The Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), a public repository run by the National Center for Biotechnology Information (NCBI), provided the transcriptomic data (GSE175759). There were 68 samples in the dataset, including 22 normal control samples and 46 samples from IgAN patients [20].

The potential targets of gallic acid (GA) were systematically identified through a network pharmacology approach. Initially, the canonical Simplified Molecular Input Line Entry System (SMILES) notation for GA was retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). This SMILES string served as the primary input for subsequent target prediction. Putative protein targets of GA were forecasted using two complementary online platforms: the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https://tcmsp-e.com/; accessed on 16 May 2024) and SwissTargetPrediction (http://www.swisstargetprediction.ch/; accessed on 19 May 2024). For the TCMSP search, the standard screening criteria of oral bioavailability (OB) ≥ 30% and drug-likeness (DL) ≥ 0.18 were applied to filter bioactive components. In the SwissTargetPrediction analysis, all predicted targets for Homo sapiens with a probability score > 0 were included. The target lists obtained from these two databases were subsequently merged, and duplicate entries were removed. Visualization was performed using Venny 2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny) to identify the overlapping genes. Furthermore, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses, alongside Gene Set Enrichment Analysis (GSEA), were performed on the relevant differentially expressed genes (DEGs). These analyses were conducted using the Weishengxin online tool (http://www.bioinformatics.com.cn/).

Polymeric IgA1 (p-IgA1) was prepared by heat-polymerizing monomeric human IgA1 antibody (ab193187, Abcam, Cambridge, MA, USA) at 63°C for 150 min using a dry bath incubator (TU-100, Shanghai Yiheng Technology). Once cooled to room temperature, the sample was centrifuged at 12,000 rpm for 5 min using a high-speed centrifuge (Sorvall Legend Micro 21R, Thermo Fisher Scientific, Waltham, MA, USA) to eliminate insoluble precipitates, and the resulting supernatant contained p-IgA1 [21].

Human mesangial cells (HMCs, CP-H067, Wuhan Procell Life Science and Technology) were cultured at 37°C under 5% CO₂ in a humidified incubator in RPMI-1640 medium (#22400-089, Thermo Fisher Scientific, Waltham, MA, USA) containing 10% FBS (A5669701, Gibco, Grand Island, NY, USA) and penicillin/streptomycin (100 U/mL, #15070063, Gibco, Grand Island, NY, USA). All tests were conducted using cells that were grown at passages 3–8. In serum-free RPMI-1640 media, EL transfection reagent (FT201-01, TransIntro, Beijing, China) was combined with plasmid DNA (empty, DUSP1 overexpression, or DUSP1 silencing vector, Sangon Biotech, Shanghai, China) and incubated for 15 min. Cells were exposed to complexes for six hours at 37°C. After that, the medium was changed to full RPMI-1640 medium for a 48-h culture period.

For experimental treatments, cells were plated at 5 × 10⁴ cells/cm² in 6-well plates. After 24-h attachment, seven treatment groups were established: the untreated control group (Ctrl); the IgAN model group (IgA), induced by treatment with 25 μg/mL p-IgA1 for 48 h; the model groups co-treated with a low or high dose (12.5 or 25 μg/mL, respectively) of GA (IgA + GA-L and IgA + GA-H groups, respectively; G7384, Sigma, St. Louis, MO, USA); and the intervention groups where, prior to modeling and drug treatment, cells were transfected with either a non-targeting control siRNA (IgA + GA-H + siCtrl group), DUSP1-targeting siRNA (IgA + GA-H + siDUSP1 group), or a DUSP1 overexpression plasmid (IgA + OE-DUSP1 group).

Using the Cell Counting Kit-8 (CCK-8; C0038, Beyotime, Shanghai, China), cell viability was assessed. Logarithmic-phase HMC cells were seeded at 10,000 cells per well in 96-well plates at 37°C with 5% CO₂ and exposed to varying concentrations of GA (G7384, Sigma, St. Louis, MO, USA) (0, 12.5, 25, 50, 75, and 100 μg/mL). Each well received 10 μL of CCK-8 reagent after 48 h of treatment, and the wells were then incubated for two hours at 37°C. Absorbance was recorded at 450 nm using a microplate reader (ReadMax 1200, Shanpu, Shanghai, China).

Following the manufacturer’s instructions, the Calcein/PI Cell Viability/Cytotoxicity Assay Kit (C2015S, Beyotime) was utilised to photograph and compare living and dead cells. The cells underwent a staining procedure using calceinacetoxymethyl ester (calcein AM) and propidium Iodide (PI) solution (1:1) following a 48-h culture period. While PI can pass through damaged cell membranes and bind to nucleic acids to create a brilliant red fluorescence (excitation/emission roughly 535/617), live cells transform calcein AM into green fluorescent calcein. Lastly, an Olympus IX71 inverted fluorescent microscope (Tokyo, Japan) was used to view the cells.

Commercialized diagnostic test kits, including glutathione/glutathione oxidized (GSH/GSSG) Assay Kit (HY-K0311, MCE, Shanghai, China), JC-1 mitochondrial membrane potential assay kit (HY-K0601, MCE), malondialdehyde (MDA) assay kit (BC1175, Solarbio, Beijing, China), were employed to assess metabolic and redox parameters in HMC. Fluorogenic probe H2DCFDA (HY-D0940, MCE, Shanghai, China) and BODIPY 581/591 C11 (HY-D1301, MCE, Shanghai, China) were used to assess the reactive oxygen species (ROS) and lipid peroxidation (LPO) of HMC, respectively. All assays were performed according to the manufacturers’ protocols with at least three technical replicates per condition and repeated in six independent biological experiments.

Utilising qRT-PCR, the mRNA expression levels of FTH1 and TFRC were examined [22,23]. The TRIzol reagent (15596026, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) was used to extract total RNA from cells. The HiScript II Reverse Transcriptase (R223, Vazyme, Nanjing, China) was used for mRNA reverse transcription following the manufacturer’s protocol, and the FastKing-RT SuperMix (KR118-02, Tiangen, Beijing, China) was used for cDNA synthesis via reverse transcription. In addition, the 2^−ΔΔCT method served to quantify relative gene expression, with GAPDH as the endogenous control. The primer sequences are provided in Table A1.

HMCs were cultured in 10-cm dishes until reaching approximately 80%–90% confluency. The cells were gently rinsed twice with 2 mL of ice-cold phosphate-buffered saline (PBS, pH 7.4). Subsequently, 1 mL of RIPA lysis buffer (containing 1× protease inhibitor cocktail and 1 mM phenylmethylsulfonyl fluoride) was added to each dish. Cells were then scraped using a chilled cell scraper, and the resulting lysates were centrifuged at 14,800 rpm for 15 min. The supernatant was collected and retained for further analysis. The protein concentration of the supernatant was determined using a BCA assay kit (P0009, Beyotime, Shanghai, China). For the detection of Fe²⁺, the supernatant was processed according to the manufacturer’s instructions of the Fe²⁺ assay kit (A039-2-1, Jiancheng, Nanjing, China). Specifically, 1.5 mL of iron chromogenic agent was added to the sample. After thorough mixing, the reaction mixture was incubated in a boiling water bath at 100°C for 5 min. Following cooling to room temperature, the mixture was centrifuged at 3500 rpm for 10 min. Finally, 200 μL of the supernatant was transferred to a 96-well plate, and its absorbance was measured at a wavelength of 520 nm using a microplate reader (ReadMax 1200, Shanpu, Shanghai, China).

Protein concentration was calculated after total protein was separated from HMC. 30 μg of total protein was extracted from each sample using 10% SDS-PAGE (PG112, Epizyme, Shanghai, China) and then put onto PVDF membranes (IPVH00010, Millipore, Billerica, MA, USA). After blocking the membranes with 5% non-fat milk for one hour at room temperature, the membranes were exposed to primary antibodies overnight at 4°C. p-p38 (Thr180/Tyr 182; 1:1000; #4511, CST, Danvers, MA, USA), p-38 (1:1000; ab170099, Abcam, Cambridge, MA, USA), DUSP1 (1:1000; MA5-32480, Thermo Fisher Scientific, Waltham, MA, USA), GPX4 (1:1000; #59735, CST, Danvers, MA, USA), SLC7A11 (1:1000; #28001, CST, Danvers, MA, USA), and α-Tubulin (1:1000; #2144, CST, Danvers, MA, USA) were the primary antibodies that were employed. The membranes were treated with HRP-conjugated goat anti-rabbit IgG (1:2000; ab97051, Abcam, Cambridge, MA, USA) for one hour at room temperature following three TBST washes. Protein bands were observed using a Bio-Rad imaging system (ChemiDoc Touch, Hercules, CA, USA) and measured using ImageJ software (version 1.53e, National Institutes of Health, Bethesda, MD, USA) after more TBST washes. The integrated density value (IDV) of each target protein band was measured and normalized to the IDV of the corresponding loading control band (α-Tubulin) from the same sample lane to obtain the relative protein expression level. Every experiment was carried out three times.

All data are shown as mean ± standard deviation (M ± SD) and were gathered from a minimum of three separate studies. For data that passed the normality test, parametric tests were used. An independent two-sample t-test compared two groups, while one-way analysis of variance (ANOVA) was applied to comparisons involving three or more groups. For all statistical studies, GraphPad Prism version 8.0.5 (GraphPad Software; La Jolla, CA, USA) was used. *p < 0.05, **p < 0.01, and ***p < 0.001 served as the thresholds for statistical significance.

To clarify the potential therapeutic mechanism of GA on IgAN, we integrated transcriptomic data and network pharmacology methods for analysis. Based on the GSE175759 dataset in the GEO database, this study included 46 renal tissue samples from IgAN patients and 22 normal control samples. In total, 1141 differentially expressed genes (DEGs) were identified, with |log₂FoldChange| ≥1 and corrected p value (p_adj) <0.05 serving as the threshold for differential expression. According to the volcano plot results, 884 genes were upregulated and 515 were downregulated in the IgAN group relative to the normal control group (Fig. 1A). The heat map of the differential gene expression pattern further showed the clustering characteristics of gene expression between samples (Fig. 1B). Subsequently, KEGG pathway enrichment analysis and GO function annotation analysis were performed on all DEGs. Analysis indicated that the differentially expressed genes were notably concentrated in the MAPK signaling pathway, implying a potential key role of this pathway in IgAN pathogenesis (Fig. 1C,D).

To identify potential targets of GA, this study retrieved its SMILES structure from the PubChem database and predicted possible targets using the SwissTargetPrediction and TCMSP platforms. After integrating the results from both platforms and removing duplicates, 109 potential GA target genes were obtained. The intersection of the target set and the above 1141 IgAN differentially expressed genes was taken, and finally, 8 common genes were screened out as candidate targets for GA intervention in IgAN, including DUSP1, SERPINE1, ADRA2A, CA5A, ALB, FASLG, FUT7, and CNR2 (Fig. 1E). Among them, DUSP1 was significantly down-regulated in the IgAN group (log₂FoldChange = −4.07, p_adj = 1.88 × 10⁻¹⁵⁸), which was the gene with the most significant difference in expression (Fig. 1F). According to the KEGG pathway enrichment analysis and GO functional annotation of the GA target set, the 515 downregulated genes in IgAN (log2FoldChange < −1, p < 0.05) showed significant enrichment in several inflammation-related pathways, including MAPK and IL-17 signalling pathways (Fig. 2A,B). The MAPK signalling pathway was frequently found in numerous enrichment results, indicating that GA may control the MAPK signalling pathway to reduce inflammation or other associated physiological and pathological processes, hence delaying the curative effect of IgAN. The samples were divided into high- and low-expression groups according to DUSP1 expression levels, and GSEA analysis was performed to further validate the functional role of DUSP1. GO and KEGG functional enrichment analyses indicated that the MAPK signalling pathway was consistently enriched in the DUSP1 low-expression group (Fig. 2C,D). These findings imply that GA may protect against IgAN by regulating the MAPK pathway by inhibiting DUSP1.

To further investigate the physiological role of DUSP1 in IgAN, HMCs were continuously stimulated with 25 μg/mL p-IgA1 for 48 h to establish an in vitro model for functional verification experiments. First, a CCK-8 assay was performed to assess the cytotoxic effects of GA on HMCs and to determine the suitable drug concentration range. As shown in Fig. 3A, GA did not show significant cytotoxicity when treated at a concentration of 12.5 μg/mL for 48 h. However, cell viability began to decrease at 25 μg/mL (Fig. 3A). Therefore, 12.5 μg/mL (GA-L) and 25 μg/mL (GA-H) were selected for subsequent experiments to ensure therapeutic effects and minimize off-target effects.

Relative to the Ctrl group, the intracellular ROS level in the IgA group rose markedly (Fig. 3B), indicating successful induction of oxidative stress. The GA treatment group, in contrast, substantially decreased ROS levels in a dose-dependent manner. To investigate whether DUSP1 is involved in this antioxidant effect, we first transfected HMC cells with siDUSP1 before p-IgA1 and GA-H treatment. The ROS level in the IgA + GA-H + siDUSP1 group was significantly increased, while transfection with siCtrl did not affect the therapeutic effect of high-dose GA. In contrast, overexpression of DUSP1 (IgA + OE-DUSP1) mimicked the protective effect of GA and inhibited p-IgA1-induced ROS accumulation even in the absence of GA treatment (Fig. 3B,C). These results together indicate that DUSP1 is essential for GA-mediated regulation of oxidative stress. The ratio of GSH/GSSG in each group of cells also showed that p-IgA1 could significantly reduce the antioxidant capacity of HMCs, and GA intervention and DUSP1 overexpression could effectively restore this ratio, while DUSP1 knockdown significantly reversed the therapeutic effect of GA (Fig. 3D).

Redox imbalance leads to the massive accumulation of lipid peroxides, a key event in ferroptosis. In this study, lipid peroxides in each group of cells were detected by the fluorescent probe C11-BODIPY™ 581/591. Lipid peroxide accumulation was significantly elevated in the model group, whereas GA treatment and DUSP1 overexpression effectively suppressed this increase (Fig. 3E). Lipid peroxidation accumulation will cause damage to organelles, eventually leading to a large number of cell deaths. JC-1 analysis revealed a decrease in the red/green fluorescence ratio in the IgA group, indicating that the reduction in membrane potential led to mitochondrial dysfunction (Fig. 3F,G). In addition, the results of live/dead cell staining confirmed that p-IgA1 induced significant cell death (Fig. 3H). Consistent with the improvement of oxidative stress, GA treatment significantly improved mitochondrial membrane potential and increased cell survival. Gain-of-function and loss-of-function experiments of DUSP1 strongly indicated that DUSP1 overexpression attenuated mitochondrial depolarization and reduced cell death. These results suggest that GA alleviates p-IgA1-induced ferroptosis in HMCs through a DUSP1-dependent mechanism, which may be achieved by regulating oxidative stress and mitochondrial integrity.

To clarify the mechanism of GA in regulating oxidative stress and ferroptosis in IgAN, we first detected the key markers of ferroptosis. Fig. 4A illustrates that the mRNA expression of the transferrin receptor TFRC in HMCs was markedly upregulated following p-IgA1 stimulation, whereas the mRNA level of the iron storage protein FTH1 was significantly downregulated, indicating disrupted cellular iron metabolism. The above molecular changes were accompanied by a large accumulation of lipid peroxidation product malondialdehyde (MDA), a decrease in reduced glutathione (GSH), and an increase in the level of free Fe²⁺, which jointly confirmed that p-IgA1 stimulation induced ferroptosis of HMCs (Fig. 4B–D). High-dose GA treatment significantly reversed the above abnormal changes, restored the expression of TFRC and FTH1, and restored redox and iron homeostasis to normal. However, in the IgA + GA-H + siDUSP1 group, the up-regulation effect of GA-H on FTH1 and GSH was partially weakened, while the levels of MDA and TFRC rebounded, aggravating the oxidative stress imbalance and ferroptosis of IgAN. In contrast, DUSP1 overexpression alone could effectively improve iron metabolism, reduce oxidative stress levels and ferroptosis damage under the p-IgA1 stimulation background, and its effect was similar to that of GA-H intervention (Fig. 4A–D). Collectively, these findings suggest that DUSP1 is crucial for GA-mediated regulation of oxidative stress and ferroptosis in IgAN.

Combined with the previous bioinformatics analysis suggesting that the MAPK pathway is the key pathway for GA intervention in IgAN, DUSP1 is the core target, and the literature shows that DUSP1 is a negative regulator of p38 MAPK, we further studied the regulatory effect of GA on DUSP1 and downstream p38MAPK by protein immunoblotting [24,25]. The results showed that p-IgA1 stimulation markedly decreased DUSP1 protein expression, whereas the phosphorylation level of p38 MAPK (p-p38) increased accordingly, with total p38 remaining unchanged. Compared with the Ctrl group, the IgA group exhibited a marked decrease in protein levels of the key ferroptosis inhibitors—glutathione peroxidase 4 (GPX4) and solute carrier family 7 member 11 (SLC7A11) (Fig. 4E,F). The data indicated that the p38 MAPK signaling pathway was aberrantly activated in the IgAN model, with its endogenous negative regulatory mechanism impaired, leading to enhanced downstream ferroptosis. High-dose GA treatment can effectively restore the expression of DUSP1, thereby inhibiting the activation of p-p38 and significantly up-regulating the expression of GPX4 and SLC7A11. In HMC cells with DUSP1 knockdown, the inhibitory effect of GA-H on p-p38 was partially reversed, and the expression of p-p38 was increased, suggesting that the inhibition of the p38 MAPK pathway by GA depends on DUSP1 (Fig. 4E,F).

In summary, this study confirmed that GA can upregulate DUSP1 expression, inhibit excessive phosphorylation of p38 MAPK, restore the protein levels of SLC7A11 and GPX4, and finally reduce the ferroptosis of HMCs induced by p-IgA1.