TMP-treated foamy macrophages (Fig. 4C and D). Furthermore, immunofluorescence demonstrated that TMP treatment elevated Shmt2 levels while concurrently reducing the intensity of the DNA oxidative damage marker 8-hydroxy-2′-deoxyguanosine (8-OHdG) (Fig. 4E). TMP significantly increased intracellular GSH and NADPH levels while decreasing GSSG in foamy macrophages (Fig. 4F–H), suggesting enhanced antioxidant capacity. These results demonstrate that TMP upregulates Shmt2 expression, promotes the generation of reducing equivalents (NADPH and GSH), and mitigates oxidative DNA damage.

Fig. 4.

TMP increases Shmt2 expression to prevent DNA oxidative damage in macrophages. A) RNA-seq results of BP, CC, MF, and KEGG pathways in BMDMs treated with MD Homo or 200 μg/ml TMP + MD Homo. B) Heat map showing differentially expressed genes between the two groups. C) WB analysis of Shmt2 protein levels in three groups of cells, with β-actin as a control, n = 3. D) Quantitative analysis of Shmt2 protein levels, n = 3. E) Representative IF images and quantification of Shmt2 (green) and 8-OHdG (red) in the three groups, scale bar = 20 μm, n = 3. F) NADPH content in the three groups, n = 3. G-H) Relative levels of GSH and GSSG in the three groups, n = 3. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.5. TMP attenuates mitochondrial damage and oxidative stress by reprogramming mitochondrial metabolism

Given TMP's impact on mitochondrial respiratory metabolism and oxidative stress, we assessed mitochondrial integrity. Mitochondrial membrane potential (Δψm) was disrupted in foamy macrophages; TMP treatment restored Δψm (Fig. 5A–F).TMP significantly reduced mtROS production (Fig. 5B), consistent with flow cytometric analysis (FCM) (Fig. 5D–H). TEM revealed that foamy macrophages exhibited reduced mitochondrial cristae density and damage to inner/outer membrane structures, both attenuated by TMP (Fig. 5C–I). FCM analysis indicated increased mPTP opening in foamy macrophages, which TMP significantly suppressed (Fig. 5E–G). Cytosolic mtDNA copy number was substantially elevated in foamy macrophages. TMP treatment significantly reduced cytosolic mtDNA levels, further demonstrating its protection of mitochondrial integrity (Fig. 5J). WB analysis showed that TMP inhibited cytosolic mtDNA-induced cGAS activation (Fig. S3A and B), confirmed by IF staining (Fig. S3C). Seahorse metabolic flux analysis revealed significant dysfunction in mitochondrial oxidative phosphorylation (OXPHOS) in foamy macrophages. TMP treatment increased basal and maximal respiration rates, enhanced ATP production, and elevated spare respiratory capacity (Fig. 5K–O), indicating improved mitochondrial energy metabolism. In an indirect co-culture model of foamy macrophages and neurons (Fig. S1C), foamy macrophages induced significant secondary neuronal apoptosis; TMP treatment markedly reduced neuronal apoptosis rates (Fig. 5P and Q). Thus, TMP enhances mitochondrial function and reduces oxidative stress by reprogramming mitochondrial metabolism, thereby attenuating secondary neuronal apoptosis.

Fig. 5.

TMP alleviates mitochondrial damage and oxidative stress by reprogramming mitochondrial metabolism. A) Representative images of JC-1 staining in BMDMs pretreated with 200 μg/ml TMP for 2 h and then stimulated with MD Homo, analyzed by confocal microscopy, scale bar = 36.8 μm, n = 3. B) Representative immunofluorescence of MitoSOX in the three groups, scale bar = 36.8 μm, n = 3. C) Representative mitochondrial structures of the three groups observed by transmission electron microscopy(TEM), scale bar = 500 nm, n = 3. D) Representative images of Flow cytometry (FCM) with MitoSOX in the three groups, n = 3. E) Representative images of FCM with mPTP in the three groups, n = 3. F) Statistical analysis of JC-1 aggregates/JC-1 monomers in the three groups, n = 3. G) Analysis of mPTP in the three groups, n = 3. H) FCM analysis of MitoSOX in the three groups, n = 3. I) Quantitative analysis of mitochondrial cristae number in the three groups by TEM, n = 3. J) qPCR analysis of mtDNA copy number in the three groups, n = 3. K) OCR detected by Seahorse XFe96 analyzer in the three groups, n = 3. L-O) Quantitative analysis of basal respiration, maximal respiratory rate, ATP production, and respiratory reserve in the three groups, n = 3. P) Representative images of FCM with apoptosis in neurons co-cultured with the three groups of cells, n = 3. Q) Statistical analysis of neuronal apoptosis, n = 3. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.6. TMP reduces macrophage DNA oxidative damage and suppresses cGAS signaling after SCI

Having validated that TMP inhibits cGAS activation caused by mtDNA leakage in vitro, we proceeded to confirm the effects in vivo. WB analysis demonstrated that TMP suppresses cGAS expression (Fig. 6A and B) and increases Shmt2 expression at 7 dpi (Fig. 6A–C). IF staining showed that TMP significantly enhances Shmt2 expression levels in foamy macrophages (Fig. 6D). Furthermore, there was a strong co-localization of the DNA oxidative damage marker 8-OHdG and F4/80 following SCI. TMP treatment reduced the expression of 8-OHdG in macrophages following SCI (Fig. 6E). IF staining also showed a strong co-localization of F4/80 and cGAS, with TMP-treated animals exhibiting suppressed cGAS activation in macrophages following SCI (Fig. 6F). These findings collectively demonstrate that TMP alleviates DNA oxidative damage and cGAS signaling activation following SCI.

Fig. 6.

TMP reduces DNA oxidative damage and cGAS signaling activation in macrophages after SCI. A) WB analysis of cGAS and Shmt2 protein levels in spinal cord tissues of Sham, SCI, and 50 mg/kg TMP-treated SCI mice at 7 dpi, with β-actin as a control, n = 6. B) Quantitative analysis of cGAS protein levels in the three groups, n = 6. C) Quantitative analysis of Shmt2 protein levels in the three groups, n = 6. D) Representative IF images and co-localization analysis curves of Shmt2 (pink) and F4/80 (green) in longitudinal sections centered on the injury core of the three groups at 7 dpi, scale bar = 20 μm, n = 6. E) Representative IF images and co-localization analysis curves of 8-OHdG (pink) and F4/80(green) in the injured area of the three groups at 7 dpi. F)Representative IF images and co-localization analysis curves of cGAS (pink) and F4/80 (green) in the injured area of the three groups at 7 dpi; scale bar = 20 μm, n = 6. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

3.7. TMP inhibits cellular senescence in foamy macrophages via Shmt2-mediated mitochondrial metabolic reprogramming

To investigate whether TMP protects mitochondrial function and suppresses senescence through Shmt2, we knocked down Shmt2 in BMDMs using LV-Shmt2KD (Fig. S3E). WB analysis demonstrated that the LV-Shmt2KD + MD + TMP group exhibited significantly restored expression of inflammatory-related proteins, including cGAS, iNOS, and COX-2, with levels comparable to the LV-NC + MD group (Fig. S4A–D). Additionally, the inhibitory effects of TMP on γH2A.X, p53, and p21 proteins were reversed in the LV-Shmt2KD group (Figs. S4A, E, F, G). PCR analysis further revealed that the inhibitory effect of TMP on SASP secretion was suppressed in the LV-Shmt2KD group (Fig. S4L). Furthermore, Shmt2 knockdown significantly increased 8-OHdG expression to levels comparable to the LV-NC + MD group (Fig. S5A), with a similar trend observed for cGAS expression (Fig. S5B). IF staining showed increased positive cells of γH2A.X (Fig. S5C and D) and p21 (Fig. S5E and F) in the LV-Shmt2KD + MD + TMP group. These results suggest that TMP exerts its anti-senescence effects through Shmt2. Reassessment of mitochondrial respiratory function via Seahorse analysis demonstrated that in the LV-Shmt2KD group, TMP treatment led to a significant decrease in improved mitochondrial respiratory metabolic indicators, including basal respiratory rate, maximum respiratory rate, ATP production, and respiratory reserve, with levels returning to those of the LV-NC + MD group (Fig. 7A–E). Reassessment of GSH and GSSH levels showed that the beneficial effects of TMP were suppressed upon Shmt2 knockdown (Fig. S4H and I). Additionally, the improvement in mitochondrial membrane potential levels by TMP was inhibited in the LV-Shmt2KD group (Fig. 7F–K). IF and FCM analysis revealed increased production of mtROS (Fig. 7G–I, M). TEM demonstrated that the protective effects of TMP on mitochondrial inner and outer membranes and cristae were almost completely lost following Shmt2 knockdown (Fig. 7H–N). Furthermore, in the LV-Shmt2KD + MD group, the opening-up mPTP could not be suppressed by TMP (Fig. 7J–L), accompanied by a re-increase in mtDNA leakage levels (Fig. 7O). When Shmt2 was knocked down, TMP failed to reduce the rate of SA-β-gal-positive foamy macrophages (Fig. S4J and K). FCM analysis showed that TMP reduced the apoptosis rate of neurons co-cultured with foamy macrophages; however, this neuroprotective effect was abolished upon Shmt2 knockdown (Fig. 7P and Q). These findings collectively indicate that TMP inhibits cellular senescence and exerts neuroprotective effects in foamy macrophages through Shmt2-mediated mitochondrial metabolic reprogramming (see Fig. 8).

Fig. 7.

TMP inhibits cellular senescence via Shmt2-mediated mitochondrial metabolic reprogramming in foamy macrophages. A) OCR detected by Seahorse XFe96 analyzer in BMDMs transfected with Shmt2 KD, pretreated with 200 μg/ml TMP, and stimulated with MD Homo for 24 h, n = 3. B-E) Quantitative analysis of basal respiration, maximal respiratory rate, ATP production, and respiratory reserve in the three groups, n = 3. F) Representative IF images of JC-1 staining in the three groups analyzed by confocal microscopy, scale bar = 36.8 μm, n = 3. G) Representative IF of MitoSOX in the three groups, scale bar = 36.8 μm, n = 3. H) Representative mitochondrial structures of the three groups observed by TEM, scale bar = 500 nm, n = 3. I) Representative images of FCM with MitoSOX in the three groups, n = 3. J) Representative images of FCM with mPTP in the three groups, n = 3. K) Statistical analysis of JC-1 aggregates/JC-1 monomers in the three groups, n = 3. L) FCM analysis of mPTP in the three groups, n = 3.M) FCM analysis of MitoSOX in the three groups, n = 3. N) Quantitative analysis of mitochondrial cristae number in the three groups by TEM, n = 3. O) qPCR analysis of mtDNA copy number in the three groups, n = 3. P) Representative images of FCM with apoptosis in neurons co-cultured with the three groups of cells, n = 3. Q) Statistical analysis of neuronal apoptosis proportion in the three groups, n = 3. Data are presented as mean ± SEM, and statistical significance was determined with one-way ANOVA followed by Tukey's post hoc test, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

Fig. 8.

TMP relieves neuroinflammation mediated by senescent foamy macrophages after SCI via mitochondrial protection. A) Representative IF images and co-localization analysis curves of Shmt2 (pink) and F4/80 (green) in longitudinal sections centered on the injury core at 7 dpi in mice pre-transfected with AAV carrying Shmt2 shRNA targeting mononuclear macrophages, at 7 dpi, scale bar = 20 μm, n = 6. B) Representative IF images and co-localization analysis curves of 8-OHdG (pink) and F4/80 (green), scale bar = 20 μm, n = 6. C)Representative immunofluorescence and co-localization analysis curves of γH2A.X (pink) and F4/80 (green), scale bar = 20 μm, n = 6. D)Representative IF images and co-localization analysis curves of p21 (pink) and F4/80 (green) in the injured area of the three groups at 7 dpi, scale bar = 20 μm, n = 6.

3.8. TMP alleviates foamy macrophage senescence-mediated neuroinflammation after SCI through mitochondrial protection

In SCI animal models, we pre-transfected mice with AAV-delivered Shmt2 shRNA targeting macrophages and confirmed the transfection efficiency (Figure S3D, F and Fig. 8A). Compared to the AAV-NC + SCI + TMP group, the Shmt2KD + TMP group exhibited a significant increase in γH2A.X, p53, and p21 protein levels at 7 dpi (Fig. S6A–D). Similarly, the expression of cGAS, iNOS, and COX-2 proteins also showed upward trends (Figs. S6A, E, F, G). Furthermore, the transcriptional levels of SASP factors were reversed in the AAV-Shmt2KD + SCI + TMP group (Fig. S6I). IF co-localization analysis revealed that the AAV-Shmt2KD + SCI + TMP group displayed increased expression of cGAS, 8-OHdG, γH2A.X, and p21 in macrophages at 7 dpi (Figure S6H and Fig. 8B–D). These findings confirm that TMP mitigates DNA oxidative damage and senescence-mediated neuroinflammation in macrophages via the Shmt2 pathway in vivo.

4. Discussion

The emergence of foamy macrophages is a critical issue that mediates secondary neuroinflammation following SCI [2,21]. Previous studies [22,23] have suggested that the proinflammatory phenotype of foamy macrophages is attributed to the engulfment of MD by BMDMs, and the absence of scavenger receptors inhibits foam formation and suppresses neuroinflammation. However, the dysfunction of organelles within foamy macrophages has been overlooked. The present study first demonstrated that mitochondrial metabolic dysfunction in foamy macrophages exacerbates senescence and neuroinflammation after SCI, while treatment with TMP suppresses mitochondrial oxidative phosphorylation dysfunction, thereby reducing mtDNA leakage-mediated inflammatory activation and mtROS-induced DNA damage in senescent foamy macrophages.

The combination of dasatinib-quercetin is currently the primary therapeutic approach targeting cellular senescence; however, its administration is associated with severe adverse effects [24]. Therefore, identifying effective anti-senescence compounds with minimal side effects has become a research priority. Taurine, as a natural amino acid, is associated with minimal side effects on the body, and its deficiency may reduce healthy lifespan [16,25]. Unfortunately, due to its sulfonic acid structure, taurine is restricted to cross the BBB [26,27]. Although numerous studies have reported that taurine promotes axonal regeneration and reduces neuroinflammation after SCI, its therapeutic potential is limited by the requirement for high dosages [28,29]. Thus, a taurine homolog, namely TMP, has garnered attention. The addition of an extra methylene group in TMP's molecular structure enhances its permeability across the BBB [30,31]. Over the past two decades, TMP has demonstrated therapeutic potential in reducing amyloid plaque deposition in Alzheimer's disease [[32], [33], [34]]. Moreover, TMP has been evaluated in Alzheimer's disease and cerebral amyloid angiopathy, where it exhibits good tolerability and minimal adverse effects, further underscoring its biological safety and translational potential [35]. Moreover, in ischemic stroke, TMP exerts robust protective actions, including dose-dependent reductions in infarct volume, attenuation of neuronal apoptosis, and suppression of excitotoxic injury [19]. These previous studies highlight TMP's broad neuroprotective and anti-injury properties, thereby supporting the plausibility of its therapeutic value in SCI and providing a strong rationale for our investigation. However, its effects on inhibiting cellular senescence remain understudied.

Our study confirmed the senescence phenotype of foamy macrophages. Administration of TMP significantly suppressed the expression of SASP, along with a reduction in the number of SA-β-gal⁺ cells. TMP treatment also decreased the production of proinflammatory mediators and cytokines in foamy macrophages. Initially, we hypothesized that TMP, similar to taurine, accelerates lipid metabolism and promotes lipid degradation [36,37], potentially reducing the formation of foamy macrophages by alleviating lipid droplet overload. Surprisingly, TMP did not eliminate lipid droplets in foamy macrophages. RNA sequencing revealed a specific mechanism underlying its anti-senescence and anti-inflammatory effects. We found that TMP treatment significantly enriched genes related to mitochondrial function and oxidative stress, while downregulating pathways associated with inflammation, cellular senescence, and the cell cycle. Mitochondria, as semi-autonomous organelles, play a central role in cellular metabolism and energy production. Their dysfunction and the leakage of mitochondrial genetic material can critically influence cellular fate [38,39]. Our previous work [40] has shown that SCI-induced mitochondrial dysfunction not only causes inflammatory activation of microglia and macrophages but also exacerbates neuronal necrosis and apoptosis. Importantly, mitochondrial dysfunction is a primary driver of cellular senescence in age-related neurodegenerative diseases [[41], [42], [43]], and our findings extend this understanding to its harmful role in cellular senescence during the acute pathogenic process of SCI.

Taurine is a natural antioxidant that plays a favorable role in repairing mitochondrial metabolism and structure, as well as maintaining mitochondrial homeostasis and translation [[44], [45], [46]]. Nevertheless, researchers seldom notice any study covering the effects of taurine analogues on mitochondrial protection. Our results, for the first time, demonstrated that TMP improves unbalanced mitochondrial membrane potential and disordered mitochondrial respiratory metabolism by rescuing Shmt2 expression. Shmt2 is a key enzyme in serine metabolism, located in mitochondria, that catalyzes the conversion of serine to glycine and the generation of GSH and NADPH [47,48]. Shmt2 not only promotes cell viability to inhibit mtROS-mediated apoptosis [49] but also alleviates mtROS-mediated DNA damage responses [50]. Furthermore, depletion of Shmt2 is crucial for the onset of immune cell senescence. A study shows that knockdown of Shmt2 expression in CD8⁺ T cells during HIV infection accelerates ROS-mediated cellular senescence [51]. We likewise identified downregulation of Shmt2 in foamy macrophages, and its expression is critical for mitochondrial dysfunction and the progression of cellular senescence. The key point is that the mitochondria-protective effect of TMP depends on Shmt2 expression, which promotes the production of GSH and NADPH and reduces mtROS [51,52]. These changes reduced the permeability of the mitochondrial membrane, prevented the activation of the mtDNA/cGAS axis, thereby counteracting inflammatory signals and senescence. It has also been confirmed that senescent cells actively release mtDNA, and senescent macrophages even exacerbate mitochondrial damage, thereby increasing the release of mtDNA [53,54]. Our previous researches also report a significant increase in the expression of type I interferon immune responses after SCI [55,56], but we did not investigate the specific reasons for this. The current study fills the gaps in our previous works and demonstrates that the activation of the cGAS/STING axis mediated by mitochondrial mtDNA leakage might be an important cause of this post-SCI immune response.

The present work has several limitations. Although we utilized viral transfection to knock down the expression of Shmt2 to verify the pharmacological effects of TMP in vitro and in vivo, it would be more convincing to examine further in the conditional knockout mouse model of Shmt2. We were unable to identify the specific transcription factor that regulates the expression of Shmt2 through the action of TMP. Previous studies have reported that ERRalpha and Eif5 bind to the transcription initiation site of Shmt2 [50,57], but it remains to be validated whether TMP directly combines with the above or some unknown factors to achieve stable protranscriptional effects of the Shmt2 gene. In addition, identifying an optimal therapeutic dose of TMP will be an important direction for future translational research. Furthermore, future investigations incorporating synaptic markers, electrophysiological assessments, and metabolic profiling will be necessary to elucidate additional impacts of senescent foamy macrophages on neuronal homeostasis and to further clarify the scope of TMP's neuroprotective actions.

In summary, our study provides a new insight and a theoretical basis for the TMP treatment in acute CNS injury, and clarifies that TMP plays a crucial pharmacological role in the inhibition of post-SCI immune senescence of foamy macrophages.

Ethical statement

All animal experimental procedures were approved by the Animal Care Committee of Nanjing Medical University (Approval No. IACUC-2409021). All the animal procedures were performed in accordance with the Guiding Principles in the Use and Care of Animals.

Consent for publication

Not applicable.

Data availability statement

All data associated with this study are included in this paper and its Supplementary Materials. Additional raw data are available from the corresponding author upon reasonable request.

Author's contributions

ZY.Q. conceived and designed the experiments. XJ.C., ZY.Q., L.Y., and JW.X. funded the research. L.Y. and X.J.C. supervised the study. CQ.W and QH.F performed the experiments. JJ.F, JL.L. and BZ.Z. performed the data analyses. J.Z., JW.X., and TY.Z. assisted with the experiments. Y.Z. draws the scheme graphs and the graphical abstract. ZY.Q., CQ.W wrote and edited the manuscript.

Declarations

Use of Generative AI in Scientific Writing statement:

No generative artificial intelligence (AI) tools were used in the writing, analysis, or preparation of this manuscript.

Funding

The work was funded by the Original Exploration Project of the National Natural Science Foundation of China (82350002, to Xiaojian Cao), the China Postdoctoral Science Foundation (2024M761521, to Zhanyang Qian), the Nantong Municipal Natural Science Foundation (JC2024027, to Zhanyang Qian), the High-level Science and Technology Cultivation Foundation of the Second Affiliated Hospital of Nantong University (YPYJJZD24002, to Zhanyang Qian), the project of Nantong Municipal Health Commission(MS2024024, to Jiawen Xu).

Declaration of competing interest

The authors have no competing interests.

Abbreviations

BMDMs	bone marrow-derived macrophages
BBB	blood–brain barrier
BMS	Basso Mouse Scale
cGAS	cyclic GMP-AMP synthase
CNS	central nervous system
DAPI	4′,6-diamidino-2-phenylindole
DNA	DeoxyriboNucleic Acid
DPI	day post injury
DMEM	Dulbecco's modified Eagle medium
EMG	Electromyography
FCM	flow cytometric
FBS	fetal bovine serum
GO	Gene Ontology
GSSG	glutathione disulfide
GSH	glutathione
HE	Hematoxylin and Eosin
IF	Immunofluorescence
KEGG	Kyoto Encyclopedia of Genes and Genomes
KD	knock down
LFB	Luxol Fast Blue
MEPs	Motor Evoked Potentials
M-CSF	macrophage colony-stimulating factor
MD Homo	myelin debris homogenate
mtROS	mitochondrial reactive oxygen species
mtDNA	mitochondrial DNA
mPTP	Mitochondrial Permeability Transition Pore
OXPHOS	mitochondrial oxidative phosphorylation
OCR	Oxygen Consumption Rate
PBS	phosphate-buffered saline
PFA	paraformaldehyde
PI	propidium iodide
PVDF	polyvinylidene fluoride
RNA-Seq	RNA Sequencing
Shmt2	serine hydroxymethyltransferase 2
SCI	Spinal cord injury
SASP	senescence-associated secretory phenotype
SDS-PAGE	sodium dodecyl sulfate - polyacrylamide gel electrophoresis
TMP	tramiprosate
TEM	Transmission electron microscopy
WB	Western Blot
8-OHdG	8-hydroxy-2′-deoxyguanosine

Appendix A. Supplementary data

The following is the Supplementary data to this article: