SIRT3-SUMO regulated Treg cell differentiation and asthma development by mediating N-glycosylation through the FAO pathway Authors: Yan Hu, Chuntao Liu Published: 2025 ABSTRACT AbstractAsthma, a chronic inflammatory disease, has a high disability rate, which greatly increases the disease burden. T cells are pivotal in the pathogenesis of asthma, and Treg cells, due to their role in maintaining immune system balance, represent a promising avenue for therapeutic intervention. Initial weighted correlation network analysis (WGCNA) analysis of asthma-related datasets indicates that N-glycosylation plays a critical role in asthma development. The establishment of an OVA-sensitized asthma model, along with the isolation of naive CD4+ T cells and subsequent in vitro induction of Treg cell differentiation, further underscores the significance of N-glycosylation in the Treg cell differentiation of asthma-related Treg cells. Employing immunofluorescence, flow cytometry, and Western blot techniques revealed that SIRT3-SUMO is instrumental in regulating N-glycosylation-mediated Treg cells development. Mechanistically, overexpression and deSUMOylation of SIRT3 enhance the expression levels of CPT1 and VLCAD to promote fatty acid oxidation (FAO), thereby increasing intracellular acetyl-CoA concentrations. Acetyl-CoA subsequently facilitates the synthesis of N-glycosylation substrates via the hexosamine biosynthetic pathway (HBP), promoting Treg cell differentiation. Ultimately, our in vivo experiments demonstrate that SIRT3-SUMO modulates asthma progression by influencing Treg cells differentiation; thus, augmenting Treg cells populations can inhibit Th2-type and non-Th2-type asthmatic developments. These findings elucidate mechanisms underlying Treg cell differentiation and provide theoretical foundations for targeted therapies aimed at enhancing Treg cells for asthma management.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10565-025-10105-8.Keywords: Asthma, N- glycosylation, Treg, Fatty acid oxidation, SIRT3-SUMO Introduction Asthma, a prevalent non-communicable disease, significantly affects the quality of life for numerous individuals. According to the Disability-Adjusted Life Year (DALY) assessment framework, asthma ranks 16th among the leading causes of disability worldwide and 28th in terms of overall disease burden, underscoring its considerable public health challenge(Resiliac and Grayson 2019). Notably, while high-income countries report higher prevalence rates of asthma, low- and middle-income nations disproportionately endure a greater share of asthma-related mortality(Stern et al. 2020). Consequently, asthma has emerged as a global public health concern, with China particularly impacted; over 45 million individuals are affected by this condition, with notable regional and gender disparities in prevalence(Cazzola et al. 2022; Kuruvilla et al. 2019). The scientific community currently recognizes that asthma is a complex and heterogeneous disorder rooted in an intricate interplay between genetic predisposition, individual physiological characteristics, and environmental exposures. These factors encompass a broad spectrum of environmental influences (including air pollution, pollen exposure, mold presence, and climatic variations), host determinants (such as obesity status, nutritional profile, infection history, and allergic tendencies), as well as genetic components (specific genetic polymorphisms associated with increased susceptibility to asthma)(Ahmed and Turner 2019). Although the precise pathophysiological mechanisms underlying asthma remain incompletely understood, key processes include airway inflammation, dysregulation of airway function regulation, and heightened bronchial reactivity. Glucocorticoid remain the most effective and preferred therapeutic option for asthma, exhibiting non-specific anti-inflammatory and immunosuppressive properties. Nevertheless, a significant number of asthma patients develop drug dependence and/or resistance during prolonged corticosteroid therapy, which poses a substantial challenge to their management(Dharmage et al. 2019). Consequently, advancing research into the pathophysiology of asthma and investigating novel therapeutic strategies is crucial for enhancing the health outcomes and quality of life for individuals with this condition. Asthma is a chronic inflammatory disease of the airways characterized by a complex interplay of immune and inflammatory responses(Chen et al. 2022). Research has demonstrated that T cells are pivotal in the development and progression of asthma(Chen et al. 2022). Antigens are captured by dendritic cells (DCs), which then process these antigens; activated DCs migrate to lymph nodes, where they influence T cell activation and Treg cell differentiation through immune synapses(Grakoui et al. 1999; Na et al. 2016). Traditionally, it has been widely accepted that an imbalance between Th1 and Th2 cells constitutes a significant pathological basis for allergic asthma(Dembele et al. 2021), with asthma being part of an inflammatory response driven by eosinophilia resulting from Th2 cell-derived chemokines(Hammad and Lambrecht 2015). However, therapies targeting only Th2 cells fail to fully mitigate airway inflammation(Yang and Shang 2019). Consequently, the pathophysiology of asthma cannot be solely attributed to this immune imbalance. Regulatory T cells (Treg cells), a pivotal inhibitory subset of CD4 + T cells, are intricately regulated by forkhead box protein 3 (Foxp3)(Wang et al. 2022). Treg cells play a pivotal role in maintaining immune homeostasis, functioning as the 'brake' system of the immune response. They effectively mitigate the overactivation of both innate and adaptive immune cells while strictly regulating their potential to release pro-inflammatory cytokines. This mechanism ensures that the immune system maintains a precise equilibrium when responding to exogenous or autoantigens(Jie et al. 2023; Ma et al. 2021). Recent studies have indicated that Treg cells cells and their secreted cytokines, interleukin-10 (IL-10) and IL-17, are implicated in the pathogenesis of asthma(Li et al. 2024; Zhou et al. 2022). Consequently, an in-depth investigation into the regulatory mechanisms governing Treg cells function is anticipated to emerge as a significant avenue for asthma treatment. Glycosylation is a biochemical process involving the transfer of sugar moieties to proteins mediated by glycosyltransferases, resulting in the formation of glycosidic bonds with amino acid residues and consequently influencing both the structure and function of proteins(Epp et al. 2016). This post-translational modification has been shown to modulate the activity and functionality of immune cells, including T cells, B cells, and macrophages(Muchowicz et al. 2024). Recent studies indicate that the biosynthesis of UDP-GlcNAc and subsequent N-glycosylation are critical determinants in establishing the Treg cells phenotype(Miao et al. 2022). As research into asthma pathogenesis advances, there is an increasing recognition of the pivotal role that glycosylation modifications play in asthma development. Thus, identifying drug targets associated with glycosylation modifications has emerged as a promising avenue for therapeutic intervention in asthma(Xie et al. 2023; Zhou et al. 2012). Treg cells is intricately linked to mitochondrial energy metabolism, with SIRT3 serving as a pivotal enzyme in the activation of mitochondria(Buck et al. 2016; He et al. 2021). Recent studies have demonstrated that SIRT3 expression is significantly downregulated in the bronchial tissues of asthmatic mice. Furthermore, upregulating SIRT3 not only mitigates apoptosis in bronchial epithelial cells from asthmatic mice but also markedly decreases cytokine levels (including TNF-α, IL-4, IL-5, and IL-13) in bronchoalveolar lavage fluid(Song and Wang 2022). Additionally, SIRT3-SUMO has been identified as playing a crucial role in asthma management(Liang et al. 2024). However, the potential for SIRT3-SUMO to modulate asthma and its relationship with N-glycosylation through Treg cells activation remains unclear. This study aims to elucidate the SUMO regulates Treg cell differentiation via N-glycosylation within an OVA-sensitized asthmatic mouse model, thereby providing new theoretical insights and research directions for asthma prevention and treatment. Methods GEO data collection In this study, gene expression data pertaining to asthma were retrieved from the GEO database (https://www.ncbi.nlm.nih.gov/geo/), and two datasets were selected using the keyword "asthma." The annotation file for all datasets utilized in this investigation is based on the Affymetrix Human Genome U133A Array (GPL96). Specifically, microarray datasets GSE147878 (comprising 60 asthma patients and 13 controls) and GSE43696 (including 88 asthma patients and 20 controls) were sourced from the GPL96 platform. The raw expression data from these microarray datasets underwent selective log2 transformation and normalization utilizing the "affy" package in R software (Version 3.48.3), guided by their numerical characteristics. Following preprocessing, batch effects across both datasets were mitigated using the "sva" software package (Version 3.40.0) to create a combined dataset. Further preprocessing was conducted with the "limma" package in R software (Version 3.48.3), employing lmFit and eBayes methods to eliminate batch effects within this combined dataset. Differential analysis was performed with thresholds set at |log2FC|> 0.1 and adjusted P-value < 0.05 between asthma samples and healthy control samples to identify differentially expressed genes. Weighted correlation network analysis (WGCNA) WGCNA analysis was performed utilizing the WGCNA package within R software. Following data preprocessing, a correlation matrix among genes was constructed. The 'pickSoftThreshold' function facilitated the calculation of the soft threshold, while the power function transformed the correlation matrix into a proximity matrix. The blockwiseModules function enabled the construction of a co-expression network for all genes and subsequent module segmentation. Module-trait correlations were assessed based on relationships between clinical features and module-specific gene expressions. Two modules exhibiting significant positive and negative correlations with clinical traits were identified, and corresponding gene information was extracted from these modules. Gene Ontology (GO) enrichment analysis was conducted using the 'clusterProfiler' package (version 4.0.5). Subsequently, KEGG pathways were visualized through an online tool provided by DAVID (https://david.ncifcrf.gov/), which also allows for secondary clustering of similar pathways. Animals All mice utilized in this study were on a C57BL/6 genetic background, with Cd4-Cre and C57BL/6 wild-type (WT) strains procured from the Jackson Laboratory. Female mice of comparable age and weight, aged 6–8 weeks, were randomly assigned to six groups of six mice each using a completely randomized design. These mice were housed in a specific pathogen-free (SPF) facility under controlled environmental conditions, including a 12 h light/dark cycle and a temperature maintained at 25 ± 1℃. The physiological status of the mice was monitored three times weekly prior to sensitization. The experimental animals were SPF-grade C57BL/6 mice.In this study, homozygous SIRT3 knock-in and SIRT3 K223R point mutant mice were generated using CRISPR/Cas9 technology and subsequently maintained through conventional mouse breeding. These strains were designated as SIRT3loxP/loxP and SIRT3K223RloxP/loxP, respectively. To achieve T cell-specific deletion of SIRT3, SIRT3loxP/loxP and SIRT3K223RloxP/loxP mice were crossed with Cd4-Cre transgenic mice to generate SIRT3loxP/+; Cd4-Cre heterozygous offspring (F1 generation). Heterozygous F1 mice were then intercrossed to obtain SIRT3loxP/loxP; Cd4-Cre homozygous mice (F2 generation), which were referred to as CD4CreSIRT3 WT and CD4CreSIRT3K223R, respectively. These conditional knockout lines were further expanded through breeding to yield a sufficient number of experimental animals for downstream analyses. Both knock-in and WT mouse cohorts were subsequently divided into control and asthma induction groups, comprising six animals per group. Mice in the control group received subcutaneous injections of phosphate-buffered saline (PBS) on day 0 followed by aerosol inhalation of PBS for 30 min on days 21, 22 and 23. In contrast, mice in the asthma-induced group were administered an intraperitoneal injection of 20 μg of Grade V ovalbumin (OVA) on Day 0, along with adjuvants consisting of either 0.2 ml of aluminum hydroxide or 1 µg of lipopolysaccharide (LPS; Pseudomonas aeruginosa 10; Sigma-Aldrich, Catalog No. L9143).On days 21, 22, and 23, mice were subjected to aerosol inhalation of 1% ovalbumin mist for 30 min to induce asthma. Mice were sacrificed 48 h after the final OVA exposure. The two asthma models are respectively represented by OVA/Alum and OVA/LPS. AAV-shCPT1, AAV-SIRT3 WT, and AAV-SIRT3 K223R were generated by Shanghai GeneChem Co., Ltd. An intranasal administration of 60 μL physiological saline containing 5 × 1010 AAV particles was performed on 4-week-old C57BL/6 mice, leading to the establishment of an asthma model within two weeks. PQ (primaquine, Sigma-Aldrich, CAT#160,393) was reconstituted in phosphate-buffered saline (pH 7.2) prior to administration. Wild-type mice were subjected to intraperitoneal injections at a dosage of 3 mg/kg for five consecutive days commencing on day 21. The mice were euthanized 48 h following the final injection for subsequent analysis. All the above experimental procedures and treatments were approved by the West China Hospital, Sichuan University, Chengdu, China Animal Experiment Research Ethics Committee. Histopathology The left lung of experimental mice was fixed in 4% paraformaldehyde for 24 h and subsequently embedded in paraffin. Following dewaxing and rehydration treatment, the left lung sections were stained with hematoxylin (Beyotime, CAT#C0105S) for 5 min, followed by eosin staining (Beyotime, CAT#C0105S) for 3 min. The cellular observations were made using a microscope (Olympus, Japan). Results were assessed based on an inflammatory scale: 0 (absence of inflammatory cells), 1 (occasional presence of inflammatory cells), 2 (1–3 layers of inflammatory cells surrounding the bronchus), 3 (4–5 layers of inflammatory cells around the bronchus or blood vessels), and 4 (more than 5 layers of inflammatory response around the bronchus or blood vessels). Periodic Acid-Schiff (PAS) staining was employed to visualize mucus-secreting cells and quantify the number of PAS-positive cells as a representation of positive cell count per unit length of epithelium. The protein expression was determined through immunohistochemical (IHC) staining. The tissues were dewaxed and hydrated, followed by removal of endogenous peroxidase activity using 3% hydrogen peroxide. Subsequently, the sections were cleaned with PBS three times, and then treated overnight at 4℃ in a buffer solution containing 5% BSA with anti-SIRT3 (Abcam, CAT#ab189860, 1:50). After that, the samples were incubated with biotin-labeled secondary antibodies at room temperature for 30 min. Horseradish peroxidase labeled streptavidin working solution was added to the sample and incubated at room temperature for 20 min. Following this step, the samples were washed with PBS three times before undergoing DAB (diaminobenzidine; Sigma) dyeing. The localization of FOXP3 and SIRT3 was detected using immunofluorescence (IF). The CPT1 and VLCAD protein expression was determined through IF. Paraffin sections were dewaxed to water and placed in a repair box filled with EDTA antigen repair buffer (pH 8.0) for antigen repair in a microwave oven. Then they were closed at 25°C with 10% serum for 45 min. The samples were subsequently incubated overnight at 4 ℃ with primary antibodies against SIRT3 (Affinity Biosciences, CAT#AF5135, 1:200), FOXP3 (Abcam, CAT#ab215106, 1:1000), anti-CPT1(Abcam, CAT#ab234111, 1:1000), and anti-VLCAD (Abcam, CAT#ab155138, 1:100). After incubation with the second antibody at room temperature in darkness for one hour, the images were collected and analyzed under confocal microscope. Cell culture and Treg cell differentiation A mouse CD4+ CD62L+ T cell isolation kit was employed to purify initial CD4+ T cells from the spleens and hilar lymph nodes of mice, utilizing magnetic beads for purification. The isolated cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) at 37 ℃ in a humidified atmosphere containing 5% CO2. Naïve CD4+T cells were treated with anti-CD3 antibody (1 μg/mL), anti-CD28 antibody (1 μg/mL), and recombinant TGF-β1 (5 ng/mL) for a duration of 72 h before being fixed onto culture dishes. Following Treg cell differentiation, the cells were harvested and analyzed via flow cytometry. Cell transfection For knockdown experiment, mouse siCPT1 (sense: 5' -GUCAAGCUGUCAAGAUAG-3'; Anti-sense: 5'- AUCUUGAACAGCUUGAACCUC–3’), siVLCAD (sense: 5'- UGGCCCGGUUCUUUGAGGAAGUGAA-3'. Anti-sense: 5'- UUCACUUCCUCAAAGAACCGGGCCA-3'); siPFK (sense: 5'-CAAGAUGUUUGCAAUCUAUGA-3'; anti-sense: 5'-AUAGAUUGCAAACAUCUUGUG-3’). siRNA or SIRT3 WT and SIRT3 K223R plasmids were transfected using Invitrogen™ Neon™ system according to manufacturer's instructions. EdU staining EdU staining experiments were conducted following the protocol provided by the EdU kit (APExBIO, CAT#K2243). Cells from the Control group, TGF-β group, and TGF-β + PNGase group were seeded into 6-well plates at a density of 2 × 105 cells per well. Subsequently, EdU reagent was added to achieve a final working concentration of 10 μM. The cells were incubated at 37℃ for 2–4 h. To clean, fix, and permeabilize the cells, DPBS, 4% PFA, and 0.5% Triton X-100 were applied in succession. Following this step, tap buffer was added to stain the nuclei with DAPI. Fluorescence microscopy was employed to capture images and perform statistical analysis. Mitochondrial separation Mitochondria were isolated through differential centrifugation. A total of 1 × 108 cells, were incubated in pre-cooled isolation buffer TD [135 mM NaCl, 5 mM KCl, 25 mM Tris–HCl (pH 7.5)], supplemented with a complete EDTA-free protease inhibitor mixture (Roche). Following centrifugation at 600 g for 10 min, the resulting pellet was washed with 10 ml of TD buffer and subjected to a second centrifugation at the same speed for an additional 10 min. The resultant particles were then resuspended in ice-cold isolation buffer MS containing 210 mM mannitol, 70 mM sucrose, 5 mM Tris–HCl (pH 7.5), and supplemented with EGTA (1 mM), BSA (0.5 mg/ml), along with the protease inhibitor mixture. Homogenization was performed using a glass Dounce homogenizer and pestle. The mitochondrial fraction was separated via differential centrifugation at first at 1300 g for five minutes followed by another spin at 17,000 g for fifteen minutes. Subsequently, the pellets were suspended in a solution of sucrose (1.5 M) containing EGTA (1 mM) and Tris–HCL (10 mM; pH 7.5). A thin layer was formed by gently overlaying this suspension with a concentrated sucrose solution (1 M). This layer was carefully removed from the tube before diluting the suspension fourfold using dilution buffer [containing EGTA (1 mM) and Tris–HCL (5 mM; pH = 7.5)] to achieve a final concentration of sucrose at approximately 250 mM sucrose. After centrifuging at 17,000 g for 15 min, purified mitochondria were resuspended in MS buffer and stored at −80℃. Western blot Use RIPA buffer (Beyotime, P0013B) containing protease and phosphatase inhibitors to lyse tissues and cells. Subsequently, collect the supernatant containing proteins and determine the protein concentration using the BCA protein assay kit (Beyotime, P0009). Separate the proteins via SDS-PAGE gel electrophoresis and transfer them onto a PVDF membrane. Block the membrane with TBST (containing 0.1% Tween-20 in Tris buffer) at room temperature for one hour. Incubate the membrane overnight at 4℃ with the appropriate primary antibody FOXP3 (Abcam, CAT#ab215106, 1:1000)、anti-SIRT3 (Abcam, CAT#ab189860, 1:1000)、SENP1(Abcam, CAT#ab236094, 1:1000)、anti-CPT1(Abcam, ab234111, 1:1000), anti-VLCAD(Abcam, CAT#ab155138, 1:1000)、pan-k-AC(Cell Signaling Technology, CAT#9441, 1:1000) diluted in TBST. Following this, incubate the membrane for two hours at 22℃ with an HRP-conjugated secondary antibody. Wash the blot three times with TBST before visualizing it using an ECL detection reagent kit; quantification can be performed utilizing the Gel-Pro analysis system. To detect N-glycoproteins, the protein transferred to a PVDF membrane was incubated with 5% BSA for 5 min to block non-specific binding sites. Subsequently, a 5 μg/mL Con A-HRP conjugate was added and incubated at 4℃ for 16 h. The membrane was then washed twice with PBS for 5 min each. Visualization was carried out using the ECL detection kit (ThermoFisher, CAT#34,579). qRT-PCR The total RNA was extracted from lung tissue, bronchoalveolar lavage fluid (BALF) and cells using Trizol reagent (Invitrogen, Cat#15596018CN). Subsequently, cDNA synthesis was performed using a reverse transcription kit (Takara, CAT#6215A). Quantitative real-time PCR (qRT-PCR) analysis was conducted employing 2xSYBR Master Mix (Applied Biosystems, CAT#4,309,155). β-Actin served as the internal reference gene for normalization purposes and the 2−ΔΔCt method was used to determine relative expression levels. Primer sequences utilized for qPCR in this study are provided in supplementary Table 1. Flow Cytometry (FCM) Following euthanasia of the mice, lungs were rapidly excised and immediately transferred into pre-cooled PBS or RPMI 1640 medium supplemented with 2% FBS to minimize cell death. The lung tissues were rinsed twice with pre-cooled PBS to remove residual blood and debris. Subsequently, the tissues were minced into fragments of approximately 1–3 mm3. The tissue fragments were then incubated in a digestion solution containing 200 U/mL collagenase IV and 0.5 U/mL DNase I at 37℃ with continuous shaking at 200 rpm for 1 h. Upon completion of digestion, the reaction was immediately terminated by adding 10 mL of PBS buffer. The digested samples were sequentially filtered through 100 μm, 70 μm, and 40 μm cell strainers to remove undigested tissue and cell aggregates. After centrifugation, the supernatant was discarded, and the cell pellet was resuspended in red blood cell lysis buffer and incubated for 3 min. The lysis reaction was stopped by adding 10 mL of PBS buffer, followed by centrifugation at 1,500 rpm at 4℃ for 5 min. Finally, the supernatant was removed, and the cells were resuspended in 1 mL of PBS buffer. The antibody was incubated in a lung single-cell suspension or in a 96-well plate of induced differentiation cells. It was initially incubated with FITC-anti-CD4 (Abcam, ab269349, 1 μg/mL) at 4 ℃ for 30 min, followed by fixation and permeabilization for 5 h. Subsequently, the cells were stained with PE-anti-Foxp3 (Abcam, CAT#ab218773, 0.25 μg) antibody for one hour and then suspended and analyzed using flow cytometry. For the n-glycosylation assay, cells were washed with PBS and subsequently incubated with PBS containing FITC-L-PHA (5 μg/mL) at 4 °C for 30 min. Following another round of PBS washing, the cells were tested on the machine. The gating strategy for the analysis of Treg cells in blood and lung tissue is as follows: Initially, the cell population is defined based on FSC-A versus SSC-A to exclude cellular debris. Subsequently, FSC-A versus FSC-W and SSC-A versus SSC-W are utilized to remove cell aggregates and identify single cells. Total CD45+ immune cells are then selected. Within the CD45+ population, Lin-DRAQ7⁻ cells are gated to isolate live cells and exclude other immune cell types. Among the CD45⁺Lin⁻DRAQ7⁻ cells, CD4⁺ T cells are identified. Finally, Treg cells are defined based on the co-expression of the key markers FoxP3 and CD25 (FoxP3⁺CD25⁺). Co-immunoprecipitation After boiling the obtained mitochondrial lysate for 5 min, it was incubated with the primary antibody overnight at 4℃. Subsequently, the mixture was incubated with 20 μL of protein A/G magnetic beads at room temperature for 6 h. The composite collected by the magnetic beads was re-suspended in 2 × SDS dosing buffer and boiled for 10 min. Finally, the samples were further analyzed using Western blotting technique with anti-SIRT3 (Abcam, CAT#ab189860, 1:100) and SUMO1 (Abcam, CAT#ab32058, 1:20). Incubate the primary antibody at room temperature with shaking for 1 h or overnight at 4℃. Incubate the primary antibody at room temperature with shaking for 1 h or overnight at 4℃. Incubate the secondary antibody at 4℃ for 2 h or at room temperature (22–25℃) with shaking for 1 h. Collection of bronchoalveolar lavage fluid (BALF) and related assay Following the sacrifice of the mice, a polyethylene tube was carefully inserted into the trachea. The trachea was accessed by making a midline incision in the neck using a 14-gauge needle, after which tracheal intubation was performed. Subsequently, 0.5 mL of cold, sterile phosphate-buffered saline (PBS) was injected into the right lung of each mouse, and the lungs were lavaged twice. The BALF obtained was centrifuged at 500 × g for 10 min. The collected BALF samples were stored at −20℃ until further analysis. Total cell counts were determined using the TC 20™ Automated Cell Counter (BIO-RAD, USA), with cells precipitated in a 4% formaldehyde solution and stained with Wright-Giemsa stain. A total of 200 BALF cells were classified and counted under microscopy. The supernatant from BALF was preserved at −80℃ for subsequent cytokine detection, specifically measuring IL-10 concentration utilizing the Mouse IL-10 ELISA Kit (Beyotime, CAT#PI522) according to manufacturer instructions. Measurement of serum IgE level Blood was collected from mice and allowed to stand at room temperature for 30 min to 2 h without vigorous agitation to prevent hemolysis, facilitating natural coagulation and serum separation. The serum was then centrifuged at 4 ℃ at a speed of 1000–2000 × g for 10 min. The resulting yellow supernatant, which constitutes the serum, was utilized for IgE level measurement using an ELISA kit (Beyotime, CAT#PI476) following the manufacturer's instructions. A volume of 100 μL of each sample was added to individual wells, which were subsequently sealed with a transparent film and incubated at room temperature for 120 min. Following five washes, 100 μL of horseradish peroxidase-labeled mouse IgE antibody was added to each well in the dark at room temperature for an additional incubation period of 20 min. After another five washes, TMB solution (100 μL) was introduced into each well and incubated in the dark at room temperature for another 20 min. Finally, a stop solution (50 μL) was added to each well and mixed thoroughly before measuring the absorbance value at A450. Airway hyperreactivity test The model mice were anesthetized with pentobarbital, intubated, and connected to a ventilator and nebulizer. Subsequently, airway reactivity was assessed across different groups using an invasive mouse pulmonary function meter (Buxco, USA), with methacholine (Mch) serving as the stimulus. The impact of Mch on airway resistance in the model mice was evaluated by measuring its effects at concentrations of 0, 3.125, 6.25, 12.5, and 25 mg/mL. Measurement of fatty acid uptake In the presence or absence of TGF-β (5 ng/mL), cells were treated with anti-CD3/CD28 for 72 h. Following collection, the cells were washed and resuspended in PBS containing BODIPY-C16 (100 nM) at 37℃ for a duration of 30 min. After incubation, the cells underwent another washing step, and the uptake of BODIPY-C16 by the cells was assessed using flow cytometry. Measurement of oxygen consumption rate (OCR) In accordance with the manufacturer's guidelines from Seahorse Bioscience (North Billerica, USA), oxygen consumption rate (OCR) was assessed using the Seahorse XF96 Bioanalyzer. The Mito Stress Test Kit was employed for this experiment. Concentrations of 1.5 μM oligomycin, 2 μM of 4-(trifluoromethoxy) phenylhydrazone (FCCP), and 0.5 μM physcion/antimycin A were administered to each well. JC-1 staining In the presence or absence of TGF-β (5 ng/mL), CD4 T cells were primed with anti-CD3/CD28 for a duration of 72 h. Following the manufacturer's protocol, the JC-1 assay kit (ThermoFish, CAT#M34152) was employed to assess alterations in mitochondrial hyperpolarization and depolarization. Briefly, the JC-1 stock solution (200 ×) was diluted to a working concentration (1 ×) and subsequently added to the cells, which were then incubated at 37℃ for 20 min. After incubation, the cells were washed with washing buffer (1 ×), and fluorescence intensity was quantified using flow cytometry. The red/green fluorescence ratio of JC-1 was calculated based on the average fluorescence intensity from each channel. Measurement of mitochondrial mass Utilize the Mito Tracker Green Kit (Beyotime, CAT#C1048) to assess mitochondrial activity. Begin by taking a small volume of the 1 mM Mito-Tracker Green storage solution and diluting it at a ratio ranging from 1:5000 to 1:50,000 with cell culture medium to prepare the working solution. After removing the existing cell culture medium, add the prewarmed Mito-Tracker Green staining working solution at 37℃ and incubate the cells for a duration of 15 to 45 min at this temperature. Subsequently, discard the staining working solution and replace it with fresh cell culture medium that has also been prewarmed to 37℃ Finally, observe the cells using a fluorescence microscope. Uridine diphosphate-N-acetylglucosamine (UDP-GlcNAc) assay Initially, 200 μL of a chloroform/water (1:1) solution was employed to collect, wash, and solubilize the cells. Subsequently, the sample was centrifuged at 4℃ for 20 min at 15,000 × g for an additional 2 min. The supernatant was carefully transferred to a new tube, followed by the addition of 10 μL of hydrochloric acid (1 N) to facilitate the hydrolysis of UDP-GlcNAc into GlcNAc. The mixture was then heated at 80℃ for 20 min and neutralized with 10 μL of potassium hydroxide (1 N). Next, 50 μL of potassium tetraborate (200 mM) was introduced and incubated at 80 ℃ for an additional 25 min. Afterward, the sample was cooled on ice for five minutes. Following this step, 150 μL of diluted acetic acid (1:2) was added to the sample and incubated at 37℃ for another twenty minutes. Finally, the sample underwent centrifugation at 15,000 × g for twenty minutes before measuring absorbance at a wavelength of 595 nm. Measurement of phosphofructokinase (PFK) and glutamine-fructose-6-phosphate transaminase (GFPT) enzymatic activity Following the methodology established by Miao et al.(Miao et al. 2022), the activity of GFPT was assessed. Cells were subjected to sonication in an extraction buffer composed of 60 mM KH2PO4 (pH 7.8), 50 mM KCL, 1 mM EDTA, and 1 mM DTT, after which the lysate was collected. The enzymatic reaction took place in a reaction buffer containing 60 mM KH2PO4 (pH 7.8), 50 mM KCL, 1 mM EDTA, and 1 mM DTT supplemented with 15 mM fructose-6-P and l-glutamine at a concentration of 15 mM for one hour at 37℃; this reaction was subsequently terminated by heating at 95℃ for 2 minutes. In the negative control group, all proteins were denatured immediately by heating at the same temperature for two minutes post-reaction to establish a baseline measurement of glutamate levels for background subtraction purposes. Following centrifugation, the supernatant was harvested and quantified using a glutamate assay kit (Solarbio, CAT#BC1580) according to manufacturer guidelines. The activity of PFK within cells was determined following protocols outlined in the PFK activity assay kit (Solarbio, CAT#BC0535). A total of 1 × 104 cells were collected and diluted with extraction solution at a ratio of one part cell suspension to five hundred parts extraction solution. After sonication, lysates underwent centrifugation at 8000 g for 10 min at four degrees Celsius; thereafter, supernatants were obtained. The working solution was preheated to thirty-seven degrees Celsius for ten minutes prior to use in a ninety-six well plate format where each well received ten microliters each of sample and reagents III and IV followed by one hundred seventy microliters of working solution. Absorbance readings were taken at 340 nm text over 20 s post-mixing; subsequent measurements occurred after incubation at 37℃. Celsius for an additional 10 min with absorbance recorded again 20 s later. Statistical analysis All data were analyzed using Prism 9.0 and are presented as means ± standard deviation (SD). First, a normality test (Shapiro–Wilk test) was conducted on the data to ensure that the data met the prerequisite conditions for analysis of variance. All groups of data conformed to a normal distribution (p > 0.05). An unpaired Student's t-test was used to compare the two groups while a one-way ANOVA was employed for comparisons among multiple groups. And Tukey's HSD test was used for multiple comparison correction. Statistical significance was established at P < 0.05. Results N-glycosylation level and Treg cells proportion were reduced in asthma. To explores the genes or signaling pathways that may play regulatory roles in the occurrence and progression of asthma, this study used the combined dataset of GSE1478784 and GSE43696 for differential expression analysis. The GEO database shows that the two datasets include a total of 33 normal samples and 148 asthma patient samples. The volcano plot shows that there are 239 differential genes, with 116 upregulated genes and 123 downregulated genes (Supplementary Fig. 1 A). Next, we used WGCNA analysis to further understand the association between key modules and different types of asthma. When the scale-free fitting index R2 was 9 and the minimum soft thresholding parameter β was 6 (Supplementary Fig. 1B), we constructed a co-expression network based on the optimal soft thresholding value. In this study, we constructed 50 gene modules (Supplementary Fig. 1C). The heatmap shows the correlation between different modules and asthma subtypes. The results show that the Midnightblue module (r = 0.3, p = 4E-05) is highly positively correlated with asthma, while the Yellow green module (r = −0.24, p = 0.001) is highly negatively correlated (Supplementary Fig. 1D). Next, to further analyze the characteristics of the module genes, we performed GO annotation and KEGG enrichment analysis on the genes in the two modules. The results show that N-glycosylation may play an important role in asthma (Supplementary Fig. 1E). Asthma is categorized into Th2-type and non-Th2-type based on neutrophil profiles. Th2-type asthma is characterized by an elevation in eosinophils, whereas non-Th2-type asthma exhibits an increase in neutrophils(Sockrider and Fussner 2020). Bioinformatics analyses indicate that N-glycosylation plays a crucial role in the pathogenesis of asthma. To validate this finding, we sensitized mice with aluminum hydroxide adjuvant and subsequently exposed their airways to ovalbumin (OVA) to induce Th2-type asthma (Fig. 1A). For the induction of non-Th2-type asthma, we employed OVA combined with lipopolysaccharide (LPS) sensitization (Supplementary Fig. 1E). Lung tissue was harvested from mice 48 h post-final OVA exposure and subjected to hematoxylin–eosin (HE) and PAS staining. HE staining revealed intact bronchial walls and minimal inflammatory cell infiltration in control group lungs; conversely, asthmatic mice exhibited thickened airway walls alongside significant inflammatory cell infiltration, increased connective tissue deposition, and pronounced alveolar septal thickening. PAS staining demonstrated a marked increase in goblet cells along with heightened mucus secretion in asthmatic mice (Fig. 1B, Supplementary Fig. 1F-G). Subsequently, we assessed N-glycosylation levels within lung tissues of asthmatic mice and observed a significant downregulation of N-glycosylation markers (Fig. 1C, Supplementary Fig. 1H), suggesting its involvement in asthma development. Treg cells, regulated by FOXP3 expression, are pivotal for maintaining immune homeostasis. Enhancing Treg cells populations has emerged as a promising strategy for managing inflammation and autoimmune disorders(Miao et al. 2022). We quantified Treg cells proportions across bronchoalveolar lavage fluid (BALF), lung tissue suspensions, and blood samples; notably, there was a significant reduction in Treg cells numbers (Fig. 1D–G, Supplementary Fig. 1I-J). Since IL-10 is the principal inflammatory factor produced by Treg cells. The determination results indicated that the content of IL-10 in the asthma model also presented a significant downregulation (Fig. 1H-J). Moreover, the proportion of IL10+FOXP3+ cells was significantly decreased in the asthma model (Fig. 1K, Supplementary Fig. 1L), which further verified that IL-10 related to Treg cells plays an important role in the process of asthma. In conclusion, N-glycosylation and Treg cells played a significant role in the development of Th2-type and non-Th2-type asthma. N-glycosylation and Treg cell differentiation influenced asthma progression. A schematic representation of the construction of an allergic asthma mouse model induced by OVA/Alum is provided. B Lung tissue H&E and PAS staining. Scale bar = 100 μm for H&E; Scale bar = 50 μm for PAS. The inflammation score and PAS score of lung tissue in asthmatic mice increased., n = 5. C The Western Blot results indicated that N-glycosylation in the lung tissue of asthmatic mice could be reduced, n = 5. D Flow cytometry analysis results showed that the number of Treg cells in the BALF, lung tissue and blood of mice was significantly reduced, n = 5. E The expression levels of FOXP3 mRNA in BALF, lung tissue were reduced, n = 5. F-G FOXP3 protein expression in lung tissues was reduced, n = 5. H-J The ELISA results showed that the content of IL-10 was decreased in BALF, lung and blood, n = 5. All data are presented as mean ± standard deviation (SD). Statistical significance was determined at *p < 0.05 compared to the Control group N-glycosylation promoted the Treg cell differentiation Treg cells are derived from naïve CD4+ T cells. To explore the relationship between N-glycosylation and Treg cell differentiation, we isolated naïve T cells from the lungslymph nodes and spleens of mice and induced their differentiation into Treg cells under TGF-β stimulation. L-PHA specifically binds to the β1,6GlcNAc branch of the N-glycan chain, serving as a highly sensitive and quantitative marker for N-glycosylation (Mkhikian et al. 2022). Flow cytometry analysis revealed that the level of N-glycosylation in Treg cells was significantly elevated compared to that in naïve T cells (Fig. 2A). This finding suggests a role for N-glycosylation in facilitating Treg cell differentiation. PNGase acts as an inhibitor of N-glycosylation; CCK8 assays demonstrated its non-toxic nature at concentrations ranging from 0 to 40 U/μL (Fig. 2B). Treatment with PNGase at a concentration of 25 U/μL markedly inhibited both Treg cell differentiation (Fig. 2C-E) and reduced mRNA expression levels of the pro-inflammatory cytokine IL-10 within these cells (Fig. 2E). EdU staining alongside apoptosis detection indicated that PNGase treatment did not significantly affect either cell proliferation or apoptosis rates (Fig. 2F-G), further supporting the notion that N-glycosylation is instrumental in promoting functional differentiation of Treg cells. N-glycosylation was implicated in the differentiation of Treg cells. A Flow cytometry results showed that the mean fluorescence intensity (MFI) of L-PHA in Treg cells was higher than that in naive CD4+ T cells. B Cell viability was evaluated using CCK8 assay. C-E Flow cytometry, WB and qRT-PCR results showed that TGF-β treatment induced Treg cell differentiation, and the combination of PNGase inhibited this effect. F-G TGF-β and PNGase had no effect on cell proliferation and apoptotic. Scale bar = 10 µm SIRT3-SUMO was involved in Treg cell differentiation through N-glycosylation Mitochondrial dynamics play a crucial role in T cell differentiation by modulating metabolic reprogramming, with mitochondrial acetylation serving as a significant regulatory mechanism for mitochondrial enzymes(Buck et al. 2016; Song and Wang 2022). Consequently, we assessed the expression of the deacetylase SIRT3 within the hilar lymph nodes induced by ovalbumin (OVA) in asthmatic mice and observed that SIRT3 was markedly downregulated in the asthma group (Fig. 3A-B). Immunofluorescence analysis revealed co-localization of SIRT3 and FOXP3 in the tracheal lymph nodes of these mice (Fig. 3C). Subsequently, mitochondria were isolated and their acetylation levels were quantified. The results revealed that the acetylation levels in Treg cells were markedly reduced compared to those in naïve CD4+ T cells (Fig. 3D). Moreover, there is an increasing trend in the expression of SIRT3 in Treg cells (Fig. 3E). The aforementioned results indicate that SIRT3-regulated acetylation may play a role in the modulation of Treg cell differentiation. SIRT3-SUMO modulated Treg cell differentiation via N-glycosylation. A-B IHC and WB results showed reduced SIRT3 expression in the pulmonary lymph nodes of asthmatic mice, n = 5, Scale bar = 200 µm. C The immunofluorescence results demonstrated that SIRT3 (red) and FOXP3 (green) exhibited co-localization in the lung tissue of asthmatic mice, Scale bar = 200 µm. D The mitochondrial acetylation level in the TGF-β treated group was lower than that in the control group, n = 5. E The expression of SIRT3 in the TGF-β treated group was higher than that in the control group, n = 5. F The binding of SIRT3 to SUMO1 was weakened in the TGF-β treated group cells, n = 5. G TGF-β treatment upregulated the expression of SENP1 in mitochondria, n = 5. H Immunofluorescence microscopy was used to visualize the localization patterns of SIRT3 (red) and SUMO1 (green) proteins, n = 5, Scale bar = 10 µm. I-K The results of flow cytometry, WB and qRT-PCR indicated that SIRT3 WT and K223R promoted the differentiation of Treg cells, n = 5. L Flow cytometry results showed that SIRT3 WT and K223R increased the N-glycosylation level of cells. M–O Flow cytometry results showed that compared with SIRT3 WT, SIRT3 K223R promoted the N-glycosylation level (M-N) of cells and the proportion of Treg cells (O), and this difference disappeared after the addition of PNGase, n = 5. The results are expressed as mean ± SD, *p < 0.05 SENP1, a SUMO-specific protease, regulates the stability of SIRT3 protein and enhances its deacetylase activity through de-SUMOylation(Wang et al. 2019). However, the SENP1-SIRT3 interaction has not yet been verified in Treg cells. However, the SENP1-SIRT3 interaction has not yet been verified in Treg cells. we observed a marked upregulation of SIRT3 SUMOylation levels in the hilar lymph nodes of mice (Supplementary Fig. 2 A), and a significant downregulation of SIRT3-SUMOylation levels in Treg cells induced in vitro (Fig. 3F). Furthermore, the expression of the de-SUMOylase SENP1 was elevated in the mitochondria of Treg cells (Fig. 3G). Immunofluorescence corroborated these results, demonstrating diminished co-localization between SIRT3 and SUMO1 within Treg cells (Fig. 3H). Additionally, we observed a marked upregulation of SIRT3 SUMOylation levels in the hilar lymph nodes of mice (Supplementary Fig. 2 A), with both SUMO-GPS (https://sumo.biocuckoo.cn/advanced.php) and SUMOplot (https://www.abcepta.com/sumoplot) identifying K223 as the site for SUMOylation (Supplementary Fig. 2B). Subsequently, we transfected wild-type SIRT3 and mutant SIRT3 K223R into naïve T cells to induce Treg cell differentiation in vitro. Immunoprecipitation analyses indicated that mutation at the K223 site significantly reduced SIRT3's SUMOylation level (Supplementary Fig. 2 C). Moreover, inhibition of SIRT3 SUMOylation markedly promoted Treg cells differentiation (Fig. 3I–K), suggesting that SIRT3-SUMO is instrumental in regulating Treg cell differentiation. Previous studies have demonstrated that N-glycosylation plays a crucial role in the differentiation of Treg cells. To explore whether SIRT3-SUMO influences the N-glycosylation process, we assessed the levels of N-glycosylation in SIRT3 WT and SIRT3 K223R cells following TGF-β induction. Our results revealed that N-glycosylation levels were significantly elevated in SIRT3 K223R cells compared to those in SIRT3 WT cells, accompanied by a marked increase in the proportion of Treg cells. Notably, upon administration of the N-glycosylation inhibitor PNGase, this difference was abolished (Fig. 3L–O). Conversely, when additional GlcNac substrate for N-glycosylation was introduced, the disparity re-emerged (Supplementary Fig. 2D–E), suggesting that the modulation of Treg cell differentiation by SIRT3-SUMO is contingent upon N-glycosylation. SIRT3-SUMO influenced asthma progression by affecting Treg cell differentiation in vivo We generated CD4Cre SIRT3 WT and CD4Cre SIRT3 K223R knock-in mice, designated as CD4CreSIRT3 WT and CD4CreSIRT3 K223R (Fig. 4A). OVA sensitization resulted in heightened inflammation and mucus production in the lungs of these mice (Fig. 4B-C, Supplementary Fig. 3 A). Notably, the introduction of SIRT3 WT significantly mitigated the increase in both inflammation and mucus levels; conversely, the knock-in of SIRT3 K223R further alleviated lung inflammation while also reducing mucus accumulation (Fig. 4B-C, Supplementary Fig. 3 A). Correspondingly, serum IgE levels exhibited a similar trend (Fig. 4D, Supplementary Fig. 3B). We assessed airway reactivity to inhaled acetylcholine using Flexvent in both non-allergic control mice and OVA-sensitized mice. The results indicated an increased airway reactivity to acetylcholine in OVA-sensitized subjects. Both SIRT3 WT and SIRT3 K223R demonstrated a degree of reduction in allergen-induced airway hyperreactivity (AHR), with SIRT3 K223R exhibiting a more pronounced effect compared to its WT counterpart (Fig. 4E, Supplementary Fig. 3 C). Furthermore, SIRT3 WT and SIRT3 K223R knock-in mice exhibited a significant reduction in the numbers of neutrophils and eosinophils present in bronchoalveolar lavage fluid (Fig. 4F, Supplementary Fig. 3D-E)., Notably, flow cytometric analysis revealed a significant reduction in the proportion of Treg cells in the asthma model. However, this decrease was markedly reversed in asthma models established in CD4Cre SIRT3 WT and CD4Cre SIRT3 K223R mice, with a significant restoration of Treg cell frequency (Fig. 4G, Supplementary Fig. 3 F). Given the established role of N-glycosylation in Treg cell differentiation, we assessed L-PHA levels—a lectin probe specific for β1–6 branched N-glycans—by flow cytometry. The results showed that changes in L-PHA signal intensity paralleled the differentiation dynamics of Treg cells, further supporting a potential link between N-glycosylation and Treg development (Fig. 4H, Supplementary Fig. 3G). The use of CD4-cre mice to generate SIRT3 WT and SIRT3 K23R knock-in models may impact not only CD4+ T cells but also other CD4-expressing immune cell populations. To evaluate potential off-target effects, we analyzed the proportions of monocytes, macrophages, dendritic cells, and iNKT cells in mouse lung tissues. Our results demonstrate that in the OVA-induced asthma model, the frequencies of these cell types were significantly elevated compared to control mice. In contrast, no significant changes in the proportions of monocytes, macrophages, dendritic cells, or iNKT cells were observed in either the CD4Cre SIRT3 WT or CD4Cre SIRT3 K223R asthma models (Supplementary Fig. 4). These findings further support a specific role for SIRT3-SUMOylation in regulating Treg cell differentiation during asthma pathogenesis, rather than exerting broad effects on innate or invariant T cell populations. In conclusion, the downregulation of SIRT3-SUMO levels results in increased SIRT3 expression, which may mitigate Th2-type and non-Th2-type asthma progression by facilitating Treg cell differentiation. SIRT3-SUMO influenced Treg cell differentiation and asthma progression via N-glycosylation. A Western blot analysis to assess SIRT3 expression. B-C HE staining, along with PAS staining of lung tissue, followed by immune scoring and PAS scoring in the OVA/Alum-induced asthma model. Scale bar = 100 µm. D-F SIRT3 WT and K223R mice showed reduced lung tissue content of IgE in serum (D), airway resistance (E), total cell counts and eosinophil numbers in BALF(F), under OVA/Alum-induced conditions. G-H Flow cytometry results showed that SIRT3 WT and SIRT3 K223R mice increased the N-glycosylation level in the lung lymph nodes (H) and the content of Treg cells in the lung tissue (G). All data are expressed as mean ± standard deviation, n = 5 FAO participates in the SIRT3-SUMO-driven differentiation of Treg cells Exogenous fatty acid β-oxidation (FAO) serves as the primary energy source for Treg cell differentiation, with its conversion to acetyl-CoA occurring in the mitochondria(Miao et al. 2022). We subsequently investigated whether SIRT3-SUMO influences the FAO process. In vitro differentiation experiments demonstrated that overexpression of SIRT3 enhanced palmitate uptake (Fig. 5A), increased mitochondrial mass, oxygen consumption rate (OCR), and mitochondrial membrane potential (Figs. 5B-D). Notably, these parameters were further elevated upon inhibition of SIRT3-SUMO. Measurement of acetyl-CoA levels, a key product of FAO, revealed a similar trend (Fig. 5E), indicating that SIRT3 promotes the FAO pathway. To elucidate the underlying molecular mechanisms, we analyzed interactions between SIRT3 and carnitine palmitoyl transferase 1 (CPT1) as well as very long-chain acyl-CoA dehydrogenase (VLCAD) using BioGRID and STRING databases (Supplementary Figs. 5A-B). Additionally, we assessed the expression levels of FAO-related enzymes in both SIRT3 WT and SIRT3 K223R cells; results indicated significant upregulation of CPT1 and VLCAD expression (Figs. 5F-J). Furthermore, protein levels of CPT1 and VLCAD were markedly downregulated in lung hilar lymph nodes from a mouse model of asthma (Supplementary Figs. 5C-D). FAO participated in SIRT3-SUMO-induced Treg cell differentiation. A The BODIPY-C16 flow cytometry result indicated that SIRT3 WT and SIRT3 K223R increased cellular palmitic acid uptake. B The immunofluorescence results indicated that SIRT3 WT and SIRT3 K223R enhanced mitochondrial viability. Scale bar = 5 µm. C The JC-1 flow cytometry results indicated that SIRT3 WT and SIRT3 K223R increased mitochondrial membrane potential. D OCR assays demonstrated that SIRT3 WT and SIRT3 K223R increased cellular respiration. E SIRT3 WT and SIRT3 K223R increase the content of acetyl-CoA in cells. F-J Expression levels of FAO-related enzymes—(F) CPT1, (G) CPT2, (H) VLCAD, (I) HADHA, and (J) HADAB—were determined by qRT-PCR analysis. All data are presented as mean ± standard deviation, *p < 0.05, n = 5 To further elucidate the role of SIRT3-SUMO in regulating Treg cell differentiation via the FAO pathway, we assessed the temporal relationship between FAO and cellular differentiation. The findings indicated that both SIRT3WT and SIRT3 K223R significantly upregulated CPT1 and VLCAD expression at 24 h, while FOXP3 expression began to increase only at 48 h (Fig. 6A-C). Additionally, we established CD4T cell-derived lines with knockdown of CPT1 and VLCAD (Fig. 6D-E) and observed that mRNA and protein levels of SIRT3 remained unchanged in these knockdown cells (Supplementary Fig. 5E-F). Under TGF-β-induced conditions of Treg cell differentiation, knockdown of CPT1 and VLCAD significantly suppressed Treg cell differentiation. Overexpression of SIRT3 and induction of de-SUMOylation were found to significantly enhance Treg cell differentiation. Notably, the promotive effect induced by SIRT3 overexpression and de-SUMOylation was significantly attenuated upon knockdown of CPT1 and VLCAD in these cells.. The expression levels of FOXP3 and IL-10 exhibited a consistent pattern of change. (Fig. 6F-G). Long-chain fatty acids are transported into the mitochondria via the action of CPT1, subsequently entering the beta-oxidation pathway; thus, CPT1 is recognized as the rate-limiting enzyme in FAO(Ngo et al. 2023). Our experimental findings further demonstrate that CPT1, acting as an upstream regulator of VLCAD, modulates both FAO and Treg cell differentiation (Supplementary Fig. 6A-B). Additionally, we observed that Etomoxir, a specific inhibitor of CPT1, significantly attenuates the enhancing effects of SIRT3 and its K223R variant on Treg cells proportions and mRNA expression levels of FOXP3 and IL-10 (Supplementary Fig. 6C-D). These results indicate that SIRT3 functions as an upstream molecule regulating FAO through CPT1 and VLCAD to influence Treg cell differentiation. SIRT3-SUMO induces Treg differentiation dependent on CPT1/VLCAD-mediated FAO. A-C qRT-PCR was employed to detect the expressions of FOXP3, CPT1, and VLCAD at various time points. D-E The expression of CPT1 was detected. G The expressions of FOXP3 and IL-10 were detected by qRT-PCR. The above data are presented as the mean ± standard deviation, *p < 0.05, n = 5 SIRT3-SUMO-mediated FAO regulated N-glycosylation through the hexosamine biosynthetic pathway (HBP) pathway We have demonstrated that SIRT3-SUMO can modulate Treg cell differentiation through the n-glycosylation pathway. Additionally, acetyl-CoA, the end product of FAO, can be utilized to synthesize UDP-GlcNAc, a glycosylated substrate involved in the HBP pathway. Therefore, we hypothesized that FAO serves as an intermediate link in SIRT3-SUMO's regulation of n-glycosylation during Treg cell differentiation. Firstly, we assessed the levels of UDP-GlcNAc and observed a significant reduction after CPT1 knockdown. Furthermore, CPT1 knockdown attenuated the enhancing effect of both SIRT3 WT and SIRT3 K223R on UDP-GlcNAc levels. Notably, this difference between the SIRT3 WT and SIRT3 K223R groups disappeared upon CPT1 knockdown (Fig. 7A). Correspondingly, changes were also observed in N-glycosylation levels (Fig. 7B-C). The level of acetyl coenzyme A is related to multiple epigenetic regulations and gene expressions in cells. We investigated which part of acetyl coenzyme A during the differentiation process of Treg cells is regulated by the FAO process and whether it exerts a biological function in Treg cell differentiation. We initially generated culture media with and without Etomoxir. The precursor substances for acetyl coenzyme A synthesis, acetic acid and citric acid, were added to the culture media for the in vitro induction of Treg cell differentiation. The results revealed that adding acetic acid and citric acid to the culture medium without Etomoxir could increase the level of UDP-GlcNAc and promote Treg cell differentiation (Fig. 7D-F), while in the culture medium with Etomoxir, the addition of acetic acid and citric acid failed to alleviate the inhibitory effect on Treg cell differentiation (Fig. 7D-F), suggesting that acetic acid and citric acid cannot synthesize acetyl CoA bypassing the FAO pathway. Additionally, adding acetic acid and citric acid to STRT3 WT cells and STRT3 K223R cells could facilitate the production of UDP-GlcNAc and Treg cell differentiation. However, when CPT1 was knocked down, the production of UDP-GlcNAc and Treg cells differentiation were significantly reduced, and the addition of acetic acid and citric acid could not alleviate this situation. It is notable that when CPT1 was knocked down in STRT3 WT and STRT3 K223R cells, the effects of acetic acid and citric acid disappeared (Supplementary Fig. 7A-B). Secondly, using the mitochondrial citrate transporter SLC25A1 inhibited by BTC and ATP citrate lyase (Acly) inhibited by SB204990 did not influence the level of UDP-GlcNAc and Treg cell differentiation (Fig. 7G-H). This indicates that the acetyl CoA generated by the SIRT3-SUMO-mediated FAO pathway can affect Treg cell differentiation through the HBP pathway. Acetyl-CoA, a product of FAO, regulated Treg differentiation via N-glycosylation. A-C SIRT3 WT and SIRT3 K223R increase the levels of UDP-GlcNac (A) and N-glycosylation (C) in cells, and this effect is counteracted when CPT1 is knocked down. D The effect of Acetate and Citrate in increasing the levels of UDP-GlcNac in cells is inhibited by Etomoxir. E–F Flow cytometry and qRT-PCR results indicated that Acetate and Citrate promoted the differentiation of Treg cells, and this effect was inhibited by Etomoxir. G-H BTC and SB204990 have no effect on the UDP-GlcNac level (G) and the proportion of Treg cells (H). All data are presented as mean ± standard deviation, *p < 0.05 Fructose-6-phosphate, generated through glycolysis, serves as a crucial substrate for the HBP. This metabolite can be further integrated into the glycolytic pathway under the regulation of phosphofructokinase (PFK) and may also enter the HBP to facilitate the synthesis of N-glycosylation substrates via GFPT enzyme activity(Araujo et al. 2017). PFK activity is modulated by citrate and ATP levels. Consequently, we investigated whether fatty acid oxidation (FAO) influences this process. Our findings revealed that knockdown of CPT1 and VLCAD did not alter GFPT or PFK expression but significantly enhanced PFK activity without affecting GFPT activity (Fig. 8A-B). Overexpression of SIRT3 markedly reduced PFK activity, while inhibition of SIRT3 SUMOylation further diminished PFK activity (Fig. 8C), an effect that was counteracted by CPT1 knockdown (Fig. 8D). Subsequently, we established a cell line with PFK knockdown (Fig. 8E-F) to induce Treg cell differentiation in vitro. The results indicated that in conditions where PFK was knocked down, the enhancement of N-glycosylation and Treg cell differentiation mediated by both wild-type SIRT3 and SIRT3 K223R was abolished (Fig. 8G-I). These observations suggest that PFK plays a pivotal role in FAO-mediated N-glycosylation processes regulated by SIRT3-SUMO dynamics. In scenarios characterized by increased FAO, SIRT3 along with its deSUMOylated form inhibits PFK activity while promoting HBP. PFK linked FAO and N-glycosylation regulated by SIRT3-SUMO. A The mRNA levels of GFPT and PFK showed no significant changes. B Knocking down CPT1 and VLCAD increases PFK activity. SIRT3 WT and SIRT3 K223R were transfected into naive CD4 T cells. C SIRT3 WT and SIRT3 K223R inhibit PFK activity. D SIRT3 WT and SIRT3 K223R regulate PFK activity through CPT1. E–F A PFK-knockdown cell line was constructed, and the expression of PFK was detected by WB and qRT-PCR. G-I The upregulation effects of SIRT3 WT and SIRT3 K223R on UDP-GlcNac content and N-glycosylation level (G-H) and the proportion of Treg cells (I) are dependent on the presence of PFK. All data are represented as the mean ± standard deviation, *p < 0.05, n = 5 Activating Treg cells could slow asthma progression We constructed AAV-SIRT3 WT, AAV-SIRT3 K223R, and AAV-shCPT1 to modulate the expression of SIRT3 and CPT1 in asthmatic mice, aiming to investigate the roles of inflammation-related SIRT3 and FAO in Treg cells generation. Notably, shCPT1 counteracted the inhibitory effects exerted by both SIRT3 WT and SIRT3 K223R on pulmonary inflammation, mucus production, serum IgE levels, and cellular counts in BALF (Fig. 9A-F; Supplementary Fig. 8A-F), The airway resistance response to the acetylcholine methacholine provocation test showed the same result (Fig. 9G). as well as enhancing Treg cells generation within lung tissue (Fig. 9H). Furthermore, primaquine (PQ), initially developed as an antimalarial agent, has recently been shown to induce FOXP3 expression thereby facilitating Treg cell differentiation and ameliorating multiple sclerosis (MS). Our experimental findings corroborate that PQ can enhance asthma outcomes through promoting Treg cell differentiation (Fig. 9I-L), underscoring the pivotal regulatory role of Treg cells in asthma pathophysiology. Additionally, inhibition of SIRT3-SUMO signaling may facilitate Treg cell differentiation via the FAO pathway to mitigate Th2-type and non-Th2-type asthma progression. Induction of Treg cells to suppress asthma progression. C57BL/6 mice were administered AAV controls, AAV-shCPT1, AAV-SIRT3, and AAV-SIRT3 K223R via intranasal administration and subsequently sensitized with OVA/Alum to establish a mouse model of asthma. A: Representative images of lung tissue stained with HE and PAS are shown. Scale bar = 100 μm. B Inflammatory scores for lung tissue in asthmatic mice. C PAS scores for lung tissue in asthmatic mice. D Measurement of IgE levels in the serum from different groups of mice. E Statistical analysis of cell counts in BALF across various groups. F Statistical analysis of eosinophil counts in BALF among different groups. G Airway pressure detection in different groups of mice. H Flow cytometric assessment of Treg cell proportions within lung tissue cell suspensions from each group. I Representative images depicting HE and PAS staining alongside quantitative immune scoring and PAS scoring results are presented. Scale bar = 100 μm. J Flow cytometric evaluation of Treg cell proportions within lung tissue cell suspensions from distinct groups is conducted again here for clarity. K Measurement of IgE levels in the serum from various mouse cohorts is reported once more for consistency. L Statistical analysis detailing cell counts present in BALF across all experimental groups is provided as well. M Schematic Illustration of the SIRT3-SUMO-Mediated Regulatory Mechanism in Treg Cell Differentiation. All data are represented as the mean ± standard deviation, *p < 0.05, n = 5 Discussion Asthma is a multifaceted inflammatory disease characterized by intricate immune responses. Recent research indicates that Treg cells may represent a promising avenue for asthma treatment(Guo et al. 2021). In this study, we demonstrate that N-glycosylation plays a crucial role in the differentiation of Treg cells, with its regulation mediated by SIRT3-SUMO pathways. Notably, the expression levels of SIRT3 in Treg cells are significantly elevated compared to naive CD4+ T cells, whereas SUMOylation levels are markedly reduced. To elucidate the molecular mechanisms underlying SIRT3-SUMO-mediated regulation of N-glycosylation, we identified the fatty acid oxidation (FAO) pathway as pivotal. A decrease in SIRT3 SUMOylation correlates with increased protein levels that enhance the expression of carnitine palmitoyltransferase 1 (CPT1) and very long-chain acyl-CoA dehydrogenase (VLCAD), both integral to FAO processes, thereby further promoting FAO activity. Concurrently, acetyl-CoA—an end product of FAO—can synthesize UDP-GlcNAc, an essential substrate for N-glycosylation via the HBP. Additionally, FAO can inhibit phosphofructokinase activity, facilitating fructose-6-phosphate's entry into HBP for N-glycosylation substrate synthesis; this cascade ultimately promotes Treg cell differentiation and mitigates Th2-type and non-Th2-type asthma progression (Fig. 9M). We performed weighted gene co-expression network analysis (WGCNA) on relevant GEO datasets and identified N-glycosylation as a biological process implicated in the development of asthma. To validate these findings, we established an ovalbumin-induced asthma model in mice. Accumulating evidence from prior studies has demonstrated that female mice tend to exhibit more pronounced allergic responses following OVA stimulation(Melgert et al. 2005). Accordingly, female mice were selected for model construction in all subsequent experiments. Utilizing an ovalbumin (OVA)-sensitized mouse model of asthma, we assessed the levels of N-glycosylation and observed a significant downregulation in the lungs of asthmatic mice compared to their healthy counterparts. Furthermore, we noted a marked reduction in the proportion of Treg cells, which are essential for maintaining immune system homeostasis by suppressing effector T cell activity and ensuring immune tolerance—critical factors for managing airway inflammation in asthma(Thomas et al. 2023). Notably, N-glycosylation plays a pivotal role in regulating the surface retention and localization of key receptors such as TCR, CD4, CD8, CD45, and CTLA-4; these processes are vital for T cell differentiation and autoimmunity(Araujo et al. 2017). In vitro experiments aimed at inducing Treg cell differentiation revealed that treatment with PNGase significantly diminished the proportion of Treg cells, suggesting that N-glycosylation adversely affects Treg cell differentiation. Mitochondrial dynamics are crucial for T cell differentiation, and alterations in mitochondrial acetylation levels significantly influence the activation of mitochondrial enzymes(Buck et al. 2016; Song and Wang 2022). SIRTs primarily function as deacetylases. To date, seven functionally distinct human SIRTs have been identified, among which SIRT2, SIRT3, SIRT4, and SIRT5 are localized in mitochondria(Liu and Shi 2022). SIRT2 has been implicated in mast cell-mediated airway inflammatory diseases(Kim et al. 2020). However, studies on SIRT4 and SIRT5 remain limited, and their potential roles in asthma pathogenesis have not yet been fully elucidated. SIRT3 is predominantly localized within mitochondria and exerts anti-proliferative effects on lung adenocarcinoma cells by scavenging reactive oxygen species (ROS). Moreover, SIRT3 has been shown to attenuate apoptosis, oxidative stress, and inflammation in bronchial epithelial cells present in bronchoalveolar lavage fluid(Song and Wang 2022; Xiao et al. 2013). These findings suggest that SIRT3 plays a role in the development of asthma and may represent a promising therapeutic target. Our study further investigated the regulatory role of SIRT3 in asthma progression and revealed a potential association between SIRT3 and Treg cells. We observed that the acetylation level in Treg cells mitochondria was markedly reduced, while the expression of the principal deacetylase SIRT3 within these organelles was notably elevated. Additionally, we found that SIRT3 exhibited high expression and co-localization with FOXP3 in the lymph nodes of OVA-sensitized mice. This implies that SIRT3 plays a role in the functional regulation of Treg cells. Protein SUMOylation is a reversible modification that plays a pivotal role in regulating gene transcription and protein stability(Geiss-Friedlander and Melchior 2007). Recent studies have implicated SUMO in asthma pathogenesis; specifically, the SUMO E3 ligase Cbx4 has been shown to facilitate Th9 cell differentiation, thereby promoting asthma development(Huang et al. 2023). Inhibition of SUMOylation can diminish pro-inflammatory factor TSLP expression, alleviating HDM-induced allergic airway inflammation(Liang et al. 2022). Notably, SIRT3 is subject to SUMOylation within mitochondria, which enhances its degradation via the proteasome pathway(Wang et al. 2019). Therefore, we first explored whether SIRT3-SUMO functions in Treg cells and thereby affects the progression of asthma. We determined that SIRT3's SUMOylation level was significantly diminished in Treg cells. Furthermore, reducing SIRT3 SUMOylation correlated with decreased inflammatory infiltration in the lungs of OVA-sensitized mice. This reduction also led to lower serum IgE levels and decreased airway resistance—indicating that diminished SIRT3-SUMO may mitigate asthma progression. Exogenous fatty acid beta-oxidation serves as a crucial energy source for Treg cells(Miao et al. 2022). Previous studies have indicated that SIRT3 plays a pivotal role in modulating fatty acid metabolism. Specifically, SIRT3 regulates the activity of long-chain acyl-CoA dehydrogenase (LCAD) and acetyl-CoA synthetase 2 (ACSS2) through its deacetylation function(Schwer et al. 2006). In acute kidney injury models, SIRT3 is implicated in the regulation of fatty acid oxidation (FAO) induced by cisplatin, with SIRT3 knockdown exacerbating FAO-related damage(Li et al. 2020). In acute myeloid leukemia (AML), SIRT3 emerges as a key regulator of FAO in leukemic stem cells (LSC). The suppression of SIRT3 leads to diminished FAO, resulting in reduced oxidative phosphorylation (OXPHOS) and ATP levels, ultimately culminating in LSC cell death(O'Brien et al. 2023). However, the potential role of SIRT3 in regulating FAO within T cells and its impact on Treg cell differentiation remains unexplored. Our study demonstrates that both wild-type SIRT3 and its deSUMOylated mutant K223R significantly enhance palmitate uptake and mitochondrial oxygen consumption rate (OCR), while also influencing the expression levels of critical FAO enzymes such as carnitine palmitoyltransferase 1 (CPT1) and very-long-chain acyl-CoA dehydrogenase (VLCAD). Notably, etomoxir—a CPT1 inhibitor—and VLCAD knockdown markedly inhibited the induction of Treg cell differentiation mediated by both SIRT3 and its deSUMOylated form. Furthermore, our in vivo experiments substantiate that inhibition of the FAO pathway attenuates the therapeutic effects exerted by SIRT3 and its deSUMOylation on asthma progression. The HBP, a branch of glycolysis, generates substrates for the N-glycosylation pathway(Crotta et al. 2023). The HBP is significantly influenced by various metabolic nutrients, including glucose, glutamine, and acetyl-CoA(Ferrer et al. 2016). Research has demonstrated that inhibition of fatty acid oxidation (FAO) adversely affects protein glycosylation(Ferrer et al. 2016). Consequently, we hypothesized that acetyl-CoA derived from the FAO pathway plays a role in SIRT3-SUMO-mediated N-glycosylation. Our findings corroborated this hypothesis and revealed that acetyl-CoA produced via FAO is crucial for sustaining the HBP. Glucose also serves as an important substrate for the HBP. The competition between fructose-6-phosphate (F6P) and F6P for phosphofructokinase 1 (PFK1) and glucosamine-6-phosphate transferase (GFPT) is pivotal in determining the extent of glucose flux towards the six-carbon sugar/branching pathway(Araujo et al. 2017). Acetyl-CoA generated through FAO can enter the tricarboxylic acid cycle for further oxidation, yielding citrate and ATP—both of which inhibit PFK activity(Kruspig et al. 2012; Zancan et al. 2007). Our study indicated that while FAO inhibition did not affect PFK1 or GFPT expression levels, it did impair PFK1 activity; this was confirmed in both SIRT3 wild-type and SIRT3 K223R cells. Furthermore, we observed that Treg cels differentiation induction in PFK1 knockdown SIRT3 WT and SIRT3 K223 cells was no longer significant compared to control cells. This suggests that SIRT3 may regulate PFK activity by enhancing FAO while promoting glucose flow toward the HBP, thereby increasing UDP-GlcNAc synthesis to facilitate N-glycosylation processes which promote Treg cell differentiation and mitigate asthma progression. Our research enhances understanding of Treg cell differentiation mechanisms and identifies a novel target for targeted Treg cells therapy in asthma. To further elucidate the role of Treg cells in modulating asthma progression, we administered PQ to asthmatic mice, as PQ has been previously demonstrated to enhance Treg cell activity(Thome et al. 2020). Our findings revealed that PQ treatment significantly alleviated asthma severity, thereby providing in vivo evidence supporting the critical involvement of Treg cells in regulating disease progression. Treg cells are known to suppress the functions of multiple effector cell types implicated in asthma pathogenesis, including Th2 cells, eosinophils, ILC2s, mast cells, and basophils(Conrad et al. 2025). This immunomodulatory effect is primarily mediated through cytokine secretion and direct cell-to-cell interactions. Specifically, IL-10 secreted by Treg cells inhibits the activation of effector T cells, while the release of cytotoxic molecules such as granzymes A and B induces apoptosis in these cells(Gondek et al. 2005; Kitani et al. 2003). Nevertheless, the precise molecular mechanisms underlying this regulatory network remain to be fully elucidated. Limitation Our study elucidates the molecular mechanism by which SIRT3-SUOMO modulates Treg cell differentiation via N-glycosylation, as demonstrated in both murine models and cellular assays. Furthermore, we have established that the augmentation of Treg cells can mitigate asthma symptoms. However, whether N-glycosylation only plays a role in regulating the differentiation of Treg cells still needs further exploration and there remains a notable deficiency in the development of targeted therapeutics and clinical investigations. Additionally, our findings indicate that SIRT3 enhances the expression of critical fatty acid oxidation (FAO) enzymes, namely CPT1 and VLCAD. It is well-established that SIRT3 functions as a deacetylase within mitochondria; thus, further research is warranted to determine whether SIRT3 regulates the activity of CPT1 and VLCAD through its deacetylation capacity. This study investigated the underlying mechanisms using a female mouse model. However, given the potential influence of sex differences in mice on the experimental outcomes, further validation in male mice is required to confirm the generalizability of the findings and minimize the risks associated with clinical translation. In addition, our research has revealed that SIRT3 can influence the mRNA expression of CPT1 and VLCAD. Although it is well established that SIRT3 possesses deacetylation activity, whether this function also plays a role in the regulation of CPT1 and VLCAD mRNA expression through deacetylation remains to be further investigated. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Fig. 1 N-glycosylation and Treg cell differentiation influenced asthma progression. A The volcano plot illustrates the differentially expressed genes (DEGs) in asthma, identified through filtering criteria of |logFC| > 0 and p < 0.05. Upregulated genes are represented in orange, while downregulated genes are depicted in blue. B The T-scale free fitting index and average connectivity across various soft threshold powers are presented. C A hierarchical clustering diagram based on the identified co-expressed genes categorizes them into distinct gene modules; different colors denote corresponding modules, with the gray module indicating genes not assigned to any network. D A heatmap depicting the relationship between each gene module and asthma subtypes is shown, where red indicates a positive correlation and green signifies a negative correlation. E Differential gene analysis of key modules associated with asthma was conducted through GO and KEGG enrichment analyses. F: A schematic representation of the OVA/LPS-induced mouse asthma model is provided. G-H The immune score and PAS score of lung tissues in OVA/LPS-induced asthma model mice were significantly higher than those in control mice. Scale bar = 100 μM. I The WB results indicated that the N-glycosylation level in the lung tissue of asthmatic mice was decreased. J Flow cytometry results showed that the proportion of Treg cells in BALF and lung tissue of asthmatic mice was decreased. K-L Flow cytometry results indicated that the proportion of Treg cells in the lymph nodes and spleen of OVA/Alum-induced asthmatic mice was decreased. All data are expressed as mean ± standard deviation (SD). Statistical significance was determined at *p< 0.05 compared to Control group, n=5. (JPG 2975 KB) Supplementary Fig. 2 SIRT3-SUMO regulated Treg cell differentiation through N-glycosylation. A The level of SIRT3-SUMO1 in the mediastinal lymph nodes of OVA-induced mice was assessed, n=5. B The SUMOylation site of SIRT3 was predicted using online databases. C Co-IP identified K223 as the SUMOylation site for SIRT3, n=5. D Flow cytometric analysis to determine N-glycosylation levels, n=5. E Flow cytometry was employed to quantify the number of Treg cells, n=5. All data are expressed as mean ± standard deviation, *p < 0.05 (JPG 918 KB) Supplementary Fig. 3 SIRT3-SUMO influenced Treg cell differentiation and asthma progression via N-glycosylation.A SIRT3 WT and K223R decreased the immune score and PAS score in the lung tissue of asthmatic mice. B-E SIRT3 WT and K223R mice showed reduced content of IgE in serum (B), airway resistance (C), F-G Flow cytometry results showed that SIRT3 WT and SIRT3 K223R mice increased the N-glycosylation level in the lung lymph nodes (G) and the content of Treg cells in the lung tissue (F). All data are expressed as mean ± standard deviation, *p<0.05, n=5 (JPG 1130 KB) Supplementary Fig. 4 SIRT3-SUMO did not influence the proportion changes of monocytes, macrophages, dendritic cells, or iNKT cells. A-H Flow cytometry was used to detect the proportions of monocytes (A-B), macrophages (C-D), dendritic cells (E-F), and iNKT cells (G-H) in the lung tissues of asthmatic mice and normal mice (JPG 1066 KB) Supplementary Fig. 5 SIRT3-SUMO regulated the FAO process. A The SIRT3 mRNA level of SIRT3 in CPT1 and VLCAD knockdown cells was detected by qRT-PCR, n=5. B The protein level of SIRT3 in CPT1 and VLCAD knockdown cells was detected by WB. C The expression of CPT1 in the hilar lymph nodes of asthma mice was detected by IF. Scale bar = 200 µm. D The expression of VLCAD in the hilar lymph nodes of asthma mice was detected by IF. Scale bar = 200 µm. All data are presented as mean ± standard deviation, *p < 0.05, n=5 (JPG 2072 KB) Supplementary Fig. 6 CPT1 acted as an upstream regulator of VLCAD and is involved in SIRT3-SUMO-mediated Treg cell differentiation. A The content of acetyl-CoA in different groups of cells was detected using a kit. B The proportion of Treg cells was detected by flow cytometry. C The proportion of Treg cells was detected by flow cytometry. D The mRNA levels of FOXP3 and IL-10 were detected by qRT-PCR. All data are presented as mean ± standard deviation, *p <0.05, n=5 (JPG 1706 KB) Supplementary Fig. 7 SIRT3-SUMO influenced Treg cell differentiation via FAO product acetyl-CoA. A The intracellular levels of UDP-GlcNac were measured. B The proportion of Treg cells was determined. All the data are presented as the mean ± standard deviation, *p<0.05, n=5 (JPG 796 KB) Supplementary Fig. 8 Induction of Treg cells to suppress asthma progression. A Inflammatory scores for lung tissue in asthmatic mice. B PAS scores for lung tissue in asthmatic mice. C Measurement of IgE levels in the serum from different groups of mice. D Statistical analysis of cell counts in BALF across various groups. E Statistical analysis of eosinophil counts in BALF among different groups. F Flow cytometric assessment of Treg cell proportions within lung tissue cell suspensions from each group. All the data are presented as the mean ± standard deviation, *p<0.05, n=5 (JPG 1201 KB) Supplementary file9 (DOCX 16 KB) Author contributions Yan Hu contributed to the conception; Yan Hu and Chuntao Liu contributed to the design; Yan Hu and Chuntao Liu analyzed and interpreted the data; Yan Hu drafted the article; Chuntao Liu critically revised it for important intellectual content. All authors approved the final version for publication. Funding None. Data availability No datasets were generated or analysed during the current study. Declarations Declarations Ethical and legal declarations All the above experimental procedures and treatments were approved by the West China Hospital, Sichuan University, Chengdu, China Animal Experiment Research Ethics Committee. Consent for publication Not applicable. Clinical Trial Number Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Fig. 1 N-glycosylation and Treg cell differentiation influenced asthma progression. A The volcano plot illustrates the differentially expressed genes (DEGs) in asthma, identified through filtering criteria of |logFC| > 0 and p < 0.05. Upregulated genes are represented in orange, while downregulated genes are depicted in blue. B The T-scale free fitting index and average connectivity across various soft threshold powers are presented. C A hierarchical clustering diagram based on the identified co-expressed genes categorizes them into distinct gene modules; different colors denote corresponding modules, with the gray module indicating genes not assigned to any network. D A heatmap depicting the relationship between each gene module and asthma subtypes is shown, where red indicates a positive correlation and green signifies a negative correlation. E Differential gene analysis of key modules associated with asthma was conducted through GO and KEGG enrichment analyses. F: A schematic representation of the OVA/LPS-induced mouse asthma model is provided. G-H The immune score and PAS score of lung tissues in OVA/LPS-induced asthma model mice were significantly higher than those in control mice. Scale bar = 100 μM. I The WB results indicated that the N-glycosylation level in the lung tissue of asthmatic mice was decreased. J Flow cytometry results showed that the proportion of Treg cells in BALF and lung tissue of asthmatic mice was decreased. K-L Flow cytometry results indicated that the proportion of Treg cells in the lymph nodes and spleen of OVA/Alum-induced asthmatic mice was decreased. All data are expressed as mean ± standard deviation (SD). Statistical significance was determined at *p< 0.05 compared to Control group, n=5. (JPG 2975 KB) Supplementary Fig. 2 SIRT3-SUMO regulated Treg cell differentiation through N-glycosylation. A The level of SIRT3-SUMO1 in the mediastinal lymph nodes of OVA-induced mice was assessed, n=5. B The SUMOylation site of SIRT3 was predicted using online databases. C Co-IP identified K223 as the SUMOylation site for SIRT3, n=5. D Flow cytometric analysis to determine N-glycosylation levels, n=5. E Flow cytometry was employed to quantify the number of Treg cells, n=5. All data are expressed as mean ± standard deviation, *p < 0.05 (JPG 918 KB) Supplementary Fig. 3 SIRT3-SUMO influenced Treg cell differentiation and asthma progression via N-glycosylation.A SIRT3 WT and K223R decreased the immune score and PAS score in the lung tissue of asthmatic mice. B-E SIRT3 WT and K223R mice showed reduced content of IgE in serum (B), airway resistance (C), F-G Flow cytometry results showed that SIRT3 WT and SIRT3 K223R mice increased the N-glycosylation level in the lung lymph nodes (G) and the content of Treg cells in the lung tissue (F). All data are expressed as mean ± standard deviation, *p<0.05, n=5 (JPG 1130 KB) Supplementary Fig. 4 SIRT3-SUMO did not influence the proportion changes of monocytes, macrophages, dendritic cells, or iNKT cells. A-H Flow cytometry was used to detect the proportions of monocytes (A-B), macrophages (C-D), dendritic cells (E-F), and iNKT cells (G-H) in the lung tissues of asthmatic mice and normal mice (JPG 1066 KB) Supplementary Fig. 5 SIRT3-SUMO regulated the FAO process. A The SIRT3 mRNA level of SIRT3 in CPT1 and VLCAD knockdown cells was detected by qRT-PCR, n=5. B The protein level of SIRT3 in CPT1 and VLCAD knockdown cells was detected by WB. C The expression of CPT1 in the hilar lymph nodes of asthma mice was detected by IF. Scale bar = 200 µm. D The expression of VLCAD in the hilar lymph nodes of asthma mice was detected by IF. Scale bar = 200 µm. All data are presented as mean ± standard deviation, *p < 0.05, n=5 (JPG 2072 KB) Supplementary Fig. 6 CPT1 acted as an upstream regulator of VLCAD and is involved in SIRT3-SUMO-mediated Treg cell differentiation. A The content of acetyl-CoA in different groups of cells was detected using a kit. B The proportion of Treg cells was detected by flow cytometry. C The proportion of Treg cells was detected by flow cytometry. D The mRNA levels of FOXP3 and IL-10 were detected by qRT-PCR. All data are presented as mean ± standard deviation, *p <0.05, n=5 (JPG 1706 KB) Supplementary Fig. 7 SIRT3-SUMO influenced Treg cell differentiation via FAO product acetyl-CoA. A The intracellular levels of UDP-GlcNac were measured. B The proportion of Treg cells was determined. All the data are presented as the mean ± standard deviation, *p<0.05, n=5 (JPG 796 KB) Supplementary Fig. 8 Induction of Treg cells to suppress asthma progression. A Inflammatory scores for lung tissue in asthmatic mice. B PAS scores for lung tissue in asthmatic mice. C Measurement of IgE levels in the serum from different groups of mice. D Statistical analysis of cell counts in BALF across various groups. E Statistical analysis of eosinophil counts in BALF among different groups. F Flow cytometric assessment of Treg cell proportions within lung tissue cell suspensions from each group. All the data are presented as the mean ± standard deviation, *p<0.05, n=5 (JPG 1201 KB) Supplementary file9 (DOCX 16 KB) Data Availability Statement No datasets were generated or analysed during the current study.