Neutrophil extracellular traps delivering lactylated S100a9 aggravate neuronal cuproptosis by regulating Ttr/Commd1/Atp7b axis after traumatic brain injury

Abstract

Traumatic brain injury (TBI) is a leading cause of disability and mortality, with secondary injury mechanisms involving neuronal death and neuroinflammation, for which effective treatments remain limited. Neutrophil extracellular traps (NETs) are implicated in post-TBI neuropathology. Cuproptosis, a copper-dependent cell death pathway characterized by mitochondrial oxidative stress, dysfunction, and disrupted dynamics, has recently been implicated in neurological disorders. This study aims to investigate whether NETs exacerbate secondary brain injury by promoting neuronal cuproptosis after TBI and to elucidate the underlying molecular mechanism. We observed elevated NET levels in brain tissues from both TBI patients and mice, which correlated with poor prognosis. Single-cell RNA sequencing revealed a significant upregulation of transthyretin (Ttr) in neurons post-TBI. Mechanistically, NETs deliver lactylated S100a9 (S100a9K26la), a glycolysis-dependent lactylated protein, to neurons. S100a9K26la translocates to the nucleus and promotes Ttr transcription. Increased neuronal Ttr protein then competes with ATPase copper transporting β (Atp7b) for binding to copper metabolism MURR1 domain-containing 1 (Commd1) at the W123 residue. This competition disrupts the Commd1-Atp7b interaction, impairing copper efflux and leading to intracellular copper accumulation, mitochondrial oxidative stress, aggregation of DLAT, loss of Fe-S cluster proteins, and ultimately neuronal cuproptosis. Neuron−specific Ttr conditional knockout ameliorated neuronal death, neuroinflammation, blood-brain barrier (BBB) disruption, and neurological deficits in a TBI model. Conversely, inhibition of cuproptosis with the copper chelator tetrathiomolybdate (TTM) yielded similar protective effects. In summary, our findings elucidate a novel pathway wherein NETs, via delivery of S100a9K26la, drive neuronal Ttr overexpression. Ttr disrupts copper homeostasis by interfering with the Commd1-Atp7b axis, ultimately triggering neuronal cuproptosis and exacerbating secondary injury after TBI. This study identifies NETosis and the Ttr/Commd1/Atp7b axis as potential therapeutic targets for mitigating TBI-induced damage.

1. Introduction

Traumatic brain injury (TBI), a predominant cause of disability and mortality among young adults, is a monumental public health challenge [1]. TBI treatment is tremendously challenging due to its heterogeneity, often causing various injuries and complex pathogeneses, which are generally classified into primary and secondary injuries. Secondary injury refers to further damage caused by complex and interrelated pathophysiological processes, such as neuronal death, neuroinflammation, and blood‒brain barrier (BBB) disruption [2,3]. The excessive neuroinflammatory response and delayed neuronal damage after TBI are key mechanisms of secondary injury, which can lead to chronic motor and cognitive deficits [4]. The treatment options for posttraumatic neuroinflammatory responses and persistent neuronal death are still limited and commonly ineffective, indicating a lack of understanding of the pathophysiological mechanisms and targeted treatment methods for secondary brain injury.

Neutrophils are the first peripheral immune cells to reach the damaged brain and are significantly associated with neuroinflammation after TBI [5]. Neutrophil extracellular traps (NETs) are extracellular reticular formations delivered by neutrophils that can exacerbate inflammation, autoimmune diseases, and thrombosis [[6], [7], [8]]. NETs are elevated and exacerbate neurological dysfunction in TBI patients [9]. NETs are composed of extracellular dsDNA and bind to various components, such as neutrophil elastase (NE), myeloperoxidase (MPO), histones, and cathepsin [10]. In this way, DNase I can degrade NETs [11]. S100 calcium binding protein A9 (S100a9) facilitates NET formation via AMPK/Nrf2/HO-1 signal-path after TBI [12]. S100a9 lactylated at lysine 26 (S100a9K26la), which can translocate to the nucleus and regulate transcription, is generated by glycolysis and accumulated in NETs [13]. Inhibition of peptidyl arginine deiminase4 (Padi4) activity is sufficient to hinder NET generation [14]. We revealed that NETs aggravate neuronal apoptosis as well as neuroinflammation via neddylation [15]. However, the potential regulatory interplay between NETs and cuproptosis after TBI remains unclear.

Cuproptosis, a copper-dependent cell death pathway characterized by mitochondrial oxidative stress, dysfunction, and disrupted dynamics, occurs through direct binding of excessive copper to lipoylated components of tricarboxylic acid (TCA) cycle, resulting in oligomerization and aggregation of lipoylated proteins, including dihydrolipoamide S-acetyltransferase (DLAT), and subsequent iron‒sulfur (Fe‒S) cluster protein disappearance, leading to mitochondrial dysfunction, proteotoxic stress and ultimately cell death; removal of lipoylation via ferredoxin 1 (FDX1) inhibition abolishes copper binding [16,17]. The copper chelator tetrathiomolybdate (TTM) inhibits cuproptosis-induced injury and the inflammatory response [18]. In addition to Wilson disease, cuproptosis is involved in the pathophysiological mechanisms of ischemia‒reperfusion injury and autoimmune diseases [19,20]. The deletion of the ATPase copper transporting β (Atp7b) leads to intracellular copper accumulation and cuproptosis [21]. Copper metabolism MURR1 domain-containing 1 (Commd1) maintains the stability of Atp7b and facilitates Atp7b-mediated copper excretion [22]. Atp7b and Commd1 play significant roles in intracellular copper homeostasis [23]. In addition, Circ-Spna2 alleviated neuronal cuproptosis after TBI via Keap1/Nrf2/Atp7b pathway in mice [24]. However, the role of cuproptosis in intercellular communication during secondary injury after TBI is not yet clear. In other words, which type of cell in the injury area causes neuronal cuproptosis deserves further investigation.

In the current study, we explored the role of cuproptosis in secondary brain injury after TBI, as well as the mechanisms underlying NET-related neuronal cuproptosis. Our findings aim to identify potential targets for treating TBI.

2. Materials and methods

2.1. Clinical specimens

Tissues from ten patients with mid-severe TBI and three controls were obtained from Department of Neurosurgery, Tianjin Medical University General Hospital. The clinical data, including Glasgow Coma Scale (GCS) and 6-month Glasgow Outcome Scale–Extended (GOSE) scores, are presented in Table S1.

2.2. Mice

Padi4^−/− mice (B6.Cg-Padi4^tm1.1Kmow/J, strain no. 030315) and Syn1-cre mice (B6.Cg-Tg(Syn1-cre)671Jxm/J, strain no. 003966) were derived from The Jackson Laboratory (Bar Harbor, USA); S100a9^−/− mice (C57BL/6Smoc-S100a9^em1Smoc, no. NM-KO-190143) were purchased from Shanghai Model Organisms Center, Inc. (Shanghai, CHN); Transthyretin (Ttr)-flox mice (C57BL/6JCya-Ttr^em1flox/Cya, no. S-CKO-06481) were purchased from Cyagen (Santa Clara, USA) and backcrossed for 8 generations on the C57BL/6J background. Syn1-cre, Ttr^fl/fl mice were generated by crossing the Syn1-cre and Ttr-flox mice. Experiments were performed using 8-week-old male mice. Male C57BL/6J mice (8-week–old, 22–24g) were purchased from Vital River Laboratory Animal Technology Co.,Ltd. (Beijing, CHN). All the mice were acclimated in house under a 12-h light/dark cycle in a humidity/temperature-controlled room (50–55%, 20–22°C) with food & water available ad libitum. All the mice were randomly assigned to experimental groups by a blinded study coordinator.

2.3. TBI model

TBI was accomplished via the application of a digital electromagnetic controlled cortical impact (CCI) device (eCCI6.3, Customs Design & Fabrication, Richmond, USA). Briefly, anesthetized with 1.5% isoflurane, the mice were placed in a stereotaxic head frame. A 3.5 mm diameter window on the right parietotemporal skull (2.5 mm lateral to the sagittal suture and 2.5 mm posterior from bregma) was created while keeping the dura mater intact. A 3 mm flat-tip impactor was placed directly in contact with the dura surface. The impact parameters were set as follows: 2.2 mm depth, 200 ms dwell time, and 5 m/s velocity. The mice were subsequently allowed to recover on a 37°C heating pad until they were fully awake. Mice in sham group were subjected to the same craniotomy without CCI. To investigate the effect of TTM on inhibiting cuproptosis, mice were gavaged with 30 mg/kg TTM (323446, Sigma‒Aldrich, St. Louis, USA) or control solvent for three days after TBI, as previously described [25].

2.4. Single-cell RNA sequencing data processing

Single-cell RNA sequencing data from cerebral cortical tissues after experimental TBI were obtained from the previously published dataset GSE290150, which includes three Sham_Normothermia and three TBI_Normothermia samples. Quality control of the cortical single-cell data was performed using the standard preprocessing pipeline in Seurat v5.1.0. Following normalization of the gene expression matrix, we retained cells that met the criteria of 500<nFeature<2500 and mitochondrial gene percentage<5%, thereby excluding low-quality or dying cells. DecontX was used for computational decontamination to exclude ambient RNA. Contaminated cells with scores higher than 0.2 were removed. Subsequently, we performed cell clustering, marker gene analysis, and cell-type annotation, and further extracted the neuronal subpopulation for downstream analyses. Differentially expressed genes (DEGs) between Sham and TBI neurons were defined as those with an adjusted P value < 0.05, expressed in at least 30% of cells, and exhibiting an average log2 fold change>0.3 between groups.

2.5. Fluoro-Jade C (FJC) staining

Nerve cell degeneration was detected applying FJC Ready-to-Dilute Staining Kit (TR-100-FJ, Biosensis, Thebarton, AU). Briefly, brain sections were sequentially incubated in sodium hydroxide working solution, 70% ethanol, distilled water, and potassium permanganate working solution and stained with FJC working solution and DAPI. Stained sections were observed via a fluorescence microscope.

2.6. Laser speckle contrast imaging (LSCI)

An LSCI system (PeriCamPSI System, Perimed AB, Stockholm, SWE) was used to estimate cerebral blood flow (CBF) changes. The mice were anesthetized, and their calvaria were exposed. The CBF was continuously measured for 30s with following settings: laser irradiation area, 1.6 × 1.6 cm; observation height, 10 cm. The entire cerebral cortex was selected as the region of interest.

2.7. Cell culture and treatment

Mouse neuronal HT-22 and human neuroblastoma SH-SY5Y cells (American Type Culture Collection, Manassas, USA) were cultured in Dulbecco's modified Eagle's medium (DMEM; 11995065, Thermo Fisher, Waltham, USA) supplemented with 10% fetal bovine serum (A5669401, Thermo Fisher) in a 37°C incubator with 5% CO₂. Cells were treated with 2-deoxy-d-glucose (2-DG; 5 mM, InvivoChem, Libertyville, USA), TTM (2 μM, 323446, Sigma‒Aldrich), Z-VAD-FMK (100 μM, A12373, Adooq, Irvine, USA), Deferoxamine (100 μM, A24229, Adooq), or CuCl₂ (2 μM, 203149, Sigma‒Aldrich) for different groups. For in vitro neuronal injury model, cells were treated with 100μΜ H₂O₂ for 24 h.

2.8. Chromatin immunoprecipitation (ChIP) assay

The ChIP assay was conducted using Simple ChIP@ Enzymatic Chromatin IP kit (9003, Cell Signaling Technology) following the manufacturer's instructions. Briefly, cells were harvested, cross-linked by 1% formaldehyde, digested chromatin with micrococcal nuclease, and yielded DNA fragments applying ultrasonic pulses. Chromatin was immunoprecipitated by either control IgG or S100a9K26la primary antibody. Eluted DNAs were quantified by qPCR. The primer sequences were listed below: Ttr primer (F:5′ TGTTGCTTTTGGACGGTTGC3′, R:5′ CGGAACACTCGCTCTACGAA3′).

2.9. Co-immunoprecipitation (co-IP) assay

We performed a co-IP assay with a Pierce Crosslink Magnetic Co-IP Kit (88805, Thermo Fisher) following the manufacturer's instructions. We obtained cell lysates and mixed them with antibody-coupled beads overnight at 4°C. After washing and denaturation, target proteins were collected for analysis. All antibodies used in the co-IP and subsequent immunoblotting assays are presented in Table S2.

2.10. Luciferase reporter assay

The promoter sequence of murine Ttr gene was subcloned into pGL3-basic vector. HT-22 cells were transfected with a firefly reporter plasmid and Renilla reporter plasmid, together with S100a9 WT or S100a9K26R plasmids with or without hypoxia. Luciferase reporter assay was conducted using the Dual-Luciferase Reporter Assay System Kit (E1910, Promega, Madison, USA). The ratio of firefly and Renilla luciferase reading was calculated to indicate the ability of S100a9 to activate Ttr-mediated luciferase expression.

2.11. Glutathione-S-transferase (GST) pull-down assay

The GST-Commd1 plasmid was transformed into Escherichia coli, and protein expression was generated with Isopropyl-β-D-thiogalactoside (I6758, Sigma‒Aldrich). GST-Commd1 was captured by Glutathione Sepharose beads (16105, Thermo Fisher). Different proportions of His-Ttr (CSB-MP025270MO, Cusabio, Wuhan, CHN) and Myc-Atp7b (TP518768, OriGene, Rockville, USA) were incubated with GST-Commd1 or GST together in TBST buffer with 5% glycerol for 2h at 4°C. Then, His (12698, Cell Signaling Technology), Myc (2272, Cell Signaling Technology) and GST (2625, Cell Signaling Technology) were used for immunoblotting assays.

2.12. Molecular docking

We obtained PDB files from Uniprot (https://www.uniprot.org/), performed docking using HDOCK server (http://hdock.phys.hust.edu.cn) and visualized using PyMOL software [26].

Explanations of evaluation metrics:

Docking Score: The docking scores are calculated by our knowledge-based iterative scoring function ITScorePP or ITScorePR. A more negative docking score means a more possible binding model.

Confidence Score: Given that the protein-protein/RNA/DNA complexes in the PDB normally have a docking score of around −200 or better, we have empirically defined a docking score-dependent confidence score to indicate the binding likeliness of two molecules as follows,

Confidence_score = 1.0/[1.0+e^0.02*^{(Docking_Score+150)}]

Roughly, when the confidence score is above 0.7, the two molecules would be very likely to bind; when the confidence score is between 0.5 and 0.7, the two molecules would be possible to bind; when the confidence score is below 0.5, the two molecules would be unlikely to bind.

2.13. SYTOX fluorescence assay

Neutrophils were seeded in 12-well plates for 1 h to measure the amount of extracellular DNA. HT-22 or SH-SY5Y cells were seeded in 12-well plates and stimulated for 2 days for cell death analysis. Following NET stimulation, cells were stained with SYTOX Green Nucleic Acid Stain (167 nM; S7020, Thermo Fisher) for 10 min at 37°C in the presence of the cell-impermeable DNA competitor, Dnase I (10U/mL, EN0521, Thermo Fisher, added 5 min prior to staining). The images were acquired via a fluorescence microscope (Olympus). SYTOX Green-positive cells were indicative of cell death.

2.14. Primary neuron culture

Primary cortical neurons were isolated from embryonic day 17–19 mouse embryos by papain-based enzyme (Primary Neuron Isolation Enzyme, 88280, Thermo Scientific) digestion and gentle trituration. Cells were plated on poly-d-lysine-coated dishes in serum-containing medium, switched to Neurobasal/B27 after 24h, and supplemented with B27 on day 3 in vitro. NETs were added to primary cortical neurons on day 7 in vitro.

2.15. Assessment of dsDNA, cuproptosis, Ttr and indicators for oxidative stress effects

dsDNA levels in plasma or NETs were detected using the Quant-iT PicoGreen dsDNA Assay Kit (P11496, Thermo Fisher). A copper colorimetric assay kit (ab272528, Abcam) was used for detecting copper levels in HT-22 cells or tissues as previously described [8]. HT-22 or tissue homogenate was added into wells of a 96-well plate, and reagents were added. After the incubation period, absorbance was measured at 359 nm. The copper concentration was normalized to protein concentration. Ttr (MBS2022393, MyBioSource, San Diego, USA), POLD1 (ABIN6958687, Antibodies-online, Limerick, USA), ACO-2 (ab109712, Abcam), SDHB (ABIN5674883, Antibodies-online), reactive oxygen species (ROS; EEA019, Thermo Fisher), ROS/RNS (STA-347, Cell Biolabs, San Diego, USA), malondialdehyde (MDA) content (EEA015, Thermo Fisher), glutathione peroxidase (GSH-Px) activity (EEA010, Thermo Fisher), and superoxide dismutase (SOD) activity (EIASODC, Thermo Fisher) were detected according to the steps of the instructions.

2.16. Statistical analysis

All data in this study were analyzed and expressed as the means±standard deviations (SDs), with dots applied to GraphPad Prism 9. Multiple comparisons were analyzed via one-way analysis of variance (ANOVA) with Tukey's multiple comparison test (equal SDs) or Brown Forsythe as well as Welch ANOVA test with Dunnett's T3 multiple comparisons test (unequal SDs). Two-way repeated-measures ANOVA with Geisser–Greenhouse correction and Tukey's multiple comparison test was performed to compare experiments with repeated measures. An unpaired t-test was performed to analyze the significance of two groups with equal SDs, whereas an unpaired t-test with Welch's correction was used to analyze the significance of two groups with unequal SDs. We used Pearson's correlation algorithm to evaluate the associations between two variables. P < 0.05 was considered statistically significant.

3. Results

3.1. NETs are elevated and are positively associated with poor prognosis after TBI

Citrullinated HistoneH3 (H3cit) is a marker of NET release after neutrophil activation [27]. The results of immunoblot assay revealed dramatically elevated H3cit expression levels in TBI patients compared with controls, which indicated that a large number of NETs were generated in patients with TBI (Fig. 1A and B). Neutrophils with high H3cit were widely present in brain tissue of TBI patients. We performed immunofluorescence (IF) assay and found co-localization of H3cit with extracellular DNA fibers (DAPI) and neutrophil granule proteins (MPO) in brain tissue of TBI patients (Fig. S1A). An emphatic negative correlation was uncovered between H3cit expression and GCS score (Fig. S1B). NETs were increased after TBI and were positively associated with poor prognosis. In addition, we uncovered a time-dependent enhancement in expression of H3cit compared with that in the sham samples, reaching a peak at day 3 after TBI (Fig. S1C and D).

Fig. 1.

NETs are elevated and are associated with Ttr overexpression and aggregation in neuron and neuronal loss after TBI. A Immunoblot analysis of H3cit in brain tissues from controls (n = 3) and TBI patients (n = 10). B Quantification of H3cit in brain tissues. C Representative images of NeuN (red) staining in the lesion at 3 days after TBI in sham, WT TBI, and Padi4^−/− TBI groups. D Quantitative analysis of NeuN-positive neurons. n = 6 per group. E Volcano plot showing DEGs between TBI neurons and sham neurons. F UMAP feature plot of Ttr expression in neuron. G Violin plot of Ttr between TBI and sham for neuron. H Bubble plot of the signaling pathways for changed genes of brain tissues compared from WT TBI to Padi4^−/− TBI. I Immunoblot analysis of S100a9, S100a9K26la and Ttr in sham, WT TBI, and Padi4^−/− TBI groups. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.2. NET inhibition alleviates the increase of S100a9K26la and Ttr expression in neuron after TBI

To assess effects of NET inhibition on neuronal damage after TBI, we conducted NeuN IF staining. NET inhibition definitely ameliorated neuronal loss in cerebral cortex (Fig. 1C and D). To evaluate the transcriptional gene expression changes in neurons after TBI, single-cell RNA sequencing exhibited that the most significant increase in Ttr expression was observed in neurons of TBI mice (Fig. 1E–G). Ttr dysregulated overexpression and aggregation after nerve injury promote neuronal hyperexcitability and pathology via NF-κB pathway, and neuron-specific Ttr ablation improves neuropathic pain and anxiety [28]. Besides, the results of single-cell RNA sequencing indicated that Ttr is also increased among other CNS cell types (such as astrocytes, microglia and epithelial cells) (Fig. S1E and F). Using magnetic cell sorting, we sequentially isolated mouse microglia, endothelial cells, astrocytes, and neurons [29]. Additionally, we dissected and collected the choroid plexus as previously described [30]. Then, we detected aggregation of Ttr among these CNS cell types. Immunoblot assay showed that the aggregation of Ttr in neurons is significantly increased, while it is not obvious in other CNS cell types (Fig. S1G). Therefore, overexpressed Ttr without aggregation in other CNS cell types might have little impact on secondary injury after TBI. NET inhibition intercepted glycolysis, which indicated NETs promoted glycolysis in TBI mice (Fig. 1H). Derived from glycolysis, S100a9K26la, which can promote neutrophil migration by regulating transcription, is accumulated in NETs [13]. In this way, we detected S100a9K26la and aggregated Ttr in WT and Padi4^−/− TBI mice. The results exhibited that S100a9K26la and aggregated Ttr were increased significantly after TBI, while NET inhibition relieved the increase of S100a9K26la and aggregated Ttr (Fig. 1I). In neuronal injury model, Ttr overexpression resulted in Ttr aggregation and aggravated neuronal death (Fig. S2A–C). We suppose that S100a9K26la delivered from NETs results in neuronal death by promoting Ttr aggregation, which is facilitated by glycolysis.

3.3. Neuron−specific Ttr conditional knockout attenuates neuronal death and microglial inflammatory activity after TBI

Deng et al. have demonstrated that Ttr dysregulated overexpression and aggregation after nerve injury promote neuronal hyperexcitability and pathology, and neuron-specific Ttr ablation improves neuropathic pain and anxiety [28]. To investigate the function of Ttr in neuron after TBI, Syn1-cre mice were crossed with Ttr-flox mice to generate a neuron−specific Ttr deletion. Considering the major primary sources of Ttr in the brain are choroid plexus epithelial cells and blood, so neuron-specific Ttr knockout cannot affect the Ttr from the choroid plexus and serum leaking to brain parenchyma after TBI [31]. We detected Ttr protein levels in the cerebrospinal fluid of conditional knockout and control mice in sham and TBI model. The results showed that the Ttr protein levels in the cerebrospinal fluid had no significant difference in different groups (Fig. S2D). In this way, we performed immunoblot assay and discovered that the Ttr in the cerebrospinal fluid was not aggregated, which suggested the Ttr in the cerebrospinal fluid was unlikely to be the major contributor to neuronal pathology (Fig. S2E). To determine whether neuronal Ttr protein reaches functional levels, we performed ELISA to measure total Ttr protein in neurons, choroid plexus, and liver, with values normalized to total protein concentration. Ttr protein levels were markedly increased in neurons after TBI compared with those in the sham group and reached levels comparable to those detected in the choroid plexus and liver under the same normalization conditions (Fig. S2F). These results provide protein-level evidence that TBI induces a substantial accumulation of Ttr within neurons, suggesting that the post TBI increase in neuronal Ttr may be functionally relevant. Importantly, neuron-specific Ttr knockout markedly attenuated TBI-induced Ttr aggregation in neurons, although a small amount of nonaggregated Ttr remained detectable (Fig. S2G). The persistence of this residual nonaggregated Ttr is consistent with the possibility that a limited amount of Ttr from non-neuronal or circulating sources may still be present in the neuron. However, this residual Ttr was insufficient to sustain the pathological Ttr aggregation observed after TBI. More importantly, co-IP assay verified that neuron specific Ttr knockout rescued the TBI elevated Ttr-Commd1 interaction in neurons (Fig. S2H). Despite the potential presence of a small amount of Ttr from non-neuronal sources after BBB disruption, depletion of neuron-derived Ttr substantially abolished both pathological Ttr aggregation and its interaction with Commd1. Collectively, these findings support neuron-derived Ttr as a major functionally relevant pool responsible for pathological Ttr aggregation and enhanced Ttr–Commd1 interaction after TBI, although a minor contribution from circulating or other non-neuronal Ttr cannot be completely excluded. Magnetic resonance imaging (MRI) scanning was performed to gauge the volume of brain edema. The results exhibited that Syn1-cre, Ttr^fl/fl mice mitigated brain edema prominently compared with Ttr^fl/fl mice at 3-day post TBI (Fig. 2A and B). FJC staining revealed that FJC-positive neurons were prominently fewer in Syn1-cre, Ttr^fl/fl TBI group than in Ttr^fl/fl TBI group. (Fig. 2C and D). Although neuron-specific Ttr deletion did not alter CSF Ttr levels, it reduced Ttr aggregation in the injured brain and attenuated neuronal injury, supporting a pathogenic contribution of neuronal Ttr aggregation after TBI. Next, we explored effects of neuron−specific Ttr conditional knockout on microglial activation and immune cell infiltration applying double IF staining. Ttr conditional knockout alleviated the increase in Iba-1⁺ microglia and CD16/32⁺ proinflammatory microglia/macrophages at 3 days post TBI, indicating that Ttr deficiency in neuron mitigated immune cell infiltration and microglial proinflammatory activation after TBI (Fig. 2E–G). Moreover, Ttr conditional knockout increased the percent of Arginase-1⁺ anti-inflammatory microglia after TBI (Fig. 2H and I). These results indicate that Ttr deficiency in neuron, which inhibits Ttr dysregulated overexpression and aggregation can improve neuronal death and neuroinflammation following TBI.

Fig. 2.

Effects of neuron−specific Ttr conditional knockout on neuronal death and neuroinflammation post-TBI. A Representative images of MRI scans in sham, Ttr^fl/fl TBI and Syn1-cre, Ttr^fl/fl TBI groups. B Quantitative analysis of brain edema volume. n = 6 per group. C Representative images of FJC (green) staining in sham, Ttr^fl/fl TBI and Syn1-cre, Ttr^fl/fl TBI groups. Scale bar = 100 μm. D Quantitative analysis of FJC-positive neurons. n = 6 per group. E Representative images of Iba-1 (green) colocalization with CD16/32 (red) in sham, Ttr^fl/fl TBI and Syn1-cre, Ttr^fl/fl TBI groups. Scale bar = 100 μm. F, G Quantitative analysis of Iba1⁺ cells (F) and Iba1⁺/CD16/32⁺ microglia (G). n = 6 per group. H Representative images of Iba-1 (green) colocalization with Arginase-1 (red) in sham, Ttr^fl/fl TBI and Syn1-cre, Ttr^fl/fl TBI groups. Scale bar = 100 μm. I Quantitative analysis of Iba1⁺/Arginase-1⁺ microglia. n = 6 per group. Nuclei were stained with DAPI (blue). ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.4. Ttr deficiency in neuron alleviates lesion size and BBB leakage

To evaluate the impact of neuron−specific Ttr conditional knockout on lesion size post TBI, we performed hematoxylin and eosin (H&E) staining on serial coronal sections. Results exhibited that neuron−specific Ttr conditional knockout reduced tissue loss after TBI (Fig. S3A and B). BBB destruction is an incipient incident of TBI with a long-term impact, which contributes to critical neurological impairment, cognitive deficits, and posttraumatic epilepsy [[32], [33], [34]]. Evans blue (EB) dye extravasation assay was performed to detect BBB permeability [15]. EB dye leakage into the brain parenchyma and EB extravasation were prominently increased in the TBI mice at day3 when compared with sham mice (Fig. S3C and D). Syn1-cre, Ttr^fl/fl TBI mice exhibited dramatically decreased EB dye leakage into the brain parenchyma and EB extravasation than did the Ttr^fl/fl TBI samples (Fig. S3C and D). We examined the effects of Ttr deficiency in neuron on cerebrovascular endothelial dysfunction 3-day after TBI using double IF staining. Our findings revealed that ZO-1⁺ vessels in contused cortex dramatically decreased 3 days post TBI (Fig. S3E and F). Interestingly, Syn1-cre, Ttr^fl/fl TBI mice alleviated ZO-1⁺ vessel decrease 3-day post-injury (Fig. S3E and F). The results of LSCI machinery exhibited that the CBF decreased dramatically from baseline 3 days after TBI (Fig. S3G and H). Ttr conditional knockout prominently improved CBF compared to that in TBI mice at 3 days post TBI (Fig. S3G and H). In summary, neuron−specific Ttr conditional knockout attenuates BBB dysfunction in mice after TBI.

3.5. Ttr deficiency in neuron ameliorates neurological deficits after TBI

To determine the ability of NET inhibition to promote short-term neurological function recovery after TBI, we performed several tests, including mNSS, rotarod, as well as corner tests. On days 3&7 post-TBI, Syn1-cre, Ttr^fl/fl TBI mice demonstrated dramatically less motor, balance and reflection dysfunction than did Ttr^fl/fl TBI group (Fig. S4A). In the rotarod test, compared with Ttr^fl/fl TBI group, Syn1-cre, Ttr^fl/fl TBI group displayed prominently enhanced motor coordination on the 3rd and 7th day (Fig. S4B). The results of corner test revealed that Ttr conditional knockout alleviated sensory‒motor dysfunction in mice at 3&7 days post TBI (Fig. S4C). MWM test indicated that Ttr^fl/fl TBI samples exhibited impaired spatial learning and memory capacity compared to sham and Syn1-cre, Ttr^fl/fl TBI samples, as evidenced via their swimming paths and heatmaps (Fig. S4D). During training phase, mice displayed a training-related reduction in escape latency, while Ttr^fl/fl TBI samples exhibited the longest escape latency among all groups (Fig. S4E). In probe test, TBI mice exhibited a remarkable increase in the first latency time and a dominant decrease in platform-crossing frequency as well as time spent in target quadrant compared with sham mice (Fig. S4F–H). Interestingly, Syn1-cre, Ttr^fl/fl TBI mice showed a significant decrease in the first latency time and noteworthy increase in platform-crossing frequencies as well as time spent in target quadrant, compared to Ttr^fl/fl TBI mice (Fig. S4F–H). No significant differences were detected in average swim speed among three groups, which demonstrated that the results were not affected by motor ability (Fig. S4I). Our findings manifest that Ttr deficiency in neuron ameliorates neurological deficits after TBI.

3.6. NET-carried S100a9K26la elevates Ttr expression, and Ttr interacted with Commd1 in neuron

To assess the considerable role of S100a9K26la in Ttr transcription, we extracted NETs from WT or S100a9^−/− mice. 2-DG was applied to inhibit glycolysis. S100a9K26la was accumulated in NETs and inhibited by 2-DG (Fig. 3A). S100a9 functions not only as a secreted cytokine but also as a nuclear transcriptional coactivator, recruited to chromatin through interactions with DNA-binding transcription factors in the nucleus [35]. Wang et al. have reported that Importin β1 can translocate lactylated S100a9 to the nucleus and regulate transcription [13]. We performed co-IP assay and found that S100a9K26la was bound to Importin β1 in HT-22 cells (Fig. S2I). Nucleocytoplasmic fractionation showed that the Importin β1 inhibitor blocked the translocation of S100a9K26la into nuclei, leading to a decrease in nuclear S100a9K26la (Fig. S2J). ChIP confirmed that S100a9K26la was associated with the Ttr promoter (Fig. 3B). NETs delivered S100a9K26la to neurons and promoted Ttr overexpression (Fig. 3C). Besides, NETs without S100a9K26la could not increase Ttr expression in neuron (Fig. 3C). We observed significantly decreased Ttr expression in S100a9^−/− TBI mice compared with WT TBI mice (Fig. 3D and E). Hypoxia promoted the lactylation of S100a9 and increase of Ttr expression, while S100a9 with a lysine-to-arginine mutation at K26 residue could not promote Ttr overexpression (Fig. S2K). Luciferase reporter assays confirmed that S100a9K26la promoted Ttr transcription, while S100a9K26R or normoxia could not promote Ttr transcription (Fig. 3F).

Fig. 3.

NET-carried S100a9K26la elevates Ttr expression, and Ttr interacted with Commd1 in neuron. A Immunoblot analysis of S100a9K26la and S100a9 in NETs derived from WT or S100a9^−/− mice. 2-DG was applied to inhibit glycolysis. B The association between S100a9K26la and Ttr promoter in HT-22 cells was confirmed by ChIP-qPCR. n = 6 per group. C Immunoblot analysis of Ttr in HT-22 cells treated by NETs derived from WT or S100a9^−/− mice. 2-DG was applied to inhibit glycolysis. D Immunoblot analysis of S100a9, S100a9K26la and Ttr in sham, WT TBI, and S100a9^−/− TBI groups. E The quantification of Ttr in brain tissues. n = 6 per group. F Luciferase reporter assays showing Ttr transcriptional activity in S100a9 WT or S100a9K26R groups with or without hypoxia. n = 6 per group. G Binding model depicting the interaction between Ttr and Commd1. Docking Score = −222.59; Confidence Score = 0.8103. Ttr is colored in orange, and Commd1 is colored in cyan. H Co-IP analysis showing the interaction between Ttr and Commd1 in HT-22 cells. I Co-IP analysis showing the interaction between Ttr and Commd1 in SH-SY5Y cells. J Co-IP assays were implemented with anti-Flag on HT-22 cell lysates transfected with HA-Ttr, along with the indicated Flag-Commd1 constructs. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

To investigate the molecular mechanisms by which Ttr induce neuronal death, we performed co-IP coupled with mass spectrometry to identify the binding partners of Ttr. Commd1, a regulator of copper homeostasis, was highly abundant in Ttr antibody-enriched samples. The molecular docking model (http://hdock.phys.hust.edu.cn) revealed potential binding sites between Ttr and Commd1 (Fig. 3G) [26]. Co-IP revealed an intense interaction between Ttr and Commd1 in neurons (Fig. 3H and I). Based on this model, we constructed three Flag-tagged Commd1 mutants. Co-IP experiments indicated that the W123 residue of Commd1 was essential for its interaction with Ttr (Fig. 3J).

3.7. NETs contribute to mitochondrial oxidative stress and neuronal cuproptosis in vitro

To further investigate the mechanism of NET-mediated neuronal death, Gene Set Enrichment Analysis (GSEA) of changed genes of brain tissues compared from WT TBI to Padi4^−/− TBI exhibited that NETs promoted oxidative stress after TBI (Fig. 4A). To further validate NETs as mitochondrial oxidative stress contributors, HT-22 and SH-SY5Y cells were treated with 0.5 μg/mL, 1 μg/mL, or 2 μg/mL (DNA content) NETs or with control media as previously described [36]. Following TBI, abundant NETs, composed largely of dsDNA, are generated within the injured brain and released into systemic circulation, with plasma dsDNA serving as a surrogate marker for local cerebral NET levels [11]. Plasma dsDNA levels in our TBI cohort ranged approximately from 0.5 to 2 μg/mL (mean, 962.18 ng/mL; Fig. 6G), providing a clinically relevant reference range for extracellular DNA exposure. Therefore, the concentration we used is reasonably representative of the pathological environment. Intracellular ROS levels were found to be elevated in HT-22 and SH-SY5Y cells treated with different concentrations of NETs compared with the control group (Fig. S5A and E). Biochemical assays exhibited that the content of lipid peroxidation marker MDA increased with increasing treatment concentration, while activities of antioxidant enzymes GSH-Px and SOD decreased with increasing treatment concentration in NET-treated HT-22 and SH-SY5Y cells compared to the control group (Fig. S5B-D, F-H). Cuproptosis is a cell death pathway characterized by mitochondrial oxidative stress [17]. We investigated lipoylated-DLAT aggregation and Fe-S cluster protein destabilization to detect cuproptosis in neuron. The Fe‒S cluster proteins include the DNA polymerase δ catalytic subunit (POLD1), ACO-2 and Succinate dehydrogenase (ubiquinone) iron‒sulfur subunit, mitochondrial (SDHB). NETs at 1 μg/mL or 2 μg/mL increased DLAT aggregation levels and decreased POLD1, ACO-2 and SDHB expression levels, indicating that NETs promoted neuronal cuproptosis (Fig. S5I–R). It is proposed that NETs activate the cuproptosis pathway, inducing mitochondrial oxidative stress in HT-22 and SH-SY5Y cells. Furthermore, we detected key defining features of cuproptosis. For copper flux, we detected copper levels in neuronal injury model from different groups as previously described [24]. The results showed that intracellular copper ion levels were elevated in neurons overexpressed Ttr compared with the control group, which was inhibited by Commd1 overexpression (Fig. S6C). For FDX1 dependence, we knocked down FDX1 and found that FDX1 knockdown significantly rescued cell death induced by Ttr overexpression in neuronal injury model, placing the mechanism upstream of this key cuproptosis node (Fig. S6D and E). To confirm specificity, we utilized inhibitors of apoptosis (Z-VAD-FMK) and ferroptosis (Deferoxamine) [37,38]. Neither inhibitor provided protection against Ttr overexpression in neuronal injury model, whereas the copper-specific chelator TTM did (Fig. S6F and G). Considering TTM is a broad chelator with anti-inflammatory and antioxidant effects, we used CuCl₂ to verify the direct effect of cuproptosis. The results showed that supplementation with copper (in the form of CuCl₂) is sufficient to reverse the protective effect of TTM, demonstrating that TTM's protection is specifically due to copper chelation (Fig. S6F and G) [16].

Fig. 4.

NETs disrupt the interaction between Atp7b and Commd1. A GSEA of changed genes of brain tissues compared from WT TBI to Padi4^−/− TBI. B Binding model depicting the interaction between Atp7b and Commd1. Docking Score = −324.75; Confidence Score = 0.9705. Atp7b is colored magenta, and Commd1 is colored cyan. C-E Co-IP analysis and quantification showing impact of NETs on the interaction between Atp7b and Commd1 in HT-22 cells. n = 3 per group. F-H Co-IP analysis and quantification showing impact of NETs on the interaction between Atp7b and Commd1 in SH-SY5Y cells. n = 3 per group. I Co-IP assays were implemented with anti-Flag on HT-22 cell lysates transfected with Myc-Atp7b, along with the Flag-Commd1 WT or Flag-Commd1 W123A. J-K Co-IP analysis and quantification detecting the interaction between Atp7b and Commd1 in control, NETs + sh-control, and NETs + sh-Ttr groups in HT-22 cells. n = 3 per group. L-M Co-IP analysis and quantification detecting the interaction between Atp7b and Commd1 in control, NETs + sh-control, and NETs + sh-Ttr groups in SH-SY5Y cells. n = 3 per group. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

Fig. 6.

Neuronal cuproptosis is increased and associated with NETs after TBI. A Bubble plot of the signaling pathways for changed proteins of brain tissues compared from TBI patient to control. B Immunoblot analysis of DLAT, POLD1, ACO-2, and SDHB in brain tissues from controls (n = 3) and TBI patients (n = 10). C Correlation between copper concentration and GOSE score (n = 20). D Correlation between POLD1 expression and GOSE score (n = 20). E Correlation between ACO-2 expression and GOSE score (n = 20). F Correlation between SDHB expression and GOSE score (n = 20). G Correlation between plasma dsDNA and intracellular copper in controls (n = 10) and TBI patients (n = 10). H Correlation between plasma dsDNA and plasma ROS/RNS in controls (n = 10) and TBI patients (n = 10). I Correlation between intracellular copper and plasma ROS/RNS in controls (n = 10) and TBI patients (n = 10). J Representative IHC images of ACO-2 in TBI tissues. Scale bar = 100 μm. K Graphic representation of scoring performed on IHC staining for ACO-2 in brain tissues of controls (n = 3) and patients with TBI (n = 10). L Immunoblot images of the time course of DLAT, POLD1, ACO-2, and SDHB after TBI. M Immunoblot analysis of cuproptosis in sham, TBI + vehicle, and TBI + TTM groups. N Immunoblot analysis of DLAT, POLD1, ACO-2, and SDHB in sham, WT TBI, and Padi4^−/− TBI groups. O, P Quantification of DLAT oligomers (O) and POLD1 (P). n = 6 per group. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.8. NETs obstruct the interaction between Atp7b and Commd1 via Ttr

The process of cellular copper efflux from cell requires participation of Atp7b transporters [39]. The combination of Commd1 with Atp7b is involved in copper efflux and regulates copper homeostasis [40,41]. The docking results revealed a high likelihood connection between Atp7b and Commd1 (Fig. 4B). To investigate the molecular mechanisms by which NETs induce neuronal cuproptosis, we performed a co-IP assay using HT-22 and SH-SY5Y cells. The results revealed a decreased interaction between Atp7b and Commd1 in NET-treated cells (Fig. 4C–H). In this way, NETs disturbed Commd1-related copper efflux via Atp7b and induced copper accumulation and cuproptosis. In addition, co-IP experiments indicated that the W123 residue of Commd1 was essential for its interaction with Atp7b, which indicated Ttr competes for the interaction between Atp7b and Commd1 W123 residue (Fig. 4I). We transfected plasmids overexpressing WT Commd1 and a Commd1 mutant with a modified W123 site to rescue Ttr overexpression-mediated cell death and found a successful rescue with WT Commd1, but not the W123A mutant (Fig. S6H and I). To investigate whether Ttr and Atp7b directly compete for binding to Commd1, we performed GST pull-down assays using purified recombinant proteins. When GST-Commd1 was incubated individually with either His-Ttr or Myc-Atp7b, each protein was pulled down, whereas the W123A mutant of Commd1 showed markedly reduced pull-down of both proteins, confirming direct binding of each protein to Commd1 through the W123 residue (Fig. S2L and M). When GST-Commd1 was incubated with increasing molar ratios of Ttr to Myc-Atp7b, the amount of Myc-Atp7b pulled down decreased in a dose-dependent manner, demonstrating direct competitive displacement of Myc-Atp7b by Ttr for binding to Commd1. (Fig. S2N). Knockdown of Ttr could alleviate interaction between Atp7b and Commd1 obstructed by Ttr in neurons (Fig. 4J–M). To confirm the interaction under physiological conditions, we conducted IF assay in mouse brain tissues obtained from sham, Ttr^fl/fl TBI and Syn1-cre, Ttr^fl/fl TBI groups. The results revealed a decreased interaction between Atp7b and Commd1 in neurons of Ttr^fl/fl TBI group, which was restored by neuron−specific Ttr conditional knockout (Fig. S6A and B). Thus, NETs increase Ttr expression and result in neuronal cuproptosis by reducing the interaction between Atp7b and Commd1.

3.9. Knockdown of Ttr alleviates neuronal cuproptosis resulting from NETs

Inhibition of Ttr alleviated NET-mediated enhancement in aggregation of DLAT, and reduction in expression of POLD1, ACO-2 and SDHB in HT-22 and SH-SY5Y cells, which demonstrated that NETs resulted in neuronal cuproptosis via Ttr overexpression (Fig. 5A–E; Fig. S7A–E). The ROS level and MDA content were elevated, while GSH-Px activity as well as SOD content were decreased in the NETs + sh-control groups compared to the control group (Fig. 5F–I; Fig. S7F–I). Knockdown of Ttr inhibited NET-mediated mitochondrial oxidative stress (Fig. 5F–I; Fig. S7F–I). SYTOX Green fluorescence assay further detected a remarkable increase in cell death at NET concentration of 1 μg/mL, indicated by presence of SYTOX⁺ cells (Fig. 5J and K; Fig. S7J and K). In addition, inhibition of Ttr alleviated NET-mediated neuronal death (Fig. 5J and K; Fig. S7J and K). The CCK-8 assay revealed that NETs inhibited the proliferative activity of HT-22 and SH-SY5Y cells, which was rescued by knockdown of Ttr (Fig. 5L; Fig. S7L). To recapitulate the major findings from cell lines and further corroborate the physiological relevance of the S100a9K26la-Ttr-Commd1-Atp7b pathway, we cultured primary neurons and treated them with NETs. Exposure to NETs derived from WT mice resulted in accumulation of S100a9K26la and increased Ttr transcription/protein expression in primary neurons, whereas exposure to NETs derived from S100a9-knockout mice did not (Fig. S7M and N). Ttr accumulation reduced Commd1-Atp7b interaction, resulting in DLAT oligomerization and loss of POLD1, ACO-2, and SDHB (Fig. S7M and O). In summary, knockdown of Ttr ameliorates NET-mediated mitochondrial oxidative stress and neuronal cuproptosis.

Fig. 5.

NETs result in Ttr-mediated cuproptosis and mitochondrial oxidative stress of HT-22 cells. A-E Immunoblot analysis of DLAT oligomers (B), POLD1 (C), ACO-2 (D), and SDHB (E) expression in control, NETs + sh-control, and NETs + sh-Ttr groups. n = 3 per group. F-I Biochemical kits to detect ROS (F), MDA content (G), GSH-Px activity (H), and SOD content (I) in HT-22 cells. n = 3 per group. J Representative images of SYTOX (green) staining for different groups of HT-22 cells. Scale bar = 50 μm. K Quantitative analysis of SYTOX positive cell proportion among different groups. n = 3 per group. L Growth curve measured by CCK-8 assay. n = 3 per group. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.10. Neuronal cuproptosis occurs in both TBI patients and mice

To investigate the molecular mechanisms underlying secondary injury in patients with TBI, quantitative proteomic and bioinformatics analyses were conducted on brain tissues from three TBI patients and three control subjects. The results revealed that proteins upregulated in TBI patients were significantly enriched in pathways associated with NET formation, stress response to copper ion, ROS process, as well as copper homeostasis (Fig. 6A). We investigated lipoylated-DLAT aggregation and Fe-S cluster protein destabilization to detect cuproptosis in brain tissues of patients with TBI. We observed dramatically elevated DLAT aggregation levels and markedly degraded POLD1, ACO-2 and SDHB expression levels in TBI patients compared with controls, which indicated that cuproptosis occurred in patients with TBI (Fig. 6B; Fig. S8A–D). Correlation analysis displayed that DLAT aggregation was positively correlated with H3cit expression levels (Fig. S8E). In addition, GCS scores showed a marginally significant negative correlation with DLAT aggregation levels, but a significant positive correlation with POLD1, ACO-2, and SDHB expression levels (Fig. S8F–I). Given that the GCS primarily reflects acute injury severity and should not be interpreted as a measure of long-term functional outcome, we further examined the associations between cuproptosis-related markers and 6-month GOSE scores. Notably, similar correlation patterns were observed: 6-month GOSE scores showed a marginally significant negative correlation with DLAT aggregation, and moderate-to-strong positive correlations with POLD1, ACO-2, and SDHB expression (Fig. S8J–M). To overcome the power limitation and enable convenient, simultaneous detection of a larger number of samples, we expanded our cohort to include 20 TBI samples (10 overlapping with the pilot cohort and 10 newly recruited) and 20 control samples (3 overlapping and 17 newly recruited). According to a recent report by Du et al. [24], copper ion concentrations, along with the expression levels of POLD1, ACO-2 and SDHB, are commonly assessed as cuproptosis-related biomarkers in TBI studies. In this expanded cohort, copper concentrations were quantified using a copper colorimetric assay kit, whereas POLD1, ACO2, and SDHB levels were determined by ELISA (DLAT aggregation was excluded, as it requires immunoblotting to detect oligomeric forms and is not amenable to large-scale ELISA-based measurements). T-tests revealed that TBI tissues exhibited significantly elevated copper concentrations and decreased POLD1/ACO-2/SDHB levels (Fig. S8N–Q). Notably, correlation analysis demonstrated that copper ion concentration was negatively correlated with 6-month GOSE scores, whereas POLD1, ACO-2 and SDHB levels showed a positive correlation (Fig. 6C–F). In other words, a higher level of cuproptosis predicted a worse prognosis. We detected intracellular copper levels, plasma dsDNA (NET marker) and ROS/RNS (oxidative stress marker) levels in TBI patients and controls. Correlation analysis exhibited that plasma dsDNA levels were positively correlated with intracellular copper and plasma ROS/RNS levels (Fig. 6G and H). Similarly, copper levels were also positively correlated with ROS/RNS levels (Fig. 6I). Intracellular copper ion levels were elevated in HT-22 cells treated with NETs compared with the control group (Fig. S8R). IHC staining revealed an emphatic decrease in ACO-2⁺ neurons in TBI patients compared with controls (Fig. 6J and K). In this way, NETs were increased after TBI and were positively associated with cuproptosis and oxidative stress in neuron.

We subjected TBI model and obtained brain cortex samples at 6 h and at 1, 3, 5, and 7 days post-injury. Our findings manifested a time-dependent enhancement in aggregation of DLAT and reduction in expression of POLD1, ACO-2 and SDHB compared with those in the sham samples (Fig. 6L; Fig. S8S–V). These results indicate that neuronal cuproptosis occurs in both TBI patients and mice. To detect the tighter temporal validation, we detected NET and cuproptosis marker levels in brain cortex samples at 2, 3, 4, and 5 days post-injury. The results showed that NET formation had peaked at 3 days and cuproptosis had peaked at 4 days after TBI, which indicated NETs induced cuproptosis after TBI (Fig. S8W). On the basis of these results, we focused on investigating TBI treatment options targeting for neuronal cuproptosis. We restrained cuproptosis induction applying the copper chelator TTM. Administration of TTM to TBI mice inhibited cuproptosis of perilesional brain (Fig. 6M; Fig. S8X–AA). In addition, TTM alleviated the decrease in ACO-2 in neurons after TBI (Fig. 7A and B). To examine whether NETs trigger cuproptosis, we developed a Padi4^−/− TBI model to constrict NET formation. The inhibition of NETs attenuated the increase in DLAT aggregation levels and decrease in POLD1, ACO-2 and SDHB expression levels in TBI mice, which indicated that NETs resulted in cuproptosis in mice after TBI (Fig. 6N–P; Fig. S8AB and AC). In addition, NET inhibition alleviated the decrease in ACO-2 in neurons after TBI (Fig. S9A and B). Thus, NETs promote neuronal cuproptosis in mice after TBI.

Fig. 7.

Effects of TTM treatment on neuronal death and neuroinflammation after TBI. A Representative images of NeuN (red) colocalization with ACO-2 (green) at 3-day after TBI. Scale bar = 100 μm. B Quantitative analysis of ACO-2⁺/NeuN⁺ cells. n = 6 per group. C Representative images of MRI scans. D Quantitative analysis of brain edema volume. n = 6 per group. E Representative images of FJC (green) staining. Scale bar = 100 μm. F Quantitative analysis of FJC-positive neurons. n = 6 per group. G Representative images of Iba-1 (green) colocalization with CD16/32 (red). Scale bar = 100 μm. H, I Quantitative analysis of Iba1⁺ cells (H) and Iba1⁺/CD16/32⁺ microglia (I). n = 6 per group. Nuclei were stained with DAPI (blue). ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.11. TTM alleviates neuronal death and microglial inflammatory activity after TBI

TBI induced noticeable brain edema on day 3 (Fig. 7C and D). TTM-treated TBI mice displayed a meaningful decrease in brain edema compared to those in TBI + vehicle mice at 3-day post-injury (Fig. 7C and D). To assess effects of TTM on neuronal death after TBI, NeuN IF and FJC staining were conducted. TTM treatment obviously ameliorated neuronal loss in cerebral cortex (Fig. S9C and D). FJC assay revealed that FJC⁺ neurons were markedly increased in cerebral cortex after TBI (Fig. 7E and F). TTM administration reversed the increase in FJC-positive neurons (Fig. 7E and F).

Next, we explored effects of TTM on microglial activation and immune cell infiltration applying double IF staining. The results exhibited that TTM mitigated the increase in Iba-1⁺ microglia and CD16/32-positive proinflammatory microglia at 3-day post TBI, which manifested that TTM alleviated immune cell infiltration and microglial proinflammatory activation after TBI (Fig. 7G–I). Moreover, TTM increased percent of Arginase-1⁺ anti-inflammatory microglia after TBI (Fig. S9E and F). These results indicate that the inhibition of cuproptosis applying TTM can alleviate neuronal death and neuroinflammation following TBI.

3.12. TTM alleviates lesion size and BBB dysfunction

To evaluate the impact of TTM on lesion size post TBI, we performed H&E staining on serial coronal sections. Results showed that TTM reduced tissue loss after TBI (Fig. 8A and B). EB dye extravasation assay was performed to detect BBB permeability. EB dye leakage into the brain parenchyma and EB extravasation were prominently greater in the TBI mice than in sham mice on day 3 (Fig. 8C and D). Compared with TBI + vehicle-treated mice, TTM-treated TBI mice exhibited dramatically decreased EB dye leakage into the brain parenchyma and EB extravasation (Fig. 8C and D). We examined the effects of TTM on cerebrovascular endothelial dysfunction 3-day after TBI using double IF staining. Our findings revealed that tight junction protein ZO-1⁺ vessels in contused cortex dramatically decreased 3 days post TBI (Fig. 8E and F). Interestingly, TTM alleviated the decrease in ZO-1⁺ vessel density in the TBI mice at 3-day post-injury (Fig. 8E and F). LSCI machinery was applied to appraise CBF in the cortex. The CBF decreased significantly from baseline 3-day after TBI (Fig. 8G and H). TTM treatment prominently improved CBF compared to that in TBI + vehicle mice at 3 days post TBI (Fig. 8G and H). In summary, inhibition of cuproptosis using TTM attenuated BBB dysfunction in mice after TBI.

Fig. 8.

Effects of TTM treatment on cerebral lesion volume and BBB disruption following TBI. A Representative images of coronal sections exhibiting cerebral loss at 21 days after TBI. B Quantitative analysis of lesion volume. n = 6 per group. C Topical and coronal images of representative brains after EB injection at 3-day post-TBI. D Quantitative analysis of EB leakage. n = 6 per group. E Representative images of IF staining of CD31 (red) and ZO-1 (green). Scale bar = 100 μm. F Quantitative analysis of ZO-1 expression of CD31⁺ vessels. n = 6 per group. G Representative images of CBF at 3-day after TBI among different groups. H Quantitative analysis of CBF among different groups. n = 6 per group. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SDs.

3.13. TTM ameliorates neurological deficits after TBI

To determine the ability of TTM treatment to promote short-term neurological function recovery after TBI, we performed several tests, including mNSS, rotarod, as well as corner tests. On days 3&7 post-TBI, TTM-treated mice demonstrated dramatically less motor, balance and reflection dysfunction than did TBI + vehicle group (Fig. 9A). In the rotarod test, compared with TBI + vehicle group, TBI + TTM group exhibited markedly enhanced motor coordination on the 3rd & 7th day (Fig. 9B). The results of corner test revealed that TTM treatment alleviated sensory‒motor dysfunction in mice at 3 and 7 days post TBI (Fig. 9C). MWM test, conducted on days 15-20 post-TBI, indicated that TBI + vehicle samples exhibited impaired spatial learning and memory capacity compared to sham and TBI + TTM samples, as evidenced via their swimming paths and heatmaps (Fig. 9D). During training phase, mice displayed a training-related reduction in escape latency, whereas the TBI + vehicle samples exhibited a significantly longer escape latency among all groups (Fig. 9E). In probe test, TBI mice exhibited a remarkable increase in the first latency time and a dominant decrease in platform-crossing frequency as well as time spent in target quadrant compared with sham mice (Fig. 9F–H). Interestingly, TBI mice treated by TTM showed a significant decrease in the first latency time and a noteworthy increase in platform-crossing frequency as well as time spent in target quadrant, compared to TBI + vehicle mice (Fig. 9F–H). No significant differences were detected in average swimming speed among three groups, which indicated that the results were not affected by motor ability (Fig. 9I). Our findings manifest that cuproptosis inhibitor TTM alleviates neurological deficits after TBI.

Fig. 9.

Effects of TTM on neurological functions after TBI. A-C mNSS score (A), rotarod test (B), and corner test (C) at 1, 3, 7, 14 days post-TBI in sham, TBI + vehicle and TBI + TTM groups. n = 6 per group. D Representative swimming tracks and heatmap of the mice in the MWM test. E Latency to platform in MWM training. n = 6 per group. F-I Latency to the platform (F), number of crossings (G), time spent in target quadrant (H), and swimming speed (I) in the MWM test. n = 6 per group. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are presented as the means ± SD.

4. Discussion

TBI triggers a cascade of secondary injury mechanisms, including neuroinflammation, oxidative stress, and blood–brain barrier disruption, yet the interplay between specific cell death pathways remains incompletely understood [42]. At present, surgical intervention is the main treatment option for addressing the initial stage, which involves mechanical damage to brain tissue [43]. Recognizing the involvement of dissimilar cell death pathways has driven efforts to develop innovative therapeutic strategies; however, their success is limited. This finding suggests that additional mechanisms yet to be deciphered during the pathogenesis of TBI may lead to secondary damage [44]. Considering copper is a cofactor for multifarious enzymes, the steady-state regulation of copper in cells is crucial. In addition, copper is involved in regulating many other cellular activities, including cell respiration, growth regulation and neurotransmitter and hormone synthesis [45,46]. Cuproptosis, a copper-dependent cell death modality driven by lipoylated protein aggregation and Fe–S cluster protein loss, has recently been implicated in neurological disorders; however, its role in TBI pathophysiology has not been explored [16]. Exploring the mechanism of cuproptosis after TBI can contribute to the prevention and treatment of secondary injuries. Neutrophils are primitive responders to invading pathogens or acute inflammatory stimuli. Concurrently, NETs are known to exacerbate post-TBI neuropathology; however, whether NETs directly modulate neuronal cuproptosis has not been reported [9].

In the current study, we confirmed that NETs resulted in neuronal death after TBI. S100a9K26la, which is generated by glycolysis and accumulated in NETs, can translocate to the nucleus and regulate transcription [13]. Single-cell RNA sequencing exhibited that the most significant increase in Ttr expression was observed in neurons of TBI mice. Within the nervous system, Ttr paradoxically exhibits both neuroprotective properties and neuropathologic processing [47]. Li et al. have demonstrated that the interaction between Ttr and Aβ reduces Aβ concentration and aggregation, which decreases Aβ toxicity in the neuron [48]. Silva et al. have discovered that Ttr proteolytic activity is required for the neuroprotective effect for Alzheimer's disease, while proteolytically inactive Ttr mutant H90A has no impact on Aβ fibrillization [49]. In this way, Ttr with interactive and proteolytic activity inhibits Aβ aggregation and fibrillization. On the contrary, Deng et al. have demonstrated that dysregulated overexpression and aggregation of Ttr after nerve injury promote neuronal hyperexcitability and pathology via NF-κB pathway, and neuron-specific Ttr ablation improves neuropathic pain and anxiety [28]. Aberrant accumulation of Ttr in the peripheral nervous system can lead to transthyretin amyloid polyneuropathy accompanied by neuropathic pain [50]. Protein misfolding and aggregation are processes that can generate because of environmental factors and elevated protein levels [51]. Importantly, our findings do not exclude the neuroprotective role of physiological soluble Ttr reported in Alzheimer's disease models. Rather, our data suggest that excessive accumulation and aggregation of Ttr under the highly oxidative and inflammatory conditions following TBI may convert Ttr from a protective molecule into a pathogenic mediator. The aggregated Ttr has been shown to activate NF-κB and a cascade of downstream effects, including endoplasmic reticulum stress, proteasome dysfunction, cell death, and oxidative stress [52]. We suggest the possibility that Ttr's role is context-dependent: protective in Alzheimer's disease but potentially maladaptive in the acute, oxidative environment of TBI. The aggregated Ttr was increased significantly after TBI, while NET inhibition relieved the increase of aggregated Ttr. To generate a neuron−specific Ttr deletion, Syn1-cre mice were crossed with Ttr-flox mice. Neuron−specific Ttr conditional knockout attenuates neuronal death and microglial inflammatory activity after TBI. Though Ttr with interactive and proteolytic activity inhibits Aβ aggregation and fibrillization, the aberrant accumulation and aggregation of Ttr in neuron after TBI is pathogenic. ChIP and luciferase reporter assay demonstrated that NET-carried S100a9K26la promoted Ttr transcription in neuron. NETs resulted in mitochondrial oxidative stress and neuronal cuproptosis, which was alleviated by Ttr inhibition. The process of cellular copper efflux from the cell requires the participation of Atp7b transporters [39]. Atp7b executes copper-related trafficking from transGolgi network to cellular membrane via vesicles to eliminate excess copper from cell [53]. The combination of Commd1 with Atp7b is involved in copper efflux and regulates copper homeostasis [40,41]. To further investigate the mechanisms of NET-mediated cuproptosis in neuron, co-IP assay exhibited that Ttr disturbed the interaction between Commd1 and Atp7b by competing Commd1 at W123 and induced copper accumulation and cuproptosis. Thus, NETs promote neuronal cuproptosis by reducing the interaction between Atp7b and Commd1. Collectively, our findings provide the first evidence that NETs can elevate neuronal cuproptosis.

NETs were increased after TBI and were positively associated with DLAT aggregation and POLD1, ACO-2 and SDHB degradation in neurons in humans and mice. Administration of TTM, a copper chelator, alleviated the aggregation of DLAT and the decrease in Fe‒S cluster proteins in brain tissue of TBI mice, which suggested that TTM treatment inhibited cuproptosis after TBI. NeuN IF and FJC staining exhibited TTM treatment obviously ameliorated neuronal degenerescence and death in cerebral cortex of TBI mice. In addition, TTM alleviated immune cell infiltration and microglial proinflammatory activation and promoted microglial anti-inflammatory activation after TBI, which indicated that the inhibition of cuproptosis using TTM could improve neuroinflammation following TBI. Furthermore, double IF staining, EB dye extravasation assay, MRI scanning, and LSCI revealed that the treatment of TTM relieved cerebrovascular endothelial dysfunction, BBB leakage, brain edema, and brain microcirculation disturbance. TTM treatment restored short-term neurological function and spatial learning and memory capacity after TBI. NETosis-deficient Padi4^−/− mice were used to examine the effect of NETs on TBI-induced pathology. The inhibition of NET formation also ameliorated neuronal death, neuroinflammation, and BBB disruption. Interestingly, NET inhibition alleviated neuronal cuproptosis in brain tissue of TBI mice.

In the present study, the SH-SY5Y cells were maintained in an undifferentiated state and were not treated with all trans retinoic acid (RA). This experimental choice was guided by two primary considerations. First, undifferentiated SH-SY5Y cells are widely employed for mechanistic study of signaling pathways. Notably, a recent study investigating neuroprotective signaling in a neuronal injury model also utilized undifferentiated SH-SY5Y cells and successfully identified mechanistically relevant findings that were subsequently validated in vivo [54]. Second, as noted in that same study, differentiated SH-SY5Y cells exhibit greater tolerance to oxidative injury and altered mitochondrial activity, which could potentially attenuate injury-related signaling responses and thus render them less suitable for detecting early mechanistic events. Nevertheless, we fully acknowledge that undifferentiated neuroblastoma-derived cells do not entirely recapitulate the phenotype of mature post-mitotic neurons. Our findings should be interpreted with caution when extrapolated to the in vivo setting. Importantly, our data from human tissues provide convergent evidence supporting the proposed mechanistic axis, thereby substantially mitigating concerns regarding cell-line specificity. Future studies employing RA-differentiated SH-SY5Y cells would be valuable to further confirm the translational relevance of our observations. In addition, a canonical cuproptosis positive control (e.g., Elesclomol-Cu) was not included in this study, as our primary objective was to investigate endogenous cuproptosis driven by pathological copper dysregulation after TBI. Nevertheless, future studies incorporating Elesclomol-Cu treatment alongside copper chelator rescue experiments would be valuable to definitively confirm the cuproptotic nature of the neuronal injury observed.

5. Conclusion

In summary, NETs lead to neuronal cuproptosis following TBI. NET-carried S100a9K26la promote Ttr transcription and inhibit interaction between Atp7b and Commd1 and disrupt copper efflux. Cu⁺ accumulation induces lipoylated DLAT aggregation and Fe‒S cluster protein loss, thereby leading to mitochondrial oxidative stress and ultimately neuronal death, which results in neuroinflammation and BBB disruption following TBI. Our findings suggest that NETosis suppression and cuproptosis inhibition might be prospective therapeutic strategies for secondary brain injury in TBI patients.

CRediT authorship contribution statement

Jianye Xu: Writing – review & editing, Writing – original draft, Visualization, Formal analysis, Data curation. Xu Zhang: Writing – review & editing, Writing – original draft, Visualization, Formal analysis, Data curation. Yang Liu: Writing – review & editing, Writing – original draft, Visualization, Formal analysis, Data curation. Bo Chen: Writing – original draft, Formal analysis, Data curation. Yao Zhang: Visualization, Formal analysis. Jinchao Wang: Formal analysis, Data curation. Yiyao Cao: Data curation. Shenghui Li: Data curation. Ruojie Wu: Formal analysis. Dongdong Sun: Formal analysis. Lei Li: Writing – original draft. Xiao Liu: Writing – original draft. Guili Yang: Writing – original draft. Zengguang Wang: Project administration, Methodology, Funding acquisition, Conceptualization. Xing Guo: Project administration, Methodology, Funding acquisition, Conceptualization. Weiguo Li: Project administration, Methodology, Funding acquisition, Conceptualization. Shu Zhang: Project administration, Methodology, Funding acquisition, Conceptualization.

Ethics declaration

Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed.

This study included organ or tissue donors. This study includes human biological material and consent was obtained by donors, or their next of kin or legal representatives, for use in this study and for publication of the article. The samples used in this research were not sourced from executed prisoners or prisoners of conscience.

This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the Ethics Committee of Tianjin Medical University General Hospital (Approval No. IRB2024-YX-536-01).

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. This study was approved by the Institutional Animal Care and Use Committee of Tianjin Medical University General Hospital (Approval No. IRB2020-DW-19).

Availability of data and materials

The data are available from the corresponding author upon reasonable request. We have deposited raw scRNA-seq processing code in a public repository (https://github.com/Xujianye1211/Re-analysis-of-GSE290150.git).

Consent for publication

Not applicable.

Funding

This work was supported by grants from National Natural Science Foundation of China (Nos. 82573138; 81930031; 81901525; 82071389; 82101440; 82301547; 82401604), Tianjin Key Medical Discipline Construction Project (No. TJYXZDXK-3-002A), China Postdoctoral Science Foundation (2025M782100), Shandong Postdoctoral Science Foundation (SDZZ-ZR-202501079); Science & Technology Planning Projects of Tianjin Municipality (No. 24ZXZSSS00030), Tianjin Educational Commission Research Project (No. 2024KJ248), Taishan Scholar Young Expert Program of Shandong Province (No. tsqn202507345), Key R&D Program of Shandong Province, China (No. 2022ZLGX03-6), National Key Research and Development Program of China (No. 2022YFC2405105), Natural Science Foundation of Shandong Province (Nos. ZR2026QC0911; ZR2023MH036), Youth Innovation Team of Colleges and Universities in Shandong Province (No.2023KJ023), and Young Elite Sponsorship Program of Shandong Provincial Medical Association.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Abbreviations

TBI: Traumatic brain injury; NETs: Neutrophil extracellular traps; BBB: Blood–brain barrier; TCA: Tricarboxylic acid; DLAT: Dihydrolipoamide S-acetyltransferase; TTM: Tetrathiomolybdate; S100a9K26la: S100a9 lactylated at lysine 26; Ttr: Transthyretin; Atp7b: ATPase copper transporting β; Commd1: Copper metabolism MURR1 domain-containing 1; dsDNA: Double-stranded DNA; NE: Neutrophil elastase; MPO: Myeloperoxidase; Padi4: Peptidyl arginine deiminase 4; GCS: Glasgow Coma Scale; GOSE: Glasgow Outcome Scale–Extended; CCI: Controlled cortical impact; IHC: Immunohistochemistry; IF: Immunofluorescence; ACO-2: Aconitase-2; FJC: Fluoro-Jade C; EB: Evans blue; LSCI: Laser speckle contrast imaging; CBF: Cerebral blood flow; mNSS: Modified neurological severity score; MWM: Morris water maze; co-IP: Co-immunoprecipitation; POLD1: DNA polymerase δ catalytic subunit; SDHB: Succinate dehydrogenase (ubiquinone) iron‒sulfur subunit, mitochondrial; H&E: hematoxylin and eosin.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

We have deposited raw scRNA-seq processing code in a public repository (https://github.com/Xujianye1211/Re-analysis-of-GSE290150.git).