Gene expression data for renal tubulointerstitial samples from patients with LN and healthy controls were obtained from the gene expression omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). A total of three datasets were included: GSE113342 (47 LN and 10 control tubulointerstitial samples), GSE200306 (14 LN and 10 control samples) and GSE127797 (44 LN samples) (28–30). Probe identifiers were annotated to gene symbols using the platform annotation file; for genes represented by multiple probes, one probe was randomly selected to avoid redundancy. Background correction and normalization were performed with the ‘limma’ package (v3.6.2) (31) in R (v4.2.1; RStudio, Inc.) to ensure data quality and comparability, yielding the final gene expression matrices. All analyses adhered to the principles of The Declaration of Helsinki (2013 revision) (32).

Differential expression analysis between LN and control tubulointerstitial samples was conducted separately for GSE113342 and GSE200306 using the ‘limma’ package (31). Genes with |log₂fold change| (|log₂FC|)>1 and adjusted P<0.05 were defined as DEGs. The intersection of DEGs from both datasets was taken to identify consistently dysregulated genes, and results were visualized using ggplot2 (v3.3.6; http://ggplot2.tidyverse.org/) and VennDiagram (v1.7.3).

To explore the potential biological roles of DEGs, gene identifiers were converted using the org.Hs.eg.db package (v3.22; http://bioconductor.org/packages/org.Hs.eg.db), and Gene Ontology (GO) (https://geneontology.org/) was used to determine relevant biological processes (BPs) and molecular functions (MFs) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (www.genome.jp/kegg/) for pathway enrichment analyses were performed with clusterProfiler, with adjusted P<0.05 considered significant (33). Enrichment results were visualized using ggplot2 (v3.3.6), igraph (v2.2.1; http://r.igraph.org/) and ggraph (v2.2.2; http://ggraph.data-imaginist.com/).

To assess the association between FRA1 expression and immune cell infiltration in LN tubulointerstitial samples, the CIBERSORT algorithm was applied, which is a linear support vector regression-based deconvolution tool that estimates immune cell proportions from microarray or RNA-seq data (34). In GSE127797, LN samples were stratified into high- and low-FRA1 expression groups based on the median, and only samples with CIBERSORT P<0.05 were retained for subsequent analyses. The correlations between FRA1 expression and infiltrating immune cell fractions and correlations between FRA1 and tubular epithelial cell-related cytokines were examined; all visualizations were generated with ggplot2 (v3.3.6).

A total of three female MRL/lpr mice and three female MRL/MPJ control mice (age, 16 weeks; body weight, 38–46 g) were obtained from Cavens Biogle (Suzhou) Model Animal Research Co. Ltd. Mice were maintained under standard housing conditions, which included a climate-controlled environment at 22±2°C and 50±10% relative humidity, a 12-h light/dark cycle and free access to food and water, for 4 weeks and then euthanized for kidney tissue collection to assess protein expression. All procedures complied with the AVMA Guidelines for the Euthanasia of Animals (2020) and approved institutional animal care protocols. Specific humane endpoints included: Weight loss ≥20% of body weight; inability to eat or drink; severe dehydration; severe respiratory distress, paralysis or continuous convulsions; uncontrolled progressive infection; or unrelievable pain or clinical signs severely compromised quality of life. Health and behaviour were routinely monitored, with weekly weighing and observation after enrolment, and immediate escalation to daily or more frequent monitoring if abnormalities arose. Mice were deeply anesthetized with pentobarbital sodium (150–200 mg/kg, intraperitoneally) until loss of the pedal withdrawal reflex, followed by cervical dislocation to ensure complete euthanasia. Death was confirmed by the absence of heartbeat and respiration, dilated pupils and lack of reflex response. No experimental intervention was performed; MRL/lpr mice, which spontaneously develop LN (35), were used as the disease model, whereas MRL/MPJ mice served as healthy controls and were used solely for terminal tissue collection. The power calculation for the mouse experiments is shown in Table SI, following the methodology described by Festing and Altman (36).

Kidney tissues were fixed in 10% neutral-buffered formalin for 24–48 h at room temperature and were then processed for paraffin embedding. Paraffin-embedded tissues were cut into 5-µm sections. The sections were dewaxed twice in xylene (15 min each), rehydrated through a descending ethanol series (100% ethanol twice, then 95, 85 and 75% ethanol, 5 min each), and rinsed in phosphate-buffered saline (PBS, pH 7.4) three times for 5 min each. Antigen retrieval was performed in citrate buffer (pH 6.0) in a microwave: Medium power for 8 min, rest for 8 min, then low power for 7 min. Subsequently, the slides were allowed to cool to room temperature for 20–40 min and then rinsed in PBS (three times, 5 min each). Endogenous peroxidase activity was quenched by incubation in 3% hydrogen peroxide at room temperature in the dark for 25 min, followed by three PBS washes (5 min each). Blocking was performed using the normal goat serum blocking solution supplied with the Vectastain ABC Elite HRP Kit (cat. no. SAP-9100; Beijing Zhongshan Jinqiao Biotechnology Co., Ltd.) according to the manufacturer's instructions (incubation at room temperature for 10–15 min). For primary antibody incubation, the sections were incubated with rabbit polyclonal anti-FRA1 (cat. no. A5372; ABclonal Biotech Co., Ltd.) at a dilution of 1:100 (diluted in PBS) and incubated overnight at 4°C. After three PBS washes (5 min each), the sections were incubated with the biotinylated goat anti-rabbit IgG provided in the Vectastain ABC Elite HRP Kit at a working dilution of 1:500 for 10–15 min at 37°C, followed by PBS rinses (3×3 min). The HRP-labeled streptavidin working solution (provided in the Vectastain ABC Elite HRP Kit) was applied and incubated for 10–15 min at 37°C according to the manufacturer's instructions, followed by PBS washes (3×3 min). Chromogenic detection was performed with DAB substrate (cat. no. ZLI-9017; Beijing Zhongshan Jinqiao Biotechnology Co., Ltd.); color development was monitored under the microscope and stopped with distilled water once an optimal signal was achieved. The sections were then counterstained with hematoxylin at room temperature for 1–2 min, differentiated in 1% hydrochloric acid-ethanol (~1 sec), rinsed in running tap water, blued in ammonia water and rinsed again. Finally, the slides were dehydrated through an ascending ethanol series (75, 85, 95 and 100%), cleared in xylene and mounted with neutral resin. The percentage of FRA1-positive area/unit tissue area was quantified using ImageJ (v2.3.0; National Institutes of Health). All stained sections were examined and images were captured using an Olympus BX40 upright light microscope (Olympus Corporation).

HK-2 cells (Pricella; Elabscience Bionovation Inc.) were maintained in DMEM/F12 (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% heat-inactivated fetal bovine serum (Beijing Baiao Leibo Technology Co., Ltd.). A total of four groups were established: FRA1 overexpression (FRA1-OE), FRA1 knockdown (FRA1-shRNA) and their respective empty-vector controls (control-OE and control-shRNA). The FRA1 overexpression vector (pCDH_CMV_MCS_EF1_copGFP-FRA1) and shRNA vector (pMAGic7.1-FRA1-shRNA) were constructed at the Clinical Laboratory, Boai Hospital of Zhongshan (Zhongshan, China); plasmid maps and sequences are provided in Fig. S1 and Table SII. For retroviral packaging, 293T cells (Pricella; Elabscience Bionovation Inc.) were co-transfected with 4 µg total plasmids using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C for 48 h. Specifically, the packaging system consisted of the packaging plasmid (psPAX2), the envelope plasmid (pMD2.G) and one of the four aforementioned transfer plasmids (FRA1-OE, FRA1-shRNA, control-OE or control-shRNA) at a mass ratio of 4:3:2. For gene manipulation, 2×10⁵ HK-2 cells were infected with retroviral particles [2×10⁷ transducing units/ml; multiplicity of infection=30; supplemented with 6 µg/ml polybrene (MedChemExpress); infection efficiency ≥85%]. The infection was performed at 37°C for 24 h, after which the medium was replaced with fresh complete medium. For the FRA1-shRNA and control-shRNA groups, stably transduced cells were selected using 3.0 µg/ml puromycin (MedChemExpress) starting 48 h post-infection for 4 days. No additional selection was performed for the FRA1-OE and control-OE groups, as the high transduction efficiency was sufficient for subsequent experiments. All groups were harvested for further analysis at 144 h post-transduction.

Kidney tissues or HK-2 cells were homogenized in RIPA buffer (Biosharp Life Sciences), and protein concentrations were determined using a BCA Protein Assay Kit (Wuhan Servicebio Technology Co., Ltd.). For HK-2 experiments, cells were harvested at 144 h post-transduction to allow time for lentiviral reverse transcription, genomic integration and establishment of stable transgene expression and for accumulation of downstream protein and secreted cytokine changes. Lentiviral-mediated expression commonly requires on the order of 5–7 days to reach stable levels, and molecular-kinetic studies indicate that reverse transcription and integration occur during the first few days post-transduction (~3 days) (37,38). Prior studies have documented progressively increased reporter expression at 48-, 96- and 144-h post-transduction, and have selected 144 h post-transduction as the endpoint for subsequent mRNA and protein analyses of FRA1 effects (24,39,40); inflammatory mediators in renal cell models are also commonly measured at ~5 days or later following perturbation (41). Equal amounts of protein (40 µg/lane) were separated by SDS-PAGE on 10% gels and were transferred to PVDF membranes (MilliporeSigma). Protein molecular weight markers (10–180 kDa; cat. no. 20350ES90; Shanghai Yeasen Biotechnology Co., Ltd.) were run alongside samples to confirm target sizes. Membranes were blocked in 5% skim milk at 37°C for 1 h, incubated with primary antibodies at 4°C overnight, then washed with TBS-0.1% Tween-20 (TBST). The primary antibodies used in the present study included: β-actin (1:5,000; cat. no. bs-0061R; BIOSS), FRA1 (1:1,000; cat. no. A5372; ABclonal Biotech Co., Ltd.), IL-1β (1:1,000; cat. no. HA601002; clone no. A7F9; HUABIO), IL-8 (1:1,000; cat. no. ab289967; clone no. EPR26511-74; Abcam), RANTES (1:1,000; cat. no. 36467; CST Biological Reagents Co., Ltd.), IL-6 (1:2,000; cat. no. DF6087; Affinity Biosciences, Ltd.), MCP-1 (1:2,000; cat. no. A7277; ABclonal Biotech Co., Ltd.), TNF-α (1:2,000; cat. no. 17590-1-AP; Proteintech Group, Inc.), and TGF-β (1:2,000; cat. no. 81746-2-RR; clone no. 230544B7; Proteintech Group, Inc.). The membranes were then incubated with HRP-conjugated secondary antibodies at a 1:50,000 dilution for 30 min at room temperature. The secondary antibodies use dincluded HRP-Goat anti Rabbit (cat. no. 5220-0336) and HRP-Goat anti Mouse (cat. no. 5220-0341; SeraCare; LGC Clinical Diagnostics). After further washing with TBST, bands were visualized using an ECL Chemiluminescence Detection Kit (Dalian Meilun Biology Technology Co., Ltd.) with the JS-1070P imaging system (Shanghai Peiqing Technology Co., Ltd.) and densitometric analysis was performed with IPWIN60 software (v6.0; Media Cybernetics, Inc.).

All analyses were performed in R (v4.0.3). For two-group comparisons of continuous variables the following tests were used: Unpaired Student's t-test for normally distributed data with equal variances; Welch's t-test for normal data with unequal variances; and the Wilcoxon rank-sum test for non-normal data. For multi-group comparisons, one-way ANOVA followed by Tukey's Honestly Significant Difference was applied to normal data with equal variances and the Kruskal-Wallis test followed by Dunn's post hoc test was applied for non-normal data. Pearson's correlation was used for normally distributed variables, and Spearman's correlation otherwise. Receiver operating characteristic curve analysis was employed to evaluate the predictive performance of upregulated DEGs for the disease. P<0.05 was considered to indicate a statistically significant difference.

Design of the present study is illustrated in Fig. 1. To determine gene expression differences between LN and healthy controls tubulointerstitial samples, the datasets GSE113342 and GSE200306 from GEO were analyzed. Using |log₂FC|>1 and adjusted P<0.05, 74 DEGs were identified in GSE113342 (Fig. 2A) and 16 DEGs were identified in GSE200306 (Fig. 2B). The intersection of these yielded 15 common DEGs (Fig. 2C), of which FRA1, C1R, CCL19, STAT1 and MX1 were upregulated in LN, whereas THY1, MME, ARG2, EGR1, KIT, IL6R, FKBP5, MRC1, ZBTB1 and RORC were downregulated. A heatmap illustrates their expression patterns across both datasets (Fig. 2D and E).

GO and KEGG enrichment analyses of the 15 shared DEGs revealed significant associations with immune and developmental processes (Fig. 3A). In GO-BP, DEGs were enriched in ‘T cell differentiation’, ‘kidney development’, ‘renal system development’ and ‘alpha-beta T cell activation’ (Fig. 3B). In GO-MF, terms included ‘histone acetyltransferase binding’, ‘transcription corepressor binding’, ‘CCR chemokine receptor binding’, ‘promoter-specific chromatin binding’ (Fig. 3B). KEGG pathways highlighted ‘Coronavirus disease-COVID-19’, ‘Hematopoietic cell lineage’, ‘Th17 cell differentiation’ and ‘Inflammatory bowel disease’ as pathways significantly associated with the DEGs (Fig. 3C).

To prioritize candidates for functional follow-up, the diagnostic performance of five upregulated DEGs was assessed, and FRA1 was identified as the top performer (AUC=0.977; Fig. 4); therefore, downstream analyses focused on FRA1. In addition, further analysis of the datasets showed that the trend of FRA1 remained stable across different subgroups and after exclusion of outliers, indicating that the upregulation of FRA1 in LN is robust and minimally affected by dataset heterogeneity (Fig. S2). Next, the relationship between FRA1 expression and immune cell infiltration in LN tubulointerstitial samples was explored using CIBERSORT on dataset GSE127797.

The results showed statistically increased proportions of CD8⁺ T cells, regulatory T cells (Tregs), monocytes, M1 macrophages and activated mast cells in the high-FRA1 group, whereas plasma cells, resting CD4⁺ memory T cells, M0/M2 macrophages, resting dendritic cells and resting mast cells were reduced (Fig. 5A-C). Correlation analyses further confirmed these associations. Specifically, Fig. 5D illustrates that FRA1 expression was significantly and positively correlated with several immune cell types, most notably naive CD4⁺ T cells, monocytes and Tregs. Conversely, FRA1 showed strong negative correlations with M0/M2 macrophages, resting CD4⁺ memory T cells and plasma cells. Furthermore, FRA1 expression was positively correlated with key tubular epithelial-related cytokines, including IL-6, IL-1β, IL-8, MCP-1, RANTES, TGF-β and TNF (Fig. 5E-K). Collectively, these findings underscore the diagnostic potential of FRA1 among upregulated DEGs and also suggest a key immunomodulatory role for FRA1 in the tubulointerstitial microenvironment of LN.

To verify the expression of FRA1 in LN kidneys, kidneys from 20-week-old MRL/lpr mice were used. Compared with the control group, H&E staining of MRL/lpr mouse kidneys showed marked inflammatory infiltration (Fig. 6A and B), and tubulointerstitial injury was a distinctive and prominent feature of the diseased kidneys. Moreover, IHC revealed significantly increased FRA1 expression in tubular epithelial cells of MRL/lpr kidneys compared with the control (Fig. 6C-I), which was corroborated by elevated protein levels in western blot analysis (Fig. 6J and K).

To investigate the functional role of FRA1 in renal tubular epithelial cells, the expression of key cytokines were measured in HK-2 cells following lentivirus-mediated FRA1 overexpression (FRA1-OE) or knockdown (FRA1-shRNA). Western blot analysis at 144 h post-infection showed that IL-6, IL-1β, IL-8, MCP-1 and RANTES were significantly upregulated in the FRA1-OE group, whereas IL-6 and RANTES levels were decreased in the FRA1-shRNA group (Fig. 7). TNF-α and TGF-β expression remained unchanged in both FRA1-OE and FRA1-shRNA cells. These results demonstrate that FRA1 modulates multiple proinflammatory cytokines in human renal tubular epithelial cells, suggesting its potential role as a regulator of the tubulointerstitial immune microenvironment in LN.