Zmiz1-Mediated SUMOylation of NLRP3 Inflammasome Regulates Satellite Glial Cell Activation and Neuronal Autophagy in Trigeminal Neuralgia

Abstract

Trigeminal neuralgia (TN) is characterized by neuroinflammation and satellite glial cell (SGC) activation, but the molecular mechanisms remain unclear. This study identifies zinc finger MIZ-type containing 1 (Zmiz1) as a key regulator that promotes SUMOylation of NLRP3 inflammasome, thereby influencing SGC activation and neuronal autophagy in TN. Using bioinformatics analysis, we identified Zmiz1 as a key SUMOylation-related gene involved in TN. Single-cell transcriptomics and co-expression network analysis revealed Zmiz1 enrichment in SGCs and neurons. Co-immunoprecipitation (Co-IP) and western blotting confirmed Zmiz1’s interaction with NLRP3 and its role in promoting NLRP3 SUMOylation. In vitro experiments assessed the impact of Zmiz1 overexpression and knockdown on SGC activation, inflammatory cytokine secretion, and neuronal autophagy. A TN rat model was established to evaluate pain behavior, neuroinflammation, and neuronal apoptosis. Zmiz1 overexpression significantly enhanced NLRP3 SUMOylation, promoting SGC activation and inflammation while inhibiting neuronal autophagy. Conversely, silencing Zmiz1 reduced neuroinflammation and improved neuronal viability. In vivo, Zmiz1 knockdown alleviated TN-associated pain hypersensitivity and neuronal apoptosis. This study unveils a novel mechanism by which Zmiz1 regulates TN via NLRP3 SUMOylation, highlighting the Zmiz1/NLRP3 axis as a potential therapeutic target for TN treatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10753-025-02330-4.

Introduction

Trigeminal neuralgia (TN) is a severely debilitating chronic pain condition characterized [1, 2]. TN significantly impairs patients’ quality of life [3–5]. Although various treatment options exist, many patients do not achieve satisfactory pain relief [6]. Therefore, a deeper understanding of the molecular mechanisms underlying TN is crucial for developing more effective therapeutic strategies [7–9].

Small ubiquitin-like modifier conjugation (SUMOylation) is a significant post-translational modification that involves covalent attachment to target proteins, affecting their stability, subcellular localization, interactions, and activity [10–12]. This modification plays a critical role in regulating various cellular processes, including gene expression, DNA repair, and the cell cycle [10, 13, 14]. Recently, increasing evidence has shown that SUMOylation also plays an important role in neurological diseases, particularly in neuroinflammation and neurodegenerative disorders [15, 16]. However, the specific mechanisms by which SUMOylation affects TN remain unclear, necessitating further research [17, 18].

Zinc finger MIZ-type containing 1 (Zmiz1) is a transcriptional regulator that has garnered significant attention in recent years due to its crucial role in inflammatory responses [19–21]. NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome is another critical molecule in inflammation regulation, capable of sensing intracellular and extracellular danger signals to activate downstream inflammatory responses. Previous studies have shown that Zmiz1 participates in transcriptional regulation through SUMOylation, and that SUMOylation of NLRP3 can suppress inflammasome activation [22, 23]. However, the specific interaction between Zmiz1 and NLRP3 in TN and its underlying mechanisms have not been systematically investigated, leaving many aspects of this regulatory axis unexplored.

Firstly, differential expression genes related to SUMOylation were identified through transcriptomic data analysis using the MsigDB and Gene Expression Omnibus (GEO) databases. Then, using Least Absolute Shrinkage and Selection Operator (LASSO) and Weighted Gene Co-expression Network Analysis (WGCNA) methods, key SUMOylation-related genes involved in the pathogenesis of TN, including Zmiz1, were further screened. Single-cell transcriptomic analysis revealed the cellular distribution of Zmiz1 in trigeminal ganglion tissues, and the “CellChat” package was employed to analyze intercellular interactions. Finally, both in vitro and in vivo experiments were conducted to elucidate the effects of Zmiz1 overexpression and silencing on satellite glial cell activation, neuronal autophagy, and pain behavior in a TN rat model.

Through bioinformatics screening and experimental validation, we have identified, for the first time, the critical role of Zmiz1 in TN and systematically evaluated its impact. Our results indicate that Zmiz1 significantly promotes the SUMOylation of NLRP3, enhancing the activation and inflammatory response of satellite glial cells while inhibiting neuronal autophagy. Additionally, in vivo experiments demonstrated that silencing Zmiz1 can reduce inflammation and pain behaviors in a TN rat model. This discovery not only provides new insights into the pathological mechanisms of TN but also identifies potential molecular targets for developing new therapeutic strategies, holding significant scientific and clinical value. A deeper understanding of the roles of Zmiz1 and NLRP3 in TN may offer more effective treatment options for TN patients, thereby improving their quality of life.

Materials and Methods

Public Data Retrieval and Processing

The transcriptome sequencing dataset GSE162284 related to TN was downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/), comprising 8 TN rat samples and 4 sham-operated control rat samples. 36 SUMOylation-related genes were obtained from the Molecular Signatures Database (MsigDB). Differential expression of SUMOylation-related genes was determined using the “edgeR” R package (v4.0.16) with thresholds set at |log2FoldChange| > 0.5 and p < 0.05.

Functional Enrichment Analysis

Gene Ontology (GO) analysis, a widely used method for large-scale functional enrichment studies, encompasses biological processes (BP), molecular functions (MF), and cellular components (CC). Kyoto Encyclopedia of Genes and Genomes (KEGG) is a widely utilized database that offers information on genomes, biological pathways, diseases, and drugs. The “clusterProfiler” R package (4.10.1) was utilized for GO and KEGG pathway enrichment analyses. Significance was established at p < 0.05.

Machine Learning-Based Selection of Key Genes

Learning Utilizing the “glmnet” R package (v4.1–8), key SUMOylation-related genes were selected through the machine learning method LASSO. The optimization of parameter selection was achieved through a 10-fold cross-valid action with partial likelihood deviance meeting the minimum criterion.

WGCNA

The gene co-expression network was constructed using the “WGCNA” R package (v1.72-5) to identify gene modules significantly associated with TN. Initially, the top 25% of genes with the highest variance were selected for analysis, followed by hierarchical clustering of samples using the hclust function. Subsequently, the optimal soft threshold was determined to calculate the similarity matrix of gene expression data across all samples, which was then transformed into a Topological Overlap Matrix (TOM). Gene modules were identified based on dynamic tree-cutting methods, and modules significantly associated with TN were determined through correlation analysis. Intersection analysis was performed between module genes and LASSO analysis results, generating a Venn diagram using online tools to identify common key regulatory factors.

Single-Cell RNA Sequencing (scRNA-seq) Data Processing and Analysis

The scRNA-seq dataset GSE186421, comprising single-cell transcriptome data from one healthy and one inflammation-induced trigeminal ganglion rat under pain conditions, was obtained from the GEO database. Data filtering using the “Seurat” R package (v5.1.0) involved the exclusion of low-quality cells (those with gene counts below 500 or UMI counts exceeding 15,000 or a mitochondrial gene proportion exceeding 10%). Subsequently, the data underwent normalization using the LogNormalize method to standardize gene expression data.

Dimensionality Reduction and Clustering Analysis

The top 3000 highly variable genes were selected using the FindVariable function, and the number of principal components (PCs) for further analysis was determined based on generated JackStraw and ElbowPlot plots. Cell clustering was performed using the FindNeighbors and FindClusters functions (with a resolution set at 0.5), then visualization using the Uniform Manifold Approximation and Projection (UMAP) method to identify distinct cellular subpopulations. Genes with FDR < 0.05 and |log2FC| > 0.5 were labeled cell cluster-specific marker genes.

Cell Type Annotation

The clustered cells were manually annotated based on known marker genes as follows: neurons (Tubb3, Uchl1, Rbfox3), myelinating Schwann cells (mSCs) (Ncmap, Prx, Nr4a2, Mpz), non-myelinating Schwann cells (nmSCs) (Scn7a, Cdh19), SGCs (Fabp7, Apoe), oligodendrocyte-like cells (OLCs) (Mog), immune cells (Ptprc, Lyz2), fibroblasts (Dcn, Lum, Col1a1), endothelial cells (Flt1, Cldn5, Pecam1), myofibroblast (Rgs5), and erythrocytes (Hbb-bt).

Cell Communication Analysis

CellChat is a tool capable of quantitatively inferring and analyzing cellular communication networks from scRNA-seq data. The “CellChat” R package (v2) was employed to analyze intercellular communication networks. The CellChatDB.mouse ligand-receptor library was initially imported to assess primary signal inputs and outputs in different cell types. Subsequently, the netVisual_circle function was utilized to depict the strength of cell-cell communication networks among different cells. Finally, the netVisual_aggregate function was used to illustrate the interaction network of significant signaling pathways among different cell types.

Cell Isolation and Culture

SGCs were isolated from the L4-L6 dorsal root ganglia (DRG) of SD rats (Slac: SD, Shanghai Slac Laboratory Animal Co., Ltd.). The isolation process involved several key steps: initially, the DRG was dissected and surrounding tissue removed to prevent fibroblast contamination. The DRG, devoid of nerves and dorsal roots, was then placed in a 1 mg/ml type II collagenase solution (17101015, Thermo Fisher Scientific) and incubated for 45 min at 37 °C under 5% CO₂. Subsequently, after washing with DMEM (C0891, Beyotime), the DRG was incubated in a 0.25% trypsin/EDTA solution (C0208, Beyotime) for 8 min at 37 °C under 5% CO₂. The enzymatic reaction was terminated with 1 ml fetal bovine serum (12483020, Gibco). The DRG cell suspension was supplemented with DMEM to a final volume of 5 ml and placed in T25 cell culture flasks (FFLK075, Beyotime) for a 2-hour incubation at 37 °C under 5% CO₂. Subsequently, the culture medium was replaced with high-glucose DMEM containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin (C0222, Beyotime). After 24 h, the medium was replaced with a medium containing 10 µM cytarabine (J65671.09, Thermo Fisher Scientific) for 24 h to eliminate residual fibroblast proliferation. Subsequent medium changes were conducted every 2 days without cytarabine. When cells reached approximately 65% confluence (10–14 days), they were ready for further experiments.Using immunofluorescence staining and confocal microscopy, the expression of the SGC-specific marker GS (ab176562, Abcam) was observed and images were collected to identify SGCs.

Rat dorsal root ganglion neurons were purchased from Wuhan Procell Life Science & Technology Co., Ltd. (CP-R126, Wuhan, China) and cultured in high-glucose DMEM containing 10% fetal bovine serum and 1% antibiotics (penicillin and streptomycin).

Lentivirus Infection and Grouping

SGC cells in the logarithmic growth phase were collected and overexpressed with Zmiz1 through lentivirus infection. The recombinant overexpression lentivirus was constructed and provided by Shanghai Genedenovo Biotechnology Co., Ltd. The infection procedure involved taking SGC cells in the logarithmic growth phase, preparing a cell suspension with a density of 5 × 10⁴ cells/mL, seeding them in a 6-well plate, with 2 mL per well, and incubating overnight. Subsequently, each well was infected with a final concentration of 1 × 10⁸ TU/mL of recombinant lentivirus, and after 48 h post-infection, further experiments were conducted.

For experiments investigating the interaction between NLPR3 and Zmiz1, the groups were defined as follows: Control group (addition of PBS to SGC); Nig + LPS (SGC treated with 100 ng/ml LPS (S1732, Beyotime) for 3 h followed by the induction of 10 µM Nigericin (Nig, purchased from bocsci, CAS No.: 28380-24-7) for 1 h); Nig + LPS + OE-NC (SGC cells infected with empty lentivirus, then treated with 100 ng/ml LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h); Nig + LPS + OE-Zmiz1 (SGC cells infected with Zmiz1 overexpression lentivirus, treated with 100 ng/ml LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h).

Experimental Grouping for Investigating the Effects of NLPR3 and Zmiz1 on SGCs and Neurons were as follows: OE-NC + PBS group (SGC cells infected with empty lentivirus, treated with 100 ng/mL LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h, then treated with PBS); OE-NC + MCC950 group (SGC cells infected with Zmiz1 overexpression lentivirus, treated with 100 ng/mL LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h, then treated with 200 µL of 10 µM MCC950 (HY-12815, MCE) NLRP3 inhibitor for 24 h); OE-Zmiz1 + PBS group (SGC cells infected with Zmiz1 overexpression lentivirus, treated with 100 ng/mL LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h, then treated with PBS); OE-Zmiz1 + MCC950 group (SGC cells infected with Zmiz1 overexpression lentivirus, treated with 100 ng/mL LPS for 3 h followed by the induction of 10 µM Nigericin for 1 h, then treated with 200 µL of 10 µM MCC950 NLRP3 inhibitor for 24 h).

Following treatment, the isolated SGCs underwent co-culture experiments in 24-well plates using inner wells with 0.4 μm pore size (FTW070, Beyotime) to physically separate SGCs and neurons. Neurons (2 × 10⁵ cells/well) were cultured in the outer wells in a 1.5 ml DMEM medium containing 10% fetal bovine serum. Subsequently, the isolated SGCs (4 × 10⁵ cells/well) were added to the 500 ml DMEM medium in the inner wells. After incubating at 37 °C for 14 days, the neurons were collected, washed with PBS, and prepared for subsequent experiments.

Western Blot (WB) Analysis

Total protein extraction from cells was performed using RIPA lysis buffer containing PMSF (P0013B, Beyotime, Shanghai), followed by quantification using the BCA protein assay kit (23225, Thermo Fisher Scientific, Rockford, IL, USA). First, 50 µg of protein was dissolved in 2× SDS loading buffer, boiled at 100 °C for 5 min, followed by SDS-PAGE gel electrophoresis. Subsequently, proteins were transferred to a PVDF membrane using the wet transfer method, blocked with 5% non-fat milk at room temperature for 1 h. The PVDF membrane was then incubated overnight at 4 °C with the following primary antibodies: Flag (AE005, abclonal), HA (AE008, abclonal), NLRP3 (A5652, abclonal), Zmiz1 (ABIN2781101, antibodies-online), IL-1β (A16288, abclonal), cleaved IL-1β (AF4006, affbiotech), caspase-1 (A16792, abclonal), cleaved caspase-1 (AF4005, affbiotech), SUMO1 (A2130, abclonal), SUMO2/3 (A2571, abclonal), GFAP (A19058, abclonal), LC3-I/LC3-II (A5618, abclonal), p62 (A19700, abclonal), and ACTIN (AC026, abclonal). The membrane was washed thrice with TBST for 10 min each, then incubated for 1 h with an HRP-conjugated goat anti-rabbit IgG H&L secondary antibody (ab97051, 1:2000, Abcam, Cambridge, UK). After TBST rinsing, the membrane was placed on a clean glass plate. Pierce™ ECL chemiluminescent substrate solutions A and B (32209, Thermo) were mixed in a dark room and then added to the membrane before capturing images using the Bio-Rad imaging system (ChemiDoc™ XRS+, BIO-RAD).

Immunoprecipitation (IP)

Flag-NLRP3 and HA-Zmiz1 were co-transfected into HEK-293T cells for the exogenous Co-IP experiment. HEK-293T cells were obtained from ATCC (CRL-3216, ATCC) and cultured in high-glucose DMEM with 10% fetal bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin.

The cells were lysed using IP cell lysis buffer (P0013, Beyotime) containing 1% protease inhibitor (P1005, Beyotime) and incubated at 4 °C for 30 min. The supernatant was collected after centrifugation at 12,000 g for 10 min at 4 °C. Subsequently, it was incubated overnight at 4 °C with 20 µL magnetic beads (P2108, Beyotime) and Flag-tag antibody (AE061, abclonal). On the following day, the beads were washed five times with TBST. The bead-protein complexes responsible for binding specific antibodies and target proteins were eluted in 1× loading buffer (P0015A, Beyotime) and denatured at 95 °C for 10 min. The samples were then briefly centrifuged at 12,000 g for 1–2 min at 4 °C before immunoblot analysis.

Similarly, SGC cells were lysed with IP lysis buffer for the endogenous Co-IP. The lysates were incubated with 20 µL magnetic beads and either NLRP3 or Zmiz1 antibodies or IgG (AC042, abclonal) overnight at 4 °C. The precipitates were collected by centrifugation at 500 g for 5 min at 4 °C, washed five times with TBST, and subjected to protein blot analysis.

Enzyme-Linked Immunosorbent Assay (ELISA) Biochemical Analysis

Cultured supernatants were centrifuged at 1500 × g for 15 min, and IL-1β, TNF-α, and Interleukin-6 (IL-6) (rat, E-UNEL-R0030, Elabscience) ELISA antibodies were utilized according to the instructions. In each well, 100 µL of diluent buffer, 100 µL of the sample, and 100 µL of the standard were added, followed by incubation at 37 °C for 90 min. Subsequently, 100 µL of biotinylated detection antibody was added to each well and incubated at 37 °C for 1 h. The wells were washed three times for 2 min each. Following this, 100 µL of horseradish peroxidase conjugate was added to each well and incubated at 37 °C for 30 min. After washing five times, substrate reagent was added, and the plates were incubated at 37 °C in the dark for 15 min before stopping the reaction. Absorbance was measured at 450 nm using the Epoch microplate spectrophotometer (Bio-Tek, Winooski, VT, USA). Each sample was set up in triplicates.

Quantitative Real-Time PCR (qRT-PCR) Analysis

Total RNA was extracted from the cells using TRIzol reagent (15596026, Invitrogen, Carlsbad, CA, USA), and its concentration and purity were assessed using a Nanodrop2000 UV spectrophotometer (1011U, Nanodrop, USA). The mRNA was then reverse transcribed into cDNA following the protocol of the PrimeScript RT reagent Kit (RR047A, Takara, Japan). Real-time PCR was conducted using an ABI 7500 qPCR system (7500, ABI, USA) with the following cycling conditions: pre-denaturation at 95 °C for 10 min, denaturation at 95 °C for 10 s, annealing at 60 °C for 20 s, and extension at 72 °C for 34 s over 40 cycles. β-actin was used as the internal reference gene. The relative transcription levels of the target gene were determined using the comparative Ct method (2^−ΔΔCT method): ΔΔCt = ΔCt experimental group - ΔCt control group, where ΔCt = Ct (target gene) - Ct (internal reference gene), and the relative transcription level of the target gene was calculated as 2^−ΔΔCt. Each experiment was repeated three times, and the primers were synthesized by TaKaRa (Table S1).

CCK-8

Cell viability was assessed using CCK-8 (C0037, Beyotime) by seeding rat neuronal cells at a density of 4 × 10³ cells per well in a 96-well plate containing 100 µL of culture medium. After a 24-hour incubation period, the cells were treated according to their respective groups. Following a 20-hour incubation at 37 °C, 10 µL of CCK-8 solution was added to each well, and the plates were further incubated for 120 min before measuring the OD 450 values.

Immunofluorescence Staining

SGCs were fixed in 4% paraformaldehyde (P0099, ABclonal) for 15 min, followed by three washes with PBS. Cell membranes were permeabilized at room temperature with PBS containing 0.5% Triton X-100 (P0096, ABclonal) for 15 min. Non-specific binding sites were blocked with 5% bovine serum albumin (ST2254, Beyotime) in PBS for 1 h, followed by three additional washes of 5 min each. SGCs were stained with GFAP (A19058, ABclonal) as a marker for glial cells, MAP2 (4542, CST) for rat neuronal cells, and LC3 (A5618, ABclonal) as a marker for neuronal cell autophagy. After washing, cells were incubated at room temperature for 1 h with secondary antibody Alexa Fluor 488 goat anti-rabbit IgG (ab150077, Abcam) and counterstained with DAPI (C1002, ABclonal) in the dark for DNA visualization. Subsequently, cells were washed with PBS. Images were captured using a Nikon microscope (AZ100; Nikon, Tokyo, Japan) equipped with fluorescence illumination (L200/D; Prior Scientific, Rockland, MA, USA) and a digital camera (DS-Vi1; Nikon) connected to a computer for further analysis.

Apoptosis Detection

Rat neuronal cells were collected post-treatment and centrifuged at 1000× g in pre-chilled PBS for 5 min. The supernatant was discarded, and this process was repeated twice. Subsequently, the cells were resuspended in 500 µL of pre-chilled PBS and stained using the V-FITC/Propidium Iodide (PI) dual dye kit (C1062S, Beyotime) for apoptosis analysis by flow cytometry (CytoFLEX, Beckman Coulter, Brea, CA, USA), with data analysis performed using FlowJo software.

Construction and Grouping of TN Rat Model

The TN rat model was established by injecting cobra venom into the nerve sheath of the infraorbital nerve (ION) in rats. Adult SD rats weighing 180–220 g were procured from Shanghai SLAC Laboratory Animal Co., Ltd. (Slac: SD). The cobra venom, derived from freeze-dried whole venom (T0195, Sigma), was dissolved in a 0.9% sterile saline solution. Throughout the experiment, each rat received 4 µL of saline solution containing 0.4 mg freeze-dried whole venom. Anesthesia was administered to the rats via intraperitoneal injection of pentobarbital sodium (P010, Sigma) at a dosage of 40 mg/kg. Subsequently, each rat was positioned on a surgical table in a supine position with the head secured. Under direct vision, a roughly 1 cm incision was made along the skin edge above the eyebrow, exposing the orbital cavity and nasal bone. The orbital contents were gently displaced using a glass rod. The ION was dissected freely within the orbit, lifted gently using a glass nerve dissector, and either saline or cobra venom was injected into the nerve sheath of the ION. The injections were performed at the anterior end of the ION, followed by a 3-day interval [24].

Thirty-si × SD rats were randomly divided into six groups: sham group (sham surgery with PBS treatment), model group (TN model induced by cobra venom), OE-NC + PBS group (TN rats induced by cobra venom and treated with empty lentivirus and PBS), OE-NC + MCC950 group (TN rats induced by cobra venom and treated with empty lentivirus and MCC950), OE-Zmiz1 + PBS group (TN rats induced by cobra venom and treated with lentivirus overexpressing Zmiz1 and PBS), and OE-Zmiz1 + MCC950 group (TN rats induced by cobra venom and treated with lentivirus overexpressing Zmiz1 and MCC950), with each group comprising 6 rats. The cobra venom, prepared from freeze-dried whole venom (T0195, Sigma), was dissolved in 0.9% sterile saline solution for the experiments. Each rat received a 4 µL saline solution containing 0.4 mg freeze-dried whole venom. Three hours after cobra venom injection, MCC950 was dissolved in sterile saline and administered to the animals via intraperitoneal injection at 50 mg/kg. Subsequently, lentivirus containing either empty vector or Zmiz1 overexpression (8 × 10⁵ TU, 1 µl) was injected into the trigeminal ganglion through the inferior orbital fissure [25, 26]. Mechanical allodynia in the rats was assessed on days 0, 3, 5, 7, and 10, with rats euthanized on day 10 for serum and demyelination analysis of the trigeminal nerve.

Mechanical Allodynia Behavior Evaluation in Rats

All procedures were conducted in a quiet environment by the same experimenter to avoid stress. Mechanical stimuli-induced periorbital withdrawal responses were evaluated using calibrated von Frey filaments (58011, Stoelting). Rats were placed on an elevated wire mesh platform and acclimated to the testing environment for at least 30 min prior to examination. A series of von Frey filaments with increasing bending forces (0.02, 0.04, 0.07, 0.16, 0.4, 1.0, and 1.4 g) were vertically applied to the periorbital skin for approximately 5 s, with enough force to slightly bend the filament.

A positive response was defined as any of the following: vigorous facial scratching with the forepaw, head withdrawal from the stimulus, or head shaking. Testing began with the 0.16 g filament. If no response was observed within 5 s, the next heavier filament was used; if a response was observed, a lighter filament was applied. Each rat was subjected to six measurements or until four consecutive positive or negative responses occurred. A minimum interval of 5 min was maintained between stimulations with different filaments. The lowest force filament eliciting a consistent response was recorded as the mechanical withdrawal threshold. Measurements were taken on days 0, 3, 5, 7, and 10 [27].

Hematoxylin and Eosin (H&E) Staining

After euthanizing the rats, trigeminal nerve demyelinated tissue samples were collected and fixed in 4% formaldehyde solution. The skin was embedded in paraffin, sectioned into 4 μm slices, and stained with H&E staining reagent (C0105M, Beyotime) for one minute, followed by rinsing with tap water until clear. Subsequently, the tissue was counterstained in Eosin solution for 15 s. After counterstaining, the sections were immediately transferred to 95% ethanol for dehydration, followed by further dehydration with 100% ethanol and xylene. The slides were finally sealed with a neutral mounting medium (C1795, Sigma) before air-drying for imaging. A 0 to 5 scale was used to quantify the degree of inflammation. The scoring criteria were as follows: 0 = no inflammation (no inflammatory cell infiltration, normal tissue structure), 1 = mild inflammation (few inflammatory cells infiltrating, localized, with normal tissue structure), 2 = mild to moderate inflammation (increased infiltration of inflammatory cells, extending to multiple areas, slight tissue disorganization), 3 = moderate inflammation (extensive infiltration of inflammatory cells, widespread, significant tissue damage, cell arrangement disorganized, possible cell degeneration or necrosis), 4 = moderate to severe inflammation (widespread inflammatory cell infiltration, almost the entire section affected, severe tissue damage, common cell degeneration and necrosis), and 5 = severe inflammation (extremely dense inflammatory cells, almost complete tissue destruction, widespread cell degeneration and necrosis, possible edema or fibrosis).

Immunofluorescence Co-localization

Fixed tissue samples were cryosectioned and subjected to immunofluorescence staining. Non-specific binding sites were blocked with 5% bovine serum albumin (BSA, ST2254, Beyotime) in PBS for 1 h at room temperature, followed by three washes (5 min each). GFAP (3670, CST) was used as a glial cell marker, and MAP2 (78667, CST) was used to label rat neurons. Both GFAP and MAP2 were co-incubated with ZMIZ1 primary antibody (ABIN2781101, antibodies-online). After washing, the sections were incubated at room temperature for 1 h with secondary antibodies: Alexa Fluor 488-conjugated goat anti-rabbit IgG (ab150077, Abcam) and Alexa Fluor 555-conjugated goat anti-mouse IgG (ab150118, Abcam). DNA was counterstained with DAPI (C1002, Abclonal) in the dark. Finally, the sections were washed in PBS. Images were acquired using a Nikon fluorescence microscope (AZ100, Nikon, Tokyo, Japan) equipped with a fluorescence illuminator (L200/D, Prior Scientific, Rockland, MA, USA) and a digital camera (DS-Vi1, Nikon) connected to a computer. Co-localization of Zmiz1 with cellular markers was analyzed using dedicated image analysis software.

TUNEL Detection

The TUNEL assay kit (C1088, Beyotime) assessed neuronal cell apoptosis. In brief, paraffin-embedded trigeminal nerve demyelinated sections were deparaffinized, treated with proteinase K (ST533, Beyotime) to remove nuclear proteins, immersed in 3% hydrogen peroxide for 10 min, and then incubated with terminal deoxynucleotidyl transferase enzyme (EP0161, Thermo Scientific) for 1 h at 37 °C, followed by incubation with biotinylated deoxyuridine triphosphate (dUTP)-FITC (S3762, Merck) for 30 min at 37 °C. Subsequently, cell apoptosis in the trigeminal nerve demyelinated tissue was examined using a confocal microscope (LMS710, ZEISS).

Statistical Analysis

The data, derived from a minimum of three independent experiments, are presented as mean ± standard deviation (Mean ± SD). A two-sample independent t-test was utilized for comparisons between the two groups. A one-way analysis of variance (ANOVA) was employed for comparisons involving three or more groups. In cases where ANOVA indicated significant differences, Tukey’s Honestly Significant Difference (HSD) post hoc tests were conducted to assess differences between individual groups. Non-normally distributed or heteroscedastic data were analyzed using the Mann-Whitney U or Kruskal-Wallis H tests. Statistical analyses were conducted using GraphPad Prism 9 (GraphPad Software, Inc.) and R software. The significance level for all tests was set at 0.05, with a two-tailed p-value less than 0.05 considered statistically significant.

Results

Transcriptome Sequencing Data Reveals Zmiz1/NLRP3 as a Key Signaling Axis in TN Progression

Recent studies have shown that SUMO modification plays an important role in various neurological disorders [28]. To comprehensively study the role of SUMOylation in TN, we identified 36 SUMOylation-related genes via MsigDB screening (Table S1). Using TN transcriptomic data from the GEO database, we found that 21 of these genes were significantly differentially expressed in TN samples (Fig. 1A). GO and KEGG functional enrichment analyses revealed that these 21 genes are mainly involved in protein modification, inflammatory responses, and cellular stress responses (Fig. 1B-C). Utilizing the LASSO machine learning method, we identified key genes most relevant to TN pathogenesis as Desi1, Senp2, Pias1, Pias2, Zmiz1, and Topors (Fig. 1D-E).

Fig. 1.

Differential expression and functional enrichment analysis of SUMOylation-related genes in TN. Note: (A) Heatmap of differentially expressed SUMOylation-related genes between TN samples and control samples; (B-C) Results of GO and KEGG enrichment analyses of differentially expressed genes; (D-E) Identification of key SUMOylation-related genes closely associated with the pathogenesis of TN using the LASSO machine learning method. * indicates p < 0.05; ** indicates p < 0.01. Control: n = 4, TN: n = 8.

Next, we performed WGCNA enrichment analysis, and hierarchical clustering (Fig. 2A) identified β = 5 (R² = 0.95) as the optimal soft threshold for establishing a scale-free network (Fig. 2B), with positive validation results (Fig. 2C). After merging modules with highly correlated feature genes, we identified 9 modules and visualized the Topological Overlap Matrix (TOM) (Fig. 2D-E). The turquoise module showed the highest correlation with TN, containing 1698 genes (Fig. 3A-B). Further intersection analysis of this module’s genes with key genes identified by LASSO revealed Zmiz1 as the only overlapping gene (Fig. 3C), which is highly expressed in TN tissues (Fig. 3D). These findings provide essential molecular mechanistic evidence to further comprehend the role of Zmiz1 in mediating SUMOylation in TN.

Fig. 2.

Construction of weighted gene co-expression networks. Note: (A) Dendrogram of samples (top) and corresponding clinical traits (bottom); (B) Selection of the optimal soft-threshold power β, with β = 5 chosen to achieve both an approximate scale-free topology fit index (R² > 0.95) and optimal mean connectivity; (C) Histogram of connectivity distribution (left) and scale-free topology (right) at β = 5; (D) Cluster dendrogram used to identify co-expression modules, with each branch representing a gene and different colors indicating distinct co-expression modules; (E) Heatmap visualization of the co-expression network, where lighter colors represent lower co-expression interconnectedness and darker colors represent higher co-expression interconnectedness. Control: n = 4, TN: n = 8.

Fig. 3.

Identification of key SUMOylation-related genes in TN. Note: (A) Module-trait correlations between co-expression modules and clinical characteristics; (B) Module-trait correlations between co-expression modules and TN; (C) Intersection of LASSO analysis results and WGCNA module genes; (D) Expression of Zmiz1 in control and TN samples, confirming its significant upregulation in TN. Zmiz1 was identified as the only overlapping gene between LASSO-selected genes and WGCNA turquoise module. * indicates p < 0.05. Control: n = 4, TN: n = 8.

Single-Cell Analysis Reveals Zmiz1’s Role in SGC-Neuron Communication and TN Pathogenesis

To explore the single-cell expression characteristics of Zmiz1 in TN, we analyzed the GEO database scRNA-seq dataset GSE186421 related to TN. After quality control (nFeature_RNA > 500, nCount_RNA < 15000, percent.mt < 10%), 6225 high-quality cells were included (Figure S1A-C), with subsequent analysis confirming good data quality.

We further selected the top 3000 most variable genes (Figure S1D), calculated cell cycle phases (Figure S1E), and performed PCA dimensionality reduction (Figure S1F). The optimal number of PCs was determined to be 40 (Figure S1G-H), and cells were clustered into 18 clusters using the UMAP algorithm (Fig. 4A). Ten cell types were successfully identified using known cell lineage-specific marker genes (Fig. 4B-C), including neurons, mSCs (myelinating Schwann cells), nmSCs (non-myelinating Schwann cells), SGCs (satellite glial cells), OLCs (oligodendrocyte-like cells), immune cells, fibroblasts, endothelial cells, myofibroblasts, and red blood cells. Zmiz1 was mainly enriched in SGCs and neurons and was significantly upregulated in the TN model (Fig. 4D), suggesting that Zmiz1 may play a key role in SGC activation and the development of TN.

Fig. 4.

Single-cell transcriptomic analysis of TN in rats. Note: (A) Visualization of cell clustering results across different samples, with each color representing a distinct cluster; (B) Visualization of cell annotation results, with each color indicating a different cell subgroup; (C) Bubble plot of known cell lineage-specific marker gene expression across different cells, where brighter colors denote higher average expression levels, and larger circles represent a greater number of cells expressing the gene; (D) Bubble plot of Zmiz1 expression in various cells of normal control and TN rats, where darker blue indicates higher average expression levels and larger circles represent a greater number of cells expressing the gene. Normal: n = 1, Pain: n = 1.

Further CellChat analysis revealed significant cell communication between SGCs and neurons (Fig. 5A-B), with neurodevelopment-related pathways (NGL signaling pathway) and inflammation-related pathways (MIF signaling pathway) being the primary communication modes between SGCs and neurons (Fig. 5C-D). These results support the critical role of Zmiz1 in the inflammatory response and pathogenesis of TN in SGCs and neurons.

Fig. 5.

Cell communication analysis. Note: (A) Network graph showing the number (top) and weight (bottom) of interactions between 10 cell types. Each node represents a cell cluster, with the size proportional to the number of cells in the cluster and the thickness of the edges indicating the number of interactions; (B) Network graph displaying the number (top) and weighted communication (bottom) of intercellular interactions among major cell types; (C-D) Network graphs illustrating the interactions of NGL and MIF signaling pathways among the 10 cell types.

Zmiz1 Promotes the SUMOylation of NLRP3 To Further Activate the NLRP3 Inflammasome

Studies have shown that the NLRP3 inflammasome responds to infection and tissue damage by rapidly escalating the intensity of inflammation by activating interleukin (IL)−1β, IL-18, and cell death via pyroptosis. The E3 SUMO ligase TRIM28 activates the NLRP3 inflammasome by promoting NLRP3 expression [29]. The E3 SUMO ligase Zmiz1 protein also acts as a transcriptional coactivator through SUMOylation [30, 31]. Subsequent analysis aimed to investigate whether Zmiz1 is involved in the SUMOylation modification process of NLRP3. Using exogenous IP, it was demonstrated that NLRP3 interacts with Zmiz1 (Fig. 6A). Since NLRP3 remains inactive in normal cells, LPS and Nigericin activated the NLRP3 inflammasome [32, 33]. Primary SGC cells with a positive rate of 97.10% were extracted from rats for further experiments (Figure S2). In endogenous IP experiments, it was found that NLRP3 interacts with Zmiz1 in SGC cells. Compared to the Nig + LPS + OE-NC group, the interaction between Zmiz1 and NLRP3 was significantly enhanced in the Nig + LPS + OE-Zmiz1 group (Fig. 6B).

Fig. 6.

Interaction between Zmiz1 and NLRP3 and the promotion of NLRP3 SUMOylation. Note: (A) Co-IP assays performed in HEK293T and primary SGC cells to assess the interaction between Zmiz1 and NLRP3; (B) Endogenous IP assay to detect the interaction between NLRP3 and Zmiz1 in SGC cells; (C) ELISA to measure the expression of pro-inflammatory factors in SGC cells; (D) RT-qPCR to assess the expression of pro-inflammatory factors in SGC cells; (E) WB analysis to detect the expression of inflammasome activation-related proteins in SGC cells; (F-G) IP to assess the SUMOylation level of NLRP3 in SGC cells. Data are presented as mean ± SD (^ns p > 0.05, * p < 0.05, ** p < 0.01, using ANOVA and Tukey’s multiple comparison test). Cell experiments were repeated three times.

ELISA and RT-qPCR analysis assessed the expression of pro-inflammatory factors in SGC cells after treatment. The results revealed that compared to the Control group, the Nig + LPS group demonstrated heightened expression of pro-inflammatory factors IL-1β, IL-6, and TNF-α, while the Nig + LPS + OE-Zmiz1 group exhibited significantly increased expression of IL-1β, IL-6, and TNF-α compared to the Nig + LPS + OE-NC group (Fig. 6C-D).WB analysis was utilized to determine the activation level of the NLRP3 inflammasome. The results indicated an upregulation in the expression levels of NLRP3 protein, IL-1β, and caspase1 in the Nig + LPS group compared to the Control group, whereas the Nig + LPS + OE-Zmiz1 group showed a significant elevation in the expression levels of NLRP3 protein, IL-1β, and caspase1 compared to the Nig + LPS + OE-NC group, suggesting that Zmiz1 overexpression enhances NLRP3 inflammasome activation (Fig. 6E). Assessment of NLRP3 SUMOylation revealed that Nig + LPS treatment induced an upregulation in NLRP3 SUMOylation compared to the Control group, and the Nig + LPS + OE-Zmiz1 group exhibited enhanced NLRP3 SUMOylation after Zmiz1 overexpression compared to the Nig + LPS + OE-NC group (Fig. 6F-G). In conclusion, the study results indicate that Zmiz1 interacts with NLRP3 and promotes NLRP3 SUMOylation.

Zmiz1 Promotes SGC Activation by Regulating the NLRP3 Inflammasome

To investigate the impact of Zmiz1 on the activation of the NLRP3 inflammasome in SGCs, we initially used lentivirus to knock down Zmiz1 expression in SGC cells. Successful Zmiz1 knockdown was confirmed through WB analysis (Fig. 7A). SGCs release cytokines and express glial fibrillary acidic protein (GFAP), which is upregulated after SGC activation. Immunofluorescence and WB analysis revealed that compared to the Control group, the LPS + Nig group showed an increase in GFAP expression, indicating that the activation of the NLRP3 inflammasome promotes SGC activation [34]. Conversely, the LPS + Nig + shZmiz1 group exhibited a decrease in GFAP expression compared to the LPS + Nig + shNC group, suggesting that Zmiz1 knockdown can suppress NLRP3 inflammasome-induced SGC activation (Fig. 7B-C). ELISA and RT-qPCR results demonstrated that compared to the Control group, the LPS + Nig group showed elevated expression of the pro-inflammatory factors IL-1β, IL-6, and TNF-α, while the LPS + Nig + shZmiz1 group displayed decreased expression of pro-inflammatory factors compared to the LPS + Nig + shNC group (Fig. 7D-E).

Fig. 7.

Regulation of NLRP3 by Zmiz1 in SGCs. Note: (A) WB analysis to evaluate the knockdown efficiency of Zmiz1 in SGC cells; (B) Immunofluorescence analysis of GFAP (green) expression in Zmiz1-knockdown SGC cells with activated NLRP3 (scale bar: 25 μm); (C) WB analysis of GFAP expression in Zmiz1-knockdown SGC cells with activated NLRP3; (D) ELISA to measure the expression of pro-inflammatory factors in Zmiz1-knockdown SGC cells with activated NLRP3; (E) RT-qPCR to assess the expression of pro-inflammatory factors in Zmiz1-knockdown SGC cells with activated NLRP3; (F) Immunofluorescence analysis of GFAP (green) expression in Zmiz1-overexpressing SGC cells with activated NLRP3 (scale bar: 25 μm); (G) WB analysis of GFAP expression in Zmiz1-overexpressing SGC cells with activated NLRP3; (H) ELISA to measure the expression of pro-inflammatory factors in Zmiz1-overexpressing SGC cells with activated NLRP3; (I) RT-qPCR to assess the expression of pro-inflammatory factors in Zmiz1-overexpressing SGC cells with activated NLRP3. MCC950 is used as a selective NLRP3 inhibitor to reverse Zmiz1-induced SGC activation. Data are presented as mean ± SD (** p < 0.01, using ANOVA and Tukey’s multiple comparison test). Cell experiments were repeated three times.

Further investigating the impact of Zmiz1 on the NLRP3 inflammasome in SGCs, we overexpressed Zmiz1 in SGC cells and simultaneously treated them with the NLRP3 inhibitor MCC950 to observe the reversal phenomenon. Immunofluorescence and WB analysis of GFAP expression in Nig and LPS-activated NLRP3 SGC cells revealed that compared to the OE-NC + PBS group, the OE-NC + MCC950 group exhibited decreased GFAP expression; the OE-Zmiz1 + PBS group showed increased GFAP expression, while the OE-Zmiz1 + MCC950 group displayed decreased GFAP expression (Fig. 7F-G).

ELISA and RT-qPCR analysis of pro-inflammatory factor expression in Nig and LPS-activated NLRP3 SGC cells showed that compared to the OE-NC + PBS group, the OE-NC + MCC950 group demonstrated decreased pro-inflammatory factor expression; the OE-Zmiz1 + PBS group exhibited increased pro-inflammatory factor expression, while the OE-Zmiz1 + MCC950 group displayed decreased pro-inflammatory factor expression (Fig. 7H-I).These results indicate that Zmiz1 overexpression enhances SGC cell activation, but treatment with the NLRP3 inhibitor reverses the enhanced SGC cell activation induced by Zmiz1 overexpression. Therefore, it is inferred that Zmiz1 activates SGC by regulating NLRP3.

Zmiz1 Regulation of NLRP3 Reduces Neuronal Viability and Inhibits Autophagy

SGCs tightly ensheathe neuronal cell bodies. Research has shown that SGCs impact neuronal cells, promoting chronic pain [35]. Consequently, we first investigated the influence of Zmiz1 knockdown on the regulatory effect of NLRP3 on SGC activation concerning neuronal cells. MAP2 is a marker of dendritic growth in neurons [36]. Immunofluorescence analysis of MAP2 expression was conducted to infer the activity of rat neuronal cells after co-culture. The results indicated that compared to the Control group, the LPS + Nig group exhibited decreased MAP2 expression, reduced neuronal numbers and lengths, suggesting hindered neuronal growth following co-culture with SGC cells activated by the NLRP3 inflammasome. In comparison to the LPS + Nig + shNC group, the LPS + Nig + shZmiz1 group showed increased MAP2 expression, and augmented neuronal numbers and lengths, indicating that Zmiz1 knockdown can reverse inhibited neuronal growth (Fig. 8A).

Fig. 8.

Zmiz1 regulates NLRP3 to decrease neuronal viability and inhibit autophagy. Note: (A) Immunofluorescence analysis of MAP2 (green) expression in rat neurons (scale bar: 25 μm); (B) CCK8 assay to assess the viability of rat neurons; (C) Flow cytometry to detect the apoptosis rate in SGC cells; (D) Immunofluorescence analysis of LC3 (green) autophagic vesicles in rat neurons (scale bar: 25 μm); (E) WB analysis to detect the expression levels of key autophagy proteins in neurons; (F) Immunofluorescence analysis of MAP2 (green) expression in rat neurons (scale bar: 25 μm); (G) CCK8 assay to assess the viability of rat neurons; (H) Flow cytometry to detect the apoptosis rate in rat neurons; (I) Immunofluorescence analysis of LC3 (green) autophagic vesicles in rat neurons (scale bar: 25 μm); (J) WB analysis to detect the expression levels of key autophagy proteins in rat neurons. Data are presented as mean ± SD (^ns p > 0.05, ** p < 0.01, using ANOVA and Tukey’s multiple comparison test). Cell experiments were repeated three times.

CCK8 assay demonstrated that the neuronal activity of rat cells after co-culture decreased in the LPS + Nig group compared to the Control group, while the LPS + Nig + shZmiz1 group showed enhanced activity compared to the LPS + Nig + shNC group (Fig. 8B). Flow cytometry analysis revealed that the apoptosis rate increased in the LPS + Nig group compared to the Control group, whereas the LPS + Nig + shZmiz1 group exhibited a reduced apoptosis rate compared to the LPS + Nig + shNC group (Fig. 8C).

Immunofluorescence was employed to detect the formation of autophagic vesicles containing the autophagy key protein LC3 in rat neuronal cells after co-culture [37]. The results revealed that compared to the Control group, the LPS + Nig group exhibited a decrease in the number of autophagic vesicles, whereas the LPS + Nig + shZmiz1 group showed an increase in autophagic vesicle numbers compared to the LPS + Nig + shNC group (Fig. 8D).

WB analysis was conducted to measure the expression levels of autophagy key proteins in rat neuronal cells after co-culture. The findings indicated that in comparison to the Control group, the LPS + Nig group showed a decrease in the LC3-II/LC3-I ratio and an increase in p62 expression. Contrarily, the LPS + Nig + shZmiz1 group exhibited an increase in the LC3-II/LC3-I ratio and a decrease in p62 expression compared to the LPS + Nig + shNC group (Fig. 8E).

Furthermore, in SGC cells overexpressing Zmiz1 and concurrently treated with the NLRP3 inhibitor MCC950, the cells were subsequently co-cultured with neurons to observe potential reversals. The results demonstrated that compared to the OE-NC + PBS group, the OE-NC + MCC950 group showed increased MAP2 expression, along with augmented neuronal numbers and lengths. In contrast, the OE-Zmiz1 + PBS group displayed decreased MAP2 expression, reduced neuronal numbers and lengths compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group exhibited increased MAP2 expression, increased neuronal numbers, and lengths relative to the OE-Zmiz1 + PBS group (Fig. 8F).

CCK-8 assay revealed that the neuronal cell activity after co-culture increased in the OE-NC + MCC950 group compared to the OE-NC + PBS group. Conversely, the cell activity decreased in the OE-Zmiz1 + PBS group compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group displayed enhanced cell activity relative to the OE-Zmiz1 + PBS group (Fig. 8G). Flow cytometry analysis indicated that the apoptosis rate of rat neuronal cells after co-culture decreased in the OE-NC + MCC950 group compared to the OE-NC + PBS group. In contrast, the apoptosis rate increased in the OE-Zmiz1 + PBS group compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group showed a reduced apoptosis rate compared to the OE-Zmiz1 + PBS group (Fig. 8H).

Immunofluorescence was utilized to investigate whether autophagic vesicles containing the autophagy key protein LC3 formed in rat neuronal cells after co-culture [37]. The results demonstrated an increase in the number of autophagic vesicles in the OE-NC + MCC950 group compared to the OE-NC + PBS group. Conversely, the OE-Zmiz1 + PBS group exhibited a reduction in the number of autophagic vesicles compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group showed an increase in the number of autophagic vesicles relative to the OE-Zmiz1 + PBS group (Fig. 8I).

After co-culture treatment, the expression levels of autophagy key proteins in rat neuronal cells were detected using WB. The results indicated that compared to the OE-NC + PBS group, the OE-NC + MCC950 group showed an increase in the ratio of LC3-II/LC3-I and a decrease in p62 expression. Conversely, compared with the OE-NC + PBS group, the OE-Zmiz1 + PBS group exhibited a decrease in the LC3-II/LC3-I ratio and increased p62 expression. Additionally, compared to the OE-Zmiz1 + PBS group, the OE-Zmiz1 + MCC950 group demonstrated an increase in the LC3-II/LC3-I ratio and decreased p62 expression (Fig. 8J). These results suggest that Zmiz1 regulates the activation of NLRP3 inflammasomes in SGCs, thereby reducing neuronal viability and suppressing autophagy.

Zmiz1 Regulates NLRP3 To Further Promote TN Symptoms

To investigate the impact of Zmiz1 and NLRP3 on the progression of TN in vivo, an SD rat model of TN was established. ZMIZ1 was found to be co-expressed with MAP2 in neurons and GFAP in glial cells (Figure S3). Mechanical hypersensitivity in each group of rats was assessed, showing a decrease in pain threshold in the Model group compared to the Sham group. Similarly, a decrease in pain threshold was observed in the shZmiz1 group compared to the shNC group. Moreover, the OE-Zmiz1 + PBS group exhibited a decreased pain threshold compared to the OE-NC + PBS group, whereas the OE-Zmiz1 + MCC950 group showed an increase in pain threshold (Fig. 9A).

Fig. 9.

Zmiz1 regulation of NLRP3 further exacerbates TN symptoms. Note: (A) Measurement of mechanical allodynia threshold in TN model rats; (B) ELISA to detect pro-inflammatory cytokine levels in TN model rats; (C) RT-qPCR to measure pro-inflammatory cytokine expression levels in TN model rats; (D) H&E staining to assess inflammation and demyelination in the trigeminal nerve tissue of TN model rats (scale bar: 100 μm); (E) WB analysis to detect the expression levels of inflammasome activation proteins in TN model rats; (F) TUNEL assay to evaluate apoptosis in demyelinated trigeminal nerve tissue of TN model rats (scale bar: 100 μm). Data are presented as mean ± SD (* p < 0.05 compared to the Sham group, $ p < 0.05 compared to the shNC group, % p < 0.05 compared to the OE-NC + PBS group, # p < 0.05 compared to the OE-Zmiz1 + PBS group, using ANOVA and Tukey’s multiple comparison test), n = 6.

ELISA and RT-qPCR was conducted to measure the expression levels of pro-inflammatory factors IL-1β, IL-6, and TNF-α in rats. The Model group demonstrated increased expression levels of pro-inflammatory factors compared to the Sham group. In contrast, the shZmiz1 group showed a decrease in pro-inflammatory factors compared to the shNC group. Additionally, the OE-Zmiz1 + PBS group exhibited an increase in pro-inflammatory factor expression compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group displayed a decrease in pro-inflammatory factor expression (Fig. 9B-C). H&E staining was performed to assess inflammation cell infiltration and patchy demyelination in the trigeminal nerve tissues of rats. The results revealed that compared to the Sham group, the Model group exhibited an increase in the inflammation score and severity of patchy demyelination. In contrast, the shZmiz1 group showed a decrease in the inflammation score and reduced severity of patchy demyelination compared to the shNC group. Additionally, the OE-Zmiz1 + PBS group demonstrated an increase in the inflammation score and severity of patchy demyelination compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group displayed a decrease in the inflammation score and severity of patchy demyelination (Fig. 9D).

WB analysis was utilized to examine the expression levels of the inflammation-related proteins NLRP3 and IL-1β in rats’ demyelinated trigeminal nerve tissues. The results indicated that the Model group showed an increase in the expression of inflammation-related proteins compared to the Sham group. Conversely, the shZmiz1 group decreased the expression of inflammation-related proteins compared to the shNC group. Moreover, the OE-Zmiz1 + PBS group displayed an increase in the expression of inflammation-related proteins compared to the OE-NC + PBS group, whereas the OE-Zmiz1 + MCC950 group showed a decrease in the expression of inflammation-related proteins when compared to the OE-Zmiz1 + PBS group (Fig. 9E).

TUNEL staining was employed to assess the level of apoptosis in neuronal cells. The results indicated that the Model group increased the apoptosis rate compared to the Sham group. In contrast, the shZmiz1 group decreased the apoptosis rate compared to the shNC group. Furthermore, the OE-Zmiz1 + PBS group showed an increase in the apoptosis rate compared to the OE-NC + PBS group, while the OE-Zmiz1 + MCC950 group demonstrated a decrease in the apoptosis rate relative to the OE-Zmiz1 + PBS group (Fig. 9F). These results suggest that Zmiz1 regulates NLRP3, further promoting symptoms of TN.

Discussion

TN is an extremely painful chronic condition in which patients often experience intense, stabbing-like episodes that severely affect their quality of life. The pathophysiology of TN is complex, involving multiple cellular and molecular pathways. Activation of SGCs is considered a crucial factor in the onset of TN, as SGCs release various inflammatory factors upon stimulation, triggering and exacerbating neuronal inflammatory responses [38]. In these inflammatory reactions, the NLRP3 inflammasome plays a pivotal role, detecting various stress signals both intra- and extracellularly and, through activation of downstream effector molecules, promoting the initiation and maintenance of inflammatory responses. Furthermore, SUMOylation modification, as a critical post-translational process, regulates protein stability, subcellular localization, and function by attaching small ubiquitin-like modifiers, thereby impacting cellular signal transduction [39, 40].

Zmiz1, emerging as a novel inflammatory regulator, has garnered increasing attention recently for its role in inflammatory responses [41]. Current studies suggest that Zmiz1 participates in inflammatory responses through various pathways, including the regulation of gene expression, protein modification, and signal transduction. Nevertheless, its specific relationship with the NLRP3 inflammasome remains unclear. Specifically, the presence of Zmiz1 not only enhances the expression levels of NLRP3 but also increases the stability and activation capability of NLRP3 through SUMOylation modification.

The activation of SGCs plays a crucial role in the pathophysiology of TN [42, 43]. Previous studies have indicated that SGC activation can promote the development of neural inflammation and exacerbate neuronal damage and pain by releasing various inflammatory factors such as IL-1β, IL-6, and TNF-α [44, 45]. However, the specific molecular mechanisms underlying SGC activation remain incompletely elucidated, particularly within TN regulation [46–48]. Experimental results have shown a significant increase in the expression levels of pro-inflammatory factors, including IL-1β, IL-6, and TNF-α, in SGCs overexpressing Zmiz1, further confirming the critical role of Zmiz1 in SGC activation.

Neuronal autophagy plays a significant role in maintaining neuronal function and regulating neuroinflammation [49, 50]. Autophagy is a crucial cellular process for clearing damaged proteins and organelles, aiding in maintaining intracellular stability [51, 52]. Prior studies have linked the inhibition of autophagy to the occurrence and progression of various neurodegenerative diseases such as Alzheimer’s, Parkinson’s, and Huntington’s diseases [52]. However, the specific role of the NLRP3 inflammasome in neuronal autophagy remains unclear [53]. Experimental findings reveal that under conditions of Zmiz1 overexpression, autophagic flux is inhibited, leading to reduced formation of autophagosomes and accumulation of damaged proteins.

Further research demonstrates that Zmiz1, through modulating the NLRP3 inflammasome, alters the activity of neuronal autophagy-related signaling pathways, thereby suppressing the autophagic process in neurons.

In vitro experiments demonstrate that overexpression of Zmiz1 significantly promotes the activation of SGCs and the release of inflammatory factors while inhibiting neuronal autophagy. In vivo experimental results further confirm this finding. Silencing Zmiz1 in a rat model shows a significant reduction in inflammatory responses in TN rats and a marked improvement in pain behavior. These results suggest that Zmiz1 plays a critical regulatory role both in vitro and in vivo, supporting its pivotal role in TN.

Although this study reveals the key regulatory role of Zmiz1 in TN and provides new insights into its pathological mechanisms, there are still some limitations. First, although single-cell transcriptome analysis provides cell-type-specific expression information, the lack of in vivo validation is crucial for confirming the expression and function of Zmiz1 in different cell types. In the future, we will consider using multiple experimental methods (such as immunofluorescence with specific markers) to further validate Zmiz1’s in vivo expression and function. Second, the experiments mainly focused on in vitro cell models and rat models, lacking validation in human patient samples. The absence of human patient sample validation could significantly affect the translation of research findings to clinical practice. In vitro and animal models cannot fully replicate the complex physiological and pathological environments of humans, which may lead to a lack of direct relevance for clinical applications, making it difficult to accurately assess their safety and efficacy. Moreover, human individual differences are vast, and these models cannot fully replicate such variations, meaning the findings may not apply to all patient populations. In future studies, we plan to collect trigeminal ganglion tissue samples from patients with TN, with appropriate ethical approval and informed consent. We will comprehensively assess the expression levels of Zmiz1 and NLRP3, their SUMOylation status, as well as key inflammatory and autophagy-related markers, and compare these findings with those from healthy controls. To this end, we will employ immunohistochemistry, Western blotting, and RT-qPCR, supplemented by immunoprecipitation and mass spectrometry to evaluate protein expression, modifications, and interactions. These analyses will facilitate the validation of our animal model findings in human samples and provide direct evidence supporting the clinical relevance of the Zmiz1–NLRP3 regulatory axis in TN pathogenesis. Furthermore, this study only explores the impact of Zmiz1 on the NLRP3 inflammasome and neuronal autophagy without fully covering other potential regulatory pathways and molecular mechanisms. The sample size limitation may also affect the results’ broad applicability and statistical reliability. Future studies should aim to validate these findings on a larger scale and employ various experimental methods to comprehensively elucidate the mechanisms of action of Zmiz1.

Looking ahead, research in this field should focus on several key aspects. Firstly, further exploration of the role of Zmiz1 in various pain models is essential to validate its regulatory mechanisms in different types of chronic pain disorders. Secondly, delving deeper into the interactions of Zmiz1 with other inflammatory factors and signaling pathways can unveil its complex network relationships in inflammatory responses. Additionally, clinical studies should intensify the development of drugs targeting Zmiz1 and its regulatory pathways to discover new strategies for effectively treating TN. Lastly, integrating cutting-edge gene-editing technologies such as CRISPR/Cas9 to precisely regulate Zmiz1 and its downstream targets can enhance treatment efficacy and safety. By designing specific guide RNAs (gRNAs) that can precisely recognize the DNA sequences of Zmiz1 and its downstream targets, the gRNA and Cas9 nuclease can be co-delivered to target cells. Cas9, guided by the gRNA, cuts the DNA, generating double-strand breaks, which are then repaired by non-homologous end joining (NHEJ) or homology-directed repair (HDR). NHEJ can introduce insertions or deletions to knock out the gene, while HDR can be used to introduce specific gene mutations or insert new gene sequences at the target site. This precise regulation can be used not only for basic research but also combined with clinical needs to develop drugs targeting Zmiz1 and its pathways, providing new strategies for TN treatment and improving both therapeutic efficacy and safety. Research in these areas will help to more comprehensively understand the pathophysiology of TN and offer more accurate and effective treatment options for patients.

Conclusion

This study unveils, for the first time, the molecular mechanism by which Zmiz1 promotes the activation of SGCs and inhibits neuronal autophagy through SUMOylation modification of NLRP3 (Fig. 10), playing a crucial role in the initiation and progression of TN. Through bioinformatics analysis and in vivo/in vitro experiments, we confirmed the significant role of Zmiz1 in TN. Zmiz1 markedly enhances the SUMOylation of NLRP3, boosting inflammatory responses and SGC activation while inhibiting neuronal autophagy, thereby exacerbating the pathological process of TN. Silencing Zmiz1 effectively alleviates inflammatory responses and pain behavior in the TN rat model, indicating Zmiz1 as a potential therapeutic target for TN. The study demonstrates the crucial role of the Zmiz1/NLRP3 signaling axis in the pathogenesis of TN, offering a fresh perspective for understanding the development of TN and laying a theoretical foundation for developing novel therapeutic strategies targeting TN. Future research can further investigate the role of Zmiz1 in other neuroinflammatory diseases and explore its potential as a therapeutic target.

Fig. 10.

Molecular mechanism figure of the Zmiz1/NLRP3 signaling axis in TN. Note: Schematic illustration summarizing the proposed mechanism: Zmiz1 promotes SUMOylation of NLRP3, enhancing SGC activation and inflammatory cytokine release, while suppressing neuronal autophagy—contributing to TN progression.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

F.J. and Q.W. designed and conducted the experiments. H.H. performed data analysis and contributed to manuscript preparation. G.C. assisted with animal modeling and behavioral assessments. H.J. and Y.L. supervised the project, acquired funding, and revised the manuscript critically for important intellectual content. All authors read and approved the final manuscript.

Funding

This study was supported by Doctoral Start-up Fund for Liaoning Provincial Natural Science Foundation Program (2023-BSBA-182).

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available due to privacy and confidentiality agreements with the participants but are available from the corresponding author on reasonable request.

Declarations

Ethics Approval and Consent to Participate

All animal experiments were approved by the Animal Ethics Committee of Liaoning Provincial People’s Hospital and were conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. Efforts were made to minimize animal suffering and reduce the number of animals used. Anesthesia and analgesia were administered during surgical procedures to ensure animal welfare, and humane endpoints were established to prevent unnecessary distress. All experimental protocols were reviewed to ensure compliance with institutional and national ethical standards for animal research.

Consent for Publication

Not applicable.

Competing interests

The authors declare no competing interests.